[{"repo":"trailhq/Graft","owner":"trailhq","name":"Graft","desc":"Turbocharge Claude Code, Cursor, Codex, Gemini & every coding agent: faster, cheaper, with contextual understanding specific to your codebase.","summaryCn":"给AI编程助手装上代码库地图，秒懂你的工程上下文的加速层。","reasonCn":"Graft是一层轻薄的加速外挂，趴在Claude Code、Cursor等主流AI编程代理上，通过构建代码库的图谱来喂给它们精准上下文。实测能省最多四倍成本、快三倍，还顺手提升正确率。它自动维护一份越用越灵的技能文件，团队协作越久，AI越懂你们的工程。安装只要一行命令，纯开发者工具。","stars":9021,"starGrowth":78,"createdAt":"2026-07-03T13:44:21Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools"],"aiDims":["AI Agent"],"zone":"AI","funScore":0,"domainTags":["编程助手增强","代码上下文理解","开发者工具"],"domainKey":"编程助手增强","pushedAt":"2026-09-23T20:11:19Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/trailhq/Graft","ts":"2026-09-23T20:11:19Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.38990921183904226,"copyOk":true},{"repo":"FareedKhan-dev/kimi-k3-in-c","owner":"FareedKhan-dev","name":"kimi-k3-in-c","desc":"A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.","summaryCn":"第一段：想象一下，平常需要上百美元的云 GPU 才能跑的大语言模型，现在只要","reasonCn":"第一段：想象一下，平常需要上百美元的云 GPU 才能跑的大语言模型，现在只要一台普通的笔记本，8 GB 内存就能完整推理，这听起来像科幻。Kimi K3-in-C 正是把这种不可能变为可能，它把原本 2.78 万亿参数的 Kimi 模型经过高效的权重量化和稀疏化，压缩到只占 8 GB，保持全精度推理","stars":8295,"starGrowth":52,"createdAt":"2026-08-01T09:29:38Z","silentRounds":0,"language":"C","topics":["deep-learning"],"aiDims":["推理引擎","大语言模型"],"zone":"AI","funScore":1,"domainTags":["大语言模型","CPU推理","权重量化","纯C实现"],"domainKey":"大语言模型","pushedAt":"2026-09-22T14:29:39Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/FareedKhan-dev/kimi-k3-in-c","ts":"2026-09-22T14:29:39Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.2622097649186257,"copyOk":false},{"repo":"DuarteSantos8/openGym","owner":"DuarteSantos8","name":"openGym","desc":"Self-hosted gym & body-weight tracker — plan routines, log workouts (supersets, warm-ups, cardio), see which muscles are trained, fatigued or detrained, import from FitNotes/Strong/Hevy, passkey login. Your data, your server.","summaryCn":"自托管的健身记录平台，支持计划、日志、肌肉分析，一键导入主流APP数据","reasonCn":"openGym 是基于 JavaScript 的自托管健身追踪系统，提供计划编排、训练日志、超集与热身管理、心率与卡路里统计等功能，并可视化显示各肌肉群的训练频率、疲劳度和退化趋势。它支持从 FitNotes、Strong、Hevy 等流行应用导入数据，使用 Passkey 登录保障安全，前端采用 React，后端 Node.js + SQLite，部署简单，适合个人或小团队自行搭建专属健身记录平台。","stars":1279,"starGrowth":64,"createdAt":"2026-07-18T20:41:24Z","language":"JavaScript","topics":["fitness"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["健身追踪","自托管","运动记录"],"domainKey":"健身追踪","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/DuarteSantos8/openGym","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.5131759186864738,"copyOk":true},{"repo":"byoungd/up","owner":"byoungd","name":"up","desc":"An advanced guide which might benefit you a lot 🎉 . 韩先凯的人生进阶指南 人生进阶指南 离谱的人生 人生进阶 AI学习 AI指南 韩先凯的AI学习指南 英语学习指南/英语学习教程/英语学习/学英语","summaryCn":"从零基础到英语流利交流的完整学习指南，58000多人收藏","reasonCn":"这是一个英语学习进阶指南仓库，涵盖词汇记忆、语法体系、口语实战、听力训练等全链路学习方法。内置大量实用技巧，如词根词缀记忆法、影子跟读训练、场景化对话练习。项目结构清晰，从入门到进阶层层递进，适合各阶段学习者系统提升英语能力，star超过5.8万说明社区认可度很高。","stars":63706,"starGrowth":218,"createdAt":"2017-05-30T07:18:52Z","language":"JavaScript","topics":["tutorial"],"aiDims":["非AI-实用","学习资源","教程"],"zone":"资源","funScore":0.5,"domainTags":["英语学习","自学方法论","词汇记忆","口语训练"],"domainKey":"英语学习","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/byoungd/up","ts":"2026-09-23T21:53:43.262Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3102659090195586,"copyOk":false},{"repo":"anywhere-labs/dsh-desktop","owner":"anywhere-labs","name":"dsh-desktop","desc":"为 DeepSeek Harness (DSH) 插件生态打造的现代化桌面端解决方案。万物皆「插件」，桌面本身也是「插件」。","summaryCn":"把本地AI网页版做成原生桌面App，手机也能远程控制。","reasonCn":"这是一款基于 DeepSeek Harness 构建的开源桌面客户端，解决了本地运行AI模型需要繁琐配置的问题。它把原本需要浏览器访问的AI界面打包成原生应用，支持 Windows 和 macOS，无需安装 Node.js 即可一键使用。最特别的是，它还集成了手机远程控制功能，让你在移动端也能管理桌面上的AI任务。","stars":28507,"starGrowth":224,"createdAt":"2026-08-13T16:30:50Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["AI应用","工具调用","开发者工具"],"zone":"AI","funScore":1,"domainTags":["桌面应用","AI插件生态","DeepSeek","Electron"],"domainKey":"桌面应用","pushedAt":"2026-09-23T18:55:19Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/anywhere-labs/dsh-desktop","ts":"2026-09-23T18:55:19Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.4853597362051426,"copyOk":true},{"repo":"undefined-ui/second-brain-os","owner":"undefined-ui","name":"second-brain-os","desc":"An AI second brain that maintains itself. Full guide, starter vault, agent skills and scripts for a self-organizing knowledge base in Claude Code and Obsidian.","summaryCn":"让AI帮你自动整理知识库，像有个人脑力一样","reasonCn":"这是一个基于Claude Code和Obsidian的AI第二大脑操作系统。它提供了完整的启动指南、初始知识库模板以及Agent技能脚本，旨在构建一个能够自我组织的知识库系统。通过自动化脚本，它可以协助用户整理笔记、梳理逻辑，让知识管理变得更加智能和高效。","stars":143,"starGrowth":80,"createdAt":"2026-09-07T15:08:13Z","silentRounds":0,"language":"HTML","topics":["knowledge-management"],"aiDims":["AI Agent","AI基础设施","AI搜索"],"zone":"资源","funScore":0.7,"domainTags":["知识管理","Obsidian","Agent脚本","第二大脑","自我组织知识库"],"domainKey":"知识管理","pushedAt":"2026-09-23T19:25:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/undefined-ui/second-brain-os","ts":"2026-09-23T19:25:17Z","category":"learning","momentum":["rising","daily"],"fromOfficial":false,"heatScore":2.8774825174825174,"copyOk":false},{"repo":"henrygd/beszel","owner":"henrygd","name":"beszel","desc":"Lightweight server monitoring with historical data, docker stats, and alerts.","summaryCn":"极简自托管服务器监控平台，内置Docker与ZFS深度统计","reasonCn":"Beszel 是一款用 Go 编写的轻量级自托管监控平台，二进制仅十兆级别，无需数据库即可开箱即用。它原生集成 Docker 容器的 CPU、内存、网络历史曲线，并能采集 ZFS 存储池容量、健康度、I/O 与数据集级用量。支持多用户隔离、OAuth 登录与自动备份，告警覆盖 CPU、内存、磁盘、带宽、温度、风扇转速、负载等十余项指标，提供 Web 界面与 API 双通道。","stars":25638,"starGrowth":39,"createdAt":"2024-07-07T21:36:28Z","silentRounds":0,"language":"Go","topics":["self-hosted","monitoring"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["服务器监控","Docker监控","自托管运维"],"domainKey":"服务器监控","pushedAt":"2026-09-23T21:07:30Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/henrygd/beszel","ts":"2026-09-23T21:07:30Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.1215635384975427,"copyOk":true},{"repo":"mvt-project/mvt","owner":"mvt-project","name":"mvt","desc":"MVT (Mobile Verification Toolkit) helps with conducting forensics of mobile devices in order to find signs of a potential compromise.","summaryCn":"手机疑似被黑？用它扫一遍就能揪出间谍软件","reasonCn":"MVT是一款移动设备取证工具，专门用于检测手机是否被植入间谍软件或恶意应用。支持Android和iOS平台，可提取设备文件、SQLite数据库、TCC权限库等信息，并结合YARA规则进行扫描检测，可识别Pegasus、Flux、Cobalt Strike等知名间谍软件留下的痕迹。","stars":14448,"starGrowth":546,"createdAt":"2021-07-16T05:50:44Z","language":"Python","topics":["mobile","security"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["移动设备取证","间谍软件检测","数字取证"],"domainKey":"移动设备取证","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/mvt-project/mvt","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.4133720930232558,"copyOk":false},{"repo":"projectdiscovery/nuclei","owner":"projectdiscovery","name":"nuclei","desc":"Nuclei is a fast, customizable vulnerability scanner powered by the global security community and built on a simple YAML-based DSL, enabling collaboration to tackle trending vulnerabilities on the internet. It helps you find vulnerabilities in your applications, APIs, networks, DNS, and cloud configurations.","summaryCn":"基于YAML模板的高性能漏洞扫描器，支持多协议并行检测。","reasonCn":"Nuclei 是一款现代化漏洞扫描器，采用简洁的 YAML 模板定义检测逻辑，由全球安全专家贡献数千个模板覆盖热门漏洞。它通过模拟真实攻击步骤验证漏洞，实现零误报，并支持超高速并行扫描与请求聚类。可无缝集成至 CI/CD 流水线，支持 HTTP、DNS、TCP、SSL、JavaScript 等多种协议，适合安全团队与开发者进行自动化安全测试与回归验证。","stars":31464,"starGrowth":37,"createdAt":"2020-04-03T18:47:11Z","language":"Go","topics":["security"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["漏洞扫描","安全测试","YAML DSL","红队工具"],"domainKey":"漏洞扫描","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/projectdiscovery/nuclei","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.11452784134248666,"copyOk":true},{"repo":"coollabsio/coolify","owner":"coollabsio","name":"coolify","desc":"An open-source, self-hostable PaaS alternative to Vercel, Heroku & Netlify that lets you easily deploy static sites, databases, full-stack applications and 280+ one-click services on your own servers.","summaryCn":"开源自托管PaaS，像用Heroku一样管理自己的服务器","reasonCn":"Coolify让你把闲置VPS或树莓派变成私有云平台，通过SSH接管服务器后，即可像用Heroku一样一键部署应用、数据库和静态站点，自动申请SSL证书并配置反向代理。所有配置数据仅存本地，即使卸载面板，运行中的容器仍保持原样可正常访问，彻底避免厂商锁定。支持Docker Compose、Node.js、Python、Go等多种应用类型，内置PostgreSQL、Redis、MySQL等一键数据库服务，适合想完全掌控数据与成本的开发者和小团队。","stars":62155,"starGrowth":28,"createdAt":"2021-01-25T20:54:01Z","silentRounds":0,"language":"PHP","topics":["self-hosted","docker"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["自托管PaaS","Docker部署","服务器管理","SSL证书","应用部署"],"domainKey":"自托管PaaS","pushedAt":"2026-09-23T21:17:22Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/coollabsio/coolify","ts":"2026-09-23T21:17:22Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08535146005952861,"copyOk":true},{"repo":"dataelement/dsh-desktop","owner":"dataelement","name":"dsh-desktop","desc":"DSHDesktop：DeepSeek Harness Desktop / DeepSeek Harness 桌面版","summaryCn":"DeepSeek大模型桌面版，本地跑AI对话无需联网","reasonCn":"基于TypeScript开发的跨平台桌面客户端，将DeepSeek大模型能力集成到本地应用中。支持对话交互、多轮上下文管理和API调用，用户无需依赖云端服务即可在本地体验DeepSeek的AI能力，特别适合注重隐私和数据安全的场景。","stars":8458,"starGrowth":201,"createdAt":"2026-08-13T13:45:50Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["AI应用","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["DeepSeek","桌面应用","本地部署","大模型客户端","隐私"],"domainKey":"DeepSeek","pushedAt":"2026-09-23T16:08:30Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/dataelement/dsh-desktop","ts":"2026-09-23T16:08:30Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.5569401749822653,"copyOk":true},{"repo":"xerj-org/xerj","owner":"xerj-org","name":"xerj","desc":"XERJ is the new way for AI to search data. Its autoindex capability activates agents to know your data without the token waste of grep and sed. One command indexes code, docs, logs and PDFs for search, RAG, security audits and agent memory, using 40x fewer tokens than grep. Elasticsearch compatible, so existing clients just work.","summaryCn":"一行命令让AI自动读懂你的代码文档日志PDF，搜索再也不靠grep","reasonCn":"XERJ是一个基于Rust构建的AI数据搜索工具，支持一键索引代码、文档、日志和PDF文件。它利用向量数据库和自动索引技术，让AI agent无需grep和sed等命令行工具就能理解你的数据。相比传统搜索方式，它能进行语义层面的精准匹配，找到真正相关的内容而非仅靠关键词。1305个star说明社区对其实用价值有认可，适合需要频繁检索项目数据的开发者和团队。","stars":2350,"starGrowth":89,"createdAt":"2026-06-30T04:32:45Z","silentRounds":0,"language":"Rust","topics":["vector-db"],"aiDims":["AI搜索","RAG与知识","向量数据库"],"zone":"AI","funScore":0.6,"domainTags":["向量数据库","代码搜索","RAG"],"domainKey":"向量数据库","pushedAt":"2026-09-23T11:30:07Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/xerj-org/xerj","ts":"2026-09-23T11:30:07Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.5709255319148936,"copyOk":false},{"repo":"Dokploy/dokploy","owner":"Dokploy","name":"dokploy","desc":"Open Source Alternative to Vercel, Netlify and Heroku.","summaryCn":"开源自托管 PaaS，一键部署应用与数据库，支持多节点集群。","reasonCn":"Dokploy 是一款开源自托管 PaaS 平台，旨在简化应用与数据库的部署运维。它支持 Node.js、Python、Go 等多语言应用一键部署，内置 MySQL、PostgreSQL、MongoDB 等数据库管理，并提供自动化备份至外部存储功能。原生兼容 Docker Compose 编排复杂服务，更支持基于 Docker Swarm 的多节点集群扩展，适合追求数据主权、不想依赖云厂商的开发者或中小团队自建类 Heroku 环境。","stars":37459,"starGrowth":28,"createdAt":"2024-04-19T07:49:30Z","silentRounds":0,"language":"TypeScript","topics":["devops","ai","agents"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["自托管PaaS","容器应用部署","DevOps平台"],"domainKey":"自托管PaaS","pushedAt":"2026-09-23T21:50:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/Dokploy/dokploy","ts":"2026-09-23T21:50:17Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08624245174724365,"copyOk":true},{"repo":"bojieli/ai-agent-book","owner":"bojieli","name":"ai-agent-book","desc":"《深入理解 AI Agent：设计原理与工程实践》（李博杰 著）开源主仓库：全书正文、编译版 PDF 与按章配套代码","summaryCn":"一本能亲手跑实验的开源AI Agent实战手册，从原理讲到工程。","reasonCn":"这是一本围绕“Agent = LLM + 上下文 + 工具”核心公式展开的开源电子书，用十章篇幅把AI Agent从设计原理讲到工程实战。全书正文、配图以及109个配套实验全部开源，读者可以边读边动手跑实验，把抽象概念落到真实代码里。适合想系统学习Agent开发、又不想只看空泛理论的开发者。","stars":50068,"starGrowth":369,"createdAt":"2025-09-09T02:41:44Z","silentRounds":0,"language":"Python","topics":["llm","rag","ai-agent"],"aiDims":[],"zone":"资源","funScore":0,"domainTags":["AI Agent教程","LLM应用","RAG实战","中文技术书"],"domainKey":"AI Agent教程","pushedAt":"2026-09-23T12:03:33Z","tags":[{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/bojieli/ai-agent-book","ts":"2026-09-23T12:03:33Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3221099304945274,"copyOk":true},{"repo":"hoppscotch/hoppscotch","owner":"hoppscotch","name":"hoppscotch","desc":"Open-Source API Development Ecosystem • https://hoppscotch.io • Offline, On-Prem & Cloud • Web, Desktop & CLI • Open-Source Alternative to Postman, Insomnia","summaryCn":"轻量开源的API调试与开发环境，支持实时HTTP请求。","reasonCn":"Hoppscotch是一款开源的API开发生态系统，旨在帮助开发者高效测试和调试HTTP请求。它提供直观的界面，支持GET、POST等多种请求方法，并能实时查看响应结果。项目由社区驱动，采用TypeScript编写，强调轻量化和极速响应，适合前端与后端开发者在日常工作中使用。","stars":80471,"starGrowth":21,"createdAt":"2019-08-21T13:15:24Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["API调试","REST","GraphQL","自托管工具"],"domainKey":"API调试","pushedAt":"2026-09-23T20:33:46Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/hoppscotch/hoppscotch","ts":"2026-09-23T20:33:46Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06378289073082229,"copyOk":true},{"repo":"huggingface/transformers","owner":"huggingface","name":"transformers","desc":"🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.","summaryCn":"一个集视觉、语音、多模态于一体的模型库，统一接口驱动万物智能。","reasonCn":"它是机器学习界的\"模型中心\"，将文本、图像、音频、视频等各类顶尖预训练模型汇聚一处，提供统一的加载、微调和部署接口。写几行代码就能调用数千个模型，无论是做聊天机器人还是图像分类，都能从这套框架轻松起步。","stars":166536,"starGrowth":52,"createdAt":"2018-10-29T13:56:00Z","silentRounds":0,"language":"Python","topics":["llm","ml"],"aiDims":["AI基础设施","大语言模型","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["机器学习框架","预训练模型","模型微调","多模态AI","自然语言处理"],"domainKey":"机器学习框架","pushedAt":"2026-09-23T21:11:38Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.98,"source":"trending","url":"https://github.com/huggingface/transformers","ts":"2026-09-23T21:11:38Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.1569367343997694,"copyOk":true},{"repo":"BerriAI/litellm","owner":"BerriAI","name":"litellm","desc":"The fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format with cost tracking, guardrails, load balancing, and logging [Bedrock, Azure, OpenAI, Anthropic, OpenAI, VertexAI, vLLM, Nvidia NIM]","summaryCn":"开源AI网关，统一100+大模型接口，支持自托管部署。","reasonCn":"LiteLLM 是一个开源 AI 网关，解决了调用百余家大模型厂商接口不统一、难管理的问题。它把 OpenAI、Anthropic、Gemini 等模型全都封装成 OpenAI 格式，开发者只需改一行代码或配置代理地址，即可无缝切换模型。它既可作为 Python SDK 直接集成，也可部署为代理服务器供团队统一使用，内置负载均衡、降级重试、预算控制与详细日志审计，满足企业级生产环境对稳定性与成本管控的严苛要求。","stars":59347,"starGrowth":82,"createdAt":"2023-07-27T00:09:52Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI基础设施","工具调用","效率工具"],"zone":"工具","funScore":0,"domainTags":["LLM网关","API聚合","成本追踪","负载均衡","多模型路由"],"domainKey":"LLM网关","pushedAt":"2026-09-23T21:51:28Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/BerriAI/litellm","ts":"2026-09-23T21:51:28Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.25014511264259354,"copyOk":true},{"repo":"gethomepage/homepage","owner":"gethomepage","name":"homepage","desc":"A highly customizable homepage (or startpage / application dashboard) with Docker and service API integrations.","summaryCn":"自托管导航页，集成百种服务，界面现代且加载飞快。","reasonCn":"这是自托管应用仪表盘，把散落在各端口、各机器的服务统一收进一个漂亮首页。支持百种服务集成，配置走YAML文件或Docker标签自动发现，API请求经代理转发保护密钥安全。静态生成保证秒开，适合折腾NAS和Docker的家庭用户、技术爱好者。","stars":32791,"starGrowth":12,"createdAt":"2022-08-24T07:29:42Z","silentRounds":0,"language":"JavaScript","topics":["self-hosted","docker"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["自托管","Dashboard","Docker","服务聚合","HomeLab"],"domainKey":"自托管","pushedAt":"2026-09-23T19:27:55Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/gethomepage/homepage","ts":"2026-09-23T19:27:55Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0370978622182916,"copyOk":true},{"repo":"usebruno/bruno","owner":"usebruno","name":"bruno","desc":"Opensource IDE For Exploring and Testing API's (lightweight alternative to Postman/Insomnia)","summaryCn":"离线优先的API客户端，用Git管理集合，数据永不上云。","reasonCn":"Bruno 是一款开源 API 客户端，主打离线优先与数据隐私。它摒弃了 Postman 等工具依赖云同步的模式，将请求集合以纯文本 Bru 格式直接存储在本地文件夹中。这意味着你可以像管理代码一样用 Git 进行版本控制、团队协作与代码审查，彻底避免敏感数据离开设备。核心功能完全免费开源，适合重视数据主权与开发体验的工程师。","stars":47140,"starGrowth":28,"createdAt":"2022-09-27T20:51:45Z","language":"JavaScript","topics":["developer-tools","automation"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["API测试","开发工具","Postman替代品","本地优先"],"domainKey":"API测试","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.52,"source":"trending","url":"https://github.com/usebruno/bruno","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08578192617734409,"copyOk":true},{"repo":"nautechsystems/nautilus_trader","owner":"nautechsystems","name":"nautilus_trader","desc":"Production-grade Rust-native trading engine with deterministic event-driven architecture","summaryCn":"用Rust打造的极速交易引擎，让量化策略回测快如闪电","reasonCn":"Rust原生高性能生产级交易引擎，采用确定性事件驱动架构，专为量化交易和策略回测设计。支持对接多个主流交易所，提供完整的订单管理、风险管理及回测框架。内置ML模块支持策略优化与预测分析，代码质量极高，性能业界顶尖，是金融科技领域不可多得的开源利器。","stars":29274,"starGrowth":27,"createdAt":"2018-06-25T06:26:16Z","silentRounds":0,"language":"Rust","topics":["ml"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["量化交易","回测引擎","算法交易","Rust"],"domainKey":"量化交易","pushedAt":"2026-09-23T21:32:04Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.52,"source":"trending","url":"https://github.com/nautechsystems/nautilus_trader","ts":"2026-09-23T21:32:04Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.08376696044271367,"copyOk":true},{"repo":"n8n-io/n8n","owner":"n8n-io","name":"n8n","desc":"Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.","summaryCn":"面向AI Agent的工作流自动化平台，支持可视化编排与代码混合开发。","reasonCn":"n8n 是一个面向 AI Agent 的工作流自动化平台，提供可视化画布与代码混合编程能力。它内置 1500 多个集成节点，支持主流大模型与开源模型无缝切换，用户可在画布上拖拽构建包含逻辑判断、工具调用、人工审批的多步骤 AI 工作流。平台支持自托管部署与云端托管两种模式，数据与执行过程完全可控，适合从原型验证到生产环境的完整生命周期管理。","stars":205716,"starGrowth":101,"createdAt":"2019-06-22T09:24:21Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI应用","AI Agent","效率工具"],"zone":"工具","funScore":0,"domainTags":["工作流自动化","低代码平台","AI集成","API集成","自托管"],"domainKey":"工作流自动化","pushedAt":"2026-09-23T21:28:37Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/n8n-io/n8n","ts":"2026-09-23T21:28:37Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3014729043924634,"copyOk":true},{"repo":"ruvnet/ruflo","owner":"ruvnet","name":"ruflo","desc":"🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, RAG integration, and native Claude Code / Codex / Hermes and many more Integrated","summaryCn":"Claude Code与Codex的智能体执行层。","reasonCn":"Ruflo是一个专为Claude Code和Codex打造的AI智能体执行层框架。它通过提供工具、记忆、循环和沙箱控制，让AI模型不仅能“写代码”，还能真正“干活”。项目核心在于通过100多个专业化智能体和协调蜂群，赋予模型协作与改进能力，并支持跨机器的联邦通信与企业级安全护栏。","stars":73076,"starGrowth":72,"createdAt":"2025-06-02T21:24:20Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["多智能体编排","Agent工作流","自主协作系统"],"domainKey":"多智能体编排","pushedAt":"2026-09-23T18:57:01Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/ruvnet/ruflo","ts":"2026-09-23T18:57:01Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.21895582681044393,"copyOk":true},{"repo":"semantica-agi/semantica","owner":"semantica-agi","name":"semantica","desc":"Graph-Native Infrastructure for Context and Accountable AI Systems","summaryCn":"把企业数据变成知识图谱，让AI决策可解释、可追溯","reasonCn":"这是一个面向开发者的AI知识基础设施，能将企业杂乱的数据自动梳理成上下文图谱。它不只做数据检索，还能在图谱上进行因果分析和推理，每一步决策都留有溯源记录。相比动辄昂贵的商业平台，它主打开发者友好和成本可控，让AI系统天然具备可解释性。","stars":13400,"starGrowth":35,"createdAt":"2025-06-25T10:41:00Z","silentRounds":0,"language":"Python","topics":["developer-tools"],"aiDims":[],"zone":"AI","funScore":0.6,"domainTags":["知识图谱","AI记忆系统","RAG","图数据库","Agent框架","上下文管理"],"domainKey":"知识图谱","pushedAt":"2026-09-23T15:41:08Z","tags":[{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/semantica-agi/semantica","ts":"2026-09-23T15:41:08Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.11283582089552238,"copyOk":true},{"repo":"juspay/hyperswitch","owner":"juspay","name":"hyperswitch","desc":"Open source, composable payments platform | PCI compliant | SaaS and Self-host options | Enables connectivity to multiple payment, payout, fraud, vault and tokenization providers | Uplifts authorization with intelligent routing and revenue recovery | Reduce payment processing costs with cost observability | Reduces payment ops with reconciliation","summaryCn":"开源支付编排平台，统一对接多家支付渠道，支持模块化按需集成。","reasonCn":"Hyperswitch 是用 Rust 编写的开源支付编排基础设施，解决企业对接多家支付网关繁琐、易被厂商锁定的问题。它将支付流程拆解为独立模块，支持路由、结算、对账等核心能力，开发者可像搭积木一样仅集成所需模块，快速构建自主可控的支付体系，避免重复造轮子。","stars":43707,"starGrowth":81,"createdAt":"2022-10-17T11:18:28Z","silentRounds":0,"language":"Rust","topics":["finance"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["支付网关","金融科技","API集成","开源软件","PCI合规"],"domainKey":"支付网关","pushedAt":"2026-09-23T20:49:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/juspay/hyperswitch","ts":"2026-09-23T20:49:06Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2485597501544375,"copyOk":true},{"repo":"Stirling-Tools/Stirling-PDF","owner":"Stirling-Tools","name":"Stirling-PDF","desc":"#1 PDF Application on GitHub that lets you edit PDFs on any device anywhere","summaryCn":"GitHub上最火的离线PDF工具箱，全能编辑处理神器","reasonCn":"这是一个功能极其强大的PDF处理工具，支持合并、拆分、旋转、加水印、OCR识别等几十种功能。全在浏览器本地运行，无需上传服务器，完美解决隐私担忧。Java开发，Docker一键部署，是企业和个人管理文档的终极利器，界面友好且性能强劲。","stars":92826,"starGrowth":70,"createdAt":"2023-01-27T18:22:42Z","silentRounds":0,"language":"Java","topics":["docker","java","pdf","pdf-converter","pdf-manipulation","pdf-merger","pdf-tools","pdf-web-apps","pdf-ocr","pdfmerger","pdf-editor","hacktoberfest","self-hosted"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":1,"domainTags":["PDF处理","本地隐私","Docker部署"],"domainKey":"PDF处理","pushedAt":"2026-09-23T21:34:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"java","source":"github","weight":0.7},{"name":"pdf","source":"github","weight":0.7},{"name":"pdf-converter","source":"github","weight":0.7},{"name":"pdf-manipulation","source":"github","weight":0.7},{"name":"pdf-merger","source":"github","weight":0.7},{"name":"pdf-tools","source":"github","weight":0.7},{"name":"pdf-web-apps","source":"github","weight":0.7},{"name":"pdf-ocr","source":"github","weight":0.7},{"name":"pdfmerger","source":"github","weight":0.7},{"name":"pdf-editor","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/Stirling-Tools/Stirling-PDF","ts":"2026-09-23T21:34:57Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.21226229720121517,"copyOk":false},{"repo":"DeusData/codebase-memory-mcp","owner":"DeusData","name":"codebase-memory-mcp","desc":"High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.","summaryCn":"给AI装上一份代码库的超级大脑，查代码快如闪电","reasonCn":"这是一个基于MCP协议的高性能代码智能服务器，能把任意代码仓库在毫秒级索引进知识图谱，支持158种编程语言。它用纯C编写，打包成单个静态二进制文件，查询延迟低于毫秒级，Token消耗减少99%。AI助手接入后能像理解自己代码一样理解别人的代码库，是构建代码理解Agent的核心基础设施。GitHub星标超3.8万，是MCP生态中最热门的代码智能项目。","stars":44514,"starGrowth":266,"createdAt":"2026-02-24T22:01:00Z","language":"C","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","知识图谱"],"zone":"AI","funScore":0.6,"domainTags":["代码知识图谱","MCP服务器","代码智能","静态二进制","多语言解析"],"domainKey":"代码知识图谱","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/DeusData/codebase-memory-mcp","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3815123331985442,"copyOk":false},{"repo":"infiniflow/ragflow","owner":"infiniflow","name":"ragflow","desc":"RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs","summaryCn":"融合Agent能力的开源RAG引擎，提供企业级文档理解与知识检索方案","reasonCn":"RAGFlow是一款开源RAG引擎，深度融合Agent技术解决复杂文档理解难题。它能自动处理PDF、表格等非结构化数据，构建高质量知识库，支持多路检索与混合排序，显著提升大模型回答准确性。项目提供Docker一键部署，适配任意规模企业场景，是构建企业级AI知识助手的核心基础设施。","stars":91173,"starGrowth":50,"createdAt":"2023-12-12T06:13:13Z","silentRounds":0,"language":"Go","topics":["ai","rag"],"aiDims":["RAG与知识","AI应用","AI Agent"],"zone":"AI","funScore":0,"domainTags":["RAG引擎","文档解析","知识图谱","Agent规划","企业知识库"],"domainKey":"知识图谱","pushedAt":"2026-09-23T15:23:46Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/infiniflow/ragflow","ts":"2026-09-23T15:23:46Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1516452239149748,"copyOk":true},{"repo":"ast-grep/ast-grep","owner":"ast-grep","name":"ast-grep","desc":"⚡A CLI tool for code structural search, lint and rewriting. Written in Rust","summaryCn":"代码结构搜索神器，比grep更聪明，支持跨语言重构和批量改写","reasonCn":"技术上，ast-grep的核心优势在于AST解析。它支持Rust、TypeScript、Python、Go等主流语言的语法解析，然后用类似正则表达式的查询语法来匹配代码模式。比如你想找到所有调用某个函数的地方，可以写一个查询模板，ast-grep会精确匹配并替换。","stars":16002,"starGrowth":13,"createdAt":"2022-07-01T19:21:00Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["代码搜索","AST解析","代码重构","CLI工具"],"domainKey":"代码搜索","pushedAt":"2026-09-23T21:49:22Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/ast-grep/ast-grep","ts":"2026-09-23T21:49:22Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04143719535058118,"copyOk":true},{"repo":"dani-garcia/vaultwarden","owner":"dani-garcia","name":"vaultwarden","desc":"Unofficial Bitwarden compatible server written in Rust, formerly known as bitwarden_rs","summaryCn":"资源占用极低的 Bitwarden 自托管替代方案。","reasonCn":"Vaultwarden 是一款用 Rust 编写的自托管密码管理服务器，它完美兼容官方 Bitwarden 客户端，但资源消耗要低得多。它非常适合那些希望将密码管理服务完全掌控在自己手中，又不想被官方庞大服务拖累的用户。通过 Docker 部署，你可以轻松搭建一个私密、安全且高效的密码中心。","stars":68026,"starGrowth":44,"createdAt":"2018-02-17T22:40:20Z","silentRounds":0,"language":"Rust","topics":["docker"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["密码管理","自托管","Bitwarden兼容","端到端加密"],"domainKey":"密码管理","pushedAt":"2026-09-23T21:22:45Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/dani-garcia/vaultwarden","ts":"2026-09-23T21:22:45Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1339404345397348,"copyOk":true},{"repo":"4ian/GDevelop","owner":"4ian","name":"GDevelop","desc":"🎮 Open-source, cross-platform 2D/3D/multiplayer game engine designed for everyone.","summaryCn":"零代码也能做游戏的开源引擎，2D到3D到多人在线都能搞","reasonCn":"GDevelop的核心技术亮点在于它的可视化事件系统。你不需要写一行代码，只需要用图形化的方式设置「当玩家碰到金币时，金币消失并加10分」这样的逻辑规则。它同时支持导入代码进行扩展，对会编程的用户也很友好。引擎支持2D和3D渲染，内置了物理引擎、动画系统、粒子效果等常用游戏","stars":26754,"starGrowth":26,"createdAt":"2014-06-29T19:58:38Z","silentRounds":0,"language":"JavaScript","topics":["game"],"aiDims":["游戏","创意工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","可视化开发","游戏开发"],"domainKey":"游戏引擎","pushedAt":"2026-09-23T18:54:06Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/4ian/GDevelop","ts":"2026-09-23T18:54:06Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.08091545189504373,"copyOk":true},{"repo":"vllm-project/vllm","owner":"vllm-project","name":"vllm","desc":"A high-throughput and memory-efficient inference and serving engine for LLMs","summaryCn":"专为 LLM 的高效推理与部署打造的极速框架。","reasonCn":"这是一个专为大语言模型设计的推理与服务库，旨在解决模型部署慢、吞吐量低的痛点。它通过创新的 PagedAttention 技术，极大地提升了内存利用率并加速了推理过程，让开发者和企业能够更轻松、快速地提供 AI 服务。项目由伯克利大学 Sky Computing 实验室发起，背后汇聚了来自全球 2000 多位贡献者的智慧，是目前开源 AI 领域最活跃的项目之一。","stars":92445,"starGrowth":80,"createdAt":"2023-02-09T11:23:20Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["推理引擎","大语言模型","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["大模型推理","推理加速","PagedAttention","模型部署","高吞吐服务"],"domainKey":"模型部署","pushedAt":"2026-09-23T21:16:52Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/vllm-project/vllm","ts":"2026-09-23T21:16:52Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.2425961382443615,"copyOk":true},{"repo":"vercel-labs/agent-browser","owner":"vercel-labs","name":"agent-browser","desc":"Browser automation CLI for AI agents","summaryCn":"为 AI 代理提供浏览器操作能力的命令行工具。","reasonCn":"这是一个用 Rust 编写的命令行工具，专门设计给 AI 智能体使用。它能帮助 AI 代理像人类一样在浏览器中执行操作，比如点击按钮、填写表单或抓取网页内容。通过原生 Rust 的性能优势，它比传统的自动化脚本更快、更可靠，非常适合集成到各种 AI 应用中，让 AI 能够真正“操作”互联网。","stars":43122,"starGrowth":70,"language":"Rust","topics":[],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["浏览器自动化","AI Agent工具","Playwright"],"domainKey":"浏览器自动化","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/vercel-labs/agent-browser","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.21486990399332126,"copyOk":true},{"repo":"frappe/frappe","owner":"frappe","name":"frappe","desc":"Low code web framework for real world applications, in Python and Javascript","summaryCn":"用Python快速搭建企业级Web应用的低代码框架","reasonCn":"这是一个基于Python和JavaScript的低代码Web框架，自带用户权限、工作流、API等完整功能栈。只需定义数据模型，就能自动生成CRUD界面和API。ERPNext、Frappe Books等知名应用都基于它构建，适合快速开发内部管理工具和企业应用。","stars":10822,"starGrowth":9,"createdAt":"2011-06-08T08:14:16Z","silentRounds":0,"language":"Python","topics":["web-framework","security"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["低代码","企业应用","Web框架","ERP"],"domainKey":"低代码","pushedAt":"2026-09-23T20:05:25Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/frappe/frappe","ts":"2026-09-23T20:05:25Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.029494917760118276,"copyOk":true},{"repo":"wwebjs/whatsapp-web.js","owner":"wwebjs","name":"whatsapp-web.js","desc":"A WhatsApp client library for NodeJS that connects through the WhatsApp Web browser app","summaryCn":"用代码操控WhatsApp的神器，自动发消息建机器人超简单","reasonCn":"whatsapp-web.js让你能用Node.js代码控制WhatsApp Web，实现自动发送消息、管理群组、接收消息等功能。它可以用来搭建客服机器人、群发消息通知，或者自动化日常操作流程。虽然不是AI项目，但结合AI可以轻松打造智能WhatsApp助手，实用价值很高。","stars":22621,"starGrowth":7,"createdAt":"2019-02-17T02:16:02Z","language":"JavaScript","topics":["api"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WhatsApp自动化","Node.js库","消息机器人开发"],"domainKey":"WhatsApp自动化","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/wwebjs/whatsapp-web.js","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.021928340922151984,"copyOk":false},{"repo":"floci-io/floci","owner":"floci-io","name":"floci","desc":"Light, fluffy, and always free - The AWS Local Emulator alternative","summaryCn":"无需账号令牌的本地AWS模拟器，Docker一键启动。","reasonCn":"Floci是一款专为开发与测试设计的本地云环境模拟器。它解决了开发者需要在本地构建云应用、运行测试却不想频繁切换环境的痛点。通过在本地部署，Floci提供了一个无需云端账户或认证令牌的AWS服务模拟环境。开发者可以像使用真实AWS一样，通过Docker Compose命令快速启动服务，从而在断网或离线状态下完成CI/CD流程和功能验证，极大地简化了本地开发环境的搭建与维护工作。","stars":25211,"starGrowth":183,"createdAt":"2026-02-18T02:28:29Z","silentRounds":0,"language":"Java","topics":["devops","docker"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["本地AWS模拟","Serverless开发","容器化模拟"],"domainKey":"本地AWS模拟","pushedAt":"2026-09-23T21:08:42Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/floci-io/floci","ts":"2026-09-23T21:08:42Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.38613145055729636,"copyOk":true},{"repo":"lidge-jun/opencodex","owner":"lidge-jun","name":"opencodex","desc":"Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code","summaryCn":"把Claude等大模型接入代码编辑器的万能胶水","reasonCn":"这是一个通用代理服务器，能把任何支持OpenAI协议的LLM（如Claude、DeepSeek）接入到各类代码开发场景。它支持Codex CLI、App、SDK和Claude Code，解决了多工具间协议不通的痛点，让开发者无需修改代码即可切换模型。","stars":15942,"starGrowth":163,"createdAt":"2026-06-18T23:03:18Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI代理","编程工具","模型适配","OpenAI兼容","开发工具"],"domainKey":"AI代理","pushedAt":"2026-09-23T12:53:44Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/lidge-jun/opencodex","ts":"2026-09-23T12:53:44Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3968084305607828,"copyOk":true},{"repo":"Comfy-Org/ComfyUI","owner":"Comfy-Org","name":"ComfyUI","desc":"The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.","summaryCn":"AI绘图的'乐高工厂'，可视化搭积木画出来","reasonCn":"ComfyUI 以节点式可视化界面著称，通过拖拽连接功能模块构建图像生成工作流，支持 Stable Diffusion、SDXL、Flux 等多种模型，可自由组合采样器、高清修复、ControlNet 等插件，实现高度定制化的创作流程，是 Stable Diffusion 生态中最灵活的 GUI 之一。","stars":134713,"starGrowth":203,"createdAt":"2023-01-17T03:15:56Z","language":"Python","topics":["ai"],"aiDims":["AI应用","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["AI绘图","可视化工作流","Stable Diffusion","节点式编程"],"domainKey":"AI绘图","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Comfy-Org/ComfyUI","ts":"2026-09-23T21:53:43.262Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3045207218308552,"copyOk":true},{"repo":"qdrant/qdrant","owner":"qdrant","name":"qdrant","desc":"Qdrant - High-performance, massive-scale Vector Database and Vector Search Engine for the next generation of AI. Also available in the cloud https://cloud.qdrant.io/","summaryCn":"高性能向量数据库，专为 AI 应用提供存储与检索服务","reasonCn":"Qdrant 是一款用 Rust 编写的向量相似度搜索引擎，专门为 AI 应用设计。它能高效存储和搜索带有额外信息的向量数据，支持复杂的过滤条件，非常适合做语义检索、推荐系统或 RAG（检索增强生成）应用。它还提供免费托管的云服务。","stars":34756,"starGrowth":19,"createdAt":"2020-05-30T21:37:01Z","silentRounds":0,"language":"Rust","topics":["ml","vector-db"],"aiDims":[],"zone":"AI","funScore":0,"domainTags":["向量数据库","向量搜索","RAG基础设施","Rust"],"domainKey":"向量数据库","pushedAt":"2026-09-23T21:31:16Z","tags":[{"name":"ml","source":"github","weight":0.7},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/qdrant/qdrant","ts":"2026-09-23T21:31:16Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05864000460352169,"copyOk":true},{"repo":"flet-dev/flet","owner":"flet-dev","name":"flet","desc":"Build realtime web, mobile and desktop apps in Python only. No frontend experience required.","summaryCn":"用Python写跨平台桌面网页App，无需前端技能。","reasonCn":"Flet是专为Python开发者设计的跨平台应用框架，它让开发者只需编写一套代码，就能生成在iOS、Android、Windows、macOS、Linux以及Web上运行的程序。该项目旨在消除前端开发的技术门槛，让不懂HTML、CSS或JavaScript的用户也能快速构建功能完备的桌面或移动应用。","stars":17051,"starGrowth":37,"createdAt":"2022-03-24T15:44:12Z","silentRounds":0,"language":"Python","topics":["desktop"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["跨平台开发","Python框架","UI开发"],"domainKey":"跨平台开发","pushedAt":"2026-09-23T16:14:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/flet-dev/flet","ts":"2026-09-23T16:14:17Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.11750988211835084,"copyOk":true},{"repo":"ublue-os/bazzite","owner":"ublue-os","name":"bazzite","desc":"Bazzite makes gaming and everyday use smoother and simpler across desktop PCs, handhelds, tablets, and home theater PCs.","summaryCn":"基于 Fedora 的游戏化 Linux 发行版，支持全平台。","reasonCn":"Bazzite 是一个基于 Fedora Atomic 技术构建的定制 Linux 发行版，它采用容器化技术，旨在将最佳 Linux 游戏体验带给所有设备，包括掌上游戏机。它解决了传统 Linux 系统在游戏兼容性和配置上的痛点，提供了一个开箱即用、适合各种硬件环境的游戏化桌面解决方案。","stars":9109,"starGrowth":3,"createdAt":"2023-02-09T19:05:47Z","silentRounds":0,"language":"Just","topics":["desktop"],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"工具","funScore":1,"domainTags":["Linux发行版","游戏","Steam Deck","桌面环境"],"domainKey":"Linux发行版","pushedAt":"2026-09-23T17:48:22Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Just","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/ublue-os/bazzite","ts":"2026-09-23T17:48:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"thedotmack/claude-mem","owner":"thedotmack","name":"claude-mem","desc":"Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More","summaryCn":"给Claude的上下文记忆补丁，自动补全跨会话信息。","reasonCn":"大语言模型虽然聪明，但每次对话都是全新的，项目信息很容易丢失。这个开源项目就像给 Claude 配了一个记忆管家，它能自动记录你用工具做了什么，把零散的操作变成有意义的摘要。下次再聊同一个项目时，它就能快速回忆起之前的细节，不再需要你反复解释背景，特别适合深度开发或需要长期跟进的复杂任务。","stars":94558,"starGrowth":96,"createdAt":"2025-08-31T20:50:03Z","language":"TypeScript","topics":["rag","ai"],"aiDims":["RAG与知识","AI Agent","AI基础设施"],"zone":"AI","funScore":1,"domainTags":["RAG","AI记忆","Claude","向量数据库"],"domainKey":"RAG","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/thedotmack/claude-mem","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.29104574969859764,"copyOk":true},{"repo":"mukul975/Anthropic-Cybersecurity-Skills","owner":"mukul975","name":"Anthropic-Cybersecurity-Skills","desc":"817 structured cybersecurity skills for AI agents · Mapped to 6 frameworks: MITRE ATT&CK, NIST CSF 2.0, MITRE ATLAS, D3FEND, NIST AI RMF & MITRE F3 (Fight Fraud) · agentskills.io standard · Works with Claude Code, GitHub Copilot, Codex CLI, Cursor, Gemini CLI & 20+ platforms · 29 security domains · Apache 2.0","summaryCn":"AI黑客攻防训练场，817个技能映射MITRE框架，让AI学会网络安全","reasonCn":"该仓库为AI代理提供了817个结构化的网络安全技能，覆盖MITRE ATT&CK、NIST CSF 2.0等6大国际标准框架。每个技能经过严格分类和标记，支持多种AI框架调用。它让开发者能像训练AI攻击手一样，系统性地提升智能体的安全操作能力，是构建AI驱动渗透测试和防御演练系统的基础设施。","stars":33279,"starGrowth":98,"createdAt":"2026-02-25T09:47:50Z","language":"Python","topics":["security","llm"],"aiDims":["AI Agent","安全工具","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["网络安全","AI代理","MITRE框架","技能包","渗透测试"],"domainKey":"网络安全","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/mukul975/Anthropic-Cybersecurity-Skills","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.36340128008654105,"copyOk":true},{"repo":"chatwoot/chatwoot","owner":"chatwoot","name":"chatwoot","desc":"Open-source live-chat, email support, omni-channel desk. An alternative to Intercom, Zendesk, Salesforce Service Cloud etc. 🔥💬","summaryCn":"企业级开源客服平台，支持多渠道对话与 AI 自动化。","reasonCn":"Chatwoot 是一款现代化的开源客户支持平台，旨在替代 Intercom、Zendesk 等商业软件。它让企业能够自托管，完全掌控客户数据，同时提供强大的跨渠道对话管理工具。项目引入了名为 Captain 的 AI 代理，能自动处理常见问题并减轻人工负担。","stars":37074,"starGrowth":33,"createdAt":"2019-08-14T06:50:51Z","silentRounds":0,"language":"Ruby","topics":["design","docker"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["多渠道客服","工单系统","实时聊天","客户支持平台","自托管"],"domainKey":"多渠道客服","pushedAt":"2026-09-23T21:51:48Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/chatwoot/chatwoot","ts":"2026-09-23T21:51:48Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10167033500566436,"copyOk":true},{"repo":"slint-ui/slint","owner":"slint-ui","name":"slint","desc":"Slint is an open-source declarative GUI toolkit to build native user interfaces for Rust, C++, JavaScript, or Python apps.","summaryCn":"用声明式语法写UI，一种代码同时跑在桌面、网页和嵌入式设备上","reasonCn":"专为资源受限设备设计的跨平台GUI框架，支持Rust、C++、Python、JavaScript多语言绑定。采用声明式UI语法，学习曲线平缓。渲染引擎轻量高效，可在树莓派、智能家电甚至工业控制器上流畅运行，是嵌入式开发者的福音。","stars":23932,"starGrowth":1,"createdAt":"2020-05-04T08:53:03Z","silentRounds":1,"language":"Rust","topics":["rust","embedded-devices","language","gui","wasm","cpp","javascript","ui","webassembly","declarative-ui","lsp-server","toolkit","user-interface","rust-lang","widgets","desktop","native"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["GUI框架","跨平台","嵌入式开发"],"domainKey":"GUI框架","pushedAt":"2026-09-23T21:47:03Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"rust","source":"github","weight":0.7},{"name":"embedded-devices","source":"github","weight":0.7},{"name":"language","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"ui","source":"github","weight":0.7},{"name":"webassembly","source":"github","weight":0.7},{"name":"declarative-ui","source":"github","weight":0.7},{"name":"lsp-server","source":"github","weight":0.7},{"name":"toolkit","source":"github","weight":0.7},{"name":"user-interface","source":"github","weight":0.7},{"name":"rust-lang","source":"github","weight":0.7},{"name":"widgets","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"native","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/slint-ui/slint","ts":"2026-09-23T21:47:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kitao/pyxel","owner":"kitao","name":"pyxel","desc":"A retro game engine for Python","summaryCn":"用Python写像素游戏的复古引擎，自带画面和音效，像搭积木一样造游戏","reasonCn":"一个用Rust写的复古像素游戏引擎，专门为Python爱好者打造。它最大的特点是自带80x80的复古屏幕渲染器和8位PCM音效，让你能像搭积木一样快速原型开发游戏。不用自己折腾复杂的渲染管线，专注游戏逻辑创意即可，非常适合制作复古风格和创意小游戏。","stars":17887,"starGrowth":9,"createdAt":"2018-06-10T04:58:54Z","silentRounds":0,"language":"Rust","topics":["game","pixel-art"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["复古游戏引擎","像素艺术创作","Python游戏开发"],"domainKey":"复古游戏引擎","pushedAt":"2026-09-23T04:33:37Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"pixel-art","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/kitao/pyxel","ts":"2026-09-23T04:33:37Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.02850947615586739,"copyOk":false},{"repo":"unslothai/unsloth","owner":"unslothai","name":"unsloth","desc":"Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4, MiniMax-H3, Gemma 4, FLUX and more.","summaryCn":"让普通显卡也能飞快微调大模型，速度翻倍显存减半","reasonCn":"技术上是针对Transformer架构的深度优化。核心包括：用FlashAttention-2替代标准注意力机制，减少显存带宽压力；优化RoPE位置编码实现长上下文支持；梯度检查点技术压缩激活值显存占用；还有针对特定模型架构的手写CUDA kernel。","stars":76592,"starGrowth":48,"createdAt":"2023-11-29T16:50:09Z","silentRounds":0,"language":"Python","topics":["llm","ai"],"aiDims":["模型与训练","大语言模型","微调"],"zone":"AI","funScore":0,"domainTags":["大模型微调","显存优化","训练加速","CUDA内核","LoRA"],"domainKey":"大模型微调","pushedAt":"2026-09-23T21:51:36Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/unslothai/unsloth","ts":"2026-09-23T21:51:36Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.14588009191560475,"copyOk":true},{"repo":"calesthio/Crucix","owner":"calesthio","name":"Crucix","desc":"Your personal intelligence agent. Watches the world from multiple data sources and pings you when something changes.","summaryCn":"像个不知疲倦的私人助理，24小时盯着全网动态，有大事第一时间通知你。","reasonCn":"这是一个24小时在线的智能监控代理，能同时聚合Twitter、Reddit、Hacker News等数十个数据源。一旦发现对你感兴趣的关键词或事件变化，立刻生成摘要并通过邮件或Telegram发送给你。无需配置复杂API，支持自定义过滤规则和通知频率，帮你节省海量浏览时间。","stars":11984,"starGrowth":58,"language":"JavaScript","topics":["ai","intelligence","osint"],"aiDims":["AI Agent","非AI-实用","效率工具"],"zone":"AI","funScore":0.6,"domainTags":["数据聚合","智能监控","社交媒体"],"domainKey":"数据聚合","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"intelligence","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/calesthio/Crucix","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.18851935914552737,"copyOk":false},{"repo":"go-gitea/gitea","owner":"go-gitea","name":"gitea","desc":"Git with a cup of tea! Painless self-hosted all-in-one software development service, including Git hosting, code review, team collaboration, package registry and CI/CD","summaryCn":"极简自托管Git服务，集成CI/CD与包管理","reasonCn":"Gitea 是一个用 Go 编写的轻量级自托管代码托管平台，单二进制即可运行，覆盖 Git 托管、代码审查、议题追踪、项目看板、Wiki、包注册表与 CI/CD 全流程。其核心优势在于原生复用 GitHub Actions 生态，迁移成本极低；跨平台支持覆盖 x86、ARM、RISC-V 等主流架构，适合从个人开发者到企业团队的各类私有化部署场景。","stars":58117,"starGrowth":10,"createdAt":"2016-11-01T02:13:26Z","silentRounds":0,"language":"Go","topics":["devops","self-hosted"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Git托管","代码审查","CI/CD","自托管","轻量级部署"],"domainKey":"Git托管","pushedAt":"2026-09-23T18:45:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.52,"source":"trending","url":"https://github.com/go-gitea/gitea","ts":"2026-09-23T18:45:57Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03051620007915068,"copyOk":true},{"repo":"rari-build/rari","owner":"rari-build","name":"rari","desc":"rari: High-performance React Server Components framework powered by a Rust runtime, delivering 18.1x faster response times (0.12ms vs 2.17ms) and 67.4x higher throughput (97,826 vs 1,452 req/sec) than Next.js with zero-config setup.","summaryCn":"基于Rust的高性能React服务端渲染框架。","reasonCn":"Rari是一个运行在Rust运行时的React服务端组件框架。它用Rust替换了传统的Node.js运行时，通过嵌入V8引擎提供极速的HTTP服务、路由和渲染能力。配合React框架的服务器操作、流式传输等功能，以及基于Rolldown的Vite构建工具链，开发者可以用标准React代码获得远超Node.js的性能表现。","stars":1175,"starGrowth":0,"createdAt":"2025-07-26T01:44:57Z","silentRounds":22,"language":"Rust","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["React Server Components","Rust运行时","Web框架","高性能渲染","服务端渲染"],"domainKey":"React Server Components","pushedAt":"2026-09-23T21:01:07Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/rari-build/rari","ts":"2026-09-23T21:01:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"readysettech/readyset","owner":"readysettech","name":"readyset","desc":"Readyset is a MySQL and Postgres wire-compatible caching layer that sits in front of existing databases to speed up queries and horizontally scale read throughput. Under the hood, ReadySet caches the results of cached select statements and incrementally updates these results over time as the underlying data changes.","summaryCn":"数据库的加速外挂，查询速度直接翻倍不夸张","reasonCn":"ReadySet是一个兼容MySQL和PostgreSQL协议的缓存层，部署在数据库前面，自动缓存查询结果，显著提升读性能。它使用Rust编写，支持水平扩展读吞吐，无需修改应用代码即可接入。核心优势在于自动识别可缓存查询、智能失效机制和零代码改动部署。项目star数超过5000，适合需要优化数据库读取性能的后端服务和架构团队。","stars":5280,"starGrowth":0,"createdAt":"2022-05-24T14:39:20Z","silentRounds":22,"language":"Rust","topics":["backend"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["数据库缓存","MySQL","PostgreSQL"],"domainKey":"数据库缓存","pushedAt":"2026-09-23T20:15:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/readysettech/readyset","ts":"2026-09-23T20:15:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"TauricResearch/TradingAgents","owner":"TauricResearch","name":"TradingAgents","desc":"TradingAgents: Multi-Agents LLM Financial Trading Framework","summaryCn":"用多智能体LLM框架自动分析数据并执行交易。","reasonCn":"这是一个面向金融交易的多智能体大语言模型框架，旨在解决复杂市场环境下数据融合与决策自动化的问题。它将宏观经济指标、社交媒体情绪等多元信息整合，通过多个专门的交易智能体进行协同分析，最终在交易策略上达成共识，实现自动化的量化交易闭环。","stars":108126,"starGrowth":134,"createdAt":"2024-12-28T03:31:08Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["多Agent系统","量化交易","金融AI","LLM应用","Agent框架"],"domainKey":"多Agent系统","pushedAt":"2026-09-23T20:11:28Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/TauricResearch/TradingAgents","ts":"2026-09-23T20:11:28Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30371788469008376,"copyOk":true},{"repo":"career-ops-hq/career-ops","owner":"career-ops-hq","name":"career-ops","desc":"Open-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)","summaryCn":"帮你自动扫遍各大招聘网站，AI给岗位打分还帮你改简历的求职神器","reasonCn":"基于AI Agent架构，能自动扫描LinkedIn、Indeed等主流招聘网站职位信息，对每个岗位生成A-H结构化评级报告和1-5分全球竞争力评分。还能根据岗位要求智能定制简历内容，全程在本地AI编程CLI中运行，支持任务状态跟踪和提醒。","stars":72444,"starGrowth":86,"createdAt":"2026-04-04T18:21:18Z","silentRounds":0,"language":"JavaScript","topics":["ai-agent","ai"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["求职自动化","简历优化","岗位评估"],"domainKey":"求职自动化","pushedAt":"2026-09-23T21:31:44Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/career-ops-hq/career-ops","ts":"2026-09-23T21:31:44Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.313873645850588,"copyOk":false},{"repo":"authelia/authelia","owner":"authelia","name":"authelia","desc":"The Single Sign-On Multi-Factor portal for web apps. OpenID Certified™ and Post-Quantum Cryptography Ready.","summaryCn":"开源身份认证服务器，为应用提供 SSO 与 2FA","reasonCn":"Authelia 是一款开源的认证授权服务器，专为反向代理场景设计。它提供统一的 Web 门户，为后端应用添加单点登录和双因子认证能力，支持 TOTP、WebAuthn 等多种 2FA 方式。通过与 Nginx、Traefik、HAProxy 等主流反向代理集成，它能在请求到达应用前完成身份验证与授权决策，允许、拒绝或重定向请求。支持 OpenID Connect 1.0 协议，可作为身份提供者对接现代应用生态。","stars":29053,"starGrowth":14,"createdAt":"2016-12-07T00:56:57Z","silentRounds":0,"language":"Go","topics":["security","docker"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["单点登录","多因子认证","身份管理","自托管","OAuth 2.0","LDAP"],"domainKey":"单点登录","pushedAt":"2026-09-23T21:51:51Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/authelia/authelia","ts":"2026-09-23T21:51:51Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.043445633841599836,"copyOk":true},{"repo":"ihhub/fheroes2","owner":"ihhub","name":"fheroes2","desc":"fheroes2 is a recreation of Heroes of Might and Magic II game engine","summaryCn":"经典策略游戏英雄无敌2的开源复刻，用现代技术让老游戏重获新生","reasonCn":"fheroes2是《魔法门之英雄无敌2》的完整开源复刻引擎，用C++重写，完美还原了原版画面、音乐和游戏逻辑。支持高分辨率显示、自定义地图编辑器、多人联机对战等功能，还兼容原版MOD和自定义内容。代码结构清晰，遵循原游戏的架构设计，是研究经典策略游戏开发的优秀参考。","stars":3486,"starGrowth":5,"createdAt":"2019-08-30T15:05:31Z","silentRounds":0,"language":"C++","topics":["game"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["英雄无敌 2","回合制策略","经典游戏复刻","C++ 游戏引擎"],"domainKey":"英雄无敌 2","pushedAt":"2026-09-23T10:08:58Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/ihhub/fheroes2","ts":"2026-09-23T10:08:58Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.019302925989672978,"copyOk":true},{"repo":"kellyjonbrazil/jc","owner":"kellyjonbrazil","name":"jc","desc":"CLI tool and python library that converts the output of popular command-line tools, file-types, and common strings to JSON, YAML, or Dictionaries. This allows piping of output to tools like jq and simplifying automation scripts.","summaryCn":"CLI工具输出JSON化，方便脚本解析处理。","reasonCn":"很多命令行工具的输出是纯文本，很难直接被程序处理。jc通过解析这些输出，把它们转换成结构化的JSON格式，让后续的数据处理变得非常简单。它支持各种常见的系统命令、文件类型和字符串，安装后直接在命令行使用即可，无需复杂的编程。开发者可以用它来编写脚本，或者配合jq等工具进行更高级的文本处理。","stars":8684,"starGrowth":2,"createdAt":"2019-10-15T22:04:52Z","silentRounds":0,"language":"Python","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.7,"domainTags":["命令行工具","JSON转换","CLI","系统运维"],"domainKey":"命令行工具","pushedAt":"2026-09-23T15:06:35Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/kellyjonbrazil/jc","ts":"2026-09-23T15:06:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zzet/gortex","owner":"zzet","name":"gortex","desc":"High-performance code-intelligence engine for AI agents and IDE, supports 257 languages, multi repositories, based on graph, with access via CLI, MCP Server, and API. AI coding agents teammate - expose only needed information, cutting token usage up to 50x. 100% local. Discord: https://discord.gg/39MFHu3J5d","summaryCn":"把代码仓库变成AI助手的强力引擎，支持257种语言。","reasonCn":"Gortex是一款基于图神经网络的高性能代码智能引擎，专为AI代理和IDE设计。它支持257种编程语言，通过多仓库分析提供代码理解、生成与调试能力。提供CLI、MCP Server和API三种访问方式，让AI编程助手能无缝接入本地或云端环境。","stars":1721,"starGrowth":71,"language":"Go","topics":["ai-tools","antigravity","claude-code","code-assistant","copilot","cursor","graphrag","kiro","knowledge-graph","mcp-server","windsurf","tokens","context-window","context-window-optimization","context-window-optimizer","skills","prompts","local-first","code-analysis","developer-tools"],"aiDims":["AI基础设施","代码助手","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["代码智能","图神经网络","MCP协议","IDE集成","多语言分析"],"domainKey":"代码智能","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-tools","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"code-assistant","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"graphrag","source":"github","weight":0.7},{"name":"kiro","source":"github","weight":0.7},{"name":"knowledge-graph","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"windsurf","source":"github","weight":0.7},{"name":"tokens","source":"github","weight":0.7},{"name":"context-window","source":"github","weight":0.7},{"name":"context-window-optimization","source":"github","weight":0.7},{"name":"context-window-optimizer","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"prompts","source":"github","weight":0.7},{"name":"local-first","source":"github","weight":0.7},{"name":"code-analysis","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/zzet/gortex","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3367652527600232,"copyOk":true},{"repo":"headroomlabs-ai/headroom","owner":"headroomlabs-ai","name":"headroom","desc":"Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.","summaryCn":"让大模型少花钱的输入压缩神器，RAG日志都能压","reasonCn":"专为大模型设计的输入压缩工具，能把工具输出、日志、RAG片段甚至JSON数据打包压缩后再喂给LLM。实测编码Agent能省20% token，JSON压缩更夸张能砍60-95%。支持三种部署模式：代码库、代理服务、MCP服务器，直接集成到你的AI工作流。","stars":73530,"starGrowth":113,"createdAt":"2026-01-07T19:58:51Z","silentRounds":0,"language":"Python","topics":["rag","llm","ai"],"aiDims":["RAG与知识","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["大模型","RAG","输入压缩","MCP"],"domainKey":"RAG","pushedAt":"2026-09-23T21:42:15Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/headroomlabs-ai/headroom","ts":"2026-09-23T21:42:15Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.36553243574051403,"copyOk":false},{"repo":"cockroachdb/cockroach","owner":"cockroachdb","name":"cockroach","desc":"CockroachDB — the cloud native, distributed SQL database designed for high availability, effortless scale, and control over data placement.","summaryCn":"能自动容错并跨数据中心扩展的云原生分布式SQL数据库。","reasonCn":"CockroachDB 是一款专为现代应用设计的云原生分布式 SQL 数据库。它基于强一致性键值存储构建，能像处理单机数据一样处理海量数据，无需人工干预即可自动应对磁盘或服务器故障。开发者可以用熟悉的 SQL 语法轻松管理数据，享受企业级的高可用性与事务处理能力，为数据密集型应用提供坚实基础。","stars":32484,"starGrowth":2,"createdAt":"2014-02-06T00:18:47Z","silentRounds":0,"language":"Go","topics":["database"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","SQL数据库","高可用存储"],"domainKey":"分布式数据库","pushedAt":"2026-09-23T15:34:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/cockroachdb/cockroach","ts":"2026-09-23T15:34:41Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tw93/Pake","owner":"tw93","name":"Pake","desc":"🤱🏻 Turn any webpage into a desktop app with one command.","summaryCn":"一键将网页转化为跨平台桌面应用，体积仅 10M。","reasonCn":"Pake 是一款基于 Rust 技术的桌面应用打包工具，旨在解决网页转桌面应用体积过大和性能差的问题。它让开发者能通过极简命令或在线平台，将任何网页封装成原生桌面程序，支持快捷键、沉浸式窗口和样式定制等功能。无论是个人项目还是企业应用，Pake 都能让网页应用拥有接近原生软件的体验。","stars":61682,"starGrowth":59,"createdAt":"2022-10-14T09:32:57Z","language":"Rust","topics":["desktop"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["桌面应用","网页打包","Rust","Electron替代","跨平台打包"],"domainKey":"桌面应用","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/tw93/Pake","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.17986955675885996,"copyOk":true},{"repo":"web-infra-dev/rsbuild","owner":"web-infra-dev","name":"rsbuild","desc":"Fast, extensible build tool for modern web development","summaryCn":"新一代Web构建工具，比Webpack快，比Vite灵活","reasonCn":"Rsbuild是Node.js团队打造的现代化构建工具，基于Rspack引擎，性能远超传统Webpack，同时兼容Vite的开发体验。它默认零配置，开箱即用，支持React、Vue等多种框架，还提供TypeScript和插件扩展能力。对于追求构建速度和开发体验的前端团队来说，是一个值得尝试的高效工具。","stars":3386,"starGrowth":2,"createdAt":"2023-10-07T12:59:03Z","silentRounds":0,"language":"TypeScript","topics":["frontend"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Web构建工具","Rspack","前端工程化","TypeScript","零配置","构建性能"],"domainKey":"Web构建工具","pushedAt":"2026-09-23T12:24:42Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/web-infra-dev/rsbuild","ts":"2026-09-23T12:24:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NousResearch/hermes-agent","owner":"NousResearch","name":"hermes-agent","desc":"The agent that grows with you","summaryCn":"会自我进化的云端AI助手，能记住你并不断变强。","reasonCn":"Hermes Agent 是 Nous Research 推出的自我改进型 AI 代理，它最特别的地方在于内置学习循环：能从与你的交互中总结经验、生成新技能，并在后续使用中不断优化。它可以在便宜的云服务器上运行，空闲时几乎不花钱，还能通过 Telegram 随时随地远程指挥。","stars":248099,"starGrowth":320,"createdAt":"2025-07-22T22:22:28Z","silentRounds":0,"language":"Python","topics":["ai-agent","ai","llm"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","智能体框架","对话记忆","工具调用","LLM"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:39:38Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/NousResearch/hermes-agent","ts":"2026-09-23T21:39:38Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3038694230932007,"copyOk":true},{"repo":"langchain-ai/langchain","owner":"langchain-ai","name":"langchain","desc":"The agent engineering platform.","summaryCn":"让AI能记住东西还会用工具的万能开发框架","reasonCn":"LangChain是业界最流行的AI应用开发框架，核心解决开发者快速搭建智能体应用的难题，无需从零造轮子。它支持数十种主流LLM即插即用，内置丰富的记忆模块和工具调用能力，能轻松连接向量数据库实现RAG，还提供了Agent编排、链式调用等高级功能。GitHub上14万+star，社区极其活跃，版本迭代速度快。","stars":146889,"starGrowth":64,"createdAt":"2022-10-17T02:58:36Z","silentRounds":0,"language":"Python","topics":["rag","llm","ai","agents"],"aiDims":["AI Agent","RAG与知识","工具调用"],"zone":"AI","funScore":0,"domainTags":["Agent框架","RAG","LLM应用开发","工具调用","记忆管理"],"domainKey":"RAG","pushedAt":"2026-09-23T21:47:51Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/langchain-ai/langchain","ts":"2026-09-23T21:47:51Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.19330710944999285,"copyOk":false},{"repo":"tw93/Mole","owner":"tw93","name":"Mole","desc":"🐹 Clean, uninstall, analyze, optimize, and monitor your Mac from the terminal.","summaryCn":"Mac 系统的清理、卸载与维护全能助手。","reasonCn":"这是一款针对 Mac 系统维护的开源工具，旨在解决磁盘空间不足、应用冗余和系统运行缓慢的问题。它提供命令行和原生应用两种形态，既能通过终端快速执行清理和维护，也能通过图形界面直观管理磁盘和应用。项目不仅关注基础的清理功能，还提供磁盘地图可视化、应用安装管理以及实时状态监控，帮助用户全面掌握系统健康状况。","stars":68112,"starGrowth":88,"createdAt":"2025-09-23T06:38:40Z","silentRounds":0,"language":"Shell","topics":["command-line"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["macOS","终端工具","系统清理","性能监控"],"domainKey":"macOS","pushedAt":"2026-09-23T04:48:20Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/tw93/Mole","ts":"2026-09-23T04:48:20Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2678759689922481,"copyOk":true},{"repo":"semaphoreui/semaphore","owner":"semaphoreui","name":"semaphore","desc":"Modern UI and powerful API for Ansible, Terraform/OpenTofu/Terragrunt, PowerShell and other DevOps tools.","summaryCn":"现代 Web 界面管理 DevOps 工具。","reasonCn":"Semaphore UI 是一款现代化的 Web 界面，用于管理 Ansible、Terraform 等流行 DevOps 工具。它解决了终端部署不再可行的问题，允许用户轻松运行剧本、脚本，并通过通知控制失败任务。该项目将复杂的命令行操作转化为直观的图形界面，显著提升了部署与配置的效率与安全性。","stars":14177,"starGrowth":6,"createdAt":"2014-08-23T22:43:31Z","silentRounds":0,"language":"Go","topics":["devops","docker"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["DevOps","Ansible","Terraform","可视化","自动化运维"],"domainKey":"DevOps","pushedAt":"2026-09-23T17:52:20Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.52,"source":"trending","url":"https://github.com/semaphoreui/semaphore","ts":"2026-09-23T17:52:20Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01926966212880017,"copyOk":true},{"repo":"bee-san/RustScan","owner":"bee-san","name":"RustScan","desc":"🤖 The Modern Port Scanner 🤖","summaryCn":"用Rust重写的极速端口扫描器，3秒扫完65535端口。","reasonCn":"RustScan 以 Rust 语言重写底层扫描引擎，配合批量发包与智能重试机制，将全端口扫描时间压缩至 3 秒级，较 Nmap 提升数十倍。它内置 Python、Lua、Shell 脚本引擎，扫描后可自动触发漏洞利用或信息收集脚本，并原生支持 Docker 部署，开箱即用。适合渗透测试人员在授权范围内快速资产梳理，也适合运维排查网络服务暴露面。","stars":20439,"starGrowth":2,"createdAt":"2020-07-11T20:00:41Z","silentRounds":0,"language":"Rust","topics":["security","docker"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["端口扫描","网络安全","渗透测试","网络扫描"],"domainKey":"端口扫描","pushedAt":"2026-09-23T09:02:57Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/bee-san/RustScan","ts":"2026-09-23T09:02:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sickn33/agentic-awesome-skills","owner":"sickn33","name":"agentic-awesome-skills","desc":"AAS Core is the local, agent-first control plane for complete catalog discovery, agent-owned selection, stack validation, and planning, backed by 2,000+ agentic skills. Includes CLI, local MCP, catalog, plugins, and Workbench.","summaryCn":"给AI Agent装上2000多件工具的万能背包","reasonCn":"这个项目就像一个给AI Agent准备的瑞士军刀包，内置了2000多种技能，涵盖了各种开发工具和常见任务。它通过本地控制平面自动发现、选择并验证这些工具，让你的Agent不再只是聊天，而是能真正动手干活，比如调接口、处理数据，实用性超强。","stars":46794,"starGrowth":73,"createdAt":"2026-01-14T17:48:09Z","silentRounds":0,"language":"Python","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["Agent技能库","MCP工具集","本地控制平面"],"domainKey":"Agent技能库","pushedAt":"2026-09-23T17:52:27Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/sickn33/agentic-awesome-skills","ts":"2026-09-23T17:52:27Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.26841610462879856,"copyOk":true},{"repo":"browser-use/video-use","owner":"browser-use","name":"video-use","desc":"Edit videos with coding agents","summaryCn":"用AI coding agent自动剪辑视频，会写代码就能剪大片","reasonCn":"这个项目是browser-use框架的视频编辑分支，用AI Agent通过生成代码来自动操作视频处理工具。你输入自然语言描述想要的效果，比如'把视频里所有红色物体去掉'，Agent就会自动生成Python代码调用视频处理库完成剪辑。相比手动剪辑，效率高得多。","stars":26446,"starGrowth":745,"language":"Python","topics":[],"aiDims":["AI Agent","AI应用","创意工具"],"zone":"AI","funScore":0,"domainTags":["视频编辑","AI Agent","自动化剪辑","自然语言编程"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/browser-use/video-use","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3845118354382515,"copyOk":true},{"repo":"iib0011/omni-tools","owner":"iib0011","name":"omni-tools","desc":"Self-hosted collection of powerful web-based tools for everyday tasks. No ads, no tracking, just fast, accessible utilities right from your browser!","summaryCn":"开源万能工具箱，浏览器里就能用的各种实用工具，还没广告","reasonCn":"这是一个自托管的Web工具集合，用TypeScript开发，内含大量日常实用工具，比如文档转换、图片处理、编码解码、颜色选择等。全部在浏览器端运行，无需上传数据到服务器，保护隐私。支持一键部署到自有服务器，没有任何广告和追踪，访问速度快，适合个人和企业自建使用。","stars":10240,"starGrowth":4,"createdAt":"2024-06-13T23:06:42Z","silentRounds":0,"language":"TypeScript","topics":["frontend"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Web 工具","前端工具","自托管","隐私"],"domainKey":"Web 工具","pushedAt":"2026-09-23T20:42:14Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/iib0011/omni-tools","ts":"2026-09-23T20:42:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lichess-org/lila","owner":"lichess-org","name":"lila","desc":"♞ lichess.org: the forever free, adless and open source chess server ♞","summaryCn":"全球千万棋手都在用的开源国际象棋平台，还能跟AI对弈，完全免费无广告","reasonCn":"这是lichess.org国际象棋网站的完整开源服务器，支持实时在线对弈、棋局分析、排名系统和在线赛事。自带AI对手功能，让普通用户也能在线和电脑下棋。完全免费开源无广告，技术栈基于Scala和Akka，用Docker部署即可搭建自己的私人棋盘平台，社区活跃star超18000。","stars":18759,"starGrowth":0,"createdAt":"2012-02-21T19:46:55Z","silentRounds":1,"language":"Scala","topics":["game"],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["国际象棋","在线对弈","实时多人游戏","开源游戏服务器","Stockfish AI"],"domainKey":"国际象棋","pushedAt":"2026-09-23T12:06:57Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Scala","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/lichess-org/lila","ts":"2026-09-23T12:06:57Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hajimehoshi/ebiten","owner":"hajimehoshi","name":"ebiten","desc":"A dead simple 2D game engine for Go","summaryCn":"一款极简的Go语言2D游戏引擎，支持十多平台开发。","reasonCn":"Ebitengine 是一款专为 Go 语言设计的 2D 游戏引擎，旨在提供简单易用的 API。它解决了传统游戏引擎学习曲线陡峭、跨平台部署复杂的问题，让开发者能够用同一段代码快速开发出可在电脑、手机甚至游戏主机上运行的游戏。该引擎通过简化开发流程，降低了游戏创作的门槛。","stars":13499,"starGrowth":4,"createdAt":"2013-06-16T15:13:01Z","silentRounds":0,"language":"Go","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"工具","funScore":0,"domainTags":["2D游戏引擎","Go语言游戏开发","跨平台图形库","WebAssembly游戏导出"],"domainKey":"2D游戏引擎","pushedAt":"2026-09-23T19:47:03Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/hajimehoshi/ebiten","ts":"2026-09-23T19:47:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"akitaonrails/ai-memory","owner":"akitaonrails","name":"ai-memory","desc":"Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors","summaryCn":"给AI Agent装上长期记忆的内存模块，解决它“转头就忘”的毛病","reasonCn":"这是一个专门为AI Agent设计的长期记忆解决方案，利用RAG技术存储对话历史和上下文，确保Agent在跨会话或跨工具调用时保持记忆连贯。项目用Rust编写，提供标准化的API接口，方便不同Agent框架集成，极大提升了Agent的可用性和实用性。","stars":8204,"starGrowth":154,"language":"Rust","topics":[],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["RAG","长期记忆","Agent","向量数据库","Rust"],"domainKey":"RAG","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/akitaonrails/ai-memory","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3563139931740614,"copyOk":false},{"repo":"agentscope-ai/agentscope","owner":"agentscope-ai","name":"agentscope","desc":"Build and run agents you can see, understand and trust.","summaryCn":"顺应大模型能力的通用智能体框架，主打生产可用与易上手。","reasonCn":"AgentScope 2.0 是一个面向生产环境的智能体开发框架，核心理念是顺应模型不断增强的推理与工具调用能力，而不是用死板的提示词和编排去限制它。它提供了构建智能体所需的基本抽象，并紧跟前沿，已支持实时语音交互和 A2A 跨智能体通信协议。","stars":32182,"starGrowth":55,"createdAt":"2024-01-12T03:41:59Z","silentRounds":0,"language":"Python","topics":["llm","chatbot"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["多智能体系统","LLM","Agent协作","Python"],"domainKey":"多智能体系统","pushedAt":"2026-09-23T08:57:58Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/agentscope-ai/agentscope","ts":"2026-09-23T08:57:58Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.17012708967745946,"copyOk":true},{"repo":"penpot/penpot","owner":"penpot","name":"penpot","desc":"Penpot: The open-source design platform for Product teams that need scalable collaboration.","summaryCn":"一款开源的在线设计协作平台，让设计师和产品团队并肩工作","reasonCn":"Penpot 是一款开源的设计平台，对标 Figma，支持矢量设计、原型制作和团队协作。采用 Web 技术栈构建，基于 Clojure 后端和 SVG 标准，无需安装插件即可在浏览器中直接编辑。提供组件系统、设计令牌等现代化设计功能，支持实时协作和多用户同时编辑同一画布，还有完善的版本管理。企业可自行部署，完全掌控设计资产。","stars":60270,"starGrowth":40,"createdAt":"2015-12-29T22:09:07Z","silentRounds":0,"language":"Clojure","topics":["design"],"aiDims":["创意工具","非AI-好玩","效率工具"],"zone":"工具","funScore":0,"domainTags":["开源设计工具","Figma替代品","实时协作","SVG格式","自托管部署","Clojure"],"domainKey":"开源设计工具","pushedAt":"2026-09-23T18:10:35Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Clojure","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/penpot/penpot","ts":"2026-09-23T18:10:35Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.12199104031856645,"copyOk":false},{"repo":"o2sh/onefetch","owner":"o2sh","name":"onefetch","desc":"Command-line Git information tool","summaryCn":"在终端用 ASCII 艺术展示代码库信息。","reasonCn":"Onefetch 是一款运行在终端的 Git 信息工具，它能在本地仓库中提取语言、贡献者、项目信息等数据，并以 ASCII 艺术画的形式直观呈现。它完全离线运行，性能出色，支持自定义输出格式，让开发者能一眼看清项目的“DNA”。","stars":12050,"starGrowth":1,"createdAt":"2018-09-14T18:51:00Z","silentRounds":7,"language":"Rust","topics":["command-line","git"],"aiDims":["非AI-好玩","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Git","命令行","ASCII","终端美学"],"domainKey":"Git","pushedAt":"2026-09-23T09:07:27Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/o2sh/onefetch","ts":"2026-09-23T09:07:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/SandDance","owner":"microsoft","name":"SandDance","desc":"Visually explore, understand, and present your data.","summaryCn":"微软出品的数据可视化神器，让枯燥表格瞬间变成可交互的 3D 魔法。","reasonCn":"这是一款基于 Web 的数据可视化探索工具，它能帮你把成千上万行的表格数据，瞬间转化为直观的 3D 视图。用户可以通过拖拽滑块、点击按钮，自由切换不同的视觉风格，比如按数值大小改变方块大小、用颜色标记类别、甚至让方块悬浮在空中旋转。这种交互方式能让你在几秒钟内发现数据中的异常点和隐藏规律，从而更轻松地讲好数据故事。它不仅是看数据的工具，更是支持决策、测试假设的利器。","stars":7148,"starGrowth":1,"createdAt":"2018-11-05T22:45:31Z","silentRounds":22,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-好玩","非AI-实用"],"zone":"工具","funScore":0.8,"domainTags":["数据可视化","交互式分析","动画"],"domainKey":"数据可视化","pushedAt":"2026-09-22T16:13:12Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/microsoft/SandDance","ts":"2026-09-22T16:13:12Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"mem0ai/mem0","owner":"mem0ai","name":"mem0","desc":"The Memory Layer for AI Agents - Drop-in memory infrastructure for AI agents and apps. Context that persists. Built for production.","summaryCn":"给AI装上会思考的大脑，自动记住你的偏好习惯，越聊越懂你","reasonCn":"Mem0为AI Agent构建统一记忆层，自动从对话中提取关键信息并持久化存储。采用向量数据库加知识图谱双层架构，支持多用户场景下的记忆隔离与共享。可与LangChain、LlamaIndex等主流框架无缝集成，几行代码即可让LLM获得长期记忆能力，真正实现越用越聪明的体验。","stars":65841,"starGrowth":51,"createdAt":"2023-06-20T08:58:36Z","silentRounds":0,"language":"Python","topics":["rag","llm","ai","agents"],"aiDims":["AI Agent","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI记忆层","长期记忆","向量数据库","知识图谱","Agent框架"],"domainKey":"向量数据库","pushedAt":"2026-09-23T19:13:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/mem0ai/mem0","ts":"2026-09-23T19:13:42Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.15532378001549185,"copyOk":true},{"repo":"MODSetter/SurfSense","owner":"MODSetter","name":"SurfSense","desc":"Open-source NotebookLM alternative. Research the open web with live data(Reddit, YT, IG, TikTok, Indeed, Google Search, Maps etc) through one platform, API or MCP server. Join our Discord: https://discord.gg/ejRNvftDp9","summaryCn":"开源版NotebookLM，整合全网实时数据做AI研究","reasonCn":"这是一个开源的NotebookLM替代品，能够抓取Reddit、YouTube、Instagram、TikTok、Indeed、Google搜索和地图等平台的实时数据。通过RAG技术和MCP服务器，让用户在一个平台内完成跨多源信息的AI辅助研究，数据鲜活且全面。","stars":16231,"starGrowth":46,"createdAt":"2024-07-30T23:00:09Z","silentRounds":0,"language":"Python","topics":["rag","agents"],"aiDims":["AI搜索","AI应用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["RAG","实时数据抓取","多源整合","研究助手","MCP"],"domainKey":"RAG","pushedAt":"2026-09-23T20:43:28Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/MODSetter/SurfSense","ts":"2026-09-23T20:43:28Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.14650224878319268,"copyOk":false},{"repo":"apache/superset","owner":"apache","name":"superset","desc":"Apache Superset is a Data Visualization and Data Exploration Platform","summaryCn":"企业级数据可视化平台，拖拽就能做出酷炫仪表盘","reasonCn":"Apache Superset是一个企业级数据可视化与探索平台，支持多种数据源连接，提供丰富的图表类型和拖拽式仪表板构建。核心特性包括SQL编辑器、即席查询、行级权限控制和可视化分享。项目star数超过7.4万，是Apache基金会的顶级项目，适合数据分析师、工程师和需要制作数据报表的业务团队。","stars":74887,"starGrowth":14,"createdAt":"2015-07-21T18:55:34Z","silentRounds":0,"language":"Python","topics":["data-visualization","data-science"],"aiDims":["非AI-实用","可视化","效率工具"],"zone":"工具","funScore":0,"domainTags":["数据可视化","商业智能","SQL查询分析"],"domainKey":"数据可视化","pushedAt":"2026-09-23T21:51:07Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.48,"source":"search","url":"https://github.com/apache/superset","ts":"2026-09-23T21:51:07Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04256084500647643,"copyOk":true},{"repo":"lissy93/dashy","owner":"lissy93","name":"dashy","desc":"🚀 A self-hostable personal dashboard built for you. Includes status-checking, widgets, themes, icon packs, a UI editor and tons more!","summaryCn":"Vue 框架构建的自托管仪表盘，定制你的个人主页。","reasonCn":"Dashy 是一款基于 Vue 的自托管仪表盘，旨在为家庭实验室和个人服务器提供统一的导航入口。它通过可视化的配置方式，帮助用户将分散的各种服务链接整合到一个页面，解决信息碎片化和入口杂乱的问题。支持 Docker 部署，能够通过配置文件快速搭建出个性化的服务总览面板。","stars":26548,"starGrowth":8,"createdAt":"2021-02-28T10:35:44Z","silentRounds":0,"language":"Vue","topics":["self-hosted","docker","awesome"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["自托管","个人仪表盘","服务监控","Vue"],"domainKey":"自托管","pushedAt":"2026-09-22T16:59:00Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/lissy93/dashy","ts":"2026-09-22T16:59:00Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.024904022901913514,"copyOk":true},{"repo":"processing/p5.js","owner":"processing","name":"p5.js","desc":"p5.js is a client-side JS platform that empowers artists, designers, students, and anyone to learn to code and express themselves creatively on the web. It is based on the core principles of Processing. Looking for p5.js 2.0? http://beta.p5js.org","summaryCn":"浏览器里的创意编程神器，零基础也能画出动态艺术作品","reasonCn":"p5.js是一套开源JavaScript库，专为创意编码设计。它将Canvas绘图、事件处理、动画循环全部封装成简洁API，灵感来自Processing。支持粒子系统、物理模拟、音乐可视化等效果，社区拥有海量示例和中文教程，对初学者极其友好。","stars":24024,"starGrowth":6,"createdAt":"2013-02-26T03:14:47Z","silentRounds":0,"language":"JavaScript","topics":["design","creative-coding"],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["创意编程","生成艺术","可视化","编程教学","Web Canvas"],"domainKey":"创意编程","pushedAt":"2026-09-22T18:54:58Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"creative-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/processing/p5.js","ts":"2026-09-22T18:54:58Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.018749250749250747,"copyOk":true},{"repo":"corsairdev/corsair","owner":"corsairdev","name":"corsair","desc":"Your Agent's Integration Layer","summaryCn":"给AI智能体装上万能插座的中间件，让LLM秒变多面手","reasonCn":"CORSAIR 是一个专为 AI Agent 设计的编排层，通过统一的工具调用接口解决大模型无法直接操作外部系统的痛点。它支持函数调用、数据流处理和自动化工作流，让 LLM 能像人类一样组合工具完成任务。项目采用 TypeScript 开发，兼容 OpenAI Function Calling 等标准协议，提供低延迟的执行引擎。","stars":11704,"starGrowth":150,"createdAt":"2025-10-14T18:50:32Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI Agent","工具调用","LLM集成","函数调用","工作流编排"],"domainKey":"AI Agent","pushedAt":"2026-09-23T16:03:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/corsairdev/corsair","ts":"2026-09-23T16:03:20Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.4061380724538619,"copyOk":false},{"repo":"obra/superpowers","owner":"obra","name":"superpowers","desc":"An agentic skills framework & software development methodology that works.","summaryCn":"面向AI编程助手的技能包与开发方法论，让Agent更听话。","reasonCn":"Superpowers 是一套专为代码编写智能体设计的开发方法论。它通过提供一套可组合的技能和初始指令，确保 AI 助手能够正确调用这些技能。这解决了智能体在执行复杂任务时可能出现的指令混乱问题，让 AI 能更精准、高效地完成代码编写工作。","stars":290647,"starGrowth":485,"createdAt":"2025-10-09T19:45:18Z","language":"Shell","topics":["ai"],"aiDims":["AI Agent","开发者工具","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI","Agent","技能框架","工作流","Shell"],"domainKey":"Agent","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/obra/superpowers","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.36600728719030295,"copyOk":true},{"repo":"Spu7Nix/obamify","owner":"Spu7Nix","name":"obamify","desc":"","summaryCn":"脑洞图像整活器，一键把任何照片变成奥巴马动漫脸，趣味满分","reasonCn":"obamify 是一款极具脑洞的图像整活工具，它能将任意输入的照片“洗”成奥巴马的动漫面孔。作者提供了一连串俏皮的设置选项，甚至提示用户通过重叠眼睛来获得更好的合成效果。这不仅是一个软件，更是一个充满幽默感的数字玩具，适合喜欢社交整活和创意表达的用户，在轻松愉快的氛围中体验算法的奇妙变形。","stars":2139,"starGrowth":0,"silentRounds":0,"language":"Rust","topics":[],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":0.74,"domainTags":["图像变形","动漫滤镜","奥巴马脸","Rust应用"],"domainKey":"图像变形","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/Spu7Nix/obamify","ts":"2026-09-22T22:51:50.275Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kubeshark/kubeshark","owner":"kubeshark","name":"kubeshark","desc":"eBPF-powered network observability for Kubernetes. Indexes L4/L7 traffic with full K8s context, decrypts TLS without keys. Queryable by AI agents via MCP and humans via dashboard.","summaryCn":"K8s网络监控神技，连TLS加密流量都能轻松解密看透。","reasonCn":"利用eBPF技术深度嵌入Kubernetes集群，对L4/L7流量进行无损抓包与索引。它不仅能还原完整的K8s上下文，还能无需私钥直接解密TLS，甚至支持通过MCP接口让AI代理直接查询网络数据，是云原生时代排查网络故障的终极利器。","stars":12085,"starGrowth":0,"createdAt":"2021-04-19T10:29:56Z","silentRounds":2,"language":"Go","topics":["devops","docker"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0.5,"domainTags":["Kubernetes网络监控","eBPF流量分析","TLS解密"],"domainKey":"Kubernetes网络监控","pushedAt":"2026-09-23T05:14:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/kubeshark/kubeshark","ts":"2026-09-23T05:14:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aquasecurity/trivy","owner":"aquasecurity","name":"trivy","desc":"Find vulnerabilities, misconfigurations, secrets, SBOM in containers, Kubernetes, code repositories, clouds and more","summaryCn":"一个功能全面的开源安全扫描器，帮你快速发现容器、代码和依赖中的漏洞。","reasonCn":"Trivy 是一款广受欢迎的开源安全扫描工具，能帮你扫描容器镜像、文件系统、代码仓库和云服务配置，找出已知漏洞、错误配置和密钥泄露。它支持多种扫描器和目标，覆盖了从开发到部署的多个环节，是 DevSecOps 实践中常用的利器。","stars":38027,"starGrowth":13,"createdAt":"2019-04-11T01:01:07Z","silentRounds":0,"language":"Go","topics":["security","docker"],"aiDims":[],"zone":"工具","funScore":0,"domainTags":["漏洞扫描","容器安全","Kubernetes安全","SBOM","供应链安全"],"domainKey":"漏洞扫描","pushedAt":"2026-09-23T07:35:52Z","tags":[{"name":"security","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/aquasecurity/trivy","ts":"2026-09-23T07:35:52Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04002558708286218,"copyOk":true},{"repo":"can1357/oh-my-pi","owner":"can1357","name":"oh-my-pi","desc":"⌥ Coding agent with the IDE wired in. Built by Stencil Labs.","summaryCn":"内置IDE的编码代理，支持60多家模型商和31种工具，开箱即用。","reasonCn":"这是一个把IDE直接接线进AI代理的编码工具，由知名游戏框架LibGDX作者出品。它内置了60多种模型供应商、31种工具，还集成了14种LSP操作和28种调试操作，核心是约8万行Rust代码。安装简单，在macOS和Linux上可直接使用。","stars":33001,"starGrowth":405,"createdAt":"2025-12-31T14:01:28Z","language":"TypeScript","topics":["llm","ai-agent"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","终端工具","代码生成","Agent框架"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/can1357/oh-my-pi","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.40418047937941276,"copyOk":true},{"repo":"crewAIInc/crewAI","owner":"crewAIInc","name":"crewAI","desc":"Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.","summaryCn":"开源多智能体框架，让AI角色自主协作完成复杂任务。","reasonCn":"crewAI 是面向生产环境的多智能体协作框架，支持为大模型分配角色、目标与工具，组建自主协作的“船员”团队。它既提供高层 Crews 抽象让智能体自主决策分工，又提供事件驱动的 Flows 实现精确流程控制。已有超十万开发者通过社区课程认证，成为构建 Agentic 自动化工作流的主流选择。","stars":58919,"starGrowth":43,"createdAt":"2023-10-27T03:26:59Z","silentRounds":0,"language":"Python","topics":["ai","agents"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["多智能体协作","角色扮演Agent","自动化工作流"],"domainKey":"多智能体协作","pushedAt":"2026-09-23T21:37:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.94,"source":"search","url":"https://github.com/crewAIInc/crewAI","ts":"2026-09-23T21:37:11Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1311894465282846,"copyOk":true},{"repo":"jesseduffield/lazygit","owner":"jesseduffield","name":"lazygit","desc":"simple terminal UI for git commands","summaryCn":"给git加了个游戏界面，终端操作从此像打电竞一样爽","reasonCn":"基于Go语言开发的终端图形界面工具，完美替代命令行git操作。支持可视化分支管理、commit历史查看、staged/unstaged文件切换、交互式rebase等高级功能。界面采用tview库渲染，响应流畅，快捷键体系完善，让日常版本控制工作变得直观高效。","stars":82582,"starGrowth":26,"createdAt":"2018-05-19T00:53:06Z","silentRounds":0,"language":"Go","topics":["git"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["git","终端UI","版本控制","开发工具","Go语言"],"domainKey":"git","pushedAt":"2026-09-23T13:29:14Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/jesseduffield/lazygit","ts":"2026-09-23T13:29:14Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07894451575403841,"copyOk":true},{"repo":"anopara/genetic-drawing","owner":"anopara","name":"genetic-drawing","desc":"","summaryCn":"Genetic Drawing 是一个诞生于 2017 年的 Python 玩具项目，其核心创意在于","reasonCn":"项目的一个独特功能是其对“用户创建采样掩码”的支持。这意味着用户可以干预算法的计算密度，指定哪些区域需要更多的笔触来达到更精细的细节。例如，在人物的面部或重要物体边缘，可以通过掩码强制算法投入更多计算资源，从而在效率和视觉效果之间取得更好的平衡。","stars":2227,"starGrowth":2,"createdAt":"2020-06-05T11:13:14Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["非AI-好玩","生成艺术","可视化"],"zone":"创意","funScore":0.8,"domainTags":["遗传算法","绘画模拟","图像生成","OpenCV","Jupyter"],"domainKey":"遗传算法","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"生成艺术","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/anopara/genetic-drawing","ts":"2026-09-22T22:51:50.275Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wailsapp/wails","owner":"wailsapp","name":"wails","desc":"Create beautiful applications using Go","summaryCn":"用Go写后端，HTML写前端，轻松打包成精美桌面应用","reasonCn":"Wails是跨平台桌面应用开发框架，基于Go语言和Chromium嵌入技术。开发者用Go处理业务逻辑和系统调用，用HTML/CSS/JS构建美观界面，可一键打包为macOS、Windows、Linux原生应用。它打通了Web技术与原生系统能力的壁垒，让前端开发者也能打造专业级桌面软件。","stars":36314,"starGrowth":12,"createdAt":"2018-12-15T23:14:06Z","silentRounds":0,"language":"Go","topics":["go","golang","vuejs","desktop-application","macos","windows","linux","angular","react","svelte","javascript"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["桌面应用","跨平台","Go框架"],"domainKey":"桌面应用","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"vuejs","source":"github","weight":0.7},{"name":"desktop-application","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"angular","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"svelte","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/wailsapp/wails","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03699135319711406,"copyOk":false},{"repo":"debpalash/VoiceStudio","owner":"debpalash","name":"VoiceStudio","desc":"VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.","summaryCn":"本地语音克隆与工作流引擎，支持视频配音与模型设计。","reasonCn":"VoiceStudio 是一款开源的语音克隆与工作流引擎，默认基于 k2-fsa/OmniVoice 构建。它允许用户在本地硬件上运行工作流，同时支持视频配音、语音克隆、语音设计及模型管理。项目提供跨平台下载，安装后即可通过图形界面快速生成语音或设计专属模型。","stars":34388,"starGrowth":474,"createdAt":"2026-04-09T21:40:26Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["非AI-实用","多模态Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["语音克隆","视频配音","语音合成","有声书制作","离线语音处理"],"domainKey":"语音克隆","pushedAt":"2026-09-23T14:04:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/debpalash/VoiceStudio","ts":"2026-09-23T14:04:00Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.4096219611492381,"copyOk":true},{"repo":"open-webui/open-webui","owner":"open-webui","name":"open-webui","desc":"User-friendly AI Interface (Supports Ollama, OpenAI API, ...)","summaryCn":"本地开源 AI 平台，支持多种模型接入。","reasonCn":"Open WebUI 是一款自托管的 AI 平台，它将多种大模型能力整合到一个界面中，让用户既能用本地的 Ollama 模型，也能调用云端 OpenAI 的接口。它专为离线运行设计，通过友好的 Web 界面让复杂的模型交互变得简单直观。","stars":152846,"starGrowth":106,"createdAt":"2023-10-06T22:08:27Z","silentRounds":0,"language":"Python","topics":["rag","llm","ai"],"aiDims":["AI应用","大语言模型","模型部署"],"zone":"AI","funScore":0.5,"domainTags":["AI聊天","自托管","RAG"],"domainKey":"RAG","pushedAt":"2026-09-23T13:47:04Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/open-webui/open-webui","ts":"2026-09-23T13:47:04Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30208052549625114,"copyOk":true},{"repo":"files-community/Files","owner":"files-community","name":"Files","desc":"A modern file manager that helps users organize their files and folders.","summaryCn":"Windows文件管理太丑？这款开源神器让你重新爱上整理文件","reasonCn":"基于C#开发的现代化文件管理器，界面对标Windows 11设计语言。支持多标签页、双窗格、文件预览、批量重命名等核心功能。深度集成Windows Shell，支持拖拽操作和自定义主题。比系统自带文件资源管理器更轻量、更现代，适合追求效率的Windows用户。","stars":45618,"starGrowth":20,"createdAt":"2019-01-04T18:19:14Z","silentRounds":0,"language":"C#","topics":["developer-tools"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文件管理器","Windows应用","开发者工具"],"domainKey":"文件管理器","pushedAt":"2026-09-23T21:44:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/files-community/Files","ts":"2026-09-23T21:44:51Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06131527028804419,"copyOk":true},{"repo":"dagger/dagger","owner":"dagger","name":"dagger","desc":"Automation engine to build, test and ship any codebase. Runs locally, in CI, or directly in the cloud","summaryCn":"用代码而非 YAML 编写、本地即可运行的可编程 CI/CD 引擎。","reasonCn":"Dagger 将软件交付流程变为可编程的代码，支持 Go、Python、TypeScript 等 8 种语言 SDK。它以容器为运行时，让管道在本地、CI、云端任意环境一致运行，并提供交互式 REPL 与丰富模块生态，彻底替代易错难调试的 YAML 脚本。","stars":16294,"starGrowth":4,"createdAt":"2019-11-20T01:31:51Z","silentRounds":0,"language":"Go","topics":["devops","agents"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["CI/CD","自动化流水线","容器编排","基础设施即代码","DevOps","构建系统"],"domainKey":"CI/CD","pushedAt":"2026-09-23T21:33:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/dagger/dagger","ts":"2026-09-23T21:33:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"reacherhq/check-if-email-exists","owner":"reacherhq","name":"check-if-email-exists","desc":"Check if an email address exists without sending any email, written in Rust. Comes with a ⚙️ HTTP backend.","summaryCn":"不发消息就能验证邮箱存不存在，防作弊利器","reasonCn":"基于Rust开发的高性能邮箱验证工具，通过DNS MX记录查询、SMTP握手探测等多种技术手段，无需发送任何实际邮件即可准确判断邮箱地址是否真实有效。提供完整的HTTP后端接口，支持单机部署或集群扩展，可轻松集成到注册系统、用户数据库维护等场景中，有效防止恶意注册和虚假信息。","stars":9971,"starGrowth":20,"language":"Rust","topics":["email","email-validation","email-verification","hacktoberfest","email-validation-api","email-verification-api","email-verifier"],"aiDims":["工具调用","安全工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["邮箱验证","DNS查询","SMTP探测","反垃圾注册"],"domainKey":"邮箱验证","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"email","source":"github","weight":0.7},{"name":"email-validation","source":"github","weight":0.7},{"name":"email-verification","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"email-validation-api","source":"github","weight":0.7},{"name":"email-verification-api","source":"github","weight":0.7},{"name":"email-verifier","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/reacherhq/check-if-email-exists","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0660174506067596,"copyOk":true},{"repo":"langchain-ai/langgraph","owner":"langchain-ai","name":"langgraph","desc":"Build resilient agents.","summaryCn":"面向大模型智能体的低级编排框架，支持状态管理与长流程控制。","reasonCn":"LangGraph 是构建复杂 AI 智能体的底层框架，专注于解决智能体状态管理和任务编排难题。它允许开发者设计长时运行的智能体，通过清晰的状态机模型来控制流程，让 AI 在多轮对话中记住上下文并自主规划下一步行动，特别适合构建需要持续交互的复杂应用场景。","stars":42149,"starGrowth":50,"createdAt":"2023-08-09T18:33:12Z","silentRounds":0,"language":"Python","topics":["rag","llm","ai","agents"],"aiDims":["AI Agent","工具调用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["Agent编排","工作流","大模型"],"domainKey":"Agent编排","pushedAt":"2026-09-23T18:02:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/langchain-ai/langgraph","ts":"2026-09-23T18:02:45Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.1535588032930793,"copyOk":true},{"repo":"Shubhamsaboo/awesome-llm-apps","owner":"Shubhamsaboo","name":"awesome-llm-apps","desc":"100+ AI Agents, Agent Skills and RAG Apps - Free and Open Source.","summaryCn":"收录百余个可直接运行的开源AI Agent与RAG应用模板，每周更新。","reasonCn":"该项目精选并验证了超过一百个生产级的LLM应用模板，涵盖智能体团队协作、RAG知识检索、自进化技能等前沿场景。所有模板均经过端到端测试，支持主流开闭源模型，开发者无需从零搭建即可克隆部署。配套的Unwind AI平台提供分步教程与周度更新，显著降低了将大模型能力落地为实际应用的门槛，适合快速原型验证与商业化交付。","stars":139449,"starGrowth":127,"createdAt":"2024-04-29T05:30:25Z","silentRounds":0,"language":"Python","topics":["agents","rag"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"资源","funScore":0,"domainTags":["AI应用","Agent","RAG","开源合集"],"domainKey":"AI应用","pushedAt":"2026-09-23T05:24:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/Shubhamsaboo/awesome-llm-apps","ts":"2026-09-23T05:24:46Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30273218165781035,"copyOk":true},{"repo":"osquery/osquery","owner":"osquery","name":"osquery","desc":"SQL powered operating system instrumentation, monitoring, and analytics.","summaryCn":"将操作系统变成数据库，用 SQL 查询系统状态。","reasonCn":"osquery 是一款 SQL 驱动的系统监控与分析框架，它能让你用数据库查询语句直接查看操作系统内的运行进程、内核模块、网络连接和文件哈希等底层信息。它将复杂的系统内部运作抽象为一张张关系型数据表，让安全运维人员无需编写代码就能像查数据库一样检索系统状态，从而快速发现异常行为或进行合规审计。","stars":23586,"starGrowth":3,"createdAt":"2014-07-29T20:27:33Z","silentRounds":0,"language":"C++","topics":["security","monitoring"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["系统监控","安全审计","SQL查询","进程管理"],"domainKey":"系统监控","pushedAt":"2026-09-22T16:26:35Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/osquery/osquery","ts":"2026-09-22T16:26:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"go-vgo/robotgo","owner":"go-vgo","name":"robotgo","desc":"RobotGo, Go Native cross-platform RPA, GUI automation, Auto test and Computer use  @vcaesar","summaryCn":"用代码操控鼠标键盘的跨平台自动化工具，写几行Go就能让电脑自己干活儿","reasonCn":"这是一个用Go语言编写的跨平台GUI自动化库，支持Windows、macOS和Linux。它能模拟鼠标点击移动、键盘输入、屏幕截图、颜色识别和窗口管理，底层通过CGO调用原生系统API，无需安装额外依赖即可使用。适合需要程序化控制电脑界面的各种场景。","stars":10816,"starGrowth":2,"createdAt":"2016-09-26T16:26:56Z","silentRounds":0,"language":"Go","topics":["automation"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["跨平台RPA","GUI自动化","Go语言库"],"domainKey":"跨平台RPA","pushedAt":"2026-09-22T17:14:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/go-vgo/robotgo","ts":"2026-09-22T17:14:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"openclaw/openclaw","owner":"openclaw","name":"openclaw","desc":"The AI that really does things. Any OS. Any Platform. The lobster way. 🦞","summaryCn":"自托管AI助手，统管二十余聊天平台，模型插件化可随意切换。","reasonCn":"OpenClaw 是一个自托管的 AI 助手网关，运行在你自己的设备上，将 Discord、Slack、iMessage、微信等二十多个聊天平台统一接入。它把模型能力（如 Claude、Codex 或本地模型）和 Agent 编排框架做成可热插拔的插件，换模型无需改动业务逻辑。同一套网关仅靠配置差异，既能作为个人助手跑在笔记本上，也能扩展为团队共享部署。所有状态、记忆与凭证均留在本地硬件，实现数据完全自主。","stars":390261,"starGrowth":39,"createdAt":"2025-11-24T10:16:47Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI Agent","AI应用","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI助手","跨平台部署","本地隐私","任务自动化","插件生态"],"domainKey":"AI助手","pushedAt":"2026-09-23T21:51:14Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/openclaw/openclaw","ts":"2026-09-23T21:51:14Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.14075975923804837,"copyOk":true},{"repo":"appsmithorg/appsmith","owner":"appsmithorg","name":"appsmith","desc":"Platform to build admin panels, internal tools, and dashboards. Integrates with 25+ databases and any API.","summaryCn":"零代码也能快速搭出超酷炫的内部管理系统，拖拽几下搞定复杂业务面板","reasonCn":"Appsmith是一个开源的低代码开发平台，支持连接25种以上数据库和任意API，通过拖拽式界面快速构建管理后台、数据仪表板和内部工具。自带权限管理和部署功能，无需编写大量代码即可完成企业级应用搭建，大幅缩短开发周期。","stars":40921,"starGrowth":0,"createdAt":"2020-06-30T04:07:36Z","silentRounds":1,"language":"TypeScript","topics":["self-hosted","developer-tools","automation"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["低代码平台","内部工具搭建","管理后台生成","可视化编程","自托管应用"],"domainKey":"低代码平台","pushedAt":"2026-09-23T14:00:46Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/appsmithorg/appsmith","ts":"2026-09-23T14:00:46Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langgenius/dify","owner":"langgenius","name":"dify","desc":"Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.","summaryCn":"可视化搭建AI应用的瑞士军刀，拖拉拽就能造自己的AI产品","reasonCn":"技术架构上，Dify提供了一套完整的应用开发栈：支持LangChain和自建的工作流引擎，可以编排多步骤的AI任务流程；内置了RAG知识库模块，支持文档上传、分段、向量化和检索；集成了多种主流大模型，包括OpenAI、Anthropic、国内各种大模型以及本地部署模型；支持Agent模式，可以让AI","stars":156879,"starGrowth":109,"createdAt":"2023-04-12T07:40:24Z","silentRounds":0,"language":"TypeScript","topics":["llm","ai"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["大语言模型应用","RAG知识库","工作流编排","Agent框架","企业级部署"],"domainKey":"大语言模型应用","pushedAt":"2026-09-23T20:26:33Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/langgenius/dify","ts":"2026-09-23T20:26:33Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30208440900311706,"copyOk":true},{"repo":"louislam/uptime-kuma","owner":"louislam","name":"uptime-kuma","desc":"A fancy self-hosted monitoring tool","summaryCn":"自托管服务器监控工具，支持 90+ 通知服务与精美状态页。","reasonCn":"Uptime Kuma 是一款开源的自托管监控工具，旨在替代老旧的 Uptime Robot。它通过网页界面直观展示服务器的运行状态，并支持 HTTP、TCP、DNS 等多种监控类型。最核心的优势在于它能与 90 多种第三方服务（如 Telegram、Discord、邮件）联动，一旦服务宕机立即推送告警。此外，它还能生成美观的状态页，方便向访客展示服务健康度，完全替代了传统的监控面板。","stars":91694,"starGrowth":42,"createdAt":"2021-07-03T01:02:42Z","silentRounds":0,"language":"JavaScript","topics":["uptime","monitoring","docker","selfhosted","self-hosted","single-page-app","webapp","responsive","websocket","socket-io","uptime-monitoring","monitor"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["服务监控","自托管","告警推送"],"domainKey":"服务监控","pushedAt":"2026-09-23T14:57:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"uptime","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"selfhosted","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"single-page-app","source":"github","weight":0.7},{"name":"webapp","source":"github","weight":0.7},{"name":"responsive","source":"github","weight":0.7},{"name":"websocket","source":"github","weight":0.7},{"name":"socket-io","source":"github","weight":0.7},{"name":"uptime-monitoring","source":"github","weight":0.7},{"name":"monitor","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/louislam/uptime-kuma","ts":"2026-09-23T14:57:57Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.12737413571226033,"copyOk":true},{"repo":"google-research/google-research","owner":"google-research","name":"google-research","desc":"Google Research","summaryCn":"Google Research这个项目堪称开源AI界的宝库。","reasonCn":"仓库按团队和研究领域分类，内容极其丰富。自然语言处理方面有BERT、T5、Albert等改变整个领域的模型代码；计算机视觉有ImageNet相关的项目；强化学习有AlphaGo、MuZero等里程碑式工作；还有大量关于AI安全、可解释性、TPU芯片设计的研究成果。","stars":38814,"starGrowth":6,"createdAt":"2018-10-04T18:42:48Z","silentRounds":0,"language":"Jupyter Notebook","topics":["ml","ai"],"aiDims":[],"zone":"AI","funScore":0.5,"domainTags":["机器学习","深度学习","模型研究"],"domainKey":"机器学习","pushedAt":"2026-09-23T12:02:29Z","tags":[{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/google-research/google-research","ts":"2026-09-23T12:02:29Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.018463750193229246,"copyOk":false},{"repo":"bookhoard/bookhoarder","owner":"bookhoard","name":"bookhoarder","desc":"Self-hosted EPUB library — no database, just object storage","summaryCn":"自己搭建的电子书库，零数据库只靠对象存储，轻松管理EPUB","reasonCn":"bookhoarder 是一款自托管的电子书管理系统，直接使用对象存储（如 S3、MinIO）保存 EPUB 文件，免去数据库配置。它提供网页界面，可浏览、搜索、标签分类，还支持一键导入导出书籍元数据，兼容常见元数据格式，开放 REST 接口方便二次开发。","stars":16,"starGrowth":0,"createdAt":"2026-08-27T17:50:12Z","silentRounds":8,"language":"TypeScript","topics":["ebook"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["电子书管理","自托管","对象存储"],"domainKey":"电子书管理","pushedAt":"2026-09-22T16:36:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/bookhoard/bookhoarder","ts":"2026-09-22T16:36:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rtk-ai/rtk","owner":"rtk-ai","name":"rtk","desc":"CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies","summaryCn":"CLI 代理压缩命令输出，减少 LLM 上下文占用。","reasonCn":"RTK 是一款专为 LLM Agent 设计的命令行代理工具，旨在解决命令输出过多导致 Token 消耗过大的问题。它通过拦截并压缩 Shell 命令的返回结果，将 Agent 需要读取的数据量削减最多达 90%，从而显著降低 API 调用成本并提升处理速度。作为单文件 Rust 程序，它兼容 100 多种常用命令，且仅带来小于 10 毫秒的系统开销，为开发者在本地或云端运行 AI 编程助手时提供了高效的输出优化方案。","stars":81452,"starGrowth":131,"createdAt":"2026-01-22T16:54:16Z","silentRounds":0,"language":"Rust","topics":["llm"],"aiDims":["AI基础设施","效率工具","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["LLM优化","命令行工具","Token压缩"],"domainKey":"LLM优化","pushedAt":"2026-09-23T20:15:39Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/rtk-ai/rtk","ts":"2026-09-23T20:15:39Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3657899130776408,"copyOk":true},{"repo":"OpenHands/OpenHands","owner":"OpenHands","name":"OpenHands","desc":"🙌 OpenHands: AI-Driven Development","summaryCn":"让AI当你的程序员搭档，自动写代码跑测试改bug","reasonCn":"OpenHands是开源AI编程Agent，支持LLM自主学习代码库、执行终端命令、操作浏览器。内置文件编辑、代码搜索、测试运行等完整工具链，使AI能像程序员一样完成开发任务。与Cursor不同，它强调全流程自动化，从需求理解到部署上线一气呵成。","stars":88861,"starGrowth":117,"createdAt":"2024-03-13T03:33:31Z","silentRounds":0,"language":"TypeScript","topics":["llm"],"aiDims":["AI Agent","代码助手","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI编程Agent","代码生成","软件工程自动化"],"domainKey":"AI编程Agent","pushedAt":"2026-09-23T20:49:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/OpenHands/OpenHands","ts":"2026-09-23T20:49:36Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30394998930914574,"copyOk":true},{"repo":"react/react-native","owner":"react","name":"react-native","desc":"A framework for building native applications using React","summaryCn":"用React技术栈轻松开发跨平台手机应用","reasonCn":"Meta开源的跨平台移动开发框架，允许你用JavaScript和React语法编写iOS和Android原生应用。一套代码双端运行，性能接近原生。社区庞大，生态丰富，是前端开发者进入移动端的最短路径，适合追求开发效率的团队和个人。","stars":126699,"starGrowth":16,"createdAt":"2015-01-09T18:10:16Z","silentRounds":0,"language":"C++","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["移动开发","React","跨平台框架","JavaScript","原生应用"],"domainKey":"移动开发","pushedAt":"2026-09-23T19:01:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/react/react-native","ts":"2026-09-23T19:01:52Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04837885066180475,"copyOk":false},{"repo":"TheAlgorithms/Python","owner":"TheAlgorithms","name":"Python","desc":"All Algorithms implemented in Python","summaryCn":"收录海量 Python 算法与数据结构实现，专为教学面试设计。","reasonCn":"该项目汇集了排序、搜索、图论、动态规划等几百种经典算法的 Python 代码实现，覆盖计算机科学核心课程与大厂面试高频考点。所有代码以教学为首要目标，注重可读性与结构清晰，而非追求工程性能，并附带详细注释帮助理解原理。它不是现成的库，而是一本可运行的「算法百科全书」，适合备战面试、补齐 CS 基础或对照标准库学习优化技巧的开发者。","stars":224880,"starGrowth":34,"createdAt":"2016-07-16T09:44:01Z","silentRounds":0,"language":"Python","topics":["interview"],"aiDims":["学习资源","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["算法实现","计算机科学","Python"],"domainKey":"算法实现","pushedAt":"2026-09-23T21:43:41Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/TheAlgorithms/Python","ts":"2026-09-23T21:43:41Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10245357524012808,"copyOk":true},{"repo":"cakephp/cakephp","owner":"cakephp","name":"cakephp","desc":"CakePHP: The Rapid Development Framework for PHP - Official Repository","summaryCn":"基于PHP的Web开发框架，用MVC模式快速构建稳健应用。","reasonCn":"CakePHP是一款PHP快速开发框架，采用MVC（模型-视图-控制器）和关联数据映射等设计模式。它的目标是提供一套结构化的框架，让不同水平的PHP开发者都能快速构建健壮的Web应用，同时保持灵活性。它通过Composer进行安装，并支持使用App Skeleton作为新项目的起点。","stars":8790,"starGrowth":0,"createdAt":"2010-05-08T14:38:22Z","silentRounds":22,"language":"PHP","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["PHP框架","Web开发","MVC","后端开发"],"domainKey":"PHP框架","pushedAt":"2026-09-23T07:53:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/cakephp/cakephp","ts":"2026-09-23T07:53:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ollama/ollama","owner":"ollama","name":"ollama","desc":"Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.","summaryCn":"把各种AI大模型装进电脑的“万能遥控器”，小白也能一键跑通本地AI","reasonCn":"Go语言打造的本地大模型运行平台，主打极简体验。一条命令即可下载并运行Qwen、GLM、DeepSeek等数十种主流开源模型，支持Mac/Linux/Windows。它屏蔽了复杂的CUDA配置和依赖冲突，让你无需专业AI背景也能在本地流畅使用大模型进行对话、创作和任务处理。","stars":181491,"starGrowth":87,"createdAt":"2023-06-26T19:39:32Z","silentRounds":0,"language":"Go","topics":["llm"],"aiDims":["大语言模型","AI应用","模型部署"],"zone":"AI","funScore":0,"domainTags":["本地大模型推理","模型部署","GGUF量化"],"domainKey":"模型部署","pushedAt":"2026-09-23T20:45:27Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/ollama/ollama","ts":"2026-09-23T20:45:27Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.2624380878390664,"copyOk":false},{"repo":"firecrawl/firecrawl","owner":"firecrawl","name":"firecrawl","desc":"The context API to search, scrape, and interact with the web at scale. 🔥","summaryCn":"帮你把散乱网页变成AI能直接用的干净数据。","reasonCn":"Firecrawl 是一个高性能的网页数据提取工具，旨在解决传统爬虫难以处理动态网页和复杂结构的痛点。它不仅能抓取静态页面，还能处理包含大量 JavaScript 的现代 Web 应用，将内容转化为结构化的 Markdown 或 JSON 数据，让 AI 模型能够直接读取和使用。这个项目特别适合需要从大量网页中快速收集信息、训练模型或构建知识库的场景。","stars":183397,"starGrowth":412,"createdAt":"2024-04-15T21:02:29Z","silentRounds":0,"language":"TypeScript","topics":["llm","ai"],"aiDims":["AI搜索","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["网页抓取","结构化数据提取","RAG数据源","AI应用开发"],"domainKey":"网页抓取","pushedAt":"2026-09-23T20:14:06Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/firecrawl/firecrawl","ts":"2026-09-23T20:14:06Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30673947774500127,"copyOk":true},{"repo":"imDema/FreeMove","owner":"imDema","name":"FreeMove","desc":"","summaryCn":"Windows 用户常面临 C 盘空间告急却不敢删除文件的困境，FreeMove 正是为这","reasonCn":"不同于命令行工具的高门槛，FreeMove 提供简洁的 GUI 界面，用户只需选择要迁移的目录即可，无需手动输入复杂的 mklink 命令。工具内置了安全检测机制，明确警告不要移动系统关键目录（如 Windows 更新或商店应用目录），防止误操作导致系统崩溃。","stars":2267,"starGrowth":10,"createdAt":"2016-10-26T13:13:32Z","silentRounds":0,"language":"C#","topics":[],"aiDims":["非AI-实用","效率工具","工具调用"],"zone":"工具","funScore":0.2,"domainTags":["Windows磁盘管理","符号链接","系统迁移","C盘清理"],"domainKey":"Windows磁盘管理","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.5,"source":"tier","url":"https://github.com/imDema/FreeMove","ts":"2026-09-22T21:05:02.747Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04323334803705337,"copyOk":false},{"repo":"bytebase/bytebase","owner":"bytebase","name":"bytebase","desc":"Database governance built for humans and agents — controlling changes and access across every major database.","summaryCn":"统一管理多数据库变更与合规的DevOps平台。","reasonCn":"Bytebase是一个开源数据库治理平台，旨在解决开发与运维中数据库变更管理混乱、权限分散的问题。它将SQL客户端、迁移脚本和工单系统整合在一起，提供统一的变更控制与访问管理。平台支持PostgreSQL、MySQL等20多种主流数据库，帮助企业降低SQL误操作风险，确保数据库变更的可追溯与合规。","stars":14503,"starGrowth":4,"createdAt":"2021-01-27T04:33:50Z","silentRounds":0,"language":"Go","topics":["devops"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["数据库治理","SQL审核","变更管理","权限控制","DevOps平台"],"domainKey":"数据库治理","pushedAt":"2026-09-23T09:26:02Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/bytebase/bytebase","ts":"2026-09-23T09:26:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wesnoth/wesnoth","owner":"wesnoth","name":"wesnoth","desc":"An open source, turn-based strategy game with a high fantasy theme.","summaryCn":"经典回合制策略游戏，免费开源的奇幻大战场","reasonCn":"游戏的亮点在于内容深度和可玩性。每个单位有成长系统，战斗中升级能变强；地形系统影响移动和防御；每个种族有独特的单位类型和特殊能力。更重要的是社区生态——玩家制作了海量自定义地图和模组，从奇幻冒险到历史战争，题材无所不包，几乎保证了永不过时的新鲜内容。","stars":6885,"starGrowth":0,"createdAt":"2013-04-14T17:13:06Z","silentRounds":3,"language":"C++","topics":["game"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0.5,"domainTags":["回合制策略","奇幻游戏","多人对战","开源游戏"],"domainKey":"回合制策略","pushedAt":"2026-09-23T17:13:56Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/wesnoth/wesnoth","ts":"2026-09-23T17:13:56Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Tencent/WeKnora","owner":"Tencent","name":"WeKnora","desc":"Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.","summaryCn":"腾讯开源的LLM驱动企业知识框架，实现文档语义理解与自动推理。","reasonCn":"WeKnora是一款面向企业级场景的LLM知识框架，旨在通过RAG和智能体技术，让静态文档转化为可检索、可推理的动态知识。它解决了传统文档管理检索效率低、难以深度利用的问题，能自动将文档转化为Wiki，并支持语义检索与自主推理，帮助企业构建高效的知识资产体系。","stars":28928,"starGrowth":426,"createdAt":"2025-07-22T08:01:23Z","silentRounds":0,"language":"Go","topics":["llm","chatbot","rag"],"aiDims":["RAG与知识","AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["知识库","RAG","AI Agent","企业知识管理"],"domainKey":"RAG","pushedAt":"2026-09-23T15:11:50Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Tencent/WeKnora","ts":"2026-09-23T15:11:50Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.34417865044247786,"copyOk":true},{"repo":"tinyhumansai/openhuman","owner":"tinyhumansai","name":"openhuman","desc":"OpenHuman is an open source agent harness with local-first memory, agent orchestration, and workflows","summaryCn":"本地优先的个人AI超级智能体，具备长期记忆与深度研究能力。","reasonCn":"OpenHuman 定位为个人 AI 超级智能，核心优势在于本地优先架构保障隐私，拥有长期记忆大脑、强大编排与深度研究能力。它不仅能记住一切，还能像专家一样协调工具完成复杂任务。早期测试版发布即连续九天霸榜 GitHub Trending，展现了开发者对本地化强 Agent 的迫切需求。","stars":40049,"starGrowth":37,"createdAt":"2026-02-18T20:01:27Z","silentRounds":0,"language":"Rust","topics":["llm"],"aiDims":["AI Agent","AI应用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["Agent框架","本地记忆","个人知识库","多智能体编排","隐私计算"],"domainKey":"Agent框架","pushedAt":"2026-09-23T21:41:30Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/tinyhumansai/openhuman","ts":"2026-09-23T21:41:30Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1365259257409673,"copyOk":true},{"repo":"semantic-release/semantic-release","owner":"semantic-release","name":"semantic-release","desc":"📦🚀 Fully automated version management and package publishing","summaryCn":"自动帮你管理版本号和发包发布的瑞士军刀，彻底告别手动发版","reasonCn":"它自动分析git提交记录，根据commit内容判断下一个版本号，还能直接发布到npm等包管理器。支持语义化版本，配置灵活，可集成到CI/CD流程中。告别手动改package.json版本号的时代，让发包变得自动化、无脑化，是开源项目维护者的必备神器。","stars":24058,"starGrowth":2,"createdAt":"2014-09-15T16:18:28Z","silentRounds":0,"language":"JavaScript","topics":["automation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["版本管理","CI/CD","npm发包","自动化","Conventional Commits"],"domainKey":"版本管理","pushedAt":"2026-09-22T07:33:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/semantic-release/semantic-release","ts":"2026-09-22T07:33:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"getsops/sops","owner":"getsops","name":"sops","desc":"Simple and flexible tool for managing secrets","summaryCn":"加密配置文件编辑器，将敏感数据安全地存储在版本控制系统中。","reasonCn":"SOPS 是一个加密文件编辑器，专门用于管理包含敏感信息的配置文件。它支持加密 YAML、JSON、ENV 等格式，并集成了 AWS、GCP、Azure、华为云等多种密钥管理服务。对于 DevOps 团队而言，它解决了将数据库密码、API 密钥等机密直接写入代码仓库的安全隐患，让你可以放心地将加密后的配置存入 Git。","stars":23192,"starGrowth":7,"createdAt":"2015-08-13T22:11:00Z","silentRounds":0,"language":"Go","topics":["security","devops"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Secrets管理","配置加密","GitOps","KMS集成","密钥轮换"],"domainKey":"Secrets管理","pushedAt":"2026-09-21T20:03:18Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/getsops/sops","ts":"2026-09-21T20:03:18Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02190548464987927,"copyOk":true},{"repo":"exacity/simplified-deeplearning","owner":"exacity","name":"simplified-deeplearning","desc":"","summaryCn":"该仓库收录了《深度学习》一书的中文笔记，基于书中章节结构进行梳理。","reasonCn":"核心亮点在于将书籍内容拆分为可直接运行的 Jupyter Notebook，既方便阅读，又可即时验证算法效果。每个 Notebook 以章节为单位，配有详细注释，说明网络结构、损失函数与训练流程。对比传统的纯文本教程，项目在代码层面提供了可复现的实验环境，让学习更具互动性。","stars":2539,"starGrowth":2,"createdAt":"2016-07-08T07:40:29Z","silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["深度学习","PyTorch","TensorFlow","学习笔记"],"domainKey":"深度学习","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/exacity/simplified-deeplearning","ts":"2026-09-21T23:17:36.060Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vivizzz007/vivi-music","owner":"vivizzz007","name":"vivi-music","desc":"Vivi-Music is an expressive Material 3–based YouTube Music client for Android.","summaryCn":"随专辑封面色彩自动变换主题的沉浸式音乐播放器。","reasonCn":"VIVI 是一款主打沉浸式体验的现代音乐播放器，它不只是播放音乐，更是把听歌变成了一种视觉享受。项目最核心的是它的响应式设计引擎，能根据你正在播放的专辑封面颜色自动调整整个界面的色调，让音乐与视觉完美融合。除了色彩匹配，它还加入了流畅的物理动效和动态画布可视化，让每一次点击和滑动都充满手感。更重要的是，它承诺 100% 隐私优先，所有数据都存储在本地数据库中，无需联网即可畅享无广告的纯净听歌环境。","stars":3098,"starGrowth":37,"createdAt":"2025-04-22T14:45:45Z","silentRounds":0,"language":"Kotlin","topics":["music"],"aiDims":["音乐","创意工具"],"zone":"创意","funScore":0.5,"domainTags":["音乐播放器","Material 3","Android","YouTube Music","动态歌词"],"domainKey":"音乐播放器","pushedAt":"2026-09-23T08:53:02Z","tags":[{"name":"音乐","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":1,"source":"search","url":"https://github.com/vivizzz007/vivi-music","ts":"2026-09-23T08:53:02Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.14682956746287928,"copyOk":true},{"repo":"anthropics/claude-code","owner":"anthropics","name":"claude-code","desc":"Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.","summaryCn":"在终端里养个AI编程助手，自动改bug写代码","reasonCn":"Anthropic官方推出的终端AI编程代理，深度集成Claude能力。能理解整个代码库上下文，自动执行重复任务、解释复杂逻辑、处理Git工作流。相比传统Copilot，它更像独立开发者，直接操作文件而非仅补全，大幅缩短从想法到代码的距离。","stars":147708,"starGrowth":195,"silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","代码库分析","终端自动化"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/anthropics/claude-code","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.30396051669510116,"copyOk":true},{"repo":"dembrandt/dembrandt","owner":"dembrandt","name":"dembrandt","desc":"Extract any website’s design system into tokens in seconds: logo, colors, typography, borders & more. One command.","summaryCn":"一键提取网页设计系统，自动生成 W3C 标准令牌。","reasonCn":"Dembrandt 能自动抓取目标网站的配色、字体、间距等设计细节，并转化为标准化的 W3C 令牌。开发者只需一条命令，就能在本地复刻任意网页的设计风格，适合前端开发、UI 模拟或竞品分析。它解决了人工提取繁琐且容易出错的问题，让设计还原变得高效。","stars":3512,"starGrowth":4,"createdAt":"2025-11-22T13:49:09Z","silentRounds":0,"language":"TypeScript","topics":["frontend","productivity"],"aiDims":[],"zone":"工具","funScore":0,"domainTags":["设计系统","前端工具","CSS提取","代码生成","自动化"],"domainKey":"设计系统","pushedAt":"2026-09-23T07:01:05Z","tags":[{"name":"frontend","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/dembrandt/dembrandt","ts":"2026-09-23T07:01:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"imba/imba","owner":"imba","name":"imba","desc":"🐤 The friendly full-stack language","summaryCn":"用类似Ruby的语法写全栈应用，前端开发不再痛苦","reasonCn":"Imba是一个独特的全栈编程语言和框架，它抛弃了繁琐的JS语法，用简洁优雅的指令式代码构建高性能前端。支持服务端渲染、实时双向绑定，且自带编译器能生成极小的JS包。对于受够了React/JSX繁琐语法的开发者来说，这是一个令人耳目一新的“另类”选择。","stars":6509,"starGrowth":0,"createdAt":"2014-06-14T18:58:32Z","silentRounds":22,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0.7,"domainTags":["全栈编程语言","前端框架","编译器优化","Ruby风格语法","响应式UI"],"domainKey":"全栈编程语言","pushedAt":"2026-09-22T09:41:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/imba/imba","ts":"2026-09-22T09:41:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dgtlmoon/changedetection.io","owner":"dgtlmoon","name":"changedetection.io","desc":"Best and simplest tool for website change detection, web page monitoring, and website change alerts. Perfect for tracking content changes, price drops, restock alerts, and website defacement monitoring—all for free or enjoy our SaaS plan!","summaryCn":"自动盯住网页变动，降价提醒、内容监控一站式搞定","reasonCn":"极简的网站变更检测工具，支持浏览器渲染、JavaScript解析和自定义触发条件。可监控商品价格、库存状态、新闻更新等，通过邮件、Webhook、Slack等多渠道通知。完全自托管，数据掌握在自己手中，适合个人和团队使用。","stars":34461,"starGrowth":23,"createdAt":"2021-01-27T16:03:30Z","silentRounds":0,"language":"Python","topics":["self-hosted"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网站监控","变更检测","价格追踪","自托管","Webhook通知"],"domainKey":"网站监控","pushedAt":"2026-09-21T16:06:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/dgtlmoon/changedetection.io","ts":"2026-09-21T16:06:16Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07100226342822322,"copyOk":true},{"repo":"HKUDS/CLI-Anything","owner":"HKUDS","name":"CLI-Anything","desc":"\"CLI-Anything: Making ALL Software Agent-Native\" -- CLI-Hub: https://clianything.cc/","summaryCn":"把任何软件都变成能听懂指令的AI智能体，让电脑自动化","reasonCn":"CLI-Anything 是一个革命性的开源框架，它通过底层驱动和模拟技术，赋予任何图形化软件（如浏览器、IDE）AI Agent 能力。它利用 Python 作为胶水语言，将大模型的自然语言指令转化为操作系统级的点击、输入和模拟操作，真正实现“软件原生 AI 化”。项目已获得超过 4.7k Star，是构建跨应用自动化工作流的神器。","stars":49880,"starGrowth":41,"language":"Python","topics":[],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI Agent","自动化","桌面控制","自然语言编程"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/HKUDS/CLI-Anything","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.12546591820368883,"copyOk":false},{"repo":"docling-project/docling","owner":"docling-project","name":"docling","desc":"Get your documents ready for gen AI","summaryCn":"把各种乱七八糟的文档变成AI能读懂的结构化数据","reasonCn":"Docling 是一个高性能的文档解析库，专为生成式AI准备数据而设计。它使用现代解析技术，能快速将PDF、Word、HTML等文档转换为结构化数据，支持Markdown和JSON输出。项目采用模块化架构，结合OCR和布局分析，解决传统工具在复杂文档处理上的瓶颈。","stars":67630,"starGrowth":73,"createdAt":"2024-07-09T07:50:26Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["文档理解","AI基础设施","AI应用"],"zone":"AI","funScore":0,"domainTags":["文档解析","PDF处理","OCR","RAG数据准备","结构化数据提取"],"domainKey":"文档解析","pushedAt":"2026-09-23T16:23:08Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/docling-project/docling","ts":"2026-09-23T16:23:08Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.22223820789590418,"copyOk":false},{"repo":"VictoriaMetrics/VictoriaMetrics","owner":"VictoriaMetrics","name":"VictoriaMetrics","desc":"VictoriaMetrics: fast, cost-effective monitoring solution and time series database","summaryCn":"高性能时序数据库，监控和指标存储的效率神器","reasonCn":"VictoriaMetrics的核心优势在于其极致的存储效率和查询性能。它采用独特的时间序列压缩算法，存储占用远低于InfluxDB和Prometheus，官方数据显示压缩比可达传统方案的五分之一。同时查询引擎经过专门优化，即使在海量数据下也能保持毫秒级响应。","stars":17754,"starGrowth":3,"createdAt":"2018-09-30T09:58:01Z","silentRounds":0,"language":"Go","topics":["database","monitoring"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["时序数据库","Prometheus","指标监控","云原生","存储压缩"],"domainKey":"时序数据库","pushedAt":"2026-09-23T18:57:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.48,"source":"search","url":"https://github.com/VictoriaMetrics/VictoriaMetrics","ts":"2026-09-23T18:57:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/monaco-editor","owner":"microsoft","name":"monaco-editor","desc":"A browser based code editor","summaryCn":"VS Code同款编辑器，在浏览器里流畅写代码的利器","reasonCn":"Monaco Editor是微软开源的浏览器端代码编辑器，与VS Code同源，提供智能补全、语法高亮、Diff对比等核心功能。轻量且可定制，适合嵌入Web应用实现在线编辑器，支持JavaScript、TypeScript、Python等数十种语言，是Web IDE的基础设施。","stars":46789,"starGrowth":4,"createdAt":"2016-06-07T16:56:31Z","silentRounds":0,"language":"JavaScript","topics":["editor"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["代码编辑器","Web IDE","语法高亮","智能补全","VS Code内核"],"domainKey":"代码编辑器","pushedAt":"2026-09-21T23:27:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/microsoft/monaco-editor","ts":"2026-09-21T23:27:43Z","category":"tool","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"bombshell-dev/clack","owner":"bombshell-dev","name":"clack","desc":"Effortlessly build beautiful command-line apps","summaryCn":"用最少代码打造漂亮命令行应用，TypeScript开发者的福音","reasonCn":"Clack的核心价值在于极简API设计。它提供了一组预构建的UI组件，包括文字提示、选择菜单、进度条、确认框等，开发者只需调用几个函数就能组合出完整的交互流程。与Ink、Blessed等方案相比，Clack的API更加直观，学习成本极低，同时内置了主题系统，轻松切换不同视觉风格。","stars":8072,"starGrowth":4,"createdAt":"2023-02-08T21:58:49Z","silentRounds":0,"language":"TypeScript","topics":["command-line"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["CLI开发","命令行界面","交互组件","TypeScript","终端UI","开发者工具"],"domainKey":"CLI开发","pushedAt":"2026-09-22T15:56:13Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/bombshell-dev/clack","ts":"2026-09-22T15:56:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vastsa/PI-Desktop","owner":"vastsa","name":"PI-Desktop","desc":"Local-first AI coding agent desktop: Electron + Rust host core + pi Agent Harness + user-installable plugins","summaryCn":"本地优先的AI编程智能体桌面工作台，自带模型直连零锁定。","reasonCn":"PI-Desktop 是一个面向 AI 编程智能体的桌面级工作台，主打本地优先与模型厂商无关。它不需要注册账号、不强制中转、不绑定编辑器，所有项目与会话数据留在本机，模型请求直连你配置的供应商端点。用户可直接打开任意本地代码库，让多个智能体并行处理重构、测试、文档生成等任务，同时保持对上下文与权限的完全掌控。","stars":5206,"starGrowth":241,"createdAt":"2023-03-22T14:23:39Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0,"domainTags":["离线编程","代码补全","插件系统","桌面应用"],"domainKey":"离线编程","pushedAt":"2026-09-23T19:44:31Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/vastsa/PI-Desktop","ts":"2026-09-23T19:44:31Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.43887821744141375,"copyOk":true},{"repo":"vectorize-io/hindsight","owner":"vectorize-io","name":"hindsight","desc":"Hindsight: Agent Memory That Learns","summaryCn":"让AI记住所有对话历史，越聊越聪明，就像有个贴心的小秘书","reasonCn":"Hindsight是一款专为AI Agent设计的记忆系统，能让智能体跨会话记住重要信息。采用向量存储和关键词索引双重机制，自动提取对话中的实体、事件和偏好，支持长期记忆和短期记忆的区分管理。安装简单，只需pip安装即可接入LangChain或LlamaIndex等主流框架。","stars":25241,"starGrowth":529,"createdAt":"2025-10-30T11:49:48Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["AI Agent","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["Agent记忆系统","长期记忆","对话历史管理","向量存储","LangChain集成"],"domainKey":"Agent记忆系统","pushedAt":"2026-09-23T19:36:48Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/vectorize-io/hindsight","ts":"2026-09-23T19:36:48Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.4354486747751674,"copyOk":false},{"repo":"cloudnative-pg/cloudnative-pg","owner":"cloudnative-pg","name":"cloudnative-pg","desc":"The most popular Kubernetes Operator for PostgreSQL.","summaryCn":"Kubernetes 上的 PostgreSQL 全自动运维平台","reasonCn":"CloudNativePG 是专为 Kubernetes 设计的 PostgreSQL 运维平台，通过 Operator 模式将部署、备份、高可用、滚动升级等全生命周期操作自动化。它原生支持声明式配置与 GitOps 工作流，让数据库像无状态应用一样纳入 CI/CD 管道，解决了有状态服务在 K8s 上运维困难、手工介入多的痛点，已成为 CNCF 毕业项目。","stars":9339,"starGrowth":5,"createdAt":"2022-03-10T11:21:32Z","silentRounds":0,"language":"Go","topics":["devops","kubernetes","database"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","PostgreSQL","数据库运维","自动备份","高可用集群"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T13:53:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.48,"source":"trending","url":"https://github.com/cloudnative-pg/cloudnative-pg","ts":"2026-09-23T13:53:27Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0166061676839062,"copyOk":true},{"repo":"getopenscreen/openscreen","owner":"getopenscreen","name":"openscreen","desc":"Record your screen, ship a demo. Free and open-source, GPU-accelerated, no watermarks, no subscriptions. Windows, macOS, Linux. Actively maintained.","summaryCn":"免费开源的录屏演示工具，无水印无订阅，支持多平台，制作产品演示超方便","reasonCn":"市面上大多数录屏工具要么收费、要么加水印、要么限制导出格式，这个项目试图用开源的方式解决这些问题。它支持GPU加速渲染，录制流畅度不错，而且完全免费使用，没有任何功能限制。项目支持Windows、macOS、Linux三大主流操作系统，对于需要跨平台协作的团队来说很友好。","stars":3167,"starGrowth":29,"createdAt":"2026-03-14T10:16:22Z","silentRounds":0,"language":"TypeScript","topics":["video-editor"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["屏幕录制","产品演示视频","跨平台GPU加速"],"domainKey":"屏幕录制","pushedAt":"2026-09-23T21:25:02Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/getopenscreen/openscreen","ts":"2026-09-23T21:25:02Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.13736495105778337,"copyOk":true},{"repo":"Jakubantalik/Libraries.dev","owner":"Jakubantalik","name":"Libraries.dev","desc":"High-crafted UI libraries for AI agents: Border beam, Orbs, Metal, Gooey, Image","summaryCn":"六款为 AI 打造的炫酷 React 特效库，让界面瞬间活过来。","reasonCn":"这个项目提供六款高度定制的 React 组件，专门为 AI 代理打造，让界面看起来充满生命力。它将 WebGL 的强大渲染能力封装成简单的组件，开发者只需几行代码就能为现有界面添加令人惊艳的动态效果。通过调整属性而非手写底层代码，就能实现如光束流动、液体融合等复杂视觉体验，极大提升了 AI 应用的交互质感。","stars":3629,"starGrowth":26,"createdAt":"2026-04-12T17:27:34Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"工具","funScore":0.5,"domainTags":["UI组件","TypeScript","前端框架","动画效果"],"domainKey":"UI组件","pushedAt":"2026-09-23T12:13:58Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/Jakubantalik/Libraries.dev","ts":"2026-09-23T12:13:58Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.11939222926426013,"copyOk":true},{"repo":"trycua/cua","owner":"trycua","name":"cua","desc":"Scale computer-use 2.0 with open-source drivers, cross-OS fleets, and benchmarks for training, evaluation, and data generation.","summaryCn":"给AI装上手脚的开源框架，让智能体跨系统操控电脑，真正替你干脏活累活","reasonCn":"trycua/cua是一个面向计算机使用（Computer Use）的开源框架，主打多操作系统下的AI智能体批量部署与基准测试。它提供标准化的驱动程序和评估协议，让大模型不仅能聊天，还能真正操控鼠标键盘完成复杂任务，是训练和评测AI Agent的关键基础设施。","stars":25981,"starGrowth":204,"createdAt":"2025-01-31T15:02:49Z","silentRounds":0,"language":"HTML","topics":["ai-agent"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI-Agent","Computer-Use","Benchmark","Automation"],"domainKey":"AI-Agent","pushedAt":"2026-09-23T20:36:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/trycua/cua","ts":"2026-09-23T20:36:40Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3235556753011816,"copyOk":false},{"repo":"CopilotKit/CopilotKit","owner":"CopilotKit","name":"CopilotKit","desc":"The Frontend Stack for Agents & Generative UI. React, Angular, Mobile, Slack, and more. Makers of the AG-UI Protocol","summaryCn":"给普通网站装上AI大脑，让用户和模型无缝对话的神器","reasonCn":"CopilotKit是一个专注于前端集成的AI Agent工具包，支持React、Angular和移动端。它通过创新的AG-UI协议，让开发者能轻松在现有应用中嵌入类似ChatGPT的对话界面。内置了工具调用和状态管理，降低了构建生成式UI的门槛。","stars":37507,"starGrowth":37,"createdAt":"2023-06-19T04:08:31Z","language":"TypeScript","topics":["ai-agent","llm","ai","agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI助手","前端框架","生成式UI"],"domainKey":"AI助手","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/CopilotKit/CopilotKit","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.11395944756978697,"copyOk":false},{"repo":"wazuh/wazuh","owner":"wazuh","name":"wazuh","desc":"Wazuh - The Open Source Security Platform. Unified XDR and SIEM protection for endpoints and cloud workloads.","summaryCn":"开源安全平台，集威胁防御、检测、响应于一体，支持跨环境部署。","reasonCn":"Wazuh 是一套免费开源的统一安全平台，用于威胁预防、检测和响应。它由部署在监控系统上的端点代理和管理服务器组成，代理收集数据，服务器分析数据。它能扫描恶意软件、rootkit 和可疑异常，常被用于入侵检测。平台还自带索引器，提供告警搜索与可视化仪表盘。对于需要跨本地、虚拟化、容器和云环境统一防护的团队来说很实用。","stars":16967,"starGrowth":9,"createdAt":"2015-08-11T17:39:03Z","silentRounds":0,"language":"C++","topics":["security"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["安全监控","SIEM","XDR","漏洞检测","合规审计"],"domainKey":"安全监控","pushedAt":"2026-09-23T21:41:46Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/wazuh/wazuh","ts":"2026-09-23T21:41:46Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.028591324335474746,"copyOk":true},{"repo":"grafana/mcp-grafana","owner":"grafana","name":"mcp-grafana","desc":"MCP server for Grafana","summaryCn":"在企业级监控环境中，常常需要部署多套 Grafana 实例来隔离业务线或地域，但","reasonCn":"在企业级监控环境中，常常需要部署多套 Grafana 实例来隔离业务线或地域，但每次新增或修改仪表盘、数据源时都要手动复制，既费时又易出错。mcp-grafana 通过一个中心化的 MCP（Metric Configuration Provider）服务器，把所有配置抽象成统一的模型，使用 HTTP/JSON 接口对外提供增","stars":3486,"starGrowth":5,"language":"Go","topics":[],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Grafana配置管理","多实例同步","Metric Configuration Provider","DevOps","Go"],"domainKey":"Grafana配置管理","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.46,"source":"trending","url":"https://github.com/grafana/mcp-grafana","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.019302925989672978,"copyOk":false},{"repo":"GraphiteEditor/Graphite","owner":"GraphiteEditor","name":"Graphite","desc":"Community-built comprehensive 2D content creation appplication for graphic design, digital art, and interactive real-time motion graphics powered by a node-based procedural graphics engine","summaryCn":"浏览器里拖节点画动画，零门槛玩转矢量设计与动效","reasonCn":"基于WebAssembly和Rust构建的节点式图形编辑器，无需安装即可在浏览器中运行。通过可视化编程实现矢量绘图与实时动效，适合非专业用户快速创作复杂动画，是开源界的Figma+After Effects平替，让图形设计变得直观有趣。","stars":27329,"starGrowth":9,"createdAt":"2020-04-26T07:29:29Z","silentRounds":0,"language":"Rust","topics":["design","creative-coding"],"aiDims":["创意工具","非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["节点编辑器","WebAssembly","矢量图形","实时动画","2D创作"],"domainKey":"节点编辑器","pushedAt":"2026-09-23T21:41:48Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"creative-coding","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/GraphiteEditor/Graphite","ts":"2026-09-23T21:41:48Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.02798796150609243,"copyOk":true},{"repo":"FareedKhan-dev/train-llm-from-scratch","owner":"FareedKhan-dev","name":"train-llm-from-scratch","desc":"A straightforward method for training your LLM, from downloading data to generating text.","summaryCn":"很多人用过大模型，但很少有人亲手从零写过一个。","reasonCn":"项目的核心亮点在于对Transformer架构的极致解构。它通过Python语言，将自注意力机制、位置编码、多层神经网络等核心模块逐一实现，让你能够清晰地看到数据是如何通过矩阵运算变成有意义的语言概率分布的。与市面上那些只会调用Hugging Face库的教程不同，这里提供的代码更具有教学","stars":10852,"starGrowth":141,"silentRounds":0,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","大语言模型","学习资源"],"zone":"资源","funScore":0,"domainTags":["LLM训练","Transformer","深度学习","教学代码"],"domainKey":"LLM训练","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/FareedKhan-dev/train-llm-from-scratch","ts":"2026-09-23T07:35:20.006Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.33897899004791743,"copyOk":false},{"repo":"HKUDS/DeepTutor","owner":"HKUDS","name":"DeepTutor","desc":"DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/.","summaryCn":"能记住你一生的AI家教，越聊越懂你，把知识变成你的长期记忆。","reasonCn":"DeepTutor 是一款主打终身个性化辅导的AI学习助手。它通过大语言模型和检索增强生成技术，构建持续进化的个人知识体系。与传统问答工具不同，它强调对话的连续性和记忆功能，能在长期交互中逐步理解学习者的薄弱点，提供量身定制的讲解、出题和复习计划。","stars":40177,"starGrowth":64,"createdAt":"2025-12-28T15:35:54Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI应用","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["智能教育","个性化辅导","知识图谱","RAG","终身学习"],"domainKey":"知识图谱","pushedAt":"2026-09-23T11:37:35Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/HKUDS/DeepTutor","ts":"2026-09-23T11:37:35Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.23613462428752768,"copyOk":true},{"repo":"lsdefine/GenericAgent","owner":"lsdefine","name":"GenericAgent","desc":"Self-evolving agent: grows skill tree from 3.3K-line seed, achieving full system control with 6x less token consumption","summaryCn":"你一定听说过AI Agent能帮你干活，但你知道怎么让Agent自己学会新技能吗？","reasonCn":"项目的核心创新在于它的自我进化机制。Agent初始只有一些基础工具，在执行任务过程中，如果某个子任务反复出现，Agent会自动将其抽象为新的技能并加入技能树。这个过程通过反思模块实现——Agent会分析自己的执行日志，识别可复用的模式，并生成新的工具定义。","stars":14234,"starGrowth":9,"createdAt":"2026-01-16T15:45:24Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":[],"zone":"AI","funScore":1,"domainTags":["AI Agent","自我进化","技能树","工具调用"],"domainKey":"AI Agent","pushedAt":"2026-09-22T15:42:24Z","tags":[{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/lsdefine/GenericAgent","ts":"2026-09-22T15:42:24Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03467623998875931,"copyOk":false},{"repo":"microsoft/PowerToys","owner":"microsoft","name":"PowerToys","desc":"Microsoft PowerToys is a collection of utilities that supercharge productivity and customization on Windows","summaryCn":"微软官方出的Windows增强小工具集，让系统更好用更好玩","reasonCn":"这是微软官方维护的开源工具集，包含FancyZones窗口管理、PowerRename批量重命名、PowerToys Run快速启动器等多个实用小工具。它能显著提升Windows用户的生产力，界面美观且功能强大，是Windows必备的效率神器，特别适合经常多任务处理的职场人和开发者。","stars":138915,"starGrowth":19,"createdAt":"2019-05-01T17:44:02Z","silentRounds":0,"language":"C","topics":["desktop"],"aiDims":["非AI-好玩","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Windows增强工具","窗口管理","批量重命名","快速启动器","系统工具"],"domainKey":"Windows增强工具","pushedAt":"2026-09-23T21:32:34Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/microsoft/PowerToys","ts":"2026-09-23T21:32:34Z","category":"tool","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.05741032285930245,"copyOk":true},{"repo":"arendst/Tasmota","owner":"arendst","name":"Tasmota","desc":"Alternative firmware for ESP8266 and ESP32 based devices with easy configuration using webUI, OTA updates, automation using timers or rules, expandability and entirely local control over MQTT, HTTP, Serial or KNX. Full documentation at","summaryCn":"给智能设备刷入的固件，支持多种协议本地控制。","reasonCn":"Tasmota 是专为 ESP8266 和 ESP32 开发的替代固件。它解决了原厂固件功能受限、配置繁琐的问题，通过网页界面即可轻松管理。项目支持 MQTT、HTTP、串口和 KNX 等多种通信协议，实现完全本地化的设备控制，非常适合想要 DIY 智能家居或改造旧设备的用户。","stars":24771,"starGrowth":2,"createdAt":"2017-01-28T13:37:08Z","silentRounds":0,"language":"C","topics":["automation","arduino"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["智能家居固件","ESP8266","MQTT自动化"],"domainKey":"智能家居固件","pushedAt":"2026-09-22T15:17:28Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"arduino","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/arendst/Tasmota","ts":"2026-09-22T15:17:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ultralytics/ultralytics","owner":"ultralytics","name":"ultralytics","desc":"Ultralytics YOLO26, YOLO11, YOLOv8 — object detection, instance segmentation, semantic segmentation, image classification, pose estimation, object tracking","summaryCn":"YOLO全家桶来了，一行代码搞定目标检测分割，谁用谁香","reasonCn":"Ultralytics推出的YOLO系列统一框架，涵盖YOLO11、YOLOv8等最新模型，支持目标检测、实例分割、语义分割、图像分类、姿态估计和目标跟踪六大任务。API设计极简，Python一行命令即可推理，同时提供CLI和SDK接口。预训练模型覆盖COCO等主流数据集，支持ONNX/TensorRT/CoreML等多格式导出，跨平台部署能力极强，从GPU服务器到手机端都能跑。社区活跃，文档完善，是计算机视觉领域最受欢迎的开源项目之一。","stars":61909,"starGrowth":40,"createdAt":"2022-09-11T16:39:45Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","模型与训练"],"zone":"AI","funScore":0.5,"domainTags":["目标检测","图像分割","姿态估计","模型部署","计算机视觉"],"domainKey":"模型部署","pushedAt":"2026-09-23T21:26:11Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/ultralytics/ultralytics","ts":"2026-09-23T21:26:11Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.12193832883748727,"copyOk":true},{"repo":"AstrBotDevs/AstrBot","owner":"AstrBotDevs","name":"AstrBot","desc":"AI Agent Assistant & development framework that integrates lots of IM platforms, LLMs, plugins and AI feature, and can be your openclaw alternative. ✨","summaryCn":"开源多模态聊天机器人平台，集成IM应用。","reasonCn":"AstrBot是一个开源的一体化智能体聊天机器人平台，旨在为个人、开发者和团队提供可靠的对话AI基础设施。它集成了主流即时通讯应用，支持大语言模型对话、多模态交互、Agent智能体、MCP工具调用和知识库等功能。用户可以快速在IM平台中构建生产级的AI应用，无论是打造个人AI助手还是企业知识库都能实现。","stars":40854,"starGrowth":53,"createdAt":"2022-12-08T13:27:46Z","silentRounds":0,"language":"Python","topics":["llm","ai","chatbot"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI助手","多平台集成","LLM框架"],"domainKey":"AI助手","pushedAt":"2026-09-23T08:09:53Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/AstrBotDevs/AstrBot","ts":"2026-09-23T08:09:53Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1628919077691291,"copyOk":true},{"repo":"OpenTTD/OpenTTD","owner":"OpenTTD","name":"OpenTTD","desc":"OpenTTD is an open source simulation game based upon Transport Tycoon Deluxe","summaryCn":"经典运输大亨开源重制版，建铁路运货物超好玩","reasonCn":"技术上，它用C++重写原游戏核心，渲染引擎支持硬件加速，界面可高度定制。游戏机制保留了经典玩法：铺设轨道、购买车辆、规划路线、管理公司财务。还支持多人联机，和朋友一起竞争或合作。Mod社区活跃，有各种地图、车辆、剧情扩展。","stars":8295,"starGrowth":7,"createdAt":"2018-04-06T18:06:22Z","silentRounds":0,"language":"C++","topics":["game","simulation"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["模拟经营","运输游戏","开源游戏","多人联机"],"domainKey":"模拟经营","pushedAt":"2026-09-23T21:51:16Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/OpenTTD/OpenTTD","ts":"2026-09-23T21:51:16Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.023531645569620254,"copyOk":false},{"repo":"Flow-Launcher/Flow.Launcher","owner":"Flow-Launcher","name":"Flow.Launcher","desc":":mag: Quick file search & app launcher for Windows with community-made plugins","summaryCn":"Windows上的极速启动神器，搜文件开应用一键搞定","reasonCn":"它的核心技术在于极速索引和插件架构。文件索引采用增量更新，实时响应文件系统变化。插件系统基于.NET开发，功能可扩展性强，已有大量社区插件支持浏览器搜索、翻译、剪贴板管理等。主题系统高度可定制，界面简洁美观。","stars":15613,"starGrowth":6,"createdAt":"2020-04-21T08:12:15Z","silentRounds":0,"language":"C#","topics":["desktop","productivity"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Windows启动器","文件搜索","插件系统","键盘驱动","效率工具"],"domainKey":"Windows启动器","pushedAt":"2026-09-23T21:48:55Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Flow-Launcher/Flow.Launcher","ts":"2026-09-23T21:48:55Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.019152885415999486,"copyOk":false},{"repo":"FunkinCrew/Funkin","owner":"FunkinCrew","name":"Funkin","desc":"A rhythm game made with HaxeFlixel","summaryCn":"一款致敬 Flash 社区文化的开源节奏游戏，在多个平台可玩。","reasonCn":"这是一个用 HaxeFlixel 开发的节奏游戏，最初为 Ludum Dare 47 制作。玩家需要跟随音乐节拍精准按键，与对手进行歌唱对战。游戏完全开源，支持模组（Mod），社区活跃，衍生出大量同人创作。你可以在网页、PC、移动端各大平台免费体验。","stars":3765,"starGrowth":2,"createdAt":"2020-10-03T06:50:43Z","silentRounds":0,"language":"Haxe","topics":["game"],"aiDims":["游戏"],"zone":"创意","funScore":1,"domainTags":["音乐游戏","节奏游戏","开源游戏","模组平台"],"domainKey":"音乐游戏","pushedAt":"2026-09-21T17:19:23Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Haxe","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/FunkinCrew/Funkin","ts":"2026-09-21T17:19:23Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"apify/apify-mcp-server","owner":"apify","name":"apify-mcp-server","desc":"The Apify MCP server enables your AI agents to extract data from social media, search engines, maps, e-commerce sites, or any other website using thousands of ready-made scrapers, crawlers, and automation tools available on the Apify Store.","summaryCn":"给AI装上数据抓取的万能插件，轻松搞定全网爬虫","reasonCn":"这是给AI Agent接入互联网能力的核心工具，基于MCP协议，允许大模型直接调用Apify平台上的海量爬虫库。它支持从社交媒体、电商平台、搜索引擎抓取结构化数据，无需编写代码。项目使用TypeScript开发，拥有3555个Star，是AI自主行动和数据处理的基础设施。","stars":8086,"starGrowth":114,"createdAt":"2025-01-02T15:43:09Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","工具调用","AI搜索"],"zone":"AI","funScore":0.6,"domainTags":["AI-Agent","Web-Scraping","MCP","Data-Extraction"],"domainKey":"AI-Agent","pushedAt":"2026-09-23T17:30:19Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/apify/apify-mcp-server","ts":"2026-09-23T17:30:19Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3422953252535246,"copyOk":true},{"repo":"code-yeongyu/oh-my-openagent","owner":"code-yeongyu","name":"oh-my-openagent","desc":"OmO: Just type \"mass ulw\" keyword with your prompt. Now you are the master of graph engineering.","summaryCn":"让AI当你的编程助理，自动帮你写代码改bug的开源神器","reasonCn":"这是一个面向复杂代码库的AI编程Agent，整合了Claude、Codex等多个主流AI模型的能力。它能自动理解项目结构、制定计划并执行任务，像一位经验丰富的程序员在帮你干活。相比传统代码补全工具，它更像能独立思考的助手。","stars":69328,"starGrowth":47,"createdAt":"2025-12-03T01:40:05Z","language":"TypeScript","topics":["ai"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","Agent框架","代码库理解","自动化编程"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/code-yeongyu/oh-my-openagent","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.17164057235171934,"copyOk":false},{"repo":"community-scripts/ProxmoxVE","owner":"community-scripts","name":"ProxmoxVE","desc":"Proxmox VE Helper-Scripts (Community Edition) ","summaryCn":"Proxmox VE 一键部署脚本集合，覆盖数百种自托管服务。","reasonCn":"这是一个为 Proxmox VE 打造的社区维护脚本库，解决了自托管服务安装配置繁琐的问题。用户只需在 Shell 中粘贴一条命令、回答几个提示，即可自动完成容器或虚拟机的创建与服务部署。项目涵盖家庭自动化、媒体服务器、网络工具、数据库、监控栈等数百种热门服务，基于 tteck 原始工作演进而来，持续由社区更新维护。","stars":29642,"starGrowth":7,"createdAt":"2024-11-01T14:06:43Z","silentRounds":0,"language":"Shell","topics":["security","self-hosted"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Proxmox","运维脚本","虚拟机备份","迁移工具","Shell"],"domainKey":"Proxmox","pushedAt":"2026-09-23T18:50:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/community-scripts/ProxmoxVE","ts":"2026-09-23T18:50:08Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.021708454220363,"copyOk":true},{"repo":"avelino/awesome-go","owner":"avelino","name":"awesome-go","desc":"A curated list of awesome Go frameworks, libraries and software","summaryCn":"Go语言开发者的宝藏清单，框架库工具一网打尽","reasonCn":"这是一个由社区维护的Go语言开源项目资源列表，收录了Go生态中各类高质量框架、库和工具，涵盖Web开发、数据处理、CLI工具、机器学习等多个领域。列表持续更新，是Go开发者快速定位合适依赖、了解生态全貌的最佳入口，省去在GitHub上盲目搜索的时间。","stars":185168,"starGrowth":122,"createdAt":"2014-07-06T13:42:15Z","silentRounds":0,"language":"Go","topics":["awesome"],"aiDims":["开发者工具","学习资源"],"zone":"资源","funScore":0,"domainTags":["Go语言生态","开发资源导航","框架库索引"],"domainKey":"Go语言生态","pushedAt":"2026-09-23T17:44:26Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/avelino/awesome-go","ts":"2026-09-23T17:44:26Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.301976583426942,"copyOk":false},{"repo":"a8hok/JavaScript-Study-Guide","owner":"a8hok","name":"JavaScript-Study-Guide","desc":"Javascript Study Guide","summaryCn":"JavaScript 学习指南，搞定面试题。","reasonCn":"内容组织得很有条理，通常先介绍概念，再配合代码示例和解释，最后给出面试题和答案。这种循序渐进的方式，非常适合自学。它还涵盖了浏览器原理、网络协议等宽泛知识，让你对前端技术栈有一个更宏观的认识。对于已经工作了一段时间的开发者来说，这也是一个很好的复盘和提升工具。","stars":39,"starGrowth":0,"createdAt":"2021-08-18T15:12:14Z","silentRounds":11,"language":"","topics":["interview"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["JavaScript","面试指南","学习资料","前端开发"],"domainKey":"JavaScript","pushedAt":"2026-09-23T00:43:42Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7}],"aiScore":0.5,"source":"search","url":"https://github.com/a8hok/JavaScript-Study-Guide","ts":"2026-09-23T00:43:42Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HKUDS/Vibe-Trading","owner":"HKUDS","name":"Vibe-Trading","desc":"\"Vibe-Trading: Your Personal Trading Agent\"","summaryCn":"AI帮你盯盘交易，用自然语言就能操作金融市场的智能代理","reasonCn":"这是一个基于大语言模型的金融交易智能代理，支持通过自然语言下达交易指令，自动连接加密货币和股票API执行买卖操作。它具备市场分析、风险评估和自动化交易能力，支持多平台接入，让普通用户无需写代码也能像专业交易员一样操作。","stars":33847,"starGrowth":59,"createdAt":"2026-04-01T09:52:20Z","silentRounds":0,"language":"Python","topics":["ai-agent","llm"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["AI交易代理","自然语言交易","金融Agent","加密货币交易","股票交易"],"domainKey":"AI交易代理","pushedAt":"2026-09-23T05:49:52Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/HKUDS/Vibe-Trading","ts":"2026-09-23T05:49:52Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.21867529766301297,"copyOk":true},{"repo":"D4Vinci/Scrapling","owner":"D4Vinci","name":"Scrapling","desc":"🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ","summaryCn":"自适应爬虫框架，轻松应对反爬虫与动态网页。","reasonCn":"Scrapling 是一个现代化的网页抓取框架，能处理从单次请求到大规模爬取的各种场景。它内置了解析器，能自动适应网站结构变化并重新定位元素；其抓取器可直接绕过 Cloudflare Turnstile 等反机器人系统。此外，它还提供代理轮换、并发爬取、智能限速等功能，帮助你在几行 Python 代码中实现高效的数据采集。","stars":82973,"starGrowth":219,"createdAt":"2024-10-13T20:29:53Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["工具调用","开发者工具","效率工具"],"zone":"工具","funScore":1,"domainTags":["网页抓取","视觉定位","数据采集","爬虫框架"],"domainKey":"网页抓取","pushedAt":"2026-09-23T11:15:33Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/D4Vinci/Scrapling","ts":"2026-09-23T11:15:33Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3079182384631145,"copyOk":true},{"repo":"d2lang/d2","owner":"d2lang","name":"d2","desc":"D2 is a modern diagram scripting language that turns text to diagrams.","summaryCn":"用文字描述就能自动生成专业图表，编程人员画流程图的神器","reasonCn":"技术上，D2是一种声明式图表语言，语法简洁直观。它支持流程图、序列图、ER图、方框图等多种类型。核心优势是版本可控——图表文件是纯文本，可以放入Git管理，随时回滚和Diff。渲染引擎支持多种后端，包括可执行的二进制文件和WebAssembly版本，输出质量媲美专业绘图工具。","stars":25488,"starGrowth":12,"createdAt":"2022-09-05T23:12:27Z","silentRounds":0,"language":"Go","topics":["developer-tools"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["图表","流程图","diagrams","文档工具","开发者工具"],"domainKey":"图表","pushedAt":"2026-09-20T16:30:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/d2lang/d2","ts":"2026-09-20T16:30:17Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03741242937853107,"copyOk":false},{"repo":"rqlite/rqlite","owner":"rqlite","name":"rqlite","desc":"The lightweight, fault-tolerant database built on SQLite. Designed to keep your data highly available with minimal effort.","summaryCn":"基于 SQLite 的轻量级分布式数据库，高可用不复杂","reasonCn":"技术上，rqlite 的核心是一个轻量的 Raft 实现，所有写操作都必须经过 leader 节点，然后复制到 follower 节点。读写分离可以通过配置实现。它提供完整的 RESTful API，你可以用 curl 直接操作数据库，支持 SQL 查询、备份、恢复等常用功能。","stars":17756,"starGrowth":5,"createdAt":"2014-08-23T04:31:18Z","silentRounds":0,"language":"Go","topics":["database"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","SQLite","Raft共识","高可用","边缘计算","REST API"],"domainKey":"分布式数据库","pushedAt":"2026-09-22T13:39:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/rqlite/rqlite","ts":"2026-09-22T13:39:26Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015844784861455283,"copyOk":true},{"repo":"Kkevsterrr/geneva","owner":"Kkevsterrr","name":"geneva","desc":"","summaryCn":"马里兰大学的AI反审查工具，利用漏洞单端突破访问限制","reasonCn":"Geneva是马里兰大学开发的AI研究项目，通过利用审查系统的bug，在无需Tor等外部协助的情况下，仅在客户端或服务器端即可实现反审查功能。它不同于常规VPN方案，专注于在单侧连接上运行，旨在为身处中国、印度等互联网限制严格地区的用户提供直达互联网的路径，是一款极具技术挑战性的网络自由工具。","stars":2142,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["非AI-好玩","安全工具"],"zone":"工具","funScore":0.5,"domainTags":["反审查","网络隐私","AI算法","网络协议"],"domainKey":"反审查","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"tier","url":"https://github.com/Kkevsterrr/geneva","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"truefoundry/trueforge","owner":"truefoundry","name":"trueforge","desc":"The open-source agent harness - the runtime layer that turns an LLM into a working agent.","summaryCn":"The open source agent harness - the runtime layer that turns an LLM into a workable agent.","reasonCn":"这个项目是一个开源的LLM代理工具，可以将LLM转换为可工作的代理。","stars":5908,"starGrowth":34,"createdAt":"2026-07-23T09:16:13Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["Agent"],"zone":"AI","funScore":0,"domainTags":["Agent框架","LLM","AI运行时"],"domainKey":"Agent框架","pushedAt":"2026-09-23T14:10:08Z","tags":[{"name":"Agent","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/truefoundry/trueforge","ts":"2026-09-23T14:10:08Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.17889708869329723,"copyOk":false},{"repo":"stablyai/orca","owner":"stablyai","name":"orca","desc":"Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.","summaryCn":"让多个 AI 编码助手并行干活并合并结果的指挥平台","reasonCn":"Orca 是一个专为 AI 编程构建的指挥中心，能把 Claude Code、Codex 等多个助手并排放在独立工作区里。它能自动监控任务进度，甚至在手机上也能随时发号施令。核心亮点是把原本孤立的 AI 工具变成了一支协同作战的团队，帮你对比不同模型的输出结果，省去了反复切换窗口的麻烦。安装方式请访问项目仓库获取，上手后直接把你的编码需求扔进工作区，Orca 就会自动调度不同助手并行生成代码。","stars":75972,"starGrowth":944,"silentRounds":0,"language":"TypeScript","topics":["claude-code","codex","ghostty","terminal","cli","cursor-agent","opencode","orchestration","worktrees","parallel-agents","pi","ade","ide","mobile-app","agent-ide","ai-agents","devtools","yc-backed"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI 代理","Agent 编排","ADE","多智能体"],"domainKey":"AI 代理","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"ghostty","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"cursor-agent","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"orchestration","source":"github","weight":0.7},{"name":"worktrees","source":"github","weight":0.7},{"name":"parallel-agents","source":"github","weight":0.7},{"name":"pi","source":"github","weight":0.7},{"name":"ade","source":"github","weight":0.7},{"name":"ide","source":"github","weight":0.7},{"name":"mobile-app","source":"github","weight":0.7},{"name":"agent-ide","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"devtools","source":"github","weight":0.7},{"name":"yc-backed","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/stablyai/orca","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.33727689148633705,"copyOk":true},{"repo":"microsoft/playwright","owner":"microsoft","name":"playwright","desc":"Playwright is a framework for Web Testing and Automation. It allows testing Chromium, Firefox and WebKit with a single API. ","summaryCn":"微软出品的浏览器自动化神器，写一次代码跑遍三大浏览器","reasonCn":"这是微软开发的Web测试和自动化框架，支持Chromium、Firefox和WebKit。它能模拟真实用户行为，做端到端测试、爬虫或自动化脚本。虽然不直接涉及AI，但它是构建AI应用前端测试和自动化工具链的基础设施，稳定且强大，开发者必备。","stars":96519,"starGrowth":46,"createdAt":"2019-11-15T18:32:42Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["浏览器自动化","Web测试","端到端测试"],"domainKey":"浏览器自动化","pushedAt":"2026-09-23T21:45:29Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/microsoft/playwright","ts":"2026-09-23T21:45:29Z","category":"tool","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.13942977030429243,"copyOk":true},{"repo":"OWASP/CheatSheetSeries","owner":"OWASP","name":"CheatSheetSeries","desc":"The OWASP Cheat Sheet Series was created to provide a concise collection of high value information on specific application security topics.","summaryCn":"网络安全从业者的救命速查手册，加密漏洞权限全搞定","reasonCn":"这个项目收集了大量应用安全主题的速查表，涵盖加密、身份认证、会话管理、服务端防护、客户端安全、日志监控等常见安全问题。每个Cheat Sheet都是精炼的实战指南，帮开发者快速查阅安全最佳实践和已知漏洞的应对方案，是安全开发人员的必备参考手册。","stars":33277,"starGrowth":8,"createdAt":"2018-12-21T14:26:43Z","silentRounds":0,"language":"Python","topics":["security"],"aiDims":["安全工具","学习资源","开发者工具"],"zone":"资源","funScore":0,"domainTags":["应用安全","速查手册","OWASP","加密","身份认证"],"domainKey":"应用安全","pushedAt":"2026-09-23T14:00:22Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/OWASP/CheatSheetSeries","ts":"2026-09-23T14:00:22Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024721218859873186,"copyOk":true},{"repo":"electricitymaps/electricitymaps-contrib","owner":"electricitymaps","name":"electricitymaps-contrib","desc":"The open source repository for Electricity Maps data parsers that powers the world's most comprehensive electricity data platform","summaryCn":"能查各国实时清洁电力的酷炫工具，环保人士必备","reasonCn":"这个项目是Electricity Maps平台的开源数据解析核心，聚合了全球电网实时排放数据。通过可视化图表，让你清晰看到家里电流是煤电还是风电。技术栈以Python为主，适合关注可持续发展、能源数据或需要绿色指标参考的开发者与环保爱好者。","stars":4034,"starGrowth":1,"createdAt":"2016-05-21T16:36:17Z","silentRounds":8,"language":"Python","topics":["data-visualization"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["电力数据可视化","电网排放","实时数据"],"domainKey":"电力数据可视化","pushedAt":"2026-09-21T15:48:48Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/electricitymaps/electricitymaps-contrib","ts":"2026-09-21T15:48:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"omnigent-ai/omnigent","owner":"omnigent-ai","name":"omnigent","desc":"Omnigent is an open-source AI agent framework and meta-harness: orchestrate Claude Code, Codex, Cursor, Pi, and custom agents — swap harnesses without rewriting, enforce policies and sandboxing, and collaborate in real time from any device.","summaryCn":"统一调度多种AI编码助手的开源枢纽，会话跨设备无缝同步。","reasonCn":"如果你在用Claude Code、Codex或Cursor等多个AI编程助手，这个项目能帮你统一管理它们。它提供了一个通用编排层，让你能自由切换或组合这些工具，还支持策略和沙箱控制。最方便的是，会话状态能跨设备同步，从手机也能接着干活。","stars":10164,"starGrowth":33,"createdAt":"2026-06-11T12:18:13Z","silentRounds":0,"language":"Python","topics":["ai-agent","agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI-agent","agent-框架","Claude","Cursor","多模型协作"],"domainKey":"AI-agent","pushedAt":"2026-09-23T21:45:49Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/omnigent-ai/omnigent","ts":"2026-09-23T21:45:49Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.13048831168831168,"copyOk":true},{"repo":"yikart/AiToEarn","owner":"yikart","name":"AiToEarn","desc":"Let's use AI to Earn!","summaryCn":"AI赚钱工具合集，教你用各种AI技能白嫖生活费","reasonCn":"基于TypeScript开发的AI变现工具集合，整合多种赚钱场景：自动化任务执行、内容创作批量生成、代码辅助接单等。支持OpenAI、Claude、Gemini及本地模型接入，提供可复用的Agent模板和教程，覆盖自由职业、副业增收等多种路径，适合想尝试AI变现的人群。","stars":26349,"starGrowth":68,"silentRounds":0,"language":"TypeScript","topics":["auto-publish","douyin","electron-app","electron-react","kuaishou","tool","xiaohongshu","kwai","published","shipinhao","douyin-api"],"aiDims":["AI应用","AI Agent","非AI-好玩"],"zone":"AI","funScore":0.5,"domainTags":["AI变现","Agent模板","内容创作","API调用","自动化任务"],"domainKey":"AI变现","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"auto-publish","source":"github","weight":0.7},{"name":"douyin","source":"github","weight":0.7},{"name":"electron-app","source":"github","weight":0.7},{"name":"electron-react","source":"github","weight":0.7},{"name":"kuaishou","source":"github","weight":0.7},{"name":"tool","source":"github","weight":0.7},{"name":"xiaohongshu","source":"github","weight":0.7},{"name":"kwai","source":"github","weight":0.7},{"name":"published","source":"github","weight":0.7},{"name":"shipinhao","source":"github","weight":0.7},{"name":"douyin-api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/yikart/AiToEarn","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.21174222930661507,"copyOk":true},{"repo":"AprilNEA/OpenLogi","owner":"AprilNEA","name":"OpenLogi","desc":"⚡️A native, local-first alternative to Logitech Options+, written in Rust 🦀 — remap buttons, DPI, and SmartShift over HID++. No account, no telemetry.","summaryCn":"给鼠标按键装上“瑞士军刀”，完全本地化操作的神器","reasonCn":"这是一个用 Rust 编写的开源替代方案，专为 Logitech Options+ 打造。它能在本地运行，无需联网或账号，支持 HID++ 协议，能精细控制鼠标的按键映射、DPI 调整和 SmartShift 智能滚轮功能。项目性能出色，Star 数超过 8600。","stars":21948,"starGrowth":80,"createdAt":"2026-05-24T10:10:36Z","silentRounds":0,"language":"Rust","topics":["privacy"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["鼠标驱动","HID++","隐私","Rust"],"domainKey":"鼠标驱动","pushedAt":"2026-09-23T15:24:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/AprilNEA/OpenLogi","ts":"2026-09-23T15:24:16Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.30112192454893383,"copyOk":false},{"repo":"GoogleContainerTools/skaffold","owner":"GoogleContainerTools","name":"skaffold","desc":"Easy and Repeatable Kubernetes Development","summaryCn":"Google出品的K8s开发利器，代码改了自动重新部署","reasonCn":"Skaffold 是 Google 开发的 Kubernetes 开发工具，简化本地到集群的部署流程。核心能力是开发循环自动化：代码修改后自动构建镜像、更新部署、同步文件，甚至自动重启 Pod。支持多种构建后端，包括 Docker、Jib、Kaniko 等，适配不同团队的技术栈。\n","stars":15892,"starGrowth":2,"createdAt":"2018-01-23T18:51:29Z","silentRounds":0,"language":"Go","topics":["developer-tools"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","开发工具","CI/CD"],"domainKey":"Kubernetes","pushedAt":"2026-09-21T12:44:18Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/GoogleContainerTools/skaffold","ts":"2026-09-21T12:44:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dgraph-io/dgraph","owner":"dgraph-io","name":"dgraph","desc":"high-performance graph database for real-time use cases","summaryCn":"高性能分布式图数据库，专为 GraphQL 设计。","reasonCn":"Dgraph 是一款基于 Go 语言开发的分布式图数据库，旨在为海量结构化数据提供高性能的查询服务。它支持 ACID 事务、一致复制和线性化读取，能够处理 PB 级数据。作为原生 GraphQL 数据库，它在底层存储上做了深度优化，能显著减少磁盘寻道和网络调用，适合对数据关系查询要求高的应用场景。","stars":21800,"starGrowth":1,"createdAt":"2015-08-25T07:15:56Z","silentRounds":4,"language":"Go","topics":["database"],"aiDims":[],"zone":"工具","funScore":0,"domainTags":["图数据库","分布式数据库","GraphQL","向量搜索"],"domainKey":"图数据库","pushedAt":"2026-09-22T18:29:14Z","tags":[{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/dgraph-io/dgraph","ts":"2026-09-22T18:29:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bytedance/deer-flow","owner":"bytedance","name":"deer-flow","desc":"An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of tasks that could take minutes to hours.","summaryCn":"字节开源的超级智能体，能自主调研、写代码、搞创作，让AI真正动手干活","reasonCn":"这是字节跳动开源的超级智能体框架，能自主完成长周期复杂任务。它内置沙箱隔离、持久化记忆、技能模块和子智能体协同，支持多轮工具调用和消息网关，配合主流大模型可像人一样拆任务、调API、写代码、查资料，适合搭建自动化工作流和AI应用开发。","stars":82862,"starGrowth":46,"createdAt":"2025-05-07T02:50:19Z","silentRounds":0,"language":"Python","topics":["llm","ai"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["Agent框架","自动化","任务规划"],"domainKey":"Agent框架","pushedAt":"2026-09-23T15:26:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/bytedance/deer-flow","ts":"2026-09-23T15:26:40Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.13966541961333304,"copyOk":false},{"repo":"laoma2053/awesome-zhuiju-free","owner":"laoma2053","name":"awesome-zhuiju-free","desc":"免费无广告的追剧资源指南，人工精选资源、每天检测资源有效性。收录在线影视、影视APP、网盘搜索、磁力BT、字幕、TVBox / 影视仓空壳软件/配置地址、IPTV直播源、会员拼团、影视相关开源项目。开源，社区共同维护。","summaryCn":"在这个会员满天飞、广告到处弹的时代，想找个能安心看剧的地方太难了。","reasonCn":"它的内容非常全面，几乎覆盖了所有主流的观影方式。你可以在这里找到各种影视APP的安装包和教程，也能找到各大网盘的搜索链接。对于喜欢用电视盒子看剧的朋友，它还提供了丰富的TVBox和影视仓的空壳软件及配置地址，插上就能用。","stars":10079,"starGrowth":367,"language":"JavaScript","topics":["awesome-list","chinese","free-streaming","iptv","magnet-search","media-player","movie-guide","subtitles","tvbox","bt-search","cloud-drive-search","movie-resources","free","movie","no-ads","tvbox-config"],"aiDims":["非AI-实用","非AI-好玩","非AI-工具"],"zone":"资源","funScore":0,"domainTags":["影视资源","追剧指南","资源聚合"],"domainKey":"影视资源","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-工具","source":"llm","weight":1},{"name":"awesome-list","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"free-streaming","source":"github","weight":0.7},{"name":"iptv","source":"github","weight":0.7},{"name":"magnet-search","source":"github","weight":0.7},{"name":"media-player","source":"github","weight":0.7},{"name":"movie-guide","source":"github","weight":0.7},{"name":"subtitles","source":"github","weight":0.7},{"name":"tvbox","source":"github","weight":0.7},{"name":"bt-search","source":"github","weight":0.7},{"name":"cloud-drive-search","source":"github","weight":0.7},{"name":"movie-resources","source":"github","weight":0.7},{"name":"free","source":"github","weight":0.7},{"name":"movie","source":"github","weight":0.7},{"name":"no-ads","source":"github","weight":0.7},{"name":"tvbox-config","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/laoma2053/awesome-zhuiju-free","ts":"2026-09-23T21:53:43.262Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.4092370274828852,"copyOk":false},{"repo":"surelle-ha/zakape","owner":"surelle-ha","name":"zakape","desc":"An open-source pixel-art and sprite-animation workbench for the desktop. It keeps the dependable parts of a traditional sprite editor—precise tools, layers, frames, onion skinning, palettes, and production exports—while adding an optional bring-your-own-model assistant that proposes reviewable pixel operations.","summaryCn":"桌面像素艺术工作站，层层帧帧，轻松做动画","reasonCn":"zakape 是一款桌面像素艺术与动画编辑器，采用 TypeScript 与 Electron，提供精确工具、分层、时间轴、洋葱皮、调色板等全套功能。其导入导出支持 PNG、GIF、Sprite Sheet、JSON、Lottie，兼容多平台。用户可通过拖拽方式快速绘制帧，实时预览动画，适合 indie 游戏、美术与教育项目。","stars":46,"starGrowth":0,"createdAt":"2026-08-31T15:19:56Z","silentRounds":4,"language":"TypeScript","topics":["pixel-art"],"aiDims":["创意工具","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["像素艺术","精灵动画","游戏美术","桌面应用","Electron"],"domainKey":"像素艺术","pushedAt":"2026-09-21T01:17:16Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.57,"source":"search","url":"https://github.com/surelle-ha/zakape","ts":"2026-09-21T01:17:16Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/fluentui-system-icons","owner":"microsoft","name":"fluentui-system-icons","desc":"Fluent System Icons are a collection of familiar, friendly and modern icons from Microsoft.","summaryCn":"微软出品的万级现代图标库，设计界的一站式宝藏","reasonCn":"Fluent System Icons是微软官方开源的图标集合，涵盖超过10000个精心设计的SVG图标，风格现代、友好、统一。这套图标遵循微软Fluent Design Design System规范，支持light/dark主题，适配各类分辨率。开发者可通过npm、CDN或直接下载SVG文件接入项目，兼容性极好，跨平台使用无压力。","stars":10858,"starGrowth":5,"createdAt":"2020-05-12T21:05:57Z","silentRounds":0,"language":"HTML","topics":["design"],"aiDims":["非AI-实用"],"zone":"资源","funScore":0,"domainTags":["图标库","SVG","Fluent Design","UI素材","设计系统"],"domainKey":"图标库","pushedAt":"2026-09-23T13:31:18Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/microsoft/fluentui-system-icons","ts":"2026-09-23T13:31:18Z","category":"learning","momentum":["daily"],"fromOfficial":true,"heatScore":0.01638146988395653,"copyOk":true},{"repo":"home-assistant/frontend","owner":"home-assistant","name":"frontend","desc":":lollipop: Frontend for Home Assistant","summaryCn":"把家里所有智能设备装进口袋，手机就是中控大屏","reasonCn":"这是智能家居开源平台Home Assistant的前端界面，支持手机App和网页操作，能统一管理灯光、空调、窗帘等上百种智能设备。界面美观流畅，支持自定义面板和自动化场景。项目完全开源免费，数据本地存储不用担心隐私泄露。适合有智能家居需求又担心隐私的用户，以及想深度定制控制界面的极客。","stars":5669,"starGrowth":3,"createdAt":"2015-07-12T23:58:06Z","silentRounds":0,"language":"TypeScript","topics":["frontend","home-automation"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["智能家居前端","本地控制面板","设备仪表盘","自动化规则","Home Assistant"],"domainKey":"智能家居前端","pushedAt":"2026-09-23T16:59:36Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"home-automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/home-assistant/frontend","ts":"2026-09-23T16:59:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"PaddlePaddle/PaddleOCR","owner":"PaddlePaddle","name":"PaddleOCR","desc":"Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.","summaryCn":"把图片和PDF里的文字一键提取出来，让AI能读懂你的文档","reasonCn":"百度开源的OCR工具，支持100多种语言文本识别，图片、PDF、扫描件都能处理。内置多种预训练模型，识别精度高。可直接接入LangChain框架，为RAG系统提供结构化数据，让大模型真正读懂你的文档内容，无需手写解析逻辑。","stars":90097,"starGrowth":76,"createdAt":"2020-05-08T10:38:16Z","language":"Python","topics":["rag"],"aiDims":["文档理解","AI应用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["OCR","文字识别","中文处理","RAG","文档数字化"],"domainKey":"RAG","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/PaddlePaddle/PaddleOCR","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.23053060590252725,"copyOk":true},{"repo":"jingyaogong/minimind","owner":"jingyaogong","name":"minimind","desc":"🧠 Train a 64M-parameter LLM from scratch in just 2h!","summaryCn":"仅需3块钱和2小时，从零训练出超小语言模型。","reasonCn":"MiniMind 是一个极简但完整的 LLM 训练复现项目，旨在让个人也能低成本训练大模型。它覆盖了从数据清洗、预训练、SFT 到 LoRA、RLHF 等全流程，主线模型体积仅是 GPT-3 的 1/2700。项目完全使用 PyTorch 原生实现，无第三方高层依赖，不仅开源了算法，更提供了可复现的教程，让开发者能快速上手体验模型训练全过程。","stars":62186,"starGrowth":153,"createdAt":"2024-07-27T04:34:12Z","silentRounds":0,"language":"Python","topics":["llm-model"],"aiDims":["模型与训练","大语言模型","教程"],"zone":"AI","funScore":0,"domainTags":["LLM训练","Transformer","教程","代码示例","参数优化"],"domainKey":"LLM训练","pushedAt":"2026-09-22T16:07:34Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/jingyaogong/minimind","ts":"2026-09-22T16:07:34Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3073810825587753,"copyOk":true},{"repo":"fission/fission","owner":"fission","name":"fission","desc":"Fast and Simple Serverless Functions for Kubernetes","summaryCn":"K8s上的免运维函数引擎，代码上传就跑，比Serverless还快","reasonCn":"Fission是运行在Kubernetes上的Serverless函数平台，支持Go、Python、Node.js等多种语言。它将你的代码打包成函数，部署到K8s集群后按需自动扩缩容，真正做到了代码即服务。相比传统Serverless方案，Fission完全开源且自托管，不受云厂商锁定。集成istio实现毫秒级冷启动，配合自定义触发器支持HTTP、事件、定时等多种调用方式，让K8s上的函数计算变得简单高效。","stars":8924,"starGrowth":1,"createdAt":"2016-08-19T03:43:56Z","silentRounds":2,"language":"Go","topics":["devops","kubernetes"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Serverless","Kubernetes","函数计算","冷启动优化","自托管"],"domainKey":"Serverless","pushedAt":"2026-09-22T01:59:19Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/fission/fission","ts":"2026-09-22T01:59:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"reflex-dev/reflex","owner":"reflex-dev","name":"reflex","desc":"🕸️ Web apps in pure Python 🐍","summaryCn":"Pythoner专属！纯Python搭Web应用，前端代码自动帮你写","reasonCn":"Reflex是一个用纯Python开发Web应用的开源框架，底层基于React和Pydantic。开发者用Python定义组件的状态和样式，框架自动编译成前端代码。支持流式响应、实时通信、状态管理等特性，一条命令即可部署到云端或本地。后端工程师不用碰HTML/CSS/JS就能构建完整的Web应用，大幅降低Web开发门槛。","stars":28901,"starGrowth":0,"createdAt":"2022-10-25T03:08:48Z","silentRounds":1,"language":"Python","topics":["developer-tools","data-visualization","data-science"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Web框架","Python全栈","React编译","内部工具开发"],"domainKey":"Web框架","pushedAt":"2026-09-23T21:50:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/reflex-dev/reflex","ts":"2026-09-23T21:50:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xDE57/SpaceProject","owner":"0xDE57","name":"SpaceProject","desc":"A 2D arcadey physics-based procedurally generated asteroid miner sandbox game using libGDX.","summaryCn":"用物理引擎和程序自动生成打造的太空采矿沙盒游戏，复古好玩。 (29字)","reasonCn":"技术上，这个项目完全基于成熟的Java游戏开发框架libGDX打造跨平台能力。它最大的亮点在于使用程序化生成技术，也就是说每一次开局，宇宙中的小行星带、矿藏分布和空间地图都是完全随机生成的，绝不重样。物理引擎精确模拟了飞船的惯性、重力和碰撞反弹，开船手感非常硬核，稍不留","stars":52,"starGrowth":0,"createdAt":"2015-07-24T23:05:04Z","silentRounds":15,"language":"Java","topics":["procedural-generation"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["太空沙盒","程序化生成","物理引擎","libGDX"],"domainKey":"太空沙盒","pushedAt":"2026-09-21T08:17:18Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/0xDE57/SpaceProject","ts":"2026-09-21T08:17:18Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pacifio/atlas","owner":"pacifio","name":"atlas","desc":"Source control for agents. Use multiple coding agents, track their changes and query them in one place","summaryCn":"像Git一样管理AI智能体的代码仓库，一站式追踪多人协作。","reasonCn":"这是一个专为多智能体协作设计的源代码管理系统，支持追踪多个AI代理的代码变更历史。通过统一的查询接口，开发者可以像查看Git日志一样查看Agent的交互记录和修改细节，有效解决了多Agent并行开发时的版本混乱与状态追踪难题。","stars":6028,"starGrowth":259,"createdAt":"2026-05-14T16:48:52Z","silentRounds":0,"language":"Rust","topics":["git"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent协作","源代码版本控制","Agent变更追踪","多智能体管理","开发工具"],"domainKey":"AI Agent协作","pushedAt":"2026-09-23T08:26:37Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/pacifio/atlas","ts":"2026-09-23T08:26:37Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.5146781685467816,"copyOk":true},{"repo":"affaan-m/ECC","owner":"affaan-m","name":"ECC","desc":"The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.","summaryCn":"让 AI 编程助手跑得更快更稳的加速神器","reasonCn":"ECC 是一个专为 Claude Code、Cursor 等主流 AI 编程助手打造的性能优化框架，支持技能扩展、记忆管理和安全控制。它采用模块化架构，让智能体在处理复杂任务时调用工具更高效、记忆更精准、交互更稳定，适合高频使用 AI 编程工具的开发者和效率派用户。","stars":265655,"starGrowth":727,"createdAt":"2026-01-18T00:51:51Z","silentRounds":0,"language":"JavaScript","topics":["llm"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","性能优化","Claude Code"],"domainKey":"AI Agent","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/affaan-m/ECC","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3698518755528787,"copyOk":false},{"repo":"nhost/nhost","owner":"nhost","name":"nhost","desc":"The Open Source Firebase Alternative with GraphQL.","summaryCn":"开源版Firebase，配好GraphQL后端，全栈开发直接起飞","reasonCn":"Nhost 是一款开源的 Firebase 替代方案，核心优势在于原生集成 GraphQL 和 PostgreSQL 数据库。开发者无需配置复杂的后端服务，就能快速构建实时应用。它支持认证、存储和服务器less 函数，特别适合希望掌控数据主权且熟悉 GraphQL 的技术团队，能显著缩短从开发到部署的周期。","stars":9297,"starGrowth":2,"createdAt":"2021-02-09T13:33:34Z","silentRounds":0,"language":"TypeScript","topics":["backend","database"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["开源Firebase替代","GraphQL后端","PostgreSQL","实时应用","自托管"],"domainKey":"开源Firebase替代","pushedAt":"2026-09-23T16:37:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/nhost/nhost","ts":"2026-09-23T16:37:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rbouteiller/airplay-esp32","owner":"rbouteiller","name":"airplay-esp32","desc":"This firmware turns a cheap ESP32 board into a wireless AirPlay 2 speaker. Plug it into an amplifier or powered speakers and it shows up on your iPhone, iPad or Mac just like a HomePod or an AirPlay TV.","summaryCn":"十几块的ESP32，让旧音箱秒变AirPlay智能音箱","reasonCn":"用一块成本仅十几元的ESP32开发板，就能把你的普通音箱升级成支持AirPlay 2的无线音箱。插上功放或音箱，iPhone、iPad、Mac就能直接投送音频。固件开源免费，硬件成本极低，技术门槛友好，玩家和音频爱好者都能轻松上手折腾。","stars":551,"starGrowth":4,"createdAt":"2026-01-22T21:56:56Z","silentRounds":0,"language":"C","topics":["esp32"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0.7,"domainTags":["ESP32固件","AirPlay 2","DIY音箱","嵌入式音频","智能家居"],"domainKey":"ESP32固件","pushedAt":"2026-09-20T22:11:08Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"esp32","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/rbouteiller/airplay-esp32","ts":"2026-09-20T22:11:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mayswind/ezbookkeeping","owner":"mayswind","name":"ezbookkeeping","desc":"A lightweight, self-hosted personal finance app with a user-friendly interface and powerful bookkeeping features.","summaryCn":"轻量级记账神器，自建账本随时掌控个人收支","reasonCn":"ezbookkeeping 是一款轻量级的个人财务管理应用，支持自建托管。界面友好，功能强大，能帮你清晰记录每一笔收支，查看统计报表。虽然不涉及 AI，但作为实用的生活工具，非常适合想要掌控财务、喜欢折腾自建服务的用户，特别是移动端体验很好。","stars":5637,"starGrowth":3,"createdAt":"2020-10-17T08:54:02Z","silentRounds":0,"language":"Go","topics":["mobile","finance"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["个人记账","自建财务","隐私账本","多账户管理","Docker部署"],"domainKey":"个人记账","pushedAt":"2026-09-23T14:43:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"finance","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/mayswind/ezbookkeeping","ts":"2026-09-23T14:43:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HBAI-Ltd/Toonflow-app","owner":"HBAI-Ltd","name":"Toonflow-app","desc":"Toonflow 是开源一站式 AI 短剧创作工具，将小说、剧本快速转化为动画短剧。集成 AI 编剧、智能分镜、角色与视频生成，跨平台桌面端轻量部署，助力创作者低成本批量产出视觉内容。Toonflow is an open-source AI tool that turns stories and scripts into animated short dramas. Features AI scriptwriting, storyboarding, character and video generation. A cross-platform desktop app for efficient content creation.","summaryCn":"开源AI短剧神器，一键将小说剧本变成动画视频，创作者福音","reasonCn":"Toonflow是一款开源的一站式AI短剧创作工具，集成AI编剧、智能分镜、角色一致性与视频生成。它将小说或剧本快速转化为动画短剧，极大降低了视频创作门槛。对于内容创作者、自媒体人和AI爱好者来说，这是一个既好玩又极具实用价值的创意工具。","stars":15909,"starGrowth":36,"createdAt":"2026-01-29T10:25:45Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["AI应用","多模态Agent","创意工具"],"zone":"创意","funScore":0.8,"domainTags":["AI视频生成","动画创作","剧本可视化","内容创作","角色一致性"],"domainKey":"AI视频生成","pushedAt":"2026-09-23T11:01:39Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.98,"source":"search","url":"https://github.com/HBAI-Ltd/Toonflow-app","ts":"2026-09-23T11:01:39Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.1377463322647558,"copyOk":false},{"repo":"sharkdp/fd","owner":"sharkdp","name":"fd","desc":"A simple, fast and user-friendly alternative to 'find'","summaryCn":"Rust 编写的现代化文件搜索工具，比 find 更直观、更快。","reasonCn":"fd 是一个用 Rust 编写的命令行工具，旨在替代传统的 find 命令。它通过使用更直观的语法和默认开启的并行搜索，显著提升了文件查找的效率。同时，它支持正则表达式和通配符，并能智能处理文件名的大小写和隐藏文件，让用户能更轻松地定位目标文件。","stars":44517,"starGrowth":10,"createdAt":"2017-05-09T21:27:10Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["文件搜索","命令行工具","Rust开发","性能优化"],"domainKey":"文件搜索","pushedAt":"2026-09-20T06:50:44Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/sharkdp/fd","ts":"2026-09-20T06:50:44Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030673899858481027,"copyOk":true},{"repo":"superset-sh/superset","owner":"superset-sh","name":"superset","desc":"Code Editor for the AI Agents Era - Run an army of Claude Code, Codex, etc. on your machine","summaryCn":"让你的电脑同时指挥多个AI编程助手协同干活的开发利器","reasonCn":"这是一个面向AI Agent时代的本地代码编辑器，支持在同一台机器上编排运行多个AI编程助手如Claude Code、OpenAI Codex等。它像是一个AI调度中枢，让你把不同专长的大模型分配给不同任务，实现代码编写、审查、优化的并行协作。","stars":14504,"starGrowth":73,"createdAt":"2025-10-21T09:16:51Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"工具","funScore":0,"domainTags":["AI编程","Agent编排","代码编辑器","多模型调度","开发工具"],"domainKey":"AI编程","pushedAt":"2026-09-23T20:12:10Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/superset-sh/superset","ts":"2026-09-23T20:12:10Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.28091913954771097,"copyOk":false},{"repo":"puppeteer/puppeteer","owner":"puppeteer","name":"puppeteer","desc":"JavaScript API for Chrome and Firefox","summaryCn":"自动化操控浏览器，爬虫测试网页抓数据全靠它","reasonCn":"Google官方维护，用JavaScript控制Chrome或Firefox，能自动填表、截图、爬取动态网页内容。API设计友好，支持无头模式运行，适合批量抓取数据和自动化测试，上手门槛低，社区文档丰富，是前端和测试工程师的必备神器。","stars":95610,"starGrowth":3,"createdAt":"2017-05-09T22:16:13Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools","automation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["浏览器自动化","无头浏览器","网页爬虫","端到端测试","Chrome DevTools Protocol"],"domainKey":"浏览器自动化","pushedAt":"2026-09-23T12:55:45Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/puppeteer/puppeteer","ts":"2026-09-23T12:55:45Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LingASDJ/Magic_Ling_Pixel_Dungeon","owner":"LingASDJ","name":"Magic_Ling_Pixel_Dungeon","desc":"The Magic Ling Pixel Dungeon began on February 12, 2021. Based on SHPD. (5 Years Old)","summaryCn":"历时五载精心打磨而成的硬核复古风像素地下城冒险游戏","reasonCn":"技术实现上，它完全采用Java语言开发，继承了其前辈作品优秀且成熟的框架体系。在长达数年的演进过程中，作者不仅修复了无数底层Bug，还加入了大量原创的职业分支、法术系统、怪物机制以及特殊的关卡事件。代码结构清晰严谨，对于想要学习如何用Java开发独立肉鸽游戏的程序员来说，","stars":86,"starGrowth":0,"createdAt":"2022-06-06T02:18:31Z","silentRounds":18,"language":"Java","topics":["roguelike"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["Roguelike","像素地牢","Java游戏","单机游戏"],"domainKey":"Roguelike","pushedAt":"2026-09-23T18:11:54Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"roguelike","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/LingASDJ/Magic_Ling_Pixel_Dungeon","ts":"2026-09-23T18:11:54Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"VoltAgent/awesome-agent-skills","owner":"VoltAgent","name":"awesome-agent-skills","desc":"A curated collection of 1000+ agent skills from official dev teams and the community, compatible with Claude Code, Codex, Gemini CLI, Cursor, and more.","summaryCn":"让 AI 智能体瞬间变强的超级技能指令大全","reasonCn":"这是一个精心搜集的 AI Agent 技能库，包含超过 1000 个高质量的技能定义，兼容 Claude Code 和 Cursor 等主流工具。通过结构化的 Prompt 定义，赋予 Agent 处理复杂开发任务、自动化运维及专业分析的能力，极大提升了工具调用精度。","stars":34745,"starGrowth":44,"createdAt":"2025-10-28T08:55:32Z","silentRounds":0,"language":"","topics":["awesome"],"aiDims":["AI Agent","提示工程","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["Agent技能","提示工程","Claude Code","Cursor","Codex"],"domainKey":"提示工程","pushedAt":"2026-09-23T07:18:28Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.95,"source":"search","url":"https://github.com/VoltAgent/awesome-agent-skills","ts":"2026-09-23T07:18:28Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.16295892934235143,"copyOk":false},{"repo":"AlexsJones/llmfit","owner":"AlexsJones","name":"llmfit","desc":"Hundreds of models & providers. One command to find what runs on your hardware.","summaryCn":"一键搞定模型适配，你的硬件能跑什么就装什么。","reasonCn":"llmfit 是一个 Rust 编写的命令行工具，旨在帮助用户在本地快速找到并运行适合自己硬件配置的 AI 模型。它内置了数百个模型和多个推理后端，通过简单的命令即可自动检测硬件能力，下载并部署合适的模型，极大地降低了大模型的使用门槛。","stars":37086,"starGrowth":95,"createdAt":"2026-02-15T16:11:31Z","language":"Rust","topics":["llm"],"aiDims":["大语言模型","模型与训练","效率工具"],"zone":"AI","funScore":0,"domainTags":["大模型本地部署","硬件适配","模型推荐","推理后端管理"],"domainKey":"大模型本地部署","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/AlexsJones/llmfit","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3512218087688076,"copyOk":false},{"repo":"micro-editor/micro","owner":"micro-editor","name":"micro","desc":"A modern and intuitive terminal-based text editor","summaryCn":"终端里的VS Code，不用鼠标也能流畅写代码，新手零门槛上手","reasonCn":"micro是一款基于终端的文本编辑器，用Go语言开发。它支持语法高亮、代码折叠、多标签页和鼠标操作。与vim/nano相比，micro学习成本极低，快捷键逻辑直观。适合想在终端高效编辑代码或配置文件，又不想深入钻研vim的用户群体。","stars":29622,"starGrowth":9,"createdAt":"2016-03-11T02:06:28Z","silentRounds":0,"language":"Go","topics":["command-line","editor"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["终端文本编辑器","语法高亮","跨平台","零配置上手","插件系统"],"domainKey":"终端文本编辑器","pushedAt":"2026-09-23T00:33:12Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"editor","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/micro-editor/micro","ts":"2026-09-23T00:33:12Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.027911484707312133,"copyOk":false},{"repo":"microsoft/AI","owner":"microsoft","name":"AI","desc":"","summaryCn":"项目中不仅明确了AI与人类智能的区别，还详尽列举了诸如自动驾驶、推荐系统","reasonCn":"该项目的核心亮点在于其宏观且结构化的知识编排，它将人工智能描述为一种能进行各种映射任务的计算工具。对于初学者而言，这份文档提供了一个清晰的视角，即理解AI并不一定非要从复杂的模型参数开始，而是可以从它所能解决的问题出发。","stars":2816,"starGrowth":1,"createdAt":"2019-09-04T22:59:06Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["AI入门","计算机视觉","自然语言处理","推荐系统"],"domainKey":"AI入门","pushedAt":"2026-09-20T18:36:02.638Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/microsoft/AI","ts":"2026-09-20T18:36:02.638Z","category":"learning","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"eddwebster/football_analytics","owner":"eddwebster","name":"football_analytics","desc":"","summaryCn":"足球数据分析爱好者的资源宝库，收录社区公开的学习材料、数据源与论文。","reasonCn":"这是一个由足球数据分析师 Edd Webster 维护的公开资源清单，把分散在社区里的学习材料、数据集、论文和博客系统地整理在一起。无论你是刚入门想找数据练手，还是想深入理解战术分析模型，都能在这里找到对应的指引。","stars":2788,"starGrowth":1,"createdAt":"2020-09-01T14:28:18Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源"],"zone":"资源","funScore":0.26,"domainTags":["足球分析","数据科学","学习资源","数据集"],"domainKey":"足球分析","pushedAt":"2026-09-20T18:36:02.638Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/eddwebster/football_analytics","ts":"2026-09-20T18:36:02.638Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"github/awesome-copilot","owner":"github","name":"awesome-copilot","desc":"Community-contributed instructions, agents, skills, and configurations to help you make the most of GitHub Copilot.","summaryCn":"GitHub Copilot的神仙攻略合集，帮你省下几十万开发费","reasonCn":"这是一个社区驱动的资源库，汇集了Copilot的最佳实践、技能包和配置文件，旨在提升开发者使用AI编程助手的效率。项目通过整理各类Agent指令和配置，帮助用户规避常见陷阱，快速掌握Copilot的高级用法，是开发者提升编程效率的实用工具箱。","stars":39318,"starGrowth":43,"createdAt":"2025-06-11T16:57:39Z","language":"JavaScript","topics":["agents","ai"],"aiDims":["代码助手","AI应用","AI基础设施"],"zone":"资源","funScore":0,"domainTags":["GitHub Copilot","技能包","提示词工程"],"domainKey":"GitHub Copilot","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/github/awesome-copilot","ts":"2026-09-23T21:53:43.262Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.13228094002746835,"copyOk":false},{"repo":"herdrdev/herdr","owner":"herdrdev","name":"herdr","desc":"the runtime your coding agents live on","summaryCn":"让AI编程助手在你电脑安家，安全可靠跑代码","reasonCn":"herdr是一个用Rust构建的本地运行时环境，专门为AI编程Agent设计。它让Claude、Cursor等编码助手能在你的机器上安全执行代码，支持沙箱隔离、权限控制和日志追踪。相比云端代理方案，herdr将执行环境完全本地化，数据不会外泄，同时提供API接口供各种Agent框架调用，是AI编程生态的基础设施级项目。","stars":40237,"starGrowth":174,"createdAt":"2026-03-27T17:54:33Z","silentRounds":0,"language":"Rust","topics":["developer-tools","ai"],"aiDims":["AI Agent","开发者工具","效率工具"],"zone":"AI","funScore":0.5,"domainTags":["AI运行时","编程Agent","沙箱隔离","Rust"],"domainKey":"AI运行时","pushedAt":"2026-09-23T20:38:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/herdrdev/herdr","ts":"2026-09-23T20:38:17Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3755677610159803,"copyOk":false},{"repo":"C4illin/ConvertX","owner":"C4illin","name":"ConvertX","desc":"💾 Self-hosted online file converter. Supports 1000+ formats ⚙️","summaryCn":"自托管的文件格式转换器，支持1000多种格式任意转换","reasonCn":"ConvertX采用自托管架构，所有文件都在你自己的服务器上处理，不会上传到任何第三方。这意味着你的文件隐私得到了保障，特别适合处理敏感文档的企业用户。它整合了多种成熟的开源转换引擎，支持超过1000种格式组合，从常见的JPG转PNG到复杂的MKV转MP4都能胜任。","stars":19039,"starGrowth":8,"createdAt":"2024-04-07T16:21:39Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted"],"aiDims":["效率工具","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文件格式转换","自托管","文件处理","数据隐私","拖拽上传"],"domainKey":"文件格式转换","pushedAt":"2026-09-23T21:27:42Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/C4illin/ConvertX","ts":"2026-09-23T21:27:42Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02526057040810967,"copyOk":false},{"repo":"Unleash/unleash","owner":"Unleash","name":"unleash","desc":"Open-source feature management platform","summaryCn":"功能开关管理神器，轻松控制线上功能的灰度发布","reasonCn":"Unleash 是一个开源的功能开关（Feature Toggle）管理平台，帮助开发团队安全地控制线上功能的发布节奏。你可以用它实现灰度发布、A/B 测试、功能下线等场景，代码写好了但不急着给所有人用，关掉开关就完全不生效，相当于给功能加了个总闸。支持多环境配置、权限管理和详细的发布数据可视化。","stars":13821,"starGrowth":1,"createdAt":"2014-09-29T12:50:46Z","silentRounds":1,"language":"TypeScript","topics":["devops"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["功能开关","灰度发布","DevOps工具"],"domainKey":"功能开关","pushedAt":"2026-09-23T15:34:45Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Unleash/unleash","ts":"2026-09-23T15:34:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pubkey/rxdb","owner":"pubkey","name":"rxdb","desc":"The local-first database that runs on every JS runtime and replicates with your existing backend - no vendor, no lock-in - https://rxdb.info/","summaryCn":"前端也能用的本地数据库，支持离线同步，数据存在浏览器里不丢","reasonCn":"RxDB是一个专为JavaScript设计的关系型数据库，最大的特点是local-first，数据优先存在本地浏览器或Node.js中。它天然支持离线使用，联网后自动与后端同步，完美解决前后端数据冲突问题。支持实时响应式查询，UI会自动更新。适合开发PWA应用、离线优先的移动端Web应用，以及需要高性能本地存储的前端项目。","stars":23391,"starGrowth":4,"createdAt":"2016-12-02T19:34:42Z","silentRounds":0,"language":"TypeScript","topics":["database"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["本地优先数据库","离线同步","响应式查询","PWA","IndexedDB"],"domainKey":"本地优先数据库","pushedAt":"2026-09-23T09:23:48Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/pubkey/rxdb","ts":"2026-09-23T09:23:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tt-a1i/archify","owner":"tt-a1i","name":"archify","desc":"Agent skill for beautiful, verifiable architecture, workflow, sequence, data-flow, and lifecycle diagrams—self-contained HTML with motion and crisp export.","summaryCn":"让代码瞬间变美，生成动图和高清架构图的神器","reasonCn":"技术上，它基于 Agent 技术构建，能识别常见的代码模式（比如 React 组件结构、数据库表关系、API 接口定义等），然后调用专门的绘图引擎渲染成 HTML。输出是单文件 HTML，这意味着你不需要依赖任何外部图表库，直接双击就能在浏览器里打开，还能通过 CSS 和 JavaScript 做各种动画","stars":69870,"starGrowth":751,"createdAt":"2026-04-15T05:27:37Z","silentRounds":0,"language":"JavaScript","topics":["developer-tools"],"aiDims":["可视化","创意工具"],"zone":"工具","funScore":1,"domainTags":["架构图","代码可视化","Agent","流程图","开发工具"],"domainKey":"架构图","pushedAt":"2026-09-23T15:03:54Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/tt-a1i/archify","ts":"2026-09-23T15:03:54Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3986947187634177,"copyOk":false},{"repo":"skyzh/tiny-llm","owner":"skyzh","name":"tiny-llm","desc":"A course of learning LLM inference serving on Apple Silicon for systems engineers: build a tiny vLLM + Qwen.","summaryCn":"让你亲手在苹果电脑上搭一套迷你版大模型推理系统","reasonCn":"这是一门面向系统工程师的实战课程，教你如何在Apple Silicon芯片上从零搭建一套轻量级LLM推理服务。项目以vLLM为蓝本，结合Qwen模型，涵盖张量并行、请求调度、GPU内存管理等核心知识点。通过实际编码，你能深入理解大模型推理的底层机制，而不仅仅是调用API。","stars":4677,"starGrowth":63,"createdAt":"2025-04-19T20:01:33Z","silentRounds":0,"language":"Python","topics":["llm-model"],"aiDims":["推理引擎","大语言模型","学习资源"],"zone":"AI","funScore":0.6,"domainTags":["LLM推理","vLLM","Apple Silicon","分布式系统","系统工程"],"domainKey":"LLM推理","pushedAt":"2026-09-23T09:06:18Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/skyzh/tiny-llm","ts":"2026-09-23T09:06:18Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.22941051956382297,"copyOk":true},{"repo":"grafana/grafana","owner":"grafana","name":"grafana","desc":"The open and composable observability and data visualization platform. Visualize metrics, logs, and traces from multiple sources like Prometheus, Loki, Elasticsearch, InfluxDB, Postgres and many more.","summaryCn":"把服务器数据变成酷炫监控大屏，一眼看懂系统健康状态","reasonCn":"Grafana是目前最流行的开源数据可视化平台，支持接入Prometheus、InfluxDB、Elasticsearch、Postgres等数十种数据源。它可以创建实时仪表盘、设置告警规则、配置钉钉/Slack/PagerDuty通知，让运维和开发团队一眼看清系统运行状况。","stars":76860,"starGrowth":15,"createdAt":"2013-12-11T15:59:56Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据可视化","监控仪表盘","时序数据库","运维监控"],"domainKey":"数据可视化","pushedAt":"2026-09-23T21:40:39Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/grafana/grafana","ts":"2026-09-23T21:40:39Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.045585480093676814,"copyOk":false},{"repo":"pshenok/server-survival","owner":"pshenok","name":"server-survival","desc":"Tower defense game that teaches cloud architecture. Build infrastructure, survive traffic, learn scaling.","summaryCn":"在 3D 云架构模拟中生存，管理流量并抵御 DDoS 攻击。","reasonCn":"这款游戏把枯燥的云架构知识变成了生存挑战。玩家扮演云架构师，必须在不断攀升的流量和 DDoS 攻击夹击下，通过搭建弹性基础设施、合理分配预算来维持服务健康。核心玩法是动态资源调度与防御策略，让人在紧张刺激中掌握云运维逻辑。","stars":6425,"starGrowth":0,"createdAt":"2025-11-21T02:36:29Z","silentRounds":2,"language":"JavaScript","topics":["game","simulation"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["塔防游戏","云架构教育","技术可视化","Web游戏","DevOps培训"],"domainKey":"塔防游戏","pushedAt":"2026-09-21T03:11:19Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/pshenok/server-survival","ts":"2026-09-21T03:11:19Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dolthub/dolt","owner":"dolthub","name":"dolt","desc":"Dolt – Git for Data","summaryCn":"像Git一样版本控制数据库，随时回滚变更像时光倒流","reasonCn":"Dolt是一款基于Go语言开发的开源版本控制系统，专为关系型数据库设计。它实现了完整的SQL引擎，支持commit、branch、merge、diff等Git风格操作，同时兼容MySQL协议，可直接用MySQL客户端连接。底层采用LSM树存储引擎，支持ACID事务和快照隔离。与liquibase等工具不同，Dolt允许在分支上并行开发数据库变更并合并，彻底改变了数据库协作方式。","stars":24501,"starGrowth":8,"createdAt":"2019-07-24T17:46:25Z","silentRounds":0,"language":"Go","topics":["database","git"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["版本控制数据库","SQL引擎","Git式数据协作","MySQL兼容","数据库分支合并"],"domainKey":"版本控制数据库","pushedAt":"2026-09-23T20:36:56Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/dolthub/dolt","ts":"2026-09-23T20:36:56Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.024979551855026327,"copyOk":false},{"repo":"Fission-AI/OpenSpec","owner":"Fission-AI","name":"OpenSpec","desc":"Spec-driven development (SDD) for AI coding assistants.","summaryCn":"用规格驱动AI编程让代码助手生成更精准更贴合需求的代码","reasonCn":"OpenSpec是一个规范驱动开发工具专门为AI编程助手设计。它让你先用自然语言写出代码规格说明然后AI助手根据这些规格来生成代码。相比直接让AI写代码这种方式能让生成结果更精准更符合预期减少反复修改的麻烦。6.7万star证明它很受欢迎。","stars":69863,"starGrowth":90,"createdAt":"2025-08-05T10:37:45Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["代码助手","AI应用","提示工程"],"zone":"AI","funScore":0,"domainTags":["提示词工程","AI编程助手","规范驱动开发","代码生成","软件工作流"],"domainKey":"提示词工程","pushedAt":"2026-09-23T21:47:33Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/Fission-AI/OpenSpec","ts":"2026-09-23T21:47:33Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.27386470664013857,"copyOk":false},{"repo":"CherryHQ/cherry-studio","owner":"CherryHQ","name":"cherry-studio","desc":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","summaryCn":"一站式AI生产力工作室，内置智能对话、自主代理及300+助手","reasonCn":"Cherry Studio是一个统一的AI生产力平台，集成ChatGPT、Claude等数十个主流大模型，支持智能对话、任务拆分与自主代理执行。内置300多个预设助手（如代码审查、文档总结、图像生成），用户可通过插件扩展功能实现工作流自动化，适合内容创作者、开发者和企业团队提升效率。","stars":52081,"starGrowth":33,"createdAt":"2024-05-24T01:56:26Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI应用","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["大模型聚合","AI客户端","自主代理","API网关"],"domainKey":"大模型聚合","pushedAt":"2026-09-23T21:33:54Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/CherryHQ/cherry-studio","ts":"2026-09-23T21:33:54Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10090088515965516,"copyOk":true},{"repo":"orf/gping","owner":"orf","name":"gping","desc":"Ping, but with a graph","summaryCn":"把枯燥的ping命令变成实时跳动的网络曲线图","reasonCn":"gping是一个基于Rust编写的命令行网络诊断工具，核心能力是将传统的文本格式ping输出转化为实时滚动的折线图。它底层调用系统ping命令，通过ncurses图形库实现动态渲染。支持Linux、macOS和Windows平台，能直观展示延迟波动、抖动和丢包情况。相比传统ping只显示一串数字，它在网络调试时特别直观，一眼就能看出网络质量变化趋势，适合开发者和运维人员排查网络问题。","stars":12689,"starGrowth":2,"createdAt":"2015-10-17T14:34:25Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"工具","funScore":0.7,"domainTags":["网络诊断","命令行","可视化"],"domainKey":"网络诊断","pushedAt":"2026-09-23T10:32:55Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/orf/gping","ts":"2026-09-23T10:32:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"octetta/k-synth","owner":"octetta","name":"k-synth","desc":"A pocket-calculator version of a synthesizer","summaryCn":"一个基于数组逻辑的合成器环境，将声音处理简化为数学向量运算。","reasonCn":"k/synth 是一项极其硬核的艺术尝试，它打破了传统 DAW（数字音频工作站）对音轨、时间轴和连线的依赖。该项目受 Arthur Whitney 作品启发，把每一个声音样本都看作一个数学向量，通过一行行简洁的表达式来控制声音的生成。它就像是一个“袖珍计算器版”的合成器，非常适合那些热衷于用数学公式去定义旋律和音色的极客，这种非传统的音乐创作方式充满了极简美感。","stars":33,"starGrowth":0,"createdAt":"2026-02-01T22:13:59Z","silentRounds":6,"language":"C","topics":["synthesizer"],"aiDims":["创意工具","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["音频合成器","数组编程音乐","表达式音频引擎","极简代码创作","电子音乐工具"],"domainKey":"音频合成器","pushedAt":"2026-09-20T18:02:31Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"synthesizer","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/octetta/k-synth","ts":"2026-09-20T18:02:31Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fleetbase/fleetbase","owner":"fleetbase","name":"fleetbase","desc":"Modular logistics and supply chain operating system (LSOS)","summaryCn":"开源物流操作系统，集调度、车队、仓储与财务于一体。","reasonCn":"Fleetbase 是一个开源的物流与供应链操作系统，解决运输管理碎片化的问题。它将调度、车队管理、实时追踪、商务、仓储和财务整合在同一个平台，并提供 REST API、webhooks 和扩展系统，方便按需定制功能。你既可以部署在自己的服务器上，也可以直接使用其云服务，适合物流行业从业者和数字化改造团队。","stars":3866,"starGrowth":19,"createdAt":"2022-09-23T09:36:08Z","silentRounds":0,"language":"JavaScript","topics":["driver-app","fleet-management","infrastructure","inventory-management","logistics","ondemand","open-source","order-management","route-optimization","supply-chain-management","transportation","warehouse-management","tracking-and-monitoring","freight-management","last-mile-delivery","transport-management-system"],"aiDims":["非AI-实用","效率工具","开发工具"],"zone":"工具","funScore":0,"domainTags":["物流管理","供应链管理","Docker","自托管"],"domainKey":"物流管理","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"driver-app","source":"github","weight":0.7},{"name":"fleet-management","source":"github","weight":0.7},{"name":"infrastructure","source":"github","weight":0.7},{"name":"inventory-management","source":"github","weight":0.7},{"name":"logistics","source":"github","weight":0.7},{"name":"ondemand","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"order-management","source":"github","weight":0.7},{"name":"route-optimization","source":"github","weight":0.7},{"name":"supply-chain-management","source":"github","weight":0.7},{"name":"transportation","source":"github","weight":0.7},{"name":"warehouse-management","source":"github","weight":0.7},{"name":"tracking-and-monitoring","source":"github","weight":0.7},{"name":"freight-management","source":"github","weight":0.7},{"name":"last-mile-delivery","source":"github","weight":0.7},{"name":"transport-management-system","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/fleetbase/fleetbase","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.07174392136575271,"copyOk":true},{"repo":"FalkorDB/FalkorDB","owner":"FalkorDB","name":"FalkorDB","desc":"A super fast Graph Database uses GraphBLAS under the hood for its sparse adjacency matrix graph representation. Our goal is to provide the best Knowledge Graph for LLM (GraphRAG).","summaryCn":"图数据库里的火箭，专门给AI提供超级快的知识图谱底座","reasonCn":"基于GraphBLAS内核构建的高性能图数据库，专为知识图谱和GraphRAG设计。采用稀疏邻接矩阵表示，在Rust语言加持下执行图遍历和推理速度极快。能轻松处理百万级节点和十亿级边的复杂关系查询，是构建大模型知识增强系统的理想基础设施。","stars":6269,"starGrowth":38,"createdAt":"2023-07-20T10:47:51Z","silentRounds":0,"language":"Rust","topics":["developer-tools"],"aiDims":["知识图谱","AI基础设施","RAG与知识"],"zone":"工具","funScore":0.7,"domainTags":["图数据库","GraphRAG","知识图谱","Rust"],"domainKey":"图数据库","pushedAt":"2026-09-23T20:22:09Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/FalkorDB/FalkorDB","ts":"2026-09-23T20:22:09Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.13218471845589408,"copyOk":true},{"repo":"1jehuang/jcode","owner":"1jehuang","name":"jcode","desc":"The most RAM efficient harness","summaryCn":"极致省内存的AI Agent框架，小设备也能玩转大模型","reasonCn":"核心亮点在于内存管理优化。它用Rust编写，利用Rust的所有权机制实现零成本抽象，内存占用可以做到极致精简。同时它内置了工具调用（Tool Use）机制，Agent可以调用外部函数、API和工具来完成复杂任务。架构上采用事件驱动模式，支持异步操作，不会阻塞主线程。","stars":20033,"starGrowth":41,"createdAt":"2026-01-05T09:43:36Z","silentRounds":0,"language":"Rust","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI Agent","内存优化","Rust"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:06:18Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/1jehuang/jcode","ts":"2026-09-23T21:06:18Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.1549678430589527,"copyOk":false},{"repo":"Tongsuo-Project/Tongsuo","owner":"Tongsuo-Project","name":"Tongsuo","desc":"铜锁/Tongsuo is a Modern Cryptographic Primitives and Protocols Library","summaryCn":"现代密码学库，C实现TLS/SM2等协议，跨平台、轻量、直接上手。","reasonCn":"Tongsuo 用 C 语言实现完整的现代密码学原语与协议栈，提供 TLS、DTLS、SM2/SM3/SM4 等算法实现，并兼容 OpenSSL 接口，能直接在嵌入式设备或服务器上替换。库采用模块化设计，支持硬件加速和可选 FIPS 模式，测试覆盖率高，文档详尽、示例丰富，帮助开发者快速集成加密功能，降低自行实现的风险与成本。","stars":1536,"starGrowth":0,"createdAt":"2020-08-25T06:44:08Z","silentRounds":3,"language":"C","topics":["cryptography"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["密码学","TLS","SM2","C语言"],"domainKey":"密码学","pushedAt":"2026-09-23T13:35:42Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cryptography","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Tongsuo-Project/Tongsuo","ts":"2026-09-23T13:35:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ArchiveBox/ArchiveBox","owner":"ArchiveBox","name":"ArchiveBox","desc":"🗃 Open source self-hosted web archiving. Takes URLs/browser history/bookmarks/Pocket/Pinboard/etc., saves HTML, JS, PDFs, media, and more...","summaryCn":"把你收藏的网页一键打包存档，再也不怕网站关闭打不开了","reasonCn":"ArchiveBox是一款开源自托管网页归档工具，支持从浏览器历史、书签、Pocket、Pinboard等来源批量导入URL，自动抓取并保存HTML、JS、PDF、图片和视频等完整内容。基于Python开发，Docker一键部署，内置Web管理界面，支持定时任务和API调用，适合个人知识库备份和内容存档。","stars":28574,"starGrowth":6,"createdAt":"2017-05-05T08:50:14Z","silentRounds":0,"language":"Python","topics":["self-hosted"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["网页归档","自托管","URL存档","WARC格式"],"domainKey":"网页归档","pushedAt":"2026-09-23T20:23:25Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/ArchiveBox/ArchiveBox","ts":"2026-09-23T20:23:25Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01862994330510254,"copyOk":false},{"repo":"jarun/nnn","owner":"jarun","name":"nnn","desc":"n³ The unorthodox terminal file manager","summaryCn":"终端里最快最极简的文件管理器，装完1分钟上手，从此告别鼠标","reasonCn":"技术上，它用C写的，没有图形界面依赖，没有python、no_deps，就一个二进制文件。它的核心思路是：键盘操作要最快，界面要最轻，所以它甚至自带了一个轻量级的图标预览模式。多面板浏览、tab管理、文件预览、批量重命名全都能做，还能通过插件扩展功能。","stars":21994,"starGrowth":6,"createdAt":"2016-08-20T14:10:14Z","silentRounds":0,"language":"C","topics":["command-line","developer-tools","raspberry-pi"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["终端文件管理器","命令行工具","效率工具"],"domainKey":"终端文件管理器","pushedAt":"2026-09-21T15:09:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/jarun/nnn","ts":"2026-09-21T15:09:26Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018818405019550784,"copyOk":true},{"repo":"jaywcjlove/awesome-mac","owner":"jaywcjlove","name":"awesome-mac","desc":" This project is dedicated to collecting high-quality macOS software and organizing them systematically by different categories for easy search and use.","summaryCn":"Mac用户必备的宝藏软件清单，帮你省下大把找软件的时间","reasonCn":"这是一个精心整理的 macOS 软件精选集，按功能分类收录了数百款高质量工具。无论是日常办公、创意设计还是开发调试，都能在这里找到最合适的利器。项目维护者持续更新，确保收录的都是当前最好用的软件，极大降低了用户寻找和试用软件的成本。","stars":114424,"starGrowth":137,"createdAt":"2016-07-17T15:33:47Z","silentRounds":0,"language":"Swift","topics":["awesome"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"资源","funScore":0,"domainTags":["macOS","软件清单","效率工具","开发者工具"],"domainKey":"macOS","pushedAt":"2026-09-22T00:59:04Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/jaywcjlove/awesome-mac","ts":"2026-09-22T00:59:04Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3035919037964063,"copyOk":true},{"repo":"longbridge/gpui-kit","owner":"longbridge","name":"gpui-kit","desc":"Rust GUI components for building fantastic cross-platform desktop application by using GPUI.","summaryCn":"用 Rust 搭跨平台桌面应用的神级 UI 库","reasonCn":"GPUI 是一个基于 Rust 的声明式 UI 框架，专为构建高性能、跨平台桌面应用设计。它采用了独特的增量渲染架构和内存安全特性，能够轻松实现复杂的交互界面。该组件库提供了丰富的预置 UI 元素，支持 macOS、Linux 和 Windows 平台，开发者可以快速搭建出既美观又流畅的应用程序。","stars":14695,"starGrowth":43,"language":"Rust","topics":["gpui","desktop-application","uikit","rust"],"aiDims":["开发者工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Rust GUI","声明式 UI","跨平台桌面","增量渲染","桌面应用开发"],"domainKey":"Rust GUI","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"gpui","source":"github","weight":0.7},{"name":"desktop-application","source":"github","weight":0.7},{"name":"uikit","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/longbridge/gpui-kit","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.13777849608710446,"copyOk":false},{"repo":"go-task/task","owner":"go-task","name":"task","desc":"A fast, cross-platform build tool inspired by Make, designed for modern workflows.","summaryCn":"告别繁琐Makefile，用YAML写任务脚本更优雅更高效","reasonCn":"现代跨平台构建工具，语法基于YAML比Makefile更直观易读。支持任务依赖、并行执行、条件分支、环境变量覆盖等高级特性，一个Taskfile.yml即可管理整个项目的构建、测试、部署流程，让开发者的日常重复操作自动化到极致。","stars":16181,"starGrowth":5,"createdAt":"2017-02-27T00:46:04Z","silentRounds":0,"language":"Go","topics":["devops"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["构建工具","DevOps","任务自动化","Go"],"domainKey":"构建工具","pushedAt":"2026-09-23T00:30:23Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/go-task/task","ts":"2026-09-23T00:30:23Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015927013163586922,"copyOk":true},{"repo":"anchore/grype","owner":"anchore","name":"grype","desc":"A vulnerability scanner for container images and filesystems","summaryCn":"扫描容器镜像漏洞的安全工具，帮你及时发现并修复安全隐患","reasonCn":"用Go语言编写的容器镜像和文件系统漏洞扫描器，能深度检测Docker镜像和本地文件系统中的已知安全漏洞。内置庞大的CVE漏洞库，支持生成SBOM软件物料清单，是CI/CD流水线中容器安全检测的关键工具，守护你的应用安全。","stars":12916,"starGrowth":3,"createdAt":"2020-05-26T13:44:38Z","silentRounds":0,"language":"Go","topics":["security","docker"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["容器漏洞扫描","SBOM生成","软件供应链安全"],"domainKey":"容器漏洞扫描","pushedAt":"2026-09-21T20:04:57Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/anchore/grype","ts":"2026-09-21T20:04:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"plausible/analytics","owner":"plausible","name":"analytics","desc":"Open source, privacy-first web analytics. Lightweight, cookie-free Google Analytics alternative. Self-hosted or cloud.","summaryCn":"不追踪用户的隐私分析工具，轻量合规就是好用","reasonCn":"技术架构上，Plausible采用极简设计，脚本代码只有几KB，加载不影响网站性能。数据存储在自有服务器上，所有分析逻辑都在后端完成。它不支持细粒度的用户行为追踪，但页面浏览量、来源页面、地理位置、设备类型等核心指标都能准确统计。","stars":29191,"starGrowth":13,"createdAt":"2018-12-04T22:33:36Z","silentRounds":0,"language":"Elixir","topics":["self-hosted","privacy"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网站流量分析","隐私保护","自托管服务","Cookie-free追踪"],"domainKey":"网站流量分析","pushedAt":"2026-09-23T18:27:14Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"Elixir","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/plausible/analytics","ts":"2026-09-23T18:27:14Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.040336028227878455,"copyOk":false},{"repo":"pytorch/pytorch","owner":"pytorch","name":"pytorch","desc":"Tensors and Dynamic neural networks in Python with strong GPU acceleration","summaryCn":"AI世界的乐高积木，谁都能用它搭出属于自己的神经网络","reasonCn":"PyTorch是目前最流行的深度学习框架之一，由Meta的AI研究院主导开发。它提供了灵活的动态计算图、强大的GPU加速能力，以及丰富的API和工具链。几乎所有前沿AI研究都建立在它之上，从图像识别到大语言模型训练都能胜任。","stars":103180,"starGrowth":24,"createdAt":"2016-08-13T05:26:41Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI基础设施","模型与训练","大语言模型"],"zone":"AI","funScore":0,"domainTags":["深度学习","神经网络","GPU加速","自动微分"],"domainKey":"深度学习","pushedAt":"2026-09-23T21:46:00Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/pytorch/pytorch","ts":"2026-09-23T21:46:00Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.07269780965303353,"copyOk":true},{"repo":"tensorflow/tensorflow","owner":"tensorflow","name":"tensorflow","desc":"An Open Source Machine Learning Framework for Everyone","summaryCn":"谷歌开源的工业级机器学习框架，训练部署一条龙","reasonCn":"技术层面，它采用数据流图作为计算核心，把复杂的数学运算转化为可优化的图结构。支持自动微分，你不需要手算梯度；支持TPU/GPU多卡分布式训练，几行代码就能把模型跑到上千张卡上；还有TF Serving专门用于生产环境的高性能模型部署。","stars":200258,"starGrowth":37,"createdAt":"2015-11-07T01:19:20Z","silentRounds":0,"language":"C++","topics":["ml"],"aiDims":["模型与训练","AI基础设施","推理引擎"],"zone":"AI","funScore":0,"domainTags":["机器学习框架","深度学习","模型训练部署","分布式训练","生产级AI"],"domainKey":"机器学习框架","pushedAt":"2026-09-23T21:44:38Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/tensorflow/tensorflow","ts":"2026-09-23T21:44:38Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.11155428497238562,"copyOk":true},{"repo":"youzan/vant","owner":"youzan","name":"vant","desc":"A lightweight, customizable Vue UI library for mobile web apps.","summaryCn":"前端开发必备的移动端组件库，让你快速搭建优雅的手机网页界面","reasonCn":"Vant是由有赞团队开源的轻量级Vue移动端组件库，提供超过80个高质量UI组件，涵盖按钮、表单、弹窗、导航等常见场景。它专注于移动端交互体验，支持按需引入和主题定制，能大幅降低H5开发成本，是构建移动端页面非常实用的工程化解决方案。","stars":24383,"starGrowth":0,"createdAt":"2017-04-19T07:55:31Z","silentRounds":2,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Vue组件库","移动端UI","前端开发","TypeScript","有赞"],"domainKey":"Vue组件库","pushedAt":"2026-09-23T09:07:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/youzan/vant","ts":"2026-09-23T09:07:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Vonng/ddia","owner":"Vonng","name":"ddia","desc":"《Designing Data-Intensive Application》DDIA 第一版 / 第二版 中文翻译","summaryCn":"数据库和分布式系统的圣经级教材，全网最全中文翻译","reasonCn":"Martin Kleppmann的《Designing Data-Intensive Application》是数据系统领域公认的经典之作，中文翻译项目由Vonng主导完成，覆盖第一版和第二版内容。这本书系统讲解数据库原理、分布式系统、消息队列等核心知识，是后端工程师必读的技术书。翻译质量高，社区持续维护，对想系统学习数据基础设施的开发者来说是一大利好。","stars":23727,"starGrowth":4,"createdAt":"2018-02-08T04:26:52Z","silentRounds":0,"language":"Python","topics":["database"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["数据库","分布式系统","技术书籍","中文翻译"],"domainKey":"数据库","pushedAt":"2026-09-23T00:42:57Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Vonng/ddia","ts":"2026-09-23T00:42:57Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sususweet/midea_auto_cloud","owner":"sususweet","name":"midea_auto_cloud","desc":"Control Midea devices via Cloud from Home Assistant","summaryCn":"不用破解路由，直接让美的空调连接Home Assistant的智能桥梁","reasonCn":"技术上，它逆向工程了美的官方App的通信协议，通过Python实现了对HTTP请求的签名和加密处理。这意味着它不依赖局域网内的设备通信，而是通过云端中转，因此即使设备不在同一WiFi下也能控制。这种架构的优势是兼容性极强，几乎支持所有能连上美的云端的设备，包括老旧型号，填补了O","stars":288,"starGrowth":0,"createdAt":"2025-09-08T15:16:09Z","silentRounds":5,"language":"Python","topics":["home-automation"],"aiDims":["非AI-实用","效率工具","安全工具"],"zone":"工具","funScore":0,"domainTags":["智能家居","Home Assistant","设备逆向","美的协议"],"domainKey":"智能家居","pushedAt":"2026-09-20T16:44:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/sususweet/midea_auto_cloud","ts":"2026-09-20T16:44:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langfuse/langfuse","owner":"langfuse","name":"langfuse","desc":"🪢 Open source AI engineering platform: LLM evals, observability, metrics, prompt management, playground, datasets. Integrates with OpenTelemetry, LangChain, OpenAI SDK, LiteLLM, and more. 🍊YC W23 ","summaryCn":"监控调试LLM应用的全能神器，从Prompt到生产上线一站式搞定","reasonCn":"这是专为LLM应用设计的开源工程化平台，提供模型调用追踪、提示词版本管理、AI评估框架和可视化Dashboard。支持OpenTelemetry集成，可以追踪每次请求的延迟、token消耗和成本，让调试AI应用像调试普通软件一样简单直观。","stars":34945,"starGrowth":35,"createdAt":"2023-05-18T17:47:09Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted","llm"],"aiDims":["AI基础设施","提示工程","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["LLM可观测性","提示词工程","AI应用监控"],"domainKey":"LLM可观测性","pushedAt":"2026-09-23T21:36:18Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/langfuse/langfuse","ts":"2026-09-23T21:36:18Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10800472170553727,"copyOk":true},{"repo":"browser-use/browser-use","owner":"browser-use","name":"browser-use","desc":"Agents that use the browser.","summaryCn":"让 AI 自动操作浏览器，像人一样上网办事。","reasonCn":"这是一款开源的浏览器自动化工具，核心是让大模型像人类一样控制浏览器。它通过 Python 和 TypeScript 实现，能自动完成填表、点击、导航等网页操作。除了本地 CLI，还提供云端浏览器服务，支持隐形模式、验证码解决和住宅代理，让 AI 能够无障碍地处理各种网页任务。","stars":115947,"starGrowth":143,"createdAt":"2024-10-31T16:00:56Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","浏览器自动化","Playwright","LLM集成","网页操作"],"domainKey":"AI Agent","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/browser-use/browser-use","ts":"2026-09-22T04:07:30.369Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30369996636394214,"copyOk":true},{"repo":"watchexec/watchexec","owner":"watchexec","name":"watchexec","desc":"Executes commands in response to file modifications","summaryCn":"文件一变就自动执行命令，开发效率直接起飞","reasonCn":"watchexec是个Rust编写的文件监听工具，能实时监测目录变化，一旦检测到修改就自动触发你设定的命令。它特别适合前端开发、文档生成或脚本自动化场景，省去了手动重新编译或运行的麻烦，支持递归监听、忽略规则和高并发，是开发者提升工作流的必备神器。","stars":7194,"starGrowth":1,"createdAt":"2016-09-18T20:41:31Z","silentRounds":3,"language":"Rust","topics":["command-line","developer-tools"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["文件监听","命令行工具","开发效率","自动化"],"domainKey":"文件监听","pushedAt":"2026-09-23T05:33:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/watchexec/watchexec","ts":"2026-09-23T05:33:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Anuken/Mindustry","owner":"Anuken","name":"Mindustry","desc":"The automation tower defense RTS","summaryCn":"既能建厂自动化又能塔防的游戏，策略党和自动化爱好者必备","reasonCn":"这是款融合工业流水线与塔防的战略游戏，你不仅要造机枪塔打敌人，还要建采矿机、冶炼厂、传送带链。它支持在线多人模式，内置Lua脚本系统可深度定制逻辑。虽然纯Java开发且没有AI功能，但它的硬核策略体验极其上瘾，是Steam上的神作，开源代码也值得游戏开发者学习。","stars":29079,"starGrowth":11,"createdAt":"2017-04-30T01:24:29Z","silentRounds":0,"language":"Java","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["塔防游戏","自动化流水线","即时战略","Lua脚本"],"domainKey":"塔防游戏","pushedAt":"2026-09-23T16:37:56Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Anuken/Mindustry","ts":"2026-09-23T16:37:56Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.034134839574951,"copyOk":false},{"repo":"sharkdp/bat","owner":"sharkdp","name":"bat","desc":"A cat(1) clone with wings.","summaryCn":"终端查看代码的神器，自带语法高亮超好用。","reasonCn":"bat 是经典 cat 命令的增强替代品，专为查看代码文件设计。它提供语法高亮、行号显示、Git 差异标注等实用功能，让枯燥的命令行文件浏览变得井井有条。虽然是纯工程工具，但对每天面对终端的开发者来说，能极大提升代码审查和日常调试的效率。","stars":60544,"starGrowth":6,"createdAt":"2018-04-21T10:52:23Z","silentRounds":0,"language":"Rust","topics":["command-line","git"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行工具","代码查看","语法高亮"],"domainKey":"命令行工具","pushedAt":"2026-09-22T18:13:30Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/sharkdp/bat","ts":"2026-09-22T18:13:30Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0182973044397463,"copyOk":false},{"repo":"TencentCloud/TencentDB-Agent-Memory","owner":"TencentCloud","name":"TencentDB-Agent-Memory","desc":"TencentDB Agent Memory is a team-level memory hub for AI Agents — turning conversations, docs, and code into four reusable memory assets (Chat Memory, Skill, LLM-Wiki, Code-Graph) that are governed, shared, and equipped across agents and frameworks.","summaryCn":"让多个AI Agent记住彼此的腾讯云智能协作记忆中枢","reasonCn":"这个项目专为解决AI智能体协作中的记忆断层问题而设计。它通过引入记忆核心、记忆中心和代理三层架构，让不同的AI Agent能够共享上下文信息，实现更连贯的对话和任务执行。无论你是开发多智能体应用，还是需要让AI助手记住复杂的历史信息，它都能提供可靠的记忆解决方案。","stars":27154,"starGrowth":39,"createdAt":"2026-04-07T06:34:26Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","llm"],"aiDims":["AI Agent","RAG与知识","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["Agent记忆","向量检索","权限管理","企业级存储"],"domainKey":"Agent记忆","pushedAt":"2026-09-23T07:22:44Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/TencentCloud/TencentDB-Agent-Memory","ts":"2026-09-23T07:22:44Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.1455705089489578,"copyOk":true},{"repo":"upscayl/upscayl","owner":"upscayl","name":"upscayl","desc":"🆙 Upscayl - #1 Free and Open Source AI Image Upscaler for Linux, MacOS and Windows.","summaryCn":"免费开源的AI图片放大神器，让模糊照片瞬间变高清","reasonCn":"Upscayl的强大之处在于它集成了多个先进的开源AI模型，比如Real-ESRGAN、SwinIR等，这些模型都是专门训练用来恢复图像细节的。它支持最高4倍、8倍的放大，而且还能保持图片的宽高比，不会把图片拉变形。对于普通用户来说，它提供了一个非常简洁的图形界面，所有的操作都在几个按钮","stars":49814,"starGrowth":200,"createdAt":"2022-07-30T17:19:20Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["创意工具","非AI-实用","非AI-好玩"],"zone":"AI","funScore":0.6,"domainTags":["图像超分辨率","AI图像增强","开源图像工具"],"domainKey":"图像超分辨率","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/upscayl/upscayl","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.31204480668085277,"copyOk":true},{"repo":"yt-dlp/yt-dlp","owner":"yt-dlp","name":"yt-dlp","desc":"A feature-rich command-line audio/video downloader","summaryCn":"一款支持数千网站的强大命令行音视频下载器。","reasonCn":"yt-dlp 是一个功能丰富的命令行音视频下载工具，支持超过数千个网站。作为 youtube-dl 的活跃分支，它解决了原项目维护停滞的问题，提供了更稳定、更频繁的更新。用户可以通过配置文件或环境变量进行高度定制，支持字幕下载、后处理和多种输出格式，是媒体归档和本地资源管理的实用助手。","stars":192954,"starGrowth":385,"createdAt":"2020-10-26T04:22:55Z","silentRounds":0,"language":"Python","topics":["youtube-dl","python","sponsorblock","yt-dlp","youtube-downloader","cli","downloader"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["视频下载","音频提取","命令行工具","多站点支持","批量下载"],"domainKey":"视频下载","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"youtube-dl","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"sponsorblock","source":"github","weight":0.7},{"name":"yt-dlp","source":"github","weight":0.7},{"name":"youtube-downloader","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"downloader","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/yt-dlp/yt-dlp","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30598588264560467,"copyOk":true},{"repo":"shardulvs/xnotes-android","owner":"shardulvs","name":"xnotes-android","desc":"A handwriting notes and sketching app for Android with stylus support","summaryCn":"支持手写绘图的安卓记事本，配合手写笔随时随地捕捉灵感。","reasonCn":"xnotes-android 是基于原生 Android 开发的手写笔记与草图应用，利用系统级触摸与压感 API 实现流畅的笔锋渲染，支持多页笔记、橡皮擦、颜色与笔粗选择，还可将手写内容导出为图片或 PDF，满足学习、会议和创意记录需求。","stars":205,"starGrowth":2,"createdAt":"2026-05-23T15:54:45Z","silentRounds":0,"language":"C","topics":["note-taking"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["手写笔记","草图绘制","Android应用"],"domainKey":"手写笔记","pushedAt":"2026-09-23T18:45:07Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/shardulvs/xnotes-android","ts":"2026-09-23T18:45:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"layabox/LayaAir","owner":"layabox","name":"LayaAir","desc":"LayaAir is a fully platform rendering engine with rich 2D/3D rendering capabilities and a mature integrated development platform","summaryCn":"跨平台2D3D游戏引擎，HTML5小游戏首选利器","reasonCn":"核心特点：2D和3D渲染能力都不错，特别是2D，性能优化做得很好；微信小游戏支持是它的强项，很多国内小游戏都是用Laya做的；配套有IDE编辑器，可视化编辑比较方便；支持TypeScript和JavaScript，前端开发者上手快。 安装使用：去LayaAir官网下载安装IDE，或者npm安装引擎包。","stars":2169,"starGrowth":14,"createdAt":"2019-09-02T07:01:45Z","silentRounds":0,"language":"TypeScript","topics":["2d-game-engine","3d-engine","html","html5","javascript","minigames","physics","typescript","webgl","webgl2","webgpu"],"aiDims":["游戏","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","2D渲染","3D渲染","微信小游戏","HTML5","TypeScript"],"domainKey":"游戏引擎","pushedAt":"2026-09-23T10:28:01Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"2d-game-engine","source":"github","weight":0.7},{"name":"3d-engine","source":"github","weight":0.7},{"name":"html","source":"github","weight":0.7},{"name":"html5","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"minigames","source":"github","weight":0.7},{"name":"physics","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"webgl2","source":"github","weight":0.7},{"name":"webgpu","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"bigbro","url":"https://github.com/layabox/LayaAir","ts":"2026-09-23T10:28:01Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.06136376210235132,"copyOk":false},{"repo":"aaif-goose/goose","owner":"aaif-goose","name":"goose","desc":"an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM","summaryCn":"不只是动嘴提建议，还能亲自动手帮你装环境改代码的AI打工人。","reasonCn":"技术亮点在于其极其强大的工具调用能力和沙盒执行环境。基于Rust语言开发，底层运行速度飞快且内存安全。它拥有开放的插件和扩展机制，可以无缝连接各种大语言模型，并赋予AI操作系统级的文件读写和命令行执行权限，让AI真正具备了“双手”。","stars":54568,"starGrowth":28,"createdAt":"2024-08-23T19:03:36Z","silentRounds":0,"language":"Rust","topics":["ai"],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI代理","编程助手","代码生成","命令行界面"],"domainKey":"AI代理","pushedAt":"2026-09-23T21:36:00Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/aaif-goose/goose","ts":"2026-09-23T21:36:00Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08553936372965841,"copyOk":true},{"repo":"Graphify-Labs/graphify","owner":"Graphify-Labs","name":"graphify","desc":"Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.","summaryCn":"把任何代码库变成 AI 能直接问的知识图谱，代码助手革命性升级","reasonCn":"Graphify 把代码库、文档、SQL  schema、PDF 全部解析成知识图谱，让 Claude Code、Cursor、Codex 等 AI 编程助手能\"理解\"整个项目结构，回答\"这个函数在哪里被调用\"、\"这个表有哪些字段\"这类跨文件问题，是 RAG 在编程领域的终极形态。","stars":120549,"starGrowth":487,"createdAt":"2026-04-03T15:49:07Z","silentRounds":0,"language":"Python","topics":["llm","rag"],"aiDims":["知识图谱","RAG与知识","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["知识图谱","代码理解","RAG","AST","AI 编程"],"domainKey":"知识图谱","pushedAt":"2026-09-23T19:25:20Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/Graphify-Labs/graphify","ts":"2026-09-23T19:25:20Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3745434636537839,"copyOk":false},{"repo":"PostgREST/postgrest","owner":"PostgREST","name":"postgrest","desc":"REST API for any Postgres database","summaryCn":"Postgres数据库秒变REST API，零代码后端自动生成","reasonCn":"PostgREST的核心思路非常简单又极其巧妙：它直接读取Postgres的schema定义，把表变成REST端点，把视图变成API，把存储过程变成自定义路由，把数据库权限系统直接映射成API的认证和授权。你只需要在Postgres里配好权限，一个完整的、符合REST规范的API就自动生成了。","stars":27677,"starGrowth":1,"createdAt":"2014-06-13T00:23:00Z","silentRounds":2,"language":"Haskell","topics":["database","api"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["PostgreSQL","REST API生成","零代码后端","数据库中间件"],"domainKey":"PostgreSQL","pushedAt":"2026-09-23T20:15:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"Haskell","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/PostgREST/postgrest","ts":"2026-09-23T20:15:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ppy/osu","owner":"ppy","name":"osu","desc":"rhythm is just a *click* away!","summaryCn":"全球最火的节奏音游，点屏幕跟着音乐打节拍","reasonCn":"由澳大利亚开发者ppy创建的开源音游社区，支持PC、手机等多平台游玩。核心玩法是根据音乐节奏点击屏幕上的圆圈、滑条和Spinner，拥有数百万用户创作的庞大谱面库。游戏强调手眼协调与音乐感知，适合放松心情或挑战极限操作。虽然是纯游戏项目，但其网络同步和反馈机制设计精良，长期保持活跃更新，是休闲放松的绝佳选择。","stars":19126,"starGrowth":7,"createdAt":"2014-07-31T04:26:30Z","silentRounds":0,"language":"C#","topics":["game"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["节奏游戏","音游","多人对战"],"domainKey":"节奏游戏","pushedAt":"2026-09-22T11:37:15Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/ppy/osu","ts":"2026-09-22T11:37:15Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.02209798180487295,"copyOk":true},{"repo":"a16z/jolt","owner":"a16z","name":"jolt","desc":"The simplest and most extensible zkVM. Fast and fully open source from a16z crypto and friends. ⚡","summaryCn":"超轻量的零知识虚拟机，帮你安全跑密码证明，代码几行即可上手。","reasonCn":"Jolt 是由 a16z crypto 团队打造的开源 zkVM，基于 Rust 实现极致轻量和可扩展性，支持自定义零知识电路和递归证明，兼容多种 zkSNARK 后端，提供简单的 SDK 让开发者快速编写、部署并验证 zk 程序，性能比同类框架提升数倍，适合区块链、隐私计算等场景。","stars":1043,"starGrowth":0,"createdAt":"2023-05-18T21:00:08Z","silentRounds":5,"language":"Rust","topics":["cryptography"],"aiDims":["安全工具","开发者工具"],"zone":"AI","funScore":0,"domainTags":["零知识证明","zkVM","区块链基础设施"],"domainKey":"零知识证明","pushedAt":"2026-09-23T21:32:18Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cryptography","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/a16z/jolt","ts":"2026-09-23T21:32:18Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bmad-code-org/BMAD-METHOD","owner":"bmad-code-org","name":"BMAD-METHOD","desc":"Breakthrough Method for Agile Ai Driven Development","summaryCn":"把敏捷全流程拆解为专用 AI Agent 协作完成开发","reasonCn":"BMAD-METHOD 将敏捷开发的各个环节抽象为 Scrum Master、架构师、开发者等专用 AI Agent，通过结构化上下文传递实现从需求分析、架构设计到代码生成的全流程自动化。它不只是生成代码，更保留了决策链路与架构意图，让 AI 协作像真实团队一样有章可循。适合想用 AI 落地复杂项目、而非只写零散代码片段的开发者。","stars":53350,"starGrowth":33,"createdAt":"2025-04-13T14:54:25Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["AI应用","AI基础设施","开发工具"],"zone":"资源","funScore":0,"domainTags":["AI开发","敏捷开发","代码生成","工作流","模板化"],"domainKey":"AI开发","pushedAt":"2026-09-23T21:36:01Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/bmad-code-org/BMAD-METHOD","ts":"2026-09-23T21:36:01Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10085567010309279,"copyOk":true},{"repo":"rohitg00/ai-engineering-from-scratch","owner":"rohitg00","name":"ai-engineering-from-scratch","desc":"Learn it. Build it. Ship it for others.","summaryCn":"从零手搓AI项目的实战秘籍，读完就能拿来用","reasonCn":"这是一个超全的AI工程化开源库，收录了上百个真实项目案例。从数据处理、模型训练、部署到Agent应用，每个项目都有完整源码和详细文档。无论你是想学习RAG、微调，还是想做视觉识别，这里都能找到从零开始的完整教程，是AI学习者的宝藏库。","stars":55269,"starGrowth":31,"createdAt":"2026-03-18T18:38:02Z","silentRounds":0,"language":"Python","topics":["ml","llm","ai","agents","transformers"],"aiDims":["学习资源","教程","模型与训练"],"zone":"资源","funScore":0,"domainTags":["机器学习","大语言模型","AI工程","教程"],"domainKey":"机器学习","pushedAt":"2026-09-23T19:13:09Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/rohitg00/ai-engineering-from-scratch","ts":"2026-09-23T19:13:09Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.11361921511154534,"copyOk":true},{"repo":"CleverRaven/Cataclysm-DDA","owner":"CleverRaven","name":"Cataclysm-DDA","desc":"Cataclysm - Dark Days Ahead. A turn-based survival game set in a post-apocalyptic world.","summaryCn":"硬核末日生存RPG，在丧尸世界里 crafting 求生体验极致的自由","reasonCn":"这是一款开源的回合制末日生存游戏，玩家在丧尸横行的世界中探索、收集资源、制作武器和工具，试图活下去。游戏以极高的自由度著称，你可以种田、养殖、驾驶载具、研究科技，甚至通过手术改造身体。每场游戏都是独特的体验，失败往往意味着重新开始，但过程极其沉浸和有趣。","stars":13218,"starGrowth":7,"createdAt":"2012-09-26T22:57:43Z","silentRounds":0,"language":"C++","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["末日生存","回合制RPG","开放世界","Crafting","独立游戏"],"domainKey":"末日生存","pushedAt":"2026-09-23T21:23:19Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/CleverRaven/Cataclysm-DDA","ts":"2026-09-23T21:23:19Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.022588742623694962,"copyOk":true},{"repo":"aimhubio/aim","owner":"aimhubio","name":"aim","desc":"Aim 💫 — An easy-to-use & supercharged open-source experiment tracker.","summaryCn":"像查Git提交记录一样管理AI实验，告别训练参数记不住的烦恼","reasonCn":"Aim是一个开源的实验跟踪工具，专门解决AI训练中“参数太多记不住、结果乱成一锅粥”的痛点。它支持可视化对比不同实验的指标曲线、超参数和代码版本，操作体验类似Git，让每次实验都有迹可循，特别适合需要频繁调参的开发者。","stars":6260,"starGrowth":1,"createdAt":"2019-05-31T18:25:07Z","silentRounds":2,"language":"Python","topics":["data-visualization","data-science"],"aiDims":["开发者工具","可视化","效率工具"],"zone":"工具","funScore":0,"domainTags":["实验跟踪","机器学习","超参数管理"],"domainKey":"实验跟踪","pushedAt":"2026-09-22T22:33:32Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/aimhubio/aim","ts":"2026-09-22T22:33:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"uswds/uswds","owner":"uswds","name":"uswds","desc":"The U.S. Web Design System helps the federal government build fast, accessible, mobile-friendly websites.","summaryCn":"美国政府官网同款UI组件库，颜值体验双在线","reasonCn":"技术上，USWDS是一套基于Sass/CSS的组件库，包含超过60种预构建UI组件，从按钮、输入框到模态框、日期选择器应有尽有。它遵循TTS（Technology Transformation Services）的设计原则，采用模块化架构，开发者可以按需引入，不必把整个库全打包进来。","stars":7198,"starGrowth":0,"createdAt":"2015-05-20T19:10:10Z","silentRounds":2,"language":"SCSS","topics":["design"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["设计系统","无障碍访问","USWDS","政府数字服务","Sass组件库"],"domainKey":"设计系统","pushedAt":"2026-09-23T00:46:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"SCSS","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/uswds/uswds","ts":"2026-09-23T00:46:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"getpaseo/paseo","owner":"getpaseo","name":"paseo","desc":"Orchestrate multiple coding agents from desktop and mobile","summaryCn":"让多个AI编程助手协同干活，手机电脑都能远程指挥","reasonCn":"这个项目让你能在桌面和手机上同时编排多个AI编程代理，每个代理负责不同任务，比如写代码、查bug、做重构。它们之间共享上下文，互相协作完成复杂开发工作流，不需要写复杂配置就能跑起来，跨平台兼容性好，上手门槛低。","stars":18096,"starGrowth":88,"createdAt":"2025-10-13T12:27:02Z","silentRounds":0,"language":"TypeScript","topics":["mobile","developer-tools","agents"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI Agent","多Agent协作","编程自动化","TypeScript"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:49:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/getpaseo/paseo","ts":"2026-09-23T21:49:20Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.33430663129973476,"copyOk":false},{"repo":"JuliusBrussee/caveman","owner":"JuliusBrussee","name":"caveman","desc":"🪨 why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.","summaryCn":"让AI当原始人说话，帮你大幅省Token费的神工具","reasonCn":"这是一个专为Claude Code设计的技能包，通过提示词工程将AI的对话风格强制转换为原始的穴居人模式。核心逻辑是减少冗余词汇和礼貌用语，只保留核心语义，从而显著压缩上下文长度。这对于频繁调用API且对成本敏感的开发场景极具吸引力，因为它能在保证基本功能理解的前提下，将Token消耗量直接砍掉约65%。虽然玩法略显搞怪，但其背后的压缩思维在长上下文应用中具有实际借鉴意义。","stars":107434,"starGrowth":233,"createdAt":"2026-04-04T10:03:00Z","silentRounds":0,"language":"Go","topics":["llm","ai"],"aiDims":["提示工程","AI Agent","大语言模型"],"zone":"AI","funScore":1,"domainTags":["Claude Code","Skill","提示工程","成本优化","效率工具"],"domainKey":"提示工程","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/JuliusBrussee/caveman","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.36780758419122433,"copyOk":false},{"repo":"dotnet/maui","owner":"dotnet","name":"maui","desc":".NET MAUI is the .NET Multi-platform App UI, a framework for building native device applications spanning mobile, tablet, and desktop.","summaryCn":"一套代码同时跑手机电脑，不用反复重写的跨平台神器","reasonCn":"微软推出的跨平台应用开发框架，用C#和XAML一套代码可发布到iOS、Android、Windows和macOS四大平台。支持热重载快速迭代，内置依赖注入和命令绑定，丰富的原生控件覆盖导航、存储、传感器等能力，配合NuGet包生态能快速接入各类第三方服务。","stars":23318,"starGrowth":0,"createdAt":"2020-05-08T18:11:05Z","silentRounds":2,"language":"C#","topics":["mobile","desktop"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["跨平台应用开发","原生渲染","XAML","C#","热重载","MVVM架构"],"domainKey":"跨平台应用开发","pushedAt":"2026-09-23T19:52:43Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/dotnet/maui","ts":"2026-09-23T19:52:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"itwanger/paicoding","owner":"itwanger","name":"paicoding","desc":"","summaryCn":"一个基于Spring全栈的开源社区系统，可一键部署文章发布与互动。","reasonCn":"paicoding 采用 Spring Boot、MyBatis‑Plus、MySQL、Redis、ElasticSearch、MongoDB、Docker 与 RabbitMQ 等主流技术，提供完整的文章、教程发布、搜索、评论、统计等全流程，并支持一键源码部署，让开发者无需自行搭建即可快速拥有功能齐全的技术社区平台，适合二次开发与实战练习。","stars":3185,"starGrowth":0,"createdAt":"2022-07-06T12:43:21Z","silentRounds":3,"language":"Java","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["社区系统","Spring Boot","MySQL","Docker","Redis"],"domainKey":"社区系统","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/itwanger/paicoding","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Atlas-OS/Atlas","owner":"Atlas-OS","name":"Atlas","desc":"🚀 An open and lightweight modification to Windows, designed to optimize performance, privacy and usability.","summaryCn":"Windows 优化利器，瘦身提速隐私保护一步到位","reasonCn":"Atlas 是一个开源的 Windows 优化脚本集合，旨在提升系统性能、保护隐私并改善用户体验。它通过禁用不必要的后台服务、移除预装软件、调整系统设置来实现。虽然不涉及 AI，但对于使用 Windows 的用户来说，这是一个非常实用的工具，能让电脑运行更流畅，同时减少隐私泄露风险。","stars":21599,"starGrowth":16,"createdAt":"2021-02-08T00:26:23Z","silentRounds":0,"language":"Batchfile","topics":["security","privacy"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Windows优化","隐私保护","系统精简","性能调优","Batchfile"],"domainKey":"Windows优化","pushedAt":"2026-09-23T17:24:45Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"Batchfile","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/Atlas-OS/Atlas","ts":"2026-09-23T17:24:45Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.05022232510764387,"copyOk":false},{"repo":"datawhalechina/hello-agents","owner":"datawhalechina","name":"hello-agents","desc":"📚 《从零开始构建智能体》——从零开始的智能体原理与实践教程","summaryCn":"Datawhale出品的从零构建AI原生智能体系统教程","reasonCn":"这是Datawhale社区发起的系统性智能体学习教程，区别于Dify、Coze等流程驱动的软件工程类Agent，专注于讲解AI Native Agent的核心原理与架构。教程从智能体基础理论出发，深入核心架构与经典范式，指导读者亲手构建多智能体应用，帮助开发者从LLM使用者转型为智能体系统构建者。","stars":80473,"starGrowth":182,"createdAt":"2025-09-07T09:50:24Z","silentRounds":0,"language":"Python","topics":["llm","rag"],"aiDims":["AI Agent","RAG与知识","教程"],"zone":"资源","funScore":0,"domainTags":["AI智能体","教程","RAG","LLM","多智能体"],"domainKey":"AI智能体","pushedAt":"2026-09-22T11:42:57Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/datawhalechina/hello-agents","ts":"2026-09-22T11:42:57Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30678488437115553,"copyOk":true},{"repo":"MakazhanAlpamys/Soup","owner":"MakazhanAlpamys","name":"Soup","desc":"Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.","summaryCn":"只需一个 YAML 文件就能微调 8B 大模型，笔记本也能跑","reasonCn":"Soup 是一款专注于简化大模型微调流程的工具，用户仅需通过一个 YAML 配置文件定义训练策略，无需编写复杂的 Python 脚本即可启动训练。它底层集成了 Layer Streaming 技术，能够将大模型拆解为多个层并行训练，有效降低了显存占用。项目声称在仅有 4GB 显存的笔记本电脑 GPU 上也能成功训练 8B 参数量的模型，极大降低了大模型微调的硬件门槛。","stars":7005,"starGrowth":34,"createdAt":"2026-02-20T11:05:01Z","silentRounds":0,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","微调","大语言模型"],"zone":"AI","funScore":0.7,"domainTags":["大模型微调","YAML配置","显存优化","Layer Streaming"],"domainKey":"大模型微调","pushedAt":"2026-09-23T18:35:48Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/MakazhanAlpamys/Soup","ts":"2026-09-23T18:35:48Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.13987323340471092,"copyOk":true},{"repo":"visgl/deck.gl","owner":"visgl","name":"deck.gl","desc":"WebGL2 powered visualization framework","summaryCn":"用浏览器就能流畅渲染百万级数据点的地图可视化神器","reasonCn":"deck.gl 是基于 WebGL2 构建的高性能地理空间数据可视化框架，专为大规模数据集优化。它内置多种图层类型，支持点、线、面、热力图等多种可视化形式，能与 Mapbox GL JS 和 Three.js 无缝集成。同时提供官方 React 绑定，Uber 内部多个产品都在用，性能表现十分出色，是可视化领域的热门选择。","stars":14604,"starGrowth":1,"createdAt":"2015-12-15T08:38:29Z","silentRounds":2,"language":"TypeScript","topics":["data-visualization","maps"],"aiDims":["可视化","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["WebGL可视化","大规模地理数据","GPU渲染","数据大屏"],"domainKey":"WebGL可视化","pushedAt":"2026-09-23T16:22:12Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"maps","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/visgl/deck.gl","ts":"2026-09-23T16:22:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"intellij-elixir/intellij-elixir","owner":"intellij-elixir","name":"intellij-elixir","desc":"Elixir plugin for JetBrain's IntelliJ Platform (including Rubymine)","summaryCn":"很多 Elixir 开发者在写代码时，需要在 IntelliJ IDEA、RubyMine 等 JetBra","reasonCn":"很多 Elixir 开发者在写代码时，需要在 IntelliJ IDEA、RubyMine 等 JetBrains 系列 IDE 与专门的编辑器之间切换，导致上下文丢失、配置繁琐。IntelliJ‑Elixir 正是为了解决这类痛点而诞生，它把 Elixir 的全部开发体验搬进熟悉的 IDE，让你不必离开工作区就能完成编码、调试和部署","stars":1896,"starGrowth":0,"createdAt":"2014-07-28T03:09:05Z","silentRounds":7,"language":"Kotlin","topics":["editor"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Elixir","IDE插件","JetBrains","LSP"],"domainKey":"Elixir","pushedAt":"2026-09-22T21:02:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/intellij-elixir/intellij-elixir","ts":"2026-09-22T21:02:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"plotly/plotly.js","owner":"plotly","name":"plotly.js","desc":"Open-source JavaScript charting library behind Plotly and Dash","summaryCn":"网页上做炫酷图表，拖拖拽拽就能画出专业级数据可视化","reasonCn":"Plotly.js是一个基于D3.js的开源JavaScript图表库，支持折线图、散点图、3D图表等数十种图表类型。它的最大特点是交互性极强，鼠标悬停能显示数据细节，支持缩放、拖拽、导出图片等功能。与ECharts相比，它在科学计算和金融行业应用更广泛，特别适合嵌入网页的复杂数据可视化场景。","stars":18338,"starGrowth":1,"createdAt":"2015-11-05T23:27:17Z","silentRounds":3,"language":"JavaScript","topics":["data-visualization"],"aiDims":["可视化"],"zone":"工具","funScore":0,"domainTags":["数据可视化","JavaScript","图表库","WebGL"],"domainKey":"数据可视化","pushedAt":"2026-09-22T19:05:29Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/plotly/plotly.js","ts":"2026-09-22T19:05:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"koala73/worldmonitor","owner":"koala73","name":"worldmonitor","desc":"Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface","summaryCn":"用AI实时追踪全球大事的动态大屏，地缘政治变化尽在掌握","reasonCn":"基于TypeScript构建的实时全球情报仪表盘，利用AI自动聚合多源新闻并识别地缘政治关键事件。整合卫星图像、基础设施数据和舆情信息，以可视化大屏形式呈现世界动态。支持自定义监控区域和事件类型，可接入RSS源和API数据。","stars":87234,"starGrowth":66,"createdAt":"2026-01-08T17:26:06Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI应用","可视化","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["数据可视化","新闻聚合","地缘政治","情报监控"],"domainKey":"数据可视化","pushedAt":"2026-09-23T19:21:47Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/koala73/worldmonitor","ts":"2026-09-23T19:21:47Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.24032370864571154,"copyOk":false},{"repo":"SenteLabsAI/OpenExecutive","owner":"SenteLabsAI","name":"OpenExecutive","desc":"AI-powered virtual executive team — a single coherent executive persona backed by 8 specialist Claude agents (FastAPI + Next.js).","summaryCn":"能模拟高管决策并自动审批的企业级AI智能体。","reasonCn":"这是一个基于大模型的AI智能体，旨在解决企业内部审批与知识管理问题。它能接入企业知识库，根据设定好的规则处理费用审批，还能作为高管的“数字双胞胎”，通过聊天窗口帮高管处理邮件、回复消息。项目强调可配置的治理能力，确保AI在互动时符合企业的合规要求。","stars":5131,"starGrowth":27,"createdAt":"2026-06-11T21:41:54Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":1,"domainTags":["多智能体","RAG","决策系统"],"domainKey":"RAG","pushedAt":"2026-09-23T21:44:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/SenteLabsAI/OpenExecutive","ts":"2026-09-23T21:44:46Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.11614367569674527,"copyOk":true},{"repo":"navidrome/navidrome","owner":"navidrome","name":"navidrome","desc":"🎧 Your Personal Streaming Service","summaryCn":"自建个人音乐电台，手机电脑随时听歌，告别版权骚扰和广告打扰","reasonCn":"Navidrome是一款开源的个人音乐流媒体服务器，用Go编写，支持各种现代播放器客户端。它能自动扫描音乐库、生成播放列表、同步歌词，并支持多用户管理和离线下载。Web界面简洁优雅，手机App体验流畅，让你的音乐收藏真正属于自己。不过项目相对小众，社区规模和知名度都比较有限。","stars":23743,"starGrowth":26,"createdAt":"2016-02-24T23:25:12Z","silentRounds":0,"language":"Go","topics":["music","raspberry-pi","self-hosted"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["音乐流媒体","自托管","Subsonic兼容","音频转码"],"domainKey":"音乐流媒体","pushedAt":"2026-09-23T21:04:53Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/navidrome/navidrome","ts":"2026-09-23T21:04:53Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.08128517878953796,"copyOk":false},{"repo":"Infisical/infisical","owner":"Infisical","name":"infisical","desc":"Infisical is the open-source platform for secrets, certificates, and privileged access management.","summaryCn":"开源密钥管理神器，帮团队安全存管密码API密钥再也不裸奔","reasonCn":"Infisical是一个开源的企业级密钥管理平台，支持客户端端到端加密和服务器端加密两种模式。它提供可视化界面和CLI工具，可管理API密钥、数据库凭证、证书等敏感信息，支持GitOps集成和细粒度权限控制，是HashiCorp Vault的轻量级开源替代方案，特别适合中小团队快速落地密钥管理。","stars":29375,"starGrowth":24,"createdAt":"2022-08-05T13:50:00Z","silentRounds":0,"language":"TypeScript","topics":["security"],"aiDims":["安全工具","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["密钥管理","安全","DevOps"],"domainKey":"密钥管理","pushedAt":"2026-09-23T21:41:09Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/Infisical/infisical","ts":"2026-09-23T21:41:09Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.07445106382978722,"copyOk":false},{"repo":"gitleaks/gitleaks","owner":"gitleaks","name":"gitleaks","desc":"Find secrets with Gitleaks 🔑","summaryCn":"代码仓库的秘密探测器，专抓泄露的API密钥和密码，安全审计必备利器。","reasonCn":"Gitleaks是一款基于正则表达式和YAML配置的开源秘密扫描工具，专为Git仓库设计。它能深度扫描整个Git历史，检测API密钥、密码、令牌等敏感信息，并支持集成到CI/CD流程中自动运行。提供丰富的规则集，可自定义规则，帮助开发者在代码提交前发现安全隐患。","stars":29432,"starGrowth":13,"createdAt":"2018-01-27T18:19:31Z","silentRounds":0,"language":"Go","topics":["llm"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["安全审计","密钥检测","CI/CD"],"domainKey":"安全审计","pushedAt":"2026-09-23T04:14:00Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/gitleaks/gitleaks","ts":"2026-09-23T04:14:00Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04032508833922262,"copyOk":false},{"repo":"sgl-project/sglang","owner":"sgl-project","name":"sglang","desc":"SGLang is a high-performance serving framework for large language models and multimodal models.","summaryCn":"专为 LLM 设计的极速推理与部署框架。","reasonCn":"SGLang 是一个高性能的大语言模型推理与部署框架，旨在解决当前模型部署速度慢、资源占用高的问题。它通过优化的调度和内存管理，显著提升推理吞吐量，支持多种模型格式与硬件加速，让开发者能更快地让大模型跑起来并投入实际应用。","stars":36326,"starGrowth":35,"createdAt":"2024-01-08T04:15:52Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["模型部署","推理引擎","大语言模型"],"zone":"AI","funScore":0,"domainTags":["大模型推理","高性能服务","连续批处理"],"domainKey":"大模型推理","pushedAt":"2026-09-23T21:50:14Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/sgl-project/sglang","ts":"2026-09-23T21:50:14Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10789049165886692,"copyOk":true},{"repo":"hugohe3/ppt-master","owner":"hugohe3","name":"ppt-master","desc":"AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He","summaryCn":"输入主题自动生成带图表动画的PPT，AI真成生产力工具了","reasonCn":"基于Python构建的AI自动化办公工具，能将文档或主题直接转换为包含原生形状、过渡动画、动态图表甚至语音旁白的完整演示文稿。它支持数据驱动的内容生成，无需手动排版，大幅降低PPT制作门槛，是职场人和内容创作者的绝佳助手。","stars":55977,"starGrowth":195,"createdAt":"2025-12-10T06:54:33Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI应用","AI写作","创意工具"],"zone":"AI","funScore":0,"domainTags":["AI生成","PPT演示文稿","办公自动化"],"domainKey":"AI生成","pushedAt":"2026-09-23T05:03:40Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/hugohe3/ppt-master","ts":"2026-09-23T05:03:40Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.37254086499812417,"copyOk":false},{"repo":"facebook/rocksdb","owner":"facebook","name":"rocksdb","desc":"A library that provides an embeddable, persistent key-value store for fast storage.","summaryCn":"Facebook出品的高性能嵌入式数据库，读写速度极快","reasonCn":"核心技术亮点在于它的LSM树（Log-Structured Merge-tree）实现。数据按批次写入WAL日志，后台线程异步合并数据文件，实现写操作的高吞吐。支持多种压缩算法（Snappy、ZSTD、LZ4），可按列设置不同的压缩策略。提供快照隔离、增量备份、Compaction调优等高级功能。","stars":32123,"starGrowth":4,"createdAt":"2012-11-30T06:16:18Z","silentRounds":0,"language":"C++","topics":["database"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["嵌入式数据库","键值存储","LSM树"],"domainKey":"嵌入式数据库","pushedAt":"2026-09-23T20:11:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/facebook/rocksdb","ts":"2026-09-23T20:11:43Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rustdesk/rustdesk","owner":"rustdesk","name":"rustdesk","desc":"An open-source remote desktop application designed for self-hosting, as an alternative to TeamViewer.","summaryCn":"自己的电脑就像在身边，远程控制任何设备的一键式远程桌面神器","reasonCn":"RustDesk是用Rust写的开源远程桌面软件，对标TeamViewer和AnyDesk。支持全平台，自建服务器完全掌控数据，不走第三方。H.264编码流畅传输，端到端加密安全，支持文件传输、多显示器、剪贴板同步。安装简单，官网下载即用，社区活跃迭代快，是企业和个人远程办公的靠谱选择。","stars":124282,"starGrowth":89,"createdAt":"2020-09-28T15:36:08Z","silentRounds":0,"language":"Rust","topics":["remote-control","remote-desktop","teamviewer","anydesk","vnc","rdp","rust","p2p","flutter","flutter-apps","dart","rust-lang","wayland","linux","android","flatpak","ios","macos","windows"],"aiDims":["效率工具","开发者工具","安全工具"],"zone":"工具","funScore":0,"domainTags":["远程桌面","远程办公","自托管","Rust","H.264编码"],"domainKey":"远程桌面","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"remote-control","source":"github","weight":0.7},{"name":"remote-desktop","source":"github","weight":0.7},{"name":"teamviewer","source":"github","weight":0.7},{"name":"anydesk","source":"github","weight":0.7},{"name":"vnc","source":"github","weight":0.7},{"name":"rdp","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"flutter","source":"github","weight":0.7},{"name":"flutter-apps","source":"github","weight":0.7},{"name":"dart","source":"github","weight":0.7},{"name":"rust-lang","source":"github","weight":0.7},{"name":"wayland","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"android","source":"github","weight":0.7},{"name":"flatpak","source":"github","weight":0.7},{"name":"ios","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/rustdesk/rustdesk","ts":"2026-09-18T10:23:42.801Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2691483400653353,"copyOk":true},{"repo":"freeCodeCamp/freeCodeCamp","owner":"freeCodeCamp","name":"freeCodeCamp","desc":"freeCodeCamp.org's open-source codebase and curriculum. Learn math, programming, and computer science for free.","summaryCn":"免费开源编程学习平台，含全栈与机器学习认证课程。","reasonCn":"freeCodeCamp 是一个非营利开源社区，提供全栈 Web 开发与机器学习的完整自学课程。平台包含数千道交互式编程挑战，涵盖从基础语法到实战项目的全链路，已助力超过十万名学员成功转行拿到第一份开发工作。所有内容完全免费、自主进度，适合零基础入门与技能进阶。","stars":455980,"starGrowth":69,"createdAt":"2014-12-24T17:49:19Z","silentRounds":0,"language":"TypeScript","topics":["learn-to-code","nonprofits","programming","nodejs","react","d3","careers","education","teachers","javascript","certification","curriculum","math","community","freecodecamp"],"aiDims":["学习资源","非AI-实用","教程"],"zone":"资源","funScore":0,"domainTags":["编程教育","在线课程","教程","前端开发","计算机科学"],"domainKey":"编程教育","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"learn-to-code","source":"github","weight":0.7},{"name":"nonprofits","source":"github","weight":0.7},{"name":"programming","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"d3","source":"github","weight":0.7},{"name":"careers","source":"github","weight":0.7},{"name":"education","source":"github","weight":0.7},{"name":"teachers","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"certification","source":"github","weight":0.7},{"name":"curriculum","source":"github","weight":0.7},{"name":"math","source":"github","weight":0.7},{"name":"community","source":"github","weight":0.7},{"name":"freecodecamp","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/freeCodeCamp/freeCodeCamp","ts":"2026-09-18T10:23:42.801Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.20745396727926663,"copyOk":true},{"repo":"dsh-tauri-desk/deepseek-harness-desktop","owner":"dsh-tauri-desk","name":"deepseek-harness-desktop","desc":"DeepSeek Harness Tauri 桌面版 | Only 5mb installer, zero environment setup, preset plugins, Windows / macOS / Linux.","summaryCn":"免环境配置的桌面版AI助手，内置插件市场与桌宠，开箱即用。","reasonCn":"这是 DeepSeek 模型的桌面客户端，用 Tauri 2 打包成原生轻量应用，下载安装即可在桌面端畅快使用 AI 对话。它自带插件面板，能管理、升级出问题的插件，甚至还能在桌面上养一只 AI 桌宠。安装后自动注册命令行工具，适合想摆脱网页版限制的普通用户和开发者。","stars":2511,"starGrowth":61,"createdAt":"2026-08-14T02:36:29Z","silentRounds":0,"language":"Rust","topics":["desktop"],"aiDims":["AI应用","非AI-实用","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["AI客户端","DeepSeek","Tauri框架","桌面应用","提示词模板"],"domainKey":"AI客户端","pushedAt":"2026-09-15T19:21:35Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/dsh-tauri-desk/deepseek-harness-desktop","ts":"2026-09-15T19:21:35Z","category":"ai","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.3838189964157706,"copyOk":true},{"repo":"QwenLM/qwen-code","owner":"QwenLM","name":"qwen-code","desc":"An open-source AI coding agent that lives in your terminal.","summaryCn":"通义千问出品的终端AI编程助手，在命令行里对话就能搞定代码","reasonCn":"qwen-code的技术亮点在于它支持多种大模型后端——你可以选择OpenAI的GPT系列、Anthropic的Claude，也可以使用通义千问自己的模型，甚至可以在本地部署开源模型。这意味着你可以根据隐私需求、成本预算和性能要求灵活选择。","stars":28071,"starGrowth":27,"createdAt":"2025-06-26T01:37:46Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","llm"],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","终端代理","多模型支持","代码生成","开发者工具"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T21:51:30Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/QwenLM/qwen-code","ts":"2026-09-23T21:51:30Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.08388554023725553,"copyOk":true},{"repo":"microsoft/DirectXTK","owner":"microsoft","name":"DirectXTK","desc":"The DirectX Tool Kit (aka DirectXTK) is a collection of helper classes for writing DirectX 11.x code in C++","summaryCn":"微软官方封装的C++图形开发神器，让3D界面开发不再痛苦","reasonCn":"DirectXTK是微软官方提供的DirectX 11.x辅助库，专为C++开发者设计。它封装了复杂的DirectX底层API，提供了纹理加载、阴影渲染、资源管理等实用工具类。相比手写几千行原生代码，它能大幅简化桌面图形应用开发流程，提高代码可读性与维护性。","stars":2854,"starGrowth":0,"createdAt":"2015-04-09T19:38:28Z","silentRounds":9,"language":"C++","topics":["desktop"],"aiDims":["开发者工具","非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["DirectX","图形开发","C++","游戏引擎"],"domainKey":"DirectX","pushedAt":"2026-09-22T17:15:59Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/microsoft/DirectXTK","ts":"2026-09-22T17:15:59Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"anitaa1990/Android-Cheat-sheet","owner":"anitaa1990","name":"Android-Cheat-sheet","desc":"","summaryCn":"这份仓库最初旨在帮助Android开发者应对技术面试，但由一位拥有十年全栈移动","reasonCn":"项目结构清晰，从初阶的数据结构编码题，逐步过渡到Kotlin高级特性如协程和Flows的实战应用，最后聚焦于当前主流的Jetpack Compose架构。这种循序渐进的设计让读者能直观理解技术演进脉络，而不仅仅是死记硬背面试八股文。","stars":2275,"starGrowth":1,"createdAt":"2018-05-29T14:08:42Z","silentRounds":1,"language":"Java","topics":[],"aiDims":["非AI-好玩","学习资源","开发者工具"],"zone":"资源","funScore":0.2,"domainTags":["Android面试","Kotlin协程","Jetpack Compose","数据结构"],"domainKey":"Android面试","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.44,"source":"tier","url":"https://github.com/anitaa1990/Android-Cheat-sheet","ts":"2026-09-22T21:05:02.747Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alitto/pond","owner":"alitto","name":"pond","desc":"","summaryCn":"在 Go 语言开发中，goroutine 的轻量级特性虽然带来了并发编程的便利，但也","reasonCn":"Pond 的核心价值在于它的“可控性”。它允许开发者精确设定同时运行的最大 goroutine 数量。当任务队列中有请求时，Worker Pool 会从线程池中分配一个空闲的 worker 来处理；处理完毕后，worker 会被释放回池中，而不是销毁。","stars":2193,"starGrowth":0,"silentRounds":0,"language":"Go","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.32,"domainTags":["Go 语言","并发任务","Worker Pool","资源限制"],"domainKey":"Go 语言","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/alitto/pond","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"usestrix/strix","owner":"usestrix","name":"strix","desc":"Open-source AI penetration testing tool to find and fix your app’s vulnerabilities.","summaryCn":"AI驱动的安全测试工具，自动帮你找App漏洞","reasonCn":"Strix 是一个开源的 AI 渗透测试工具，利用大语言模型自动发现和修复应用安全漏洞。它支持多种 Web 框架，内置多种攻击向量，能自动生成详细的漏洞报告和修复建议。支持交互式 CLI 和 API 调用，可集成到 CI/CD 流水线中实现安全左移。GitHub star 超过 5 万，是 AI 安全测试领域的明星项目，让安全测试门槛大幅降低。","stars":64212,"starGrowth":146,"createdAt":"2025-08-05T21:28:30Z","silentRounds":0,"language":"Python","topics":["security","agents"],"aiDims":["安全工具","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI渗透测试","安全漏洞扫描","自动化工具","Web安全"],"domainKey":"AI渗透测试","pushedAt":"2026-09-23T05:09:24Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/usestrix/strix","ts":"2026-09-23T05:09:24Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3068211549243132,"copyOk":true},{"repo":"QuantumNous/new-api","owner":"QuantumNous","name":"new-api","desc":"A unified AI model hub for aggregation & distribution. It supports cross-converting various LLMs into OpenAI-compatible, Claude-compatible, or Gemini-compatible formats. A centralized gateway for personal and enterprise model management.","summaryCn":"把所有AI模型API统一成一个接口，一个网关搞定所有调用","reasonCn":"这是一个AI模型的聚合网关系统，支持将各家LLM平台统一转换成OpenAI、Claude或Gemini兼容格式。底层支持数十种模型供应商，提供统一的API接口管理、流量控制和计费统计。支持跨平台API密钥管理、路由策略配置和详细的用量监控，是搭建多模型AI应用的后端基础设施。","stars":48739,"starGrowth":94,"silentRounds":0,"language":"Go","topics":["claude","gemini","openai","rerank","ai-gateway","deepseek","newapi"],"aiDims":["AI基础设施","工具调用"],"zone":"工具","funScore":0,"domainTags":["AI API网关","多模型聚合","密钥管理","流量路由","OpenAI兼容"],"domainKey":"AI API网关","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"rerank","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"newapi","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/QuantumNous/new-api","ts":"2026-09-23T07:35:20.006Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.28778592092574734,"copyOk":false},{"repo":"amusi/ICCV2025-Papers-with-Code","owner":"amusi","name":"ICCV2025-Papers-with-Code","desc":"","summaryCn":"作为计算机视觉领域最顶级的学术会议之一，ICCV 2025 的论文集不仅代表了当","reasonCn":"在内容组织架构上，该项目展现了极强的针对性与细致度，没有简单罗列，而是将论文细分为 3DGS（高斯泼溅）、Agent、Avatars、ClIP、Mamba、世界模型、扩散模型等数十个细分标签。这种分类方式直接呼应了当下 AI 视觉最火的技术栈，例如在 3DGS 方向，你可以迅速找到关于实时辐射场","stars":2882,"starGrowth":0,"createdAt":"2021-07-23T00:47:16Z","silentRounds":4,"language":"","topics":[],"aiDims":["论文","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["ICCV 2025","高斯泼溅","视觉Transformer"],"domainKey":"ICCV 2025","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"论文","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/amusi/ICCV2025-Papers-with-Code","ts":"2026-09-19T22:19:04.205Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"habitat-sh/habitat","owner":"habitat-sh","name":"habitat","desc":"","summaryCn":"一款将部署、更新与扩缩容自动化逻辑与应用一同打包的开源运维工具。","reasonCn":"Habitat 重新定义了现代软件的打包与部署模式。它允许开发者在编写应用的同时定义其自动化运维行为，并将程序本体与运行所需的策略、更新机制、故障应对和扩缩容逻辑完全捆绑在一起。无论将构建好的产物部署到云端、物理机还是容器中，应用都能具备自主运行和管理的能力。","stars":2754,"starGrowth":0,"createdAt":"2015-05-22T22:47:14Z","silentRounds":1,"language":"Rust","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["应用打包","自动化部署","运维工具","生命周期管理"],"domainKey":"应用打包","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/habitat-sh/habitat","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"caddyserver/caddy","owner":"caddyserver","name":"caddy","desc":"Fast and extensible multi-platform HTTP/1-2-3 web server with automatic HTTPS","summaryCn":"自动给你配HTTPS的Web服务器，省掉一堆麻烦","reasonCn":"Caddy是一款现代化的Web服务器，最大的亮点是自动获取和续期HTTPS证书。配置文件极简，几行代码就能跑起一个支持HTTP/2和HTTP/3的服务器。内置反向代理、负载均衡等功能，比Nginx配置简单太多，是个人开发者和小型团队的首选。","stars":76009,"starGrowth":76,"createdAt":"2015-01-13T19:45:03Z","silentRounds":0,"language":"Go","topics":["privacy","security"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web服务器","反向代理","自动HTTPS"],"domainKey":"Web服务器","pushedAt":"2026-09-23T20:09:48Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/caddyserver/caddy","ts":"2026-09-23T20:09:48Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.23099964477890775,"copyOk":false},{"repo":"upstash/context7","owner":"upstash","name":"context7","desc":"Context7 Platform -- Up-to-date code documentation for LLMs and AI code editors","summaryCn":"让AI和代码编辑器能查到最新代码文档的神器","reasonCn":"Context7专为LLM和AI代码编辑器打造，能实时获取GitHub最新代码文档。传统AI模型知识存在滞后，这个工具通过RAG技术从GitHub拉取文档，让AI始终参考最新API。支持主流框架语言，与Claude Code等工具集成，一行命令即可使用，解决AI写代码时引用过期文档的痛点。","stars":62338,"starGrowth":41,"createdAt":"2025-03-26T23:40:39Z","silentRounds":0,"language":"TypeScript","topics":["llm"],"aiDims":["工具调用","AI基础设施","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["RAG","代码文档","LLM工具","向量检索"],"domainKey":"RAG","pushedAt":"2026-09-23T14:10:55Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/upstash/context7","ts":"2026-09-23T14:10:55Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.12497311431229746,"copyOk":false},{"repo":"milvus-io/milvus","owner":"milvus-io","name":"milvus","desc":"Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search","summaryCn":"给AI装上记忆库，支持十亿级向量的极速相似度搜索","reasonCn":"技术架构上采用云原生设计，计算和存储分离，支持水平扩展。内置多种高效向量索引算法，包括 HNSW、IVF-PQ、DiskANN 等，能在十亿级数据规模下保持毫秒级检索响应。支持纯内存、混合内存磁盘和全磁盘部署模式，适应不同成本需求。","stars":46219,"starGrowth":29,"createdAt":"2019-09-16T06:43:43Z","silentRounds":0,"language":"Go","topics":["vector-db","llm","rag"],"aiDims":["向量数据库","AI基础设施","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["向量数据库","ANN搜索","云原生","RAG基础设施","Go"],"domainKey":"向量数据库","pushedAt":"2026-09-23T20:57:35Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/milvus-io/milvus","ts":"2026-09-23T20:57:35Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08888234275947121,"copyOk":true},{"repo":"OpenRCT2/OpenRCT2","owner":"OpenRCT2","name":"OpenRCT2","desc":"An open source re-implementation of RollerCoaster Tycoon 2 🎢","summaryCn":"开源复刻经典过山车大亨2，还能在上面跑AI算法做自动化运营","reasonCn":"一款高质量的RollerCoaster Tycoon 2开源重制版，支持跨平台运行。它不仅保留了原版游戏的乐趣，还开放了数据接口，吸引了不少AI爱好者用它来训练强化学习算法进行乐园自动化管理，兼具怀旧游戏和AI实验场双重属性。","stars":16245,"starGrowth":1,"createdAt":"2014-04-01T23:32:25Z","silentRounds":1,"language":"C++","topics":["game","simulation"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["模拟经营游戏","过山车大亨","强化学习环境"],"domainKey":"模拟经营游戏","pushedAt":"2026-09-23T21:47:08Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/OpenRCT2/OpenRCT2","ts":"2026-09-23T21:47:08Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"qunitjs/qunit","owner":"qunitjs","name":"qunit","desc":"🔮 An easy-to-use JavaScript unit testing framework.","summaryCn":"前端代码找茬神器，让JavaScript单元测试变得像呼吸一样自然顺畅","reasonCn":"这是一款老牌且极其稳定的JavaScript单元测试框架，核心旨在简化测试编写流程。它提供简洁的API和内置的异步测试支持，无需复杂配置即可集成到项目中。拥有独立的测试运行器界面，能直观展示测试通过或失败状态，极大降低了入门门槛。","stars":4030,"starGrowth":0,"createdAt":"2009-07-24T16:26:23Z","silentRounds":10,"language":"JavaScript","topics":["testing"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["单元测试","JavaScript","测试框架","前端工程化"],"domainKey":"单元测试","pushedAt":"2026-09-22T13:06:08Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/qunitjs/qunit","ts":"2026-09-22T13:06:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"skypjack/entt","owner":"skypjack","name":"entt","desc":"Gaming meets modern C++ - a fast and reliable entity component system (ECS) and much more","summaryCn":"C++游戏开发利器，ECS架构让游戏引擎跑得飞快如闪电","reasonCn":"技术上的最大亮点是它的数据导向设计。entt把组件数据组织成连续内存块，CPU缓存命中率极高。它的注册表系统支持高效的类型擦除和动态组件添加，不需要重新编译就能修改游戏逻辑。与Unity的DOTS或Unreal的ECS实现相比，entt更轻量、更纯粹，没有绑定特定游戏引擎，可以在任何C++项","stars":13132,"starGrowth":3,"createdAt":"2017-03-27T13:15:36Z","silentRounds":0,"language":"C++","topics":["architectural-patterns","cpp","cpp17","cpp20","data-oriented","data-oriented-design","ecs","ecs-framework","entity-component-system","entt","game-dev","game-development","game-engine","game-programming","gamedev","header-only","modern-cpp","no-dependencies","production-ready","reflection"],"aiDims":["游戏","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["ECS架构","C++","游戏开发","数据导向设计"],"domainKey":"ECS架构","pushedAt":"2026-09-23T17:24:32Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"architectural-patterns","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"cpp17","source":"github","weight":0.7},{"name":"cpp20","source":"github","weight":0.7},{"name":"data-oriented","source":"github","weight":0.7},{"name":"data-oriented-design","source":"github","weight":0.7},{"name":"ecs","source":"github","weight":0.7},{"name":"ecs-framework","source":"github","weight":0.7},{"name":"entity-component-system","source":"github","weight":0.7},{"name":"entt","source":"github","weight":0.7},{"name":"game-dev","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"game-programming","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"header-only","source":"github","weight":0.7},{"name":"modern-cpp","source":"github","weight":0.7},{"name":"no-dependencies","source":"github","weight":0.7},{"name":"production-ready","source":"github","weight":0.7},{"name":"reflection","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"bigbro","url":"https://github.com/skypjack/entt","ts":"2026-09-23T17:24:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"elder-plinius/CL4R1T4S","owner":"elder-plinius","name":"CL4R1T4S","desc":"LEAKED SYSTEM PROMPTS FOR CHATGPT, CLAUDE, GEMINI, GROK, PERPLEXITY, CURSOR, LOVABLE, REPLIT, AND MORE! - AI SYSTEMS TRANSPARENCY FOR ALL! 👐","summaryCn":"各大AI模型的提示词泄露合集，窥探系统提示的黑科技","reasonCn":"这个仓库汇集了大量主流AI模型（ChatGPT、Claude、Gemini等）的系统提示词泄露。系统提示是AI行为的核心指令，泄露后能帮你窥探模型底层逻辑，甚至尝试绕过安全限制。兼具技术研究与趣味性，是AI安全探索者和好奇者的必备资料库。","stars":50232,"starGrowth":111,"createdAt":"2025-03-04T20:03:16Z","silentRounds":0,"language":"","topics":["agents","ai","chatgpt","gemini","google","grok","hacking","leak","leaked","openai","prompt","prompt-engineering","prompts","red-team","red-teaming","system","system-info","system-prompts","tools","transparency"],"aiDims":["提示工程","安全工具","非AI-好玩"],"zone":"资源","funScore":1,"domainTags":["系统提示词","AI安全","提示词注入","开源情报"],"domainKey":"系统提示词","pushedAt":"2026-09-22T21:44:42Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"google","source":"github","weight":0.7},{"name":"grok","source":"github","weight":0.7},{"name":"hacking","source":"github","weight":0.7},{"name":"leak","source":"github","weight":0.7},{"name":"leaked","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"prompt","source":"github","weight":0.7},{"name":"prompt-engineering","source":"github","weight":0.7},{"name":"prompts","source":"github","weight":0.7},{"name":"red-team","source":"github","weight":0.7},{"name":"red-teaming","source":"github","weight":0.7},{"name":"system","source":"github","weight":0.7},{"name":"system-info","source":"github","weight":0.7},{"name":"system-prompts","source":"github","weight":0.7},{"name":"tools","source":"github","weight":0.7},{"name":"transparency","source":"github","weight":0.7}],"aiScore":0.85,"source":"bigbro","url":"https://github.com/elder-plinius/CL4R1T4S","ts":"2026-09-22T21:44:42Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30662924032489247,"copyOk":false},{"repo":"openai/codex-plugin-cc","owner":"openai","name":"codex-plugin-cc","desc":"Use Codex from Claude Code to review code or delegate tasks.","summaryCn":"让Claude变成超级编程助手，代码审查和任务委托一把抓","reasonCn":"这是一个用于Claude Code的插件，旨在增强代码审查和任务委托能力。它允许开发者将复杂的编程任务分配给Codex，同时利用Claude本身强大的上下文理解能力进行代码审查。这结合了两种AI模型的优点，能显著提升开发效率和代码质量。","stars":33499,"starGrowth":55,"silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["代码助手","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["代码审查","任务委托","Claude插件"],"domainKey":"代码审查","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/openai/codex-plugin-cc","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.16992552016478105,"copyOk":false},{"repo":"kestra-io/kestra","owner":"kestra-io","name":"kestra","desc":"Event Driven Orchestration & Scheduling Platform for Mission Critical Applications","summaryCn":"可视化编排平台，让你像搭积木一样调度复杂的后端任务流程","reasonCn":"Kestra是一个开源的工作流编排引擎，通过YAML定义任务依赖关系，支持定时触发、事件驱动等多种模式。它的Web界面可以直观看到整个流程的运行状态和日志，还能对接Kubernetes、Docker等基础设施，适合数据管道、ETL和自动化运维场景，让复杂流程变得可视可控。","stars":28264,"starGrowth":53,"createdAt":"2019-08-24T13:56:15Z","silentRounds":0,"language":"Java","topics":["devops","automation"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["工作流编排","任务调度","ETL"],"domainKey":"工作流编排","pushedAt":"2026-09-23T19:24:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/kestra-io/kestra","ts":"2026-09-23T19:24:57Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.1646255307104444,"copyOk":true},{"repo":"matomo-org/matomo","owner":"matomo-org","name":"matomo","desc":"Empowering People Ethically 🚀 — Matomo is hiring! Join us → https://matomo.org/jobs Matomo is the leading open-source alternative to Google Analytics, giving you complete control and built-in privacy. Easily collect, visualise, and analyse data from websites & apps. Star us on GitHub ⭐️  – Pull Requests welcome! ","summaryCn":"开源分析平台替代GA，你的网站数据自己说了算","reasonCn":"Matomo是一款开源的网站分析工具，你可以把它理解为Google Analytics的开源替代品。它的核心理念很简单：你的网站数据应该属于你，而不是被Google拿去用于广告生态。所以Matomo让你把所有数据存在自己服务器上，完全掌控。","stars":21892,"starGrowth":1,"createdAt":"2011-03-30T21:18:17Z","silentRounds":3,"language":"PHP","topics":["security","mobile","privacy"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["网站分析","隐私合规","GDPR"],"domainKey":"网站分析","pushedAt":"2026-09-23T21:10:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"mobile","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/matomo-org/matomo","ts":"2026-09-23T21:10:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"endless-sky/endless-sky","owner":"endless-sky","name":"endless-sky","desc":"Space exploration, trading, and combat game.","summaryCn":"太空贸易战斗游戏，自由探索星域，适合喜欢星战的玩家。","reasonCn":"技术上，它基于C++开发，保证了流畅的游戏体验，支持Windows、macOS和Linux。游戏引擎是自定义的，但注重2D图形和物理模拟，让战斗和贸易系统真实可感。与同类太空游戏如《星际公民》相比，它更轻量级，开源且免费，社区能贡献自定义内容，如新飞船设计或任务脚本，扩展性很强。","stars":7579,"starGrowth":2,"createdAt":"2015-03-14T20:44:39Z","silentRounds":0,"language":"C++","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["太空游戏","贸易模拟","战斗游戏","开源游戏","C++游戏"],"domainKey":"太空游戏","pushedAt":"2026-09-23T17:44:10Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/endless-sky/endless-sky","ts":"2026-09-23T17:44:10Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zhayujie/CowAgent","owner":"zhayujie","name":"CowAgent","desc":"Open-source super AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-model, multi-channel. Lightweight, extensible, one-line install. (formerly chatgpt-on-wechat)","summaryCn":"能自己想办法、还能自动干活儿的超级AI助手，装好就能让它帮你跑任务。","reasonCn":"这是一个开源的AI Agent框架，主打‘自主规划+工具调用’。它能像真人一样拆解任务、调用各种工具（搜网页、跑代码、查数据库等），还自带记忆和知识积累。支持多模型切换，安装极简，适合想快速搭建自动化助手的开发者。","stars":47086,"starGrowth":22,"createdAt":"2022-08-07T08:33:41Z","silentRounds":0,"language":"Python","topics":["ai-agent","llm","ai"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent框架","任务规划","工具调用","记忆管理","多模型支持"],"domainKey":"AI Agent框架","pushedAt":"2026-09-23T16:09:57Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/zhayujie/CowAgent","ts":"2026-09-23T16:09:57Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06740169052372255,"copyOk":false},{"repo":"activepieces/activepieces","owner":"activepieces","name":"activepieces","desc":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","summaryCn":"400多个AI工具随意组合，搭自动化工作流太爽了","reasonCn":"开源AI工作流自动化平台，内置400多个MCP服务器连接各类AI模型和外部工具，支持AI Agent自主执行多步骤任务。通过可视化拖拽界面即可搭建从简单通知到复杂AI Agent的各种自动化流程，无需复杂编程。支持多种大模型接入，自托管保障数据安全，是AI时代普通用户也能上手的自动化神器。","stars":24669,"starGrowth":23,"createdAt":"2022-12-03T02:52:46Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["工作流自动化","AI Agent","MCP协议","低代码平台","自托管"],"domainKey":"AI Agent","pushedAt":"2026-09-23T20:46:27Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/activepieces/activepieces","ts":"2026-09-23T20:46:27Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.07179703271312174,"copyOk":true},{"repo":"elalish/manifold","owner":"elalish","name":"manifold","desc":"","summaryCn":"在3D建模与制造领域，非流形网格导致的崩溃和错误一直是痛点，Manifold 正是","reasonCn":"除了可靠性，该库在性能上下足功夫，不仅支持多线程并行计算，还在单线程极限下利用流水线技术优化算法效率。对于需要进行复杂结构分析或大规模网格处理的应用来说，这意味着更短的等待时间和更流畅的交互体验。库还支持任意顶点属性映射，方便开发者在渲染着色或物理模拟中附加自","stars":2285,"starGrowth":20,"createdAt":"2019-02-27T05:01:59Z","silentRounds":0,"language":"C++","topics":[],"aiDims":["非AI-实用","可视化","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["3D几何","流形网格","CAD建模","制造业"],"domainKey":"3D几何","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/elalish/manifold","ts":"2026-09-22T21:05:02.747Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.08625820568927789,"copyOk":false},{"repo":"Alex-Rachel/TEngine","owner":"Alex-Rachel","name":"TEngine","desc":"Unity 商用级别开发框架，原生内置 AI 工作流支持，集成 HybridCLR 高性能热更、Obfuz 代码混淆加固、YooAssets 企业级资源管理方案，构建高效、安全、可扩展的工业化开发底座。","summaryCn":"用Unity造游戏的工业化神器，自带AI工作流还能热更","reasonCn":"TEngine是基于Unity的商用级游戏开发框架，采用ECS架构设计，内置AI工作流支持。集成HybridCLR实现C#热更新无需重打包，Obfuz提供代码混淆加固，YooAssets负责企业级资源管理。模块化架构让团队协作更高效，适合中大型游戏项目的工业化生产流程。","stars":2354,"starGrowth":3,"createdAt":"2022-05-17T06:23:11Z","silentRounds":0,"language":"C#","topics":["csharp","ecs","framework","game","game-development","gameframework","hotupdate","huatuo","hybridclr","obfuz","unity","unity3d","unity3d-hotfix","unityframework"],"aiDims":["游戏","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["游戏开发","游戏引擎","ECS 架构"],"domainKey":"游戏开发","pushedAt":"2026-09-22T05:06:09Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"ecs","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"game","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"gameframework","source":"github","weight":0.7},{"name":"hotupdate","source":"github","weight":0.7},{"name":"huatuo","source":"github","weight":0.7},{"name":"hybridclr","source":"github","weight":0.7},{"name":"obfuz","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"unity3d-hotfix","source":"github","weight":0.7},{"name":"unityframework","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"bigbro","url":"https://github.com/Alex-Rachel/TEngine","ts":"2026-09-22T05:06:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Murmele/Gittyup","owner":"Murmele","name":"Gittyup","desc":"Understand your Git history!","summaryCn":"把 Git 历史变成可视化的时间轴，一目了然看懂代码演变。","reasonCn":"核心上，它直接读取 Git 仓库内部的对象数据库，不需要联网，也不需要重新拉取代码。它提取每个提交的哈希值、作者、日期以及父子关系，然后在一个终端窗口里画出时间轴。分支合并的地方会画出叉号，不同颜色的线条代表不同的开发路径，这让查看代码演变历史变得像看漫画一样轻松。","stars":2279,"starGrowth":0,"createdAt":"2019-08-25T08:41:10Z","silentRounds":2,"language":"C++","topics":["git"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Git","可视化","终端工具","代码历史"],"domainKey":"Git","pushedAt":"2026-09-23T20:27:04Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Murmele/Gittyup","ts":"2026-09-23T20:27:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"earendil-works/pi","owner":"earendil-works","name":"pi","desc":"AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI","summaryCn":"让终端也能跑AI智能体，统一调用各大模型的瑞士军刀","reasonCn":"基于TypeScript构建的AI智能体工具包，提供统一LLM API接口，支持OpenAI、Anthropic等主流模型无缝切换。内置智能体循环机制，能自动拆解复杂任务并规划执行链。配备交互式终端界面，还集成代码助手功能，让开发者在命令行就能调度AI能力。","stars":108526,"starGrowth":311,"createdAt":"2025-08-09T14:03:50Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","命令行","LLM API","编程助手","TypeScript"],"domainKey":"AI Agent","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/earendil-works/pi","ts":"2026-09-22T04:07:30.369Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3085970182260472,"copyOk":false},{"repo":"promptfoo/promptfoo","owner":"promptfoo","name":"promptfoo","desc":"Test your prompts, agents, and RAGs. Red teaming/pentesting/vulnerability scanning for AI. Compare performance of GPT, Claude, Gemini, DeepSeek, and more. Simple declarative configs with command line and CI/CD integration. Used by OpenAI and Anthropic.","summaryCn":"AI提示词测试与红队安全评估神器，让大模型输出更稳定可靠","reasonCn":"核心架构基于YAML/JSON声明式配置，你可以描述测试用例、预期输出、评分规则，然后框架自动在多个模型（GPT-4、Claude、Gemini、DeepSeek等）上运行并对比结果。支持确定性评分和LLM-as-judge两种评估方式。独特的是它的红队测试功能，能自动生成对抗性提示词来探测模型的安全漏洞","stars":25372,"starGrowth":28,"createdAt":"2023-04-28T15:48:49Z","silentRounds":0,"language":"TypeScript","topics":["rag","llm"],"aiDims":["提示工程","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["提示词工程","LLM评测","红队测试","AI安全","CI/CD集成"],"domainKey":"提示词工程","pushedAt":"2026-09-23T19:14:23Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/promptfoo/promptfoo","ts":"2026-09-23T19:14:23Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.08731073624467918,"copyOk":true},{"repo":"goauthentik/authentik","owner":"goauthentik","name":"authentik","desc":"The authentication glue you need.","summaryCn":"一站式身份认证管理神器，统一管所有系统的登录，告别密码焦虑","reasonCn":"技术上，Authentik基于Python和Go构建，后端用PostgreSQL存储数据，前端用React。它支持SAML 2.0、OAuth2、OIDC、LDAP等多种协议，意味着几乎所有主流系统都能接入。内置的用户流程引擎支持多因素认证、密码重置、社交登录等完整身份生命周期管理。","stars":25686,"starGrowth":20,"createdAt":"2019-12-30T09:19:48Z","silentRounds":0,"language":"Python","topics":["security","kubernetes"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["身份认证","安全","单点登录"],"domainKey":"身份认证","pushedAt":"2026-09-23T21:37:14Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/goauthentik/authentik","ts":"2026-09-23T21:37:14Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.06233590282644242,"copyOk":false},{"repo":"fallintodusk/alis","owner":"fallintodusk","name":"alis","desc":"Open-source Unreal Engine 5 world-generation framework and survival-game prototype. Real geography rebuilt as playable, streamed worlds through a modular, data-driven pipeline.","summaryCn":"用 UE5 实现的地理真实世界生成框架，支持流式可玩世界与生存游戏原型","reasonCn":"基于 Unreal Engine 5 的模块化地理生成管线，实时流式加载真实地理数据，生成可玩世界。框架整合地形、植被、天气等系统，支持快速原型。配套的生存游戏原型演示采集、建造与生存玩法，展示从数据到玩法的完整闭环。","stars":18,"starGrowth":0,"createdAt":"2026-03-11T16:13:06Z","silentRounds":6,"language":"C++","topics":["procedural-generation"],"aiDims":["游戏","创意工具"],"zone":"工具","funScore":0.7,"domainTags":["游戏开发","程序生成","地理数据","Unreal Engine"],"domainKey":"游戏开发","pushedAt":"2026-09-21T12:36:15Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.48,"source":"search","url":"https://github.com/fallintodusk/alis","ts":"2026-09-21T12:36:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"warp-tech/warpgate","owner":"warp-tech","name":"warpgate","desc":"Fully transparent SSH, HTTPS, Kubernetes, database and RDP/VNC bastion/PAM that doesn't need additional client-side software","summaryCn":"零配置的SSH和数据库堡垒机，不用装客户端也能安全运维","reasonCn":"Warpgate是一个现代化的Bastion主机（堡垒机），支持SSH、HTTPS、Kubernetes、MySQL和PostgreSQL协议。它的最大亮点是“透明性”——用户无需安装任何客户端软件，通过浏览器即可安全访问后端资源。基于Rust编写，兼顾了高性能与安全性，是运维团队的神器。","stars":7952,"starGrowth":8,"createdAt":"2022-02-05T13:30:51Z","silentRounds":0,"language":"Rust","topics":["bastion","bastion-host","infrastructure","ssh","ssh-server","proxy","rust","https","https-proxy","mysql","mysql-proxy","postgresql-proxy","pam","privileged-access-management","kubectl","kubernetes","rdp","rdp-access","rdp-gateway","vnc-server"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["SSH","运维","堡垒机"],"domainKey":"SSH","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"bastion","source":"github","weight":0.7},{"name":"bastion-host","source":"github","weight":0.7},{"name":"infrastructure","source":"github","weight":0.7},{"name":"ssh","source":"github","weight":0.7},{"name":"ssh-server","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"https","source":"github","weight":0.7},{"name":"https-proxy","source":"github","weight":0.7},{"name":"mysql","source":"github","weight":0.7},{"name":"mysql-proxy","source":"github","weight":0.7},{"name":"postgresql-proxy","source":"github","weight":0.7},{"name":"pam","source":"github","weight":0.7},{"name":"privileged-access-management","source":"github","weight":0.7},{"name":"kubectl","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"rdp","source":"github","weight":0.7},{"name":"rdp-access","source":"github","weight":0.7},{"name":"rdp-gateway","source":"github","weight":0.7},{"name":"vnc-server","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/warp-tech/warpgate","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02701810865191147,"copyOk":true},{"repo":"citrolabs/ego-lite","owner":"citrolabs","name":"ego-lite","desc":"The fastest browser for AI agents to run browser automation, built for sharing your logged-in browser state with your AI agents, like Codex or Claude Code, without disturbing you. Zero cost, zero config.","summaryCn":"让AI代理开浏览器搞事情，还能共享你的登录状态，完全不打扰你干活。","reasonCn":"这个项目专门解决AI Agent控制浏览器的痛点，核心亮点是能无缝共享用户已登录的浏览器状态。这意味着AI助手可以直接用你的账号操作网页，比如自动填写表单、抓取数据，而不需要你手动登录或提供敏感Token，既安全又高效，特别适合处理复杂的网页自动化任务。","stars":16381,"starGrowth":38,"createdAt":"2026-04-16T12:51:09Z","silentRounds":0,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","效率工具"],"zone":"AI","funScore":0,"domainTags":["浏览器自动化","AI代理","会话共享"],"domainKey":"浏览器自动化","pushedAt":"2026-09-23T07:16:19Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/citrolabs/ego-lite","ts":"2026-09-23T07:16:19Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.14515113851413222,"copyOk":false},{"repo":"addyosmani/agent-skills","owner":"addyosmani","name":"agent-skills","desc":"Production-grade engineering skills for AI coding agents.","summaryCn":"为AI编程助手提供生产级工程技能包，标准化高质量开发流程。","reasonCn":"该项目将资深工程师的最佳实践封装为技能包，包含9个覆盖全生命周期的Slash命令。核心功能/build auto支持一次审批后自主完成规划与实现，每个任务仍独立测试驱动并提交，遇风险自动暂停。它解决了AI编程助手流程混乱、质量不可控的问题，让AI按统一标准交付生产级代码。","stars":98442,"starGrowth":253,"createdAt":"2026-02-15T20:20:26Z","silentRounds":0,"language":"JavaScript","topics":["agent-skills","antigravity","claude-code","cursor","skills","codex"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["AI编程助手","代码审查","重构","单元测试","Cursor"],"domainKey":"AI编程助手","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/addyosmani/agent-skills","ts":"2026-09-22T04:07:30.369Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.36925214847321264,"copyOk":true},{"repo":"allinurl/goaccess","owner":"allinurl","name":"goaccess","desc":"GoAccess is a real-time web log analyzer and interactive viewer that runs in a terminal in *nix systems or through your browser.","summaryCn":"终端里的网页日志分析神器，实时查看谁在访问你的网站","reasonCn":"一款开源的实时Web日志分析工具，支持Apache、Nginx等主流格式，能在终端里直接展示访问统计、流量趋势和错误分析。无需复杂配置，一行命令即可启动交互式仪表盘，还能导出HTML报告分享给团队。相比商业方案，它零依赖、内存占用极低，特别适合服务器资源有限的小团队和开发者日常监控使用。","stars":20943,"starGrowth":4,"createdAt":"2013-07-14T03:42:08Z","silentRounds":0,"language":"C","topics":["command-line","monitoring","privacy"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Web日志分析","终端仪表盘","实时流量统计"],"domainKey":"Web日志分析","pushedAt":"2026-09-19T02:33:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/allinurl/goaccess","ts":"2026-09-19T02:33:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dair-ai/ML-Notebooks","owner":"dair-ai","name":"ML-Notebooks","desc":"","summaryCn":"“The notebooks are meant to be minimal, easily reusable, and extendabl","reasonCn":"为了彻底消除环境配置这一最大痛点，项目深度整合了 GitHub Codespaces。用户无需在本地安装 Anaconda、调试依赖冲突或下载巨大的数据集，只需点击仓库页面的代码按钮，即可在云端浏览器中拥有一个预装了所有必要依赖的 Jupyter 环境。","stars":3439,"starGrowth":2,"createdAt":"2022-03-26T21:39:49Z","silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0.2,"domainTags":["机器学习","Jupyter 笔记本","教学示例"],"domainKey":"机器学习","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/dair-ai/ML-Notebooks","ts":"2026-09-18T22:32:16.581Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SBoudrias/Inquirer.js","owner":"SBoudrias","name":"Inquirer.js","desc":"A collection of common interactive command line user interfaces.","summaryCn":"你有没有用过一些命令行工具，运行之后不是直接出一堆选项让你选，就是给你","reasonCn":"Inquirer.js之所以强大，是因为它把常见的命令行交互模式都封装成了简单易用的API。你可以用它创建单选多选菜单、文本输入、密码输入、确认提示、进度条、列表选择……而且支持自定义模板、验证规则、动态选项。很多知名项目的CLI都是用它搭建的，比如Vue CLI、Create React App等。","stars":21626,"starGrowth":0,"createdAt":"2013-05-17T01:41:36Z","silentRounds":3,"language":"TypeScript","topics":["command-line"],"aiDims":[],"zone":"工具","funScore":0,"domainTags":["命令行界面","交互式UI","Node.js","TypeScript","CLI工具"],"domainKey":"命令行界面","pushedAt":"2026-09-20T23:23:33Z","tags":[{"name":"command-line","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/SBoudrias/Inquirer.js","ts":"2026-09-20T23:23:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"multica-ai/multica","owner":"multica-ai","name":"multica","desc":"Make humans and AI agents work as one team — open-source and self-hostable.","summaryCn":"多AI编程代理协作工作区，像管理团队一样分配任务。","reasonCn":"Multica 是一个自托管的 AI 编程代理协作平台，解决了开发者需同时操作多个终端、反复解释上下文的痛点。它把 GitHub Issue 变成任务看板，支持 26 种主流 CLI 代理（如 Claude Code、Codex）像队友一样自主认领任务、汇报进度、提交代码供审查。会话上下文持久化，避免重复沟通，让人类从繁琐的“保姆式”监管中解脱出来。","stars":51135,"starGrowth":101,"createdAt":"2026-01-13T17:59:46Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"AI","funScore":0.7,"domainTags":["AI 编程","多智能体","GitHub"],"domainKey":"AI 编程","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/multica-ai/multica","ts":"2026-09-22T04:07:30.369Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.367110589615723,"copyOk":true},{"repo":"lobehub/lobehub","owner":"lobehub","name":"lobehub","desc":"🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.","summaryCn":"把AI代理招进公司当员工，全自动轮班干活，你只管收报告","reasonCn":"LobeHub是一个AI Agent管理平台，能把多个AI代理像员工一样统一管理。支持招聘、排班、任务调度、工作汇报等完整流程，实现7×24小时全天候运转。支持多种模型接入，可通过Web界面轻松配置不同代理的职责和协作关系，大幅降低多Agent系统的管理复杂度，适合需要自动化工作流的技术团队。","stars":82763,"starGrowth":30,"createdAt":"2023-05-21T07:19:12Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI Agent","效率工具","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent编排","多Agent协作管理","可视化工作流"],"domainKey":"AI Agent编排","pushedAt":"2026-09-23T18:59:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/lobehub/lobehub","ts":"2026-09-23T18:59:36Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09108744245617002,"copyOk":false},{"repo":"kubernetes-sigs/kueue","owner":"kubernetes-sigs","name":"kueue","desc":"Kubernetes-native Job Queueing","summaryCn":"Kubernetes 批处理作业调度与资源管理控制器。","reasonCn":"Kueue 是一个基于 Kubernetes 的作业队列控制器，它负责决定何时让作业开始运行以及何时停止。它通过优先级队列和资源切片来管理作业，支持 Fair Sharing（公平共享）和抢占等高级策略，确保集群资源被高效利用。它整合了 Kubernetes、Kubeflow、Ray 等多种工作负载，为用户提供了统一的资源调度视角。","stars":2996,"starGrowth":3,"createdAt":"2022-02-16T05:52:19Z","silentRounds":0,"language":"Go","topics":["k8s-sig-scheduling","k8s","kubernetes"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","任务队列","GPU调度"],"domainKey":"Kubernetes","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"k8s-sig-scheduling","source":"github","weight":0.7},{"name":"k8s","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/kubernetes-sigs/kueue","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"TeamAmaze/AmazeFileManager","owner":"TeamAmaze","name":"AmazeFileManager","desc":"Material design file manager for Android","summaryCn":"安卓Material设计文件管理器，界面清爽整理文件超顺手","reasonCn":"基于Kotlin开发的Material Design风格Android文件管理器，支持文件浏览、ZIP解压压缩、FTP/SMB网络共享访问、根目录权限操作。内置多种颜色主题、手势快捷操作和丰富的文件预览能力，覆盖图片音视频文档等多种格式，轻量高效。","stars":6399,"starGrowth":2,"createdAt":"2014-07-14T19:47:53Z","silentRounds":0,"language":"Kotlin","topics":["mobile","frontend"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Android文件管理","Material Design","本地存储","网络文件传输"],"domainKey":"Android文件管理","pushedAt":"2026-09-22T02:53:06Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/TeamAmaze/AmazeFileManager","ts":"2026-09-22T02:53:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nodejs/node","owner":"nodejs","name":"node","desc":"Node.js JavaScript runtime ✨🐢🚀✨","summaryCn":"用JavaScript写后端服务器，让全栈开发变得超简单","reasonCn":"Node.js是基于Google V8引擎的JavaScript运行时，让前端代码能在服务器端直接运行。采用事件驱动和非阻塞I/O架构，天然适合高并发场景，无论是搭建REST API、WebSocket实时应用还是微服务网关，都能以极低的资源消耗处理数千个并发连接。","stars":122052,"starGrowth":33,"language":"JavaScript","topics":["nodejs","javascript","node","js","runtime","mit","linux","macos","windows"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript","服务器端","事件驱动","非阻塞I/O","后端开发"],"domainKey":"JavaScript","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"nodejs","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"node","source":"github","weight":0.7},{"name":"js","source":"github","weight":0.7},{"name":"runtime","source":"github","weight":0.7},{"name":"mit","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"trending","url":"https://github.com/nodejs/node","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09981112968243044,"copyOk":true},{"repo":"mudler/LocalAI","owner":"mudler","name":"LocalAI","desc":"LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.","summaryCn":"在自家电脑上跑各种AI模型的开源引擎，不用GPU也能玩转大模型","reasonCn":"技术上它采用C++和Go混合开发，核心推理引擎来自llama.cpp等项目，经过深度优化。支持GGUF格式的量化模型，4位量化后普通CPU也能跑起来。兼容OpenAI的API接口格式，意味着你现有的AI应用几乎不用改代码就能切换成本地部署。","stars":49225,"starGrowth":22,"createdAt":"2023-03-18T22:58:02Z","silentRounds":0,"language":"Go","topics":["agents","ai","llm"],"aiDims":["推理引擎","模型部署","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["本地大模型推理","GGUF量化","OpenAI兼容API","CPU推理优化","多模态模型部署"],"domainKey":"本地大模型推理","pushedAt":"2026-09-23T21:12:39Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/mudler/LocalAI","ts":"2026-09-23T21:12:39Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06734078212290504,"copyOk":true},{"repo":"maximhq/bifrost","owner":"maximhq","name":"bifrost","desc":"Fastest enterprise AI gateway (50x faster than LiteLLM) with adaptive load balancer, cluster mode, guardrails, 1000+ models support & <100 µs overhead at 5k RPS.","summaryCn":"企业级 AI 网关，比传统方案快50倍，支持1000+模型。","reasonCn":"Bifrost 是一款高性能的 AI 网关，专为解决企业级 AI 部署的痛点设计。它提供了自适应负载均衡、集群模式、护栏等功能，能同时支持 1000+ 大模型，处理 5000 RPS 时延迟低至 100 微秒，是连接企业应用与 AI 模型的核心基础设施。","stars":8286,"starGrowth":34,"language":"Go","topics":["gateway","llm","llm-gateway","mcp-client","mcp-server","guardrails","mcp-gateway","ai-gateway","gateway-services","generative-ai","llm-cost","llm-observability","llmops","load-balancing","token-management","model-router"],"aiDims":["AI基础设施","模型部署","效率工具"],"zone":"AI","funScore":0,"domainTags":["API Gateway","Load Balancing","LLM"],"domainKey":"API Gateway","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"gateway","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"llm-gateway","source":"github","weight":0.7},{"name":"mcp-client","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"guardrails","source":"github","weight":0.7},{"name":"mcp-gateway","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"gateway-services","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"llm-cost","source":"github","weight":0.7},{"name":"llm-observability","source":"github","weight":0.7},{"name":"llmops","source":"github","weight":0.7},{"name":"load-balancing","source":"github","weight":0.7},{"name":"token-management","source":"github","weight":0.7},{"name":"model-router","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/maximhq/bifrost","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.11430992034757423,"copyOk":false},{"repo":"Axorax/awesome-free-apps","owner":"Axorax","name":"awesome-free-apps","desc":"Curated list of the best free apps for PC and mobile","summaryCn":"免费软件大全：精选PC和手机实用工具，从系统到创作一应俱全","reasonCn":"这是一个精心整理的免费软件合集项目，收录了PC和移动平台上各类实用免费工具，从系统优化到图像编辑、从开发工具到办公套件，分类清晰、更新频繁。虽然和AI没有直接关系，但作为资源导航站非常实用，适合经常找免费替代付费软件的技术爱好者，目前已获七千多Star认可。","stars":7718,"starGrowth":0,"createdAt":"2024-10-24T03:42:18Z","silentRounds":2,"language":"JavaScript","topics":["mobile","awesome"],"aiDims":["非AI-实用"],"zone":"资源","funScore":0,"domainTags":["免费软件","资源导航","awesome list","软件推荐"],"domainKey":"免费软件","pushedAt":"2026-09-23T16:54:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Axorax/awesome-free-apps","ts":"2026-09-23T16:54:21Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jrnl-org/jrnl","owner":"jrnl-org","name":"jrnl","desc":"Collect your thoughts and notes without leaving the command line.","summaryCn":"命令行里记笔记的神器，极简高效让灵感瞬间沉淀","reasonCn":"这是一个运行在终端里的极简日记本，支持Markdown语法和多种标签。它无需图形界面，所有操作通过命令行完成，速度极快且不受打扰。支持加密存储保护隐私，还能导出为JSON或Markdown格式。对于习惯键盘操作的开发者或追求高效记录的用户来说，它是替代传统笔记软件的绝佳选择，真正实现随时随地捕捉想法。","stars":7322,"starGrowth":1,"createdAt":"2012-03-29T17:23:58Z","silentRounds":3,"language":"Python","topics":["command-line","productivity","note-taking"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["命令行笔记","终端日记","Markdown","本地加密存储","Python"],"domainKey":"命令行笔记","pushedAt":"2026-09-22T02:25:34Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"note-taking","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/jrnl-org/jrnl","ts":"2026-09-22T02:25:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"slimtoolkit/slim","owner":"slimtoolkit","name":"slim","desc":"Slim(toolkit): Don't change anything in your container image and minify it by up to 30x (and for compiled languages even more) making it secure too! (free and open source)","summaryCn":"Docker镜像一键瘦身30倍，安全扫描也顺手做了","reasonCn":"基于OCI标准开发的容器镜像优化工具，支持多阶段构建、分层裁剪和重复层去重。能自动检测镜像漏洞并生成安全报告，无需修改Dockerfile即可原地瘦身。对Go、Java等编译型语言的压缩效果尤为显著，可显著提升镜像传输和部署效率。","stars":23415,"starGrowth":1,"createdAt":"2015-09-09T17:39:11Z","silentRounds":4,"language":"Go","topics":["security"],"aiDims":["开发者工具","安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Docker镜像","容器优化","安全扫描"],"domainKey":"Docker镜像","pushedAt":"2026-09-19T23:40:59Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/slimtoolkit/slim","ts":"2026-09-19T23:40:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rmyndharis/OpenWA","owner":"rmyndharis","name":"OpenWA","desc":"Free, Open Source, Self-Hosted WhatsApp API Gateway","summaryCn":"把WhatsApp变成智能客服机器人，完全自己控制数据","reasonCn":"OpenWA 是一个开源免费的 WhatsApp API 网关，允许你在本地自建服务，替代昂贵的官方 API。基于 TypeScript 开发，支持 Node.js，提供完整的 REST API 和 WebSocket 接口。它不仅能接收消息，还能通过简单的配置触发 AI Agent 或自动化脚本，实现智能回复、客户咨询、自动通知等场景。无需依赖第三方云服务，数据完全掌握在自己手中，是构建企业级消息应用的最佳底座。","stars":14487,"starGrowth":29,"createdAt":"2026-02-02T12:14:45Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted","api"],"aiDims":["AI应用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["WhatsApp","API网关","消息路由","自托管","自动化"],"domainKey":"WhatsApp","pushedAt":"2026-09-23T13:00:58Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/rmyndharis/OpenWA","ts":"2026-09-23T13:00:58Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.11160646096500311,"copyOk":false},{"repo":"HKUDS/nanobot","owner":"HKUDS","name":"nanobot","desc":"Ultra-lightweight, open-source, self-hosted personal AI agent framework in Python with WebUI, tools, memory, MCP, multi-agent workflows, automation, and chat apps","summaryCn":"超轻量本地AI助手，自带记忆工具和自动化能力","reasonCn":"主打轻量级和个人使用的AI Agent框架，完全支持本地部署。内置WebUI、记忆管理、MCP协议支持和多智能体工作流。Python开发，一键启动即可拥有自己的私人AI助理，能调用各种工具、处理多模态任务，且注重隐私安全，极具可玩性。","stars":48492,"starGrowth":23,"createdAt":"2026-02-01T07:16:15Z","silentRounds":0,"language":"Python","topics":["ai-agent","chatbot"],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","本地部署","模型上下文协议"],"domainKey":"AI Agent","pushedAt":"2026-09-23T19:44:27Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/HKUDS/nanobot","ts":"2026-09-23T19:44:27Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08450749814402377,"copyOk":false},{"repo":"Tencent-TDS/KuiklyUI","owner":"Tencent-TDS","name":"KuiklyUI","desc":"A Kotlin Multiplatform UI framework from Tencent TDS — high-performance, one codebase for six platforms, with dynamic delivery.","summaryCn":"一套代码搞定六端，腾讯打造的高效跨平台UI框架","reasonCn":"腾讯TDS团队推出的Kotlin Multiplatform跨平台UI框架，支持iOS、Android、Web、Desktop、macOS、Windows六大平台，一套代码生成全端应用。采用声明式UI范式，性能接近原生，支持动态下发能力，适合追求高效开发的移动开发者。","stars":3528,"starGrowth":2,"createdAt":"2025-04-24T03:50:37Z","silentRounds":0,"language":"Kotlin","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kotlin","UI框架","跨平台","移动端开发"],"domainKey":"Kotlin","pushedAt":"2026-09-23T09:26:25Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Tencent-TDS/KuiklyUI","ts":"2026-09-23T09:26:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"theonedev/onedev","owner":"theonedev","name":"onedev","desc":"The Unified and Autonomous Development Platform","summaryCn":"一站式代码托管加CI平台，GitLab的轻量级替代方案","reasonCn":"The One Dev是一个开源的DevOps平台，集成了代码托管、CI/CD、问题追踪等功能于一体。相比GitLab更轻量，界面简洁，上手门槛低。支持Git仓库管理、流水线自动化构建部署、代码审查和团队协作文档，适合中小团队快速搭建私有开发平台。自带智能代码补全和自动化工作流，提升开发效率。","stars":15257,"starGrowth":0,"createdAt":"2018-11-06T02:57:01Z","silentRounds":2,"language":"Java","topics":["devops","self-hosted"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["代码托管","CI/CD","DevOps平台","问题追踪","代码审查"],"domainKey":"代码托管","pushedAt":"2026-09-23T13:05:52Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/theonedev/onedev","ts":"2026-09-23T13:05:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"BurntSushi/ripgrep","owner":"BurntSushi","name":"ripgrep","desc":"ripgrep recursively searches directories for a regex pattern while respecting your gitignore","summaryCn":"代码搜索神器，比grep快十倍搜索体验拉满","reasonCn":"ripgrep是一款超快的命令行文本搜索工具，用Rust编写，支持递归搜索目录、忽略.gitignore中的文件、支持正则表达式，性能远超传统grep。它专为开发者设计，集成到IDE和命令行中都能大幅提升搜索效率。项目star数超过6.7万，是Rust生态中最受欢迎的命令行工具之一，几乎所有程序员都值得拥有。","stars":68512,"starGrowth":21,"createdAt":"2016-03-11T02:02:33Z","silentRounds":0,"language":"Rust","topics":["ripgrep","recursively-search","search","regex","gitignore","grep","command-line-tool","command-line","cli","rust"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行搜索","正则表达式","代码搜索","开发者工具","文本处理"],"domainKey":"命令行搜索","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ripgrep","source":"github","weight":0.7},{"name":"recursively-search","source":"github","weight":0.7},{"name":"search","source":"github","weight":0.7},{"name":"regex","source":"github","weight":0.7},{"name":"gitignore","source":"github","weight":0.7},{"name":"grep","source":"github","weight":0.7},{"name":"command-line-tool","source":"github","weight":0.7},{"name":"command-line","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/BurntSushi/ripgrep","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06391954694068193,"copyOk":true},{"repo":"opensandbox-group/OpenSandbox","owner":"opensandbox-group","name":"OpenSandbox","desc":"Secure, Fast, and Extensible Sandbox runtime for AI agents.","summaryCn":"给AI Agent装上隔离舱，代码跑崩了也不怕炸机","reasonCn":"这个项目为AI Agent提供了一个安全的沙箱运行时环境，让Agent在执行代码、调用工具时不会把主机搞崩。它支持多种编程语言隔离执行，能精细控制文件读写、网络访问等权限。对于经常让AI写代码执行的工具来说，这相当于给Agent穿了防弹衣，既能让它放开手脚干活，又保证了主机安全。","stars":15472,"starGrowth":29,"createdAt":"2025-12-17T08:41:09Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","安全工具","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI Agent沙箱","代码执行隔离","安全运行时"],"domainKey":"AI Agent沙箱","pushedAt":"2026-09-23T15:25:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/opensandbox-group/OpenSandbox","ts":"2026-09-23T15:25:20Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.11114767321613235,"copyOk":false},{"repo":"huggingface/speech-to-speech","owner":"huggingface","name":"speech-to-speech","desc":"Build voice agents with open-source models","summaryCn":"想让你的程序变得像人一样能实时听懂你说的话，并且还能用自然的声音秒回你","reasonCn":"核心架构采用了端到端的流式处理技术，这意味着它不需要等待整段话讲完才开始响应，而是能一边听、一边思考、一边组织语言，从而把响应速度压到极致。它兼容HuggingFace上众多的开源语音模型（如Whisper、Parler-TTS等），这意味着你可以根据自己的需求，灵活更换声音风格和识别准","stars":13305,"starGrowth":23,"createdAt":"2024-08-07T15:32:09Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["多模态Agent","模型与训练"],"zone":"AI","funScore":0,"domainTags":["语音交互","实时对话","流式处理","开源AI","HuggingFace"],"domainKey":"语音交互","pushedAt":"2026-09-23T13:15:49Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/huggingface/speech-to-speech","ts":"2026-09-23T13:15:49Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.07418602029312289,"copyOk":false},{"repo":"GoogleChrome/lighthouse","owner":"GoogleChrome","name":"lighthouse","desc":"Automated auditing, performance metrics, and best practices for the web.","summaryCn":"网页质量与性能审计工具，自动检测并评分。","reasonCn":"Lighthouse 是一个用于网页审计的开源工具，它通过自动化手段分析网页的性能、可访问性、SEO 等多个维度，并提供优化建议。开发者可以将其嵌入 Chrome 浏览器、作为命令行工具使用，或者直接在 Node.js 环境中运行，从而快速了解网站当前的健康状况和改进方向。","stars":30798,"starGrowth":0,"createdAt":"2016-03-08T01:03:11Z","silentRounds":1,"language":"JavaScript","topics":["developer-tools"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网页性能审计","前端工程","SEO检测","CI/CD集成"],"domainKey":"网页性能审计","pushedAt":"2026-09-20T12:06:48Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/GoogleChrome/lighthouse","ts":"2026-09-20T12:06:48Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"csprance/gecs","owner":"csprance","name":"gecs","desc":"Godot Entity Component System - GECS","summaryCn":"专为 Godot 4.6+ 设计的 ECS 引擎，让游戏开发更清晰。","reasonCn":"GECS 是专为 Godot 4.6 及以上版本打造的实体组件系统（ECS）。它通过将数据与逻辑分离，帮助开发者构建可扩展且易维护的游戏项目。项目不仅完美融入 Godot 的节点体系，还提供了强大的查询过滤能力，让游戏架构更加整洁。","stars":612,"starGrowth":0,"createdAt":"2024-10-30T03:38:08Z","silentRounds":2,"language":"GDScript","topics":["godot"],"aiDims":["游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Godot引擎","ECS架构","游戏开发框架","GDScript","实体组件系统"],"domainKey":"Godot引擎","pushedAt":"2026-09-19T00:46:10Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/csprance/gecs","ts":"2026-09-19T00:46:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"cloudflare/pingora","owner":"cloudflare","name":"pingora","desc":"A library for building fast, reliable and evolvable network services.","summaryCn":"Cloudflare造的瑞士军刀，帮你写出又快又稳的网络服务程序","reasonCn":"Rust编写的高性能网络服务库，专为极端流量场景设计。采用异步架构，支持热更新、负载均衡和自定义协议，很多核心组件已用于生产环境。对追求极致性能的系统开发者来说，是构建下一代网关、代理或API服务的神器，学习曲线稍陡但回报巨大。","stars":27512,"starGrowth":8,"createdAt":"2023-05-05T20:03:08Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["开发者工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["网络服务","网关","代理","负载均衡","Rust"],"domainKey":"网络服务","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/cloudflare/pingora","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02487234661238732,"copyOk":false},{"repo":"SillyTavern/SillyTavern","owner":"SillyTavern","name":"SillyTavern","desc":"LLM Frontend for Power Users.","summaryCn":"把大模型变成二次元老婆或虚拟角色的究极前端","reasonCn":"核心亮点在于其强大的角色管理系统和“世界书”功能。世界书就像一个动态的百科全书，当你提到某个特定词汇时，系统会自动将相关的背景设定发送给 AI，确保 AI 不会在对话中出现逻辑漏洞。此外，它支持多种前端扩展，比如可以给角色配上会动的立绘，甚至接入语音合成，让虚拟角色不仅","stars":33687,"starGrowth":16,"createdAt":"2023-02-09T10:19:24Z","silentRounds":0,"language":"JavaScript","topics":["llm","ai"],"aiDims":["AI应用","提示工程","非AI-好玩"],"zone":"AI","funScore":1,"domainTags":["LLM","角色扮演","虚拟形象","语音合成","前端交互"],"domainKey":"LLM","pushedAt":"2026-09-23T21:19:13Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/SillyTavern/SillyTavern","ts":"2026-09-23T21:19:13Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.049424882001959214,"copyOk":true},{"repo":"getzep/graphiti","owner":"getzep","name":"graphiti","desc":"Build Real-Time Knowledge Graphs for AI Agents","summaryCn":"为AI Agent构建实时知识图谱，让智能体更懂你。","reasonCn":"Graphiti是一个开源库，帮助AI Agent动态构建和维护知识图谱。它整合检索增强生成（RAG）技术，让智能体能从多源数据中抽取实体和关系，实时更新图谱，增强上下文理解与推理能力。代码友好，支持主流LLM和向量数据库，适合开发智能对话系统。","stars":31081,"starGrowth":21,"createdAt":"2024-08-08T22:08:30Z","silentRounds":0,"language":"Python","topics":["rag","agents"],"aiDims":["AI Agent","知识图谱","RAG与知识"],"zone":"AI","funScore":0.6,"domainTags":["知识图谱","Agent","RAG","LLM","Neo4j"],"domainKey":"知识图谱","pushedAt":"2026-09-23T13:17:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/getzep/graphiti","ts":"2026-09-23T13:17:40Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06502696180946559,"copyOk":true},{"repo":"healthchecks/healthchecks","owner":"healthchecks","name":"healthchecks","desc":"Open-source cron job and background task monitoring service, written in Python & Django","summaryCn":"盯紧你的定时任务和后台作业，别再等崩了才发现","reasonCn":"healthchecks是一个开源的定时任务和后台作业监控服务，用Python+Django开发。它帮你监控cron任务、后台工作流和CI/CD流水线是否正常执行，任务失败或超时时会通过邮件、Slack、微信等平台发送告警。自带Web仪表盘，历史运行状态一目了然，支持REST API自定义集成，适合中小团队快速部署使用。","stars":10354,"starGrowth":3,"createdAt":"2015-06-25T18:38:33Z","silentRounds":0,"language":"Python","topics":["devops","monitoring"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["定时任务监控","后台作业告警","Cron健康检查"],"domainKey":"定时任务监控","pushedAt":"2026-09-21T15:56:36Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/healthchecks/healthchecks","ts":"2026-09-21T15:56:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yihong0618/running_page","owner":"yihong0618","name":"running_page","desc":"Make your own running home page","summaryCn":"自动生成炫酷跑步主页，让每一次奔跑都有专属展示空间","reasonCn":"一款创意十足的自动化工具，通过GitHub Actions自动同步Strava或Apple Watch等平台的跑步数据，生成精美的个人跑步主页。支持自定义主题、地图轨迹展示和详细运动统计，将枯燥的运动记录变成可分享的视觉作品。适合喜欢跑步且乐于折腾的极客用户。","stars":4528,"starGrowth":2,"createdAt":"2020-09-17T05:59:22Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-好玩"],"zone":"工具","funScore":1,"domainTags":["跑步数据","数据可视化","GitHub Actions","Strava","个人主页"],"domainKey":"跑步数据","pushedAt":"2026-09-23T06:23:05Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/yihong0618/running_page","ts":"2026-09-23T06:23:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/WindowsAppSDK","owner":"microsoft","name":"WindowsAppSDK","desc":"The Windows App SDK empowers all Windows desktop apps with modern Windows UI, APIs, and platform features, including back-compat support, shipped via NuGet.","summaryCn":"微软官方出的Windows应用升级工具箱，让你用现代方式重新包装老程序","reasonCn":"微软Windows团队维护的桌面应用开发框架，通过NuGet包分发，帮助开发者将Win32、WPF、WinForms等传统应用迁移到现代化Windows平台。支持最新的Fluent UI、WinUI 3控件，封装Windows Runtime API，兼容Windows 10/11所有版本，是微软官方推荐的桌面应用现代化升级方案。","stars":4690,"starGrowth":1,"createdAt":"2020-04-15T22:16:56Z","silentRounds":3,"language":"C++","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Windows开发","桌面应用框架","WinUI3","Fluent UI","NuGet包管理"],"domainKey":"Windows开发","pushedAt":"2026-09-23T21:27:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/microsoft/WindowsAppSDK","ts":"2026-09-23T21:27:57Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"superplanehq/superplane","owner":"superplanehq","name":"superplane","desc":"Open source factory for one-shot engineering","summaryCn":"给AI Agent装上大脑的中控台，让干活更丝滑","reasonCn":"这是一个用Go语言编写的开源控制平面，专门为AI Agent编排设计。它提供了统一的API接口，让Agent可以调用各种外部工具和API，同时具备自动重试、限流熔断和审计日志等企业级特性。它通过声明式配置管理Agent的工作流，极大降低了构建复杂Agent系统的难度。","stars":7548,"starGrowth":18,"createdAt":"2025-05-07T17:48:39Z","silentRounds":0,"language":"Go","topics":["devops","automation"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent编排","控制平面","工具调用管理","企业级Agent框架"],"domainKey":"AI Agent编排","pushedAt":"2026-09-23T21:25:02Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/superplanehq/superplane","ts":"2026-09-23T21:25:02Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.06115421303656598,"copyOk":false},{"repo":"matrixorigin/matrixone","owner":"matrixorigin","name":"matrixone","desc":"AI-native HTAP database with Git-for-Data and built-in vector search, serving as the data and memory backbone for intelligent agents and applications.","summaryCn":"AI应用的幕后大脑，存数据还能做语义搜索的数据库","reasonCn":"MatrixOne是一个AI原生数据库，支持HTAP混合负载，可在同一引擎中同时处理事务和分析查询。亮点是内置向量搜索能力，无需额外部署向量数据库，直接支持语义检索和相似性查询，让AI Agent和应用可以直接用它作为记忆和知识存储。","stars":1987,"starGrowth":25,"createdAt":"2021-02-18T09:07:17Z","silentRounds":0,"language":"Go","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI Agent"],"zone":"AI","funScore":0,"domainTags":["HTAP数据库","向量检索","AI原生数据库","Git-for-Data","混合事务分析"],"domainKey":"HTAP数据库","pushedAt":"2026-09-23T20:14:27Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/matrixorigin/matrixone","ts":"2026-09-23T20:14:27Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.1127453447408153,"copyOk":true},{"repo":"lissy93/web-check","owner":"lissy93","name":"web-check","desc":"🕵️‍♂️ All-in-one OSINT tool for analysing any website","summaryCn":"扒开网站底裤的全能侦探，一键分析任意网站的全部底细","reasonCn":"GitHub Star 34000+的开源Web安全分析神器，本地部署即可对任意网站进行全方位指纹识别。支持检测DNS记录、SSL证书、HTTP头、技术栈、端口扫描、Whois信息、CMS识别等20+项指标，可视化界面直观呈现，是渗透测试和网站审计的必备工具。","stars":34899,"starGrowth":11,"createdAt":"2023-06-25T14:32:11Z","silentRounds":0,"language":"TypeScript","topics":["security","privacy"],"aiDims":["安全工具","非AI-好玩","非AI-实用"],"zone":"工具","funScore":1,"domainTags":["OSINT","网站指纹","安全审计","DNS查询","SSL检测"],"domainKey":"OSINT","pushedAt":"2026-09-20T21:08:24Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/lissy93/web-check","ts":"2026-09-20T21:08:24Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03394558583340497,"copyOk":false},{"repo":"biolab/orange3","owner":"biolab","name":"orange3","desc":"🍊 :bar_chart: :bulb: Orange: Interactive data analysis","summaryCn":"不用写代码，拖拖拽拽就能做数据可视化分析的神器","reasonCn":"Orange3是一款基于Python的开源数据可视化与分析工具，采用节点式图形界面工作流设计。用户无需编写代码，通过拖拽连接各种小部件即可完成数据清洗、特征选择、聚类、分类、回归等任务。内置500多个可视化组件，支持从CSV、Excel到数据库等多种数据源，是数据科学家和爱好者探索数据的利器。","stars":5708,"starGrowth":3,"createdAt":"2013-02-22T12:52:56Z","silentRounds":0,"language":"Python","topics":["data-visualization","data-science"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["数据可视化","拖拽式分析","机器学习","节点工作流","Python"],"domainKey":"数据可视化","pushedAt":"2026-09-22T10:08:26Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/biolab/orange3","ts":"2026-09-22T10:08:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Chimahon/chimahon","owner":"Chimahon","name":"chimahon","desc":"Mihon immersion fork with native dictionary lookup (Yomitan), Mokuro manga and anime support, Novel EPUB reader, and instant Anki card mining.","summaryCn":"一站式看动漫、阅读轻小说，还能秒生成背单词卡的移动神器。","reasonCn":"此项目基于Mihon的沉浸式阅读客户端，内置本地词典查询，支持Mokuro漫画、动漫章节以及EPUB小说，阅读时可一键提取生词并自动生成Anki卡片，省去手动记忆的繁琐。界面简洁，支持离线缓存，兼容Android平台。并可通过插件扩展功能，如自动翻译或同步至云笔记。","stars":206,"starGrowth":0,"createdAt":"2026-03-27T13:58:01Z","silentRounds":3,"language":"Kotlin","topics":["anime"],"aiDims":["非AI-好玩","非AI-实用","学习资源"],"zone":"工具","funScore":0.6,"domainTags":["漫画阅读器","轻小说阅读","语言学习辅助","Anki制卡"],"domainKey":"漫画阅读器","pushedAt":"2026-09-22T15:31:37Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/Chimahon/chimahon","ts":"2026-09-22T15:31:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Ontos-AI/knowhere","owner":"Ontos-AI","name":"knowhere","desc":"Knowhere extracts, parses, and outputs structured chunks ready for AI Agents and RAG.","summaryCn":"把杂乱文档变成向量库，喂给AI搞知识问答。","reasonCn":"技术上，Knowhere集成了多种业界领先的向量索引算法，如IVF、HNSW等，并针对ARM架构做了深度优化，在CPU上的检索速度远超同类Python库。它还支持自定义Schema解析，能把PDF、Word、Markdown等不同格式的文档智能拆分成结构化的数据块，直接喂给向量引擎。","stars":3449,"starGrowth":20,"createdAt":"2026-04-30T03:13:10Z","silentRounds":0,"language":"Python","topics":["vector-db"],"aiDims":["向量数据库","RAG与知识","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["向量检索","RAG","索引算法","知识库"],"domainKey":"RAG","pushedAt":"2026-09-23T13:18:26Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Ontos-AI/knowhere","ts":"2026-09-23T13:18:26Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.09287561612061467,"copyOk":false},{"repo":"simonlin1212/TradingAgents-astock","owner":"simonlin1212","name":"TradingAgents-astock","desc":"A股多Agent投研框架 — 适配A股数据源(龙虎榜/游资/解禁等)，7位分析师基于A股规则的辩论决策，基于TradingAgents深度改造，适配大A。A-share multi-agent investment research framework — 7 AI analysts, bull/bear debate, risk assessment。","summaryCn":"七个 AI 分析师为你激烈辩论 A 股怎么买","reasonCn":"专门为 A 股市场定制的多 Agent 投研框架。内置 7 位扮演不同角色的 AI 分析师，针对龙虎榜、游资动向、解禁等特有数据源进行多轮辩论。通过牛熊对峙和风险评估机制，将碎片化的 A 股数据转化为结构化的投资决策参考。","stars":3518,"starGrowth":92,"createdAt":"2026-05-13T00:22:30Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["A股量化投研","多Agent辩论","金融AI应用"],"domainKey":"A股量化投研","pushedAt":"2026-09-21T00:30:21Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/simonlin1212/TradingAgents-astock","ts":"2026-09-21T00:30:21Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.42534440022740194,"copyOk":true},{"repo":"flutter/flutter","owner":"flutter","name":"flutter","desc":"Flutter makes it easy and fast to build beautiful apps for mobile and beyond","summaryCn":"一套代码同时生成iOS和安卓App，告别两头适配的烦恼","reasonCn":"它的技术架构很有意思，底层用的是 Skia 图形引擎，和 Chrome 浏览器同款，不依赖系统原生控件，所以渲染速度极快且视觉高度一致。核心抽象是 Widget——一切皆组件，按钮、布局、动画全部由 Widget 组合而成。另一个杀手锏是热重载（Hot Reload），改完代码点保存，界面几乎瞬间刷新","stars":179047,"starGrowth":12,"createdAt":"2015-03-06T22:54:58Z","silentRounds":0,"language":"Dart","topics":["mobile","desktop"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["跨平台UI框架","移动应用开发","Dart语言","Skia渲染引擎","热重载开发"],"domainKey":"跨平台UI框架","pushedAt":"2026-09-23T21:35:52Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/flutter/flutter","ts":"2026-09-23T21:35:52Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03620106452495713,"copyOk":false},{"repo":"sqlitebrowser/sqlitebrowser","owner":"sqlitebrowser","name":"sqlitebrowser","desc":"Official home of the DB Browser for SQLite (DB4S) project. Previously known as \"SQLite Database Browser\" and \"Database Browser for SQLite\". Website at: ","summaryCn":"不用写代码就能管理SQLite数据库，图形界面比敲命令爽多了","reasonCn":"SQLite数据库的官方桌面管理工具，提供完整的可视化操作界面。支持浏览表结构、执行SQL查询、导入导出数据、设计表模型等功能。跨平台支持Windows、Mac和Linux，是数据库开发者的瑞士军刀，尤其适合轻量级项目快速操作数据库。","stars":24616,"starGrowth":2,"createdAt":"2014-05-04T01:54:44Z","silentRounds":0,"language":"C++","topics":["database"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["SQLite","数据库管理","可视化查询","跨平台桌面应用","Qt"],"domainKey":"SQLite","pushedAt":"2026-09-23T00:40:18Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/sqlitebrowser/sqlitebrowser","ts":"2026-09-23T00:40:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fschutt/azul","owner":"fschutt","name":"azul","desc":"Desktop GUI Framework","summaryCn":"用Rust写的现代桌面GUI框架，性能强跨平台支持好","reasonCn":"Azul是一款用Rust语言开发的现代桌面GUI框架，采用软件渲染架构，不依赖系统原生控件，因此跨平台表现一致。它基于事件驱动架构，提供声明式的UI构建方式，支持硬件加速渲染和响应式布局。Azul专注于性能和内存安全，适合需要高性能桌面应用开发的场景，填补了Rust生态中成熟GUI框架的空白。","stars":6117,"starGrowth":0,"createdAt":"2018-01-16T20:57:07Z","silentRounds":21,"language":"Rust","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["GUI","Rust","Desktop","Cross-Platform"],"domainKey":"GUI","pushedAt":"2026-09-23T19:00:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/fschutt/azul","ts":"2026-09-23T19:00:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"harry0703/MoneyPrinterTurbo","owner":"harry0703","name":"MoneyPrinterTurbo","desc":"利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.","summaryCn":"AI一键生成短视频脚本素材，自动合成带字幕配乐。","reasonCn":"MoneyPrinterTurbo 是一款一站式 AI 短视频生成工具，它解决了创作者找素材、写脚本、配字幕的繁琐流程。只需输入主题，它就能自动产出脚本、匹配画面、生成字幕并配上背景音乐，最后合成高清视频，极大降低了短视频制作门槛。","stars":125146,"starGrowth":173,"createdAt":"2024-03-11T02:57:34Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["AI 视频生成","短视频制作","大语言模型","自动化工作流"],"domainKey":"大语言模型","pushedAt":"2026-09-21T12:52:09Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/harry0703/MoneyPrinterTurbo","ts":"2026-09-21T12:52:09Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3041471561216499,"copyOk":true},{"repo":"kubernetes/kubernetes","owner":"kubernetes","name":"kubernetes","desc":"Production-Grade Container Scheduling and Management","summaryCn":"容器编排界的神，让服务器集群像搭积木一样管理。","reasonCn":"Kubernetes是目前最流行的容器编排系统，解决了大规模容器部署、调度和管理的难题。它采用声明式API，通过YAML文件定义资源状态，自动处理故障恢复和负载均衡。在AI领域，它常被用于构建高性能的GPU调度集群，确保算力资源被高效分配给训练任务。","stars":127909,"starGrowth":29,"createdAt":"2014-06-06T22:56:04Z","silentRounds":0,"language":"Go","topics":["kubernetes"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["容器编排","云原生","DevOps","Kubernetes"],"domainKey":"容器编排","pushedAt":"2026-09-23T21:42:17Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/kubernetes/kubernetes","ts":"2026-09-23T21:42:17Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08768017105911233,"copyOk":false},{"repo":"azerothcore/azerothcore-wotlk","owner":"azerothcore","name":"azerothcore-wotlk","desc":"Complete Open Source and Modular solution for MMO","summaryCn":"搭一个属于自己的魔兽世界服务器，免费畅玩经典怀旧服","reasonCn":"AzothCore是魔兽世界WotLK版本的开源服务器框架，完整复刻了经典游戏核心玩法。提供数据库、认证、游戏服务器等全套模块，支持Docker一键部署。社区活跃持续维护，众多成熟私服都基于此搭建。无论是想怀旧重温还是二次开发，这里都是理想起点。","stars":8959,"starGrowth":11,"createdAt":"2016-06-26T08:40:05Z","silentRounds":0,"language":"C++","topics":["game","emulator"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["MMO服务器","魔兽世界","游戏模拟器"],"domainKey":"MMO服务器","pushedAt":"2026-09-23T21:30:45Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"emulator","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/azerothcore/azerothcore-wotlk","ts":"2026-09-23T21:30:45Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.03668344681326041,"copyOk":true},{"repo":"Clooos/Bubble-Card","owner":"Clooos","name":"Bubble-Card","desc":"Bubble Card is a minimalist and customizable card collection for Home Assistant, featuring modern pop-ups and an integrated Module Store with over 100 community-made modules.","summaryCn":"让智能家居界面颜值翻倍的卡片组件库，100多个模块随便挑","reasonCn":"专为Home Assistant打造的极简风格卡片集合，支持现代弹出交互和主题自定义。内置Module Store集成100+社区模块，可自由组合灯光、开关、传感器等控件，彻底告别原生界面千篇一律的单调感，打造个性化智能中控面板。","stars":4595,"starGrowth":5,"createdAt":"2023-08-18T11:22:10Z","silentRounds":0,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","效率工具","可视化"],"zone":"工具","funScore":0.5,"domainTags":["智能家居","用户界面","卡片组件"],"domainKey":"智能家居","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/Clooos/Bubble-Card","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018264417845484222,"copyOk":false},{"repo":"zeromicro/go-zero","owner":"zeromicro","name":"go-zero","desc":"A cloud-native Go microservices framework with cli tool for productivity.","summaryCn":"Go语言写的云原生微服务框架，快速搭建企业级后端系统","reasonCn":"这是一个用Go语言开发的云原生微服务框架，内置代码生成工具和CLI，帮助开发者快速构建高并发、高性能的后端服务。支持gRPC、RESTful API、服务注册发现、限流熔断等微服务治理能力，代码生成器可以根据接口定义自动生成全部业务代码。虽然和AI没有直接关系，但作为Go后端开发者的实用工具，star超过3万，社区活跃，是国产微服务框架中的佼佼者。","stars":33346,"starGrowth":4,"createdAt":"2020-08-07T15:37:57Z","silentRounds":0,"language":"Go","topics":["web-framework"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["微服务框架","Go语言","代码生成","服务治理"],"domainKey":"微服务框架","pushedAt":"2026-09-22T01:31:38Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/zeromicro/go-zero","ts":"2026-09-22T01:31:38Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"agentscope-ai/QwenPaw","owner":"agentscope-ai","name":"QwenPaw","desc":"Your Personal AI Assistant; easy to install, deploy on your own machine or on the cloud; supports multiple chat apps with easily extensible capabilities.","summaryCn":"把通义千问装进电脑，随时随地和你的AI助手聊天","reasonCn":"技术上，QwenPaw采用Agent架构设计，核心是一个可扩展的对话引擎，支持多种大语言模型后端切换。它内置了工具调用框架，可以通过插件机制接入各种能力，比如搜索、代码执行、文件操作等。项目采用了模块化设计，插件机制让能力扩展变得非常灵活。","stars":35215,"starGrowth":22,"createdAt":"2026-02-24T03:42:56Z","silentRounds":0,"language":"Python","topics":["ai-agent","chatbot"],"aiDims":["AI Agent","AI应用","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["Agent框架","大模型部署","聊天机器人","插件生态"],"domainKey":"Agent框架","pushedAt":"2026-09-23T10:59:31Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/agentscope-ai/QwenPaw","ts":"2026-09-23T10:59:31Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08144904160159024,"copyOk":false},{"repo":"mnfst/awesome-free-llm-apis","owner":"mnfst","name":"awesome-free-llm-apis","desc":"List of Permanent Free LLM API (API Keys)","summaryCn":"整理了好多永久免费的AI大模型API，省去翻遍全网找资源的烦恼。","reasonCn":"它的核心价值在于“筛选”和“聚合”。开发者不用再去官网一个个注册申请，也不用担心有些API已经下线或者限流严格，这里都整理好了状态。它支持文本生成、对话、代码补全等多种功能，对想要快速验证想法或者做Demo的个人开发者非常友好。","stars":8192,"starGrowth":91,"language":"JavaScript","topics":["awesome"],"aiDims":["AI基础设施","大语言模型","AI应用"],"zone":"资源","funScore":0,"domainTags":["LLM","API","免费资源","开发者工具"],"domainKey":"LLM","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/mnfst/awesome-free-llm-apis","ts":"2026-09-23T21:53:43.262Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.3063251953125,"copyOk":false},{"repo":"justcallmekoko/ESP32Marauder","owner":"justcallmekoko","name":"ESP32Marauder","desc":"A suite of WiFi/Bluetooth offensive and defensive tools for the ESP32","summaryCn":"把网络攻防工具装进开发板，即插即用的黑科技黑客玩具。","reasonCn":"这是一个基于 ESP32 开发板的 WiFi 和蓝牙攻防工具集，把复杂的网络安全技术变成了可以握在手里的硬件玩具。它通过开源固件实现了信号嗅探、欺骗和中间人攻击等实用功能，让普通用户也能在本地网络环境中体验网络安全攻防的乐趣，非常适合对网络安全感兴趣的极客动手尝试。","stars":12426,"starGrowth":11,"createdAt":"2019-09-13T18:37:40Z","silentRounds":0,"language":"C++","topics":["command-line","arduino"],"aiDims":["非AI-好玩","安全工具","可视化"],"zone":"创意","funScore":0.9,"domainTags":["网络安全","WiFi","蓝牙","硬件开发","渗透测试"],"domainKey":"网络安全","pushedAt":"2026-09-17T02:33:59Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"arduino","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/justcallmekoko/ESP32Marauder","ts":"2026-09-17T02:33:59Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.035655721873491066,"copyOk":true},{"repo":"seemoo-lab/nexmon","owner":"seemoo-lab","name":"nexmon","desc":"","summaryCn":"nexmon是一个开源的固件补丁框架，使用C语言编写，专门针对Broadcom和Cypre","reasonCn":"这个项目最吸引人的地方在于它扩展了WiFi芯片的能力边界。官方还提供了一个名为nexmon/jammer的附加项目，它把一个普通的Wi-Fi芯片变成了软件定义无线电设备。你可以用它在空中生成任意的干扰信号，覆盖正在进行的帧传输。","stars":2888,"starGrowth":0,"createdAt":"2016-10-27T09:31:28Z","silentRounds":2,"language":"C","topics":[],"aiDims":["安全工具"],"zone":"工具","funScore":0.8,"domainTags":["WiFi固件","无线安全","软件定义无线电","帧注入"],"domainKey":"WiFi固件","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5,"source":"tier","url":"https://github.com/seemoo-lab/nexmon","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"johnkerl/miller","owner":"johnkerl","name":"miller","desc":"Miller is like awk, sed, cut, join, and sort for name-indexed data such as CSV, TSV, and tabular JSON","summaryCn":"命令行处理CSV和JSON数据的瑞士军刀，一行搞定","reasonCn":"Miller是一个命令行数据处理工具，支持CSV、TSV和JSON等多种格式，提供类似awk/sed/cut/join/sort的功能。它能在管道中高效处理表格数据，支持按列名筛选、聚合统计、数据转换等操作。跨平台运行，无多余依赖，是数据处理和ETL流程的理想选择，适合数据分析师、运维人员和后端开发者日常使用。","stars":10024,"starGrowth":0,"createdAt":"2015-05-03T23:00:37Z","silentRounds":2,"language":"Go","topics":["devops","command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["CSV处理","JSON处理","数据清洗"],"domainKey":"CSV处理","pushedAt":"2026-09-21T11:35:30Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/johnkerl/miller","ts":"2026-09-21T11:35:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ArcReel/ArcReel","owner":"ArcReel","name":"ArcReel","desc":"AI Agent 驱动的开源可自部署视频工作台：将小说与剧本转为角色、场景、道具资产、分镜、视频和剪映草稿，支持跨镜头一致性、多供应商与费用追踪 | Self-hosted AI video workspace for stories, storyboards and short-form video production","summaryCn":"把小说秒变精美短视频的智能工作台，连剪映草稿都帮你自动生成了","reasonCn":"技术上，它深度利用了现代AI Agent架构，通过多个智能体分工合作来解决创意生成中的痛点。最厉害的是它能维持“角色一致性”——也就是说，主角在第一幕和第五幕里长得不会判若两人，这在以前的AI生图里是个老大难问题。","stars":5136,"starGrowth":17,"createdAt":"2026-02-07T08:17:10Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","多模态Agent","创意工具"],"zone":"AI","funScore":0.6,"domainTags":["AI视频生成","角色一致性","小说改编","分镜生成"],"domainKey":"AI视频生成","pushedAt":"2026-09-23T16:29:16Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/ArcReel/ArcReel","ts":"2026-09-23T16:29:16Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0731158878504673,"copyOk":false},{"repo":"ohmyzsh/ohmyzsh","owner":"ohmyzsh","name":"ohmyzsh","desc":"🙃   A delightful community-driven (with 2,500+ contributors) framework for managing your zsh configuration. Includes 300+ optional plugins (rails, git, macOS, hub, docker, homebrew, node, php, python, etc), 140+ themes to spice up your morning, and an auto-update tool that makes it easy to keep up with the latest updates from the community.","summaryCn":"让命令行终端变成超级英雄的基地，自动补全和主题皮肤爽到停不下来","reasonCn":"专为zshshell打造的配置管理框架，自带300多个实用插件。它能让你敲命令时自动补全、智能跳转目录、高亮显示错误，还能切换酷炫主题。对于每天面对终端的开发者来说，这是提升幸福感最直接、成本最低的投资，装完就回不去了。","stars":189881,"starGrowth":19,"createdAt":"2009-08-28T18:15:37Z","silentRounds":0,"language":"Shell","topics":["productivity"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行","zsh","终端","开发工具"],"domainKey":"命令行","pushedAt":"2026-09-23T00:57:31Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/ohmyzsh/ohmyzsh","ts":"2026-09-23T00:57:31Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05730018801249203,"copyOk":false},{"repo":"coder/ghostty-web","owner":"coder","name":"ghostty-web","desc":"","summaryCn":"Ghostty 是近期受到关注的桌面终端模拟器，ghostty-web 项目则是由 Coder 团","reasonCn":"对于前端开发者而言，最具吸引力的一点是它的 API 与 xterm.js 保持兼容。xterm.js 是目前 Web 终端的事实标准库，如果你现有的项目正在用它，迁移到 ghostty-web 只需要修改一行 import 代码，无需重构整个终端组件。","stars":2912,"starGrowth":10,"createdAt":"2025-11-10T18:24:06Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.44,"domainTags":["终端模拟器","WebAssembly","xterm.js","VT100"],"domainKey":"终端模拟器","pushedAt":"2026-09-20T18:36:02.638Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.44,"source":"tier","url":"https://github.com/coder/ghostty-web","ts":"2026-09-20T18:36:02.638Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04836263736263736,"copyOk":false},{"repo":"blacklanternsecurity/bbot","owner":"blacklanternsecurity","name":"bbot","desc":"The recursive internet scanner for hackers. 🧡","summaryCn":"自动递归扫描互联网，帮你摸清目标所有资产的神器","reasonCn":"bbot是一个递归式互联网扫描工具，采用模块化插件架构，自动发现并枚举目标域名、IP、端口、服务、漏洞等资产信息。通过链式扫描机制，发现新目标后自动加入扫描队列，内置子域名枚举、端口扫描、服务识别、漏洞检测等模块，可集成nuclei、httpx等工具，适合渗透测试和安全研究。","stars":10615,"starGrowth":8,"createdAt":"2022-03-12T02:21:59Z","silentRounds":0,"language":"Python","topics":["automation"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["攻击面测绘","递归扫描","资产发现"],"domainKey":"攻击面测绘","pushedAt":"2026-09-23T21:36:36Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/blacklanternsecurity/bbot","ts":"2026-09-23T21:36:36Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.026260951483749413,"copyOk":true},{"repo":"whittlem/pycryptobot","owner":"whittlem","name":"pycryptobot","desc":"","summaryCn":"PyCryptobot 提供了一个完全免费的加密货币交易自动化方案，其社区维护者特","reasonCn":"该机器人的核心优势在于其开源和持续开发的特性，作者不仅提供代码，还通过 Telegram 建立了一个活跃的讨论区，用户可以在主频道中提问并获得支持，而不是被迫进入私人聊天。这种透明的社区运营模式，让初学者能够快速找到解决方案，也让高级用户能够分享自定义策略，形成了一个正","stars":2059,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["加密货币","自动交易"],"zone":"工具","funScore":0.2,"domainTags":["加密交易","Python脚本","机器人","API对接"],"domainKey":"加密交易","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"加密货币","source":"llm","weight":1},{"name":"自动交易","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/whittlem/pycryptobot","ts":"2026-09-23T18:23:13.861Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mostlygeek/llama-swap","owner":"mostlygeek","name":"llama-swap","desc":"Reliable model swapping for any local OpenAI/Anthropic compatible server - llama.cpp, vllm, etc","summaryCn":"给各种本地大模型装个万能翻译器，轻松混用不同开源项目","reasonCn":"它解决的是本地模型生态碎片化的问题，让不同推理框架和模型格式能无缝对接。通过模拟OpenAI接口，它能把llama.cpp、vLLM等后端统一到同一个HTTP服务上，实现模型的快速切换和兼容。核心优势是无需改动应用代码，只需配置转发规则即可。","stars":5738,"starGrowth":18,"language":"Go","topics":["golang","llama","llamacpp","localllama","localllm","openai","openai-api","vllm"],"aiDims":["工具调用","AI基础设施","模型部署"],"zone":"工具","funScore":0,"domainTags":["本地推理","模型兼容","API转发","推理框架"],"domainKey":"本地推理","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"golang","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"llamacpp","source":"github","weight":0.7},{"name":"localllama","source":"github","weight":0.7},{"name":"localllm","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"openai-api","source":"github","weight":0.7},{"name":"vllm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/mostlygeek/llama-swap","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.06341094457999302,"copyOk":false},{"repo":"davila7/claude-code-templates","owner":"davila7","name":"claude-code-templates","desc":"CLI tool for configuring and monitoring Claude Code","summaryCn":"为终端编程助手提供百余种开箱即用的智能体、扩展指令与钩子配置。","reasonCn":"如果你在使用 Anthropic 推出的 AI 终端工具，该开源模板库能大幅提升你的开发效率。它汇集了智能体、自定义命令、系统钩子及 MCP 外部扩展全套配置。开发者无需手动编写脚本，只需在 Web 页面挑选即可一键安装百余组件，快速赋予终端助手联网检索、数据抓取和自动化开发等综合实战能力。","stars":31458,"starGrowth":393,"language":"Python","topics":["anthropic","anthropic-claude","claude","claude-code"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["Claude Code","AI助手","模板库","MCP扩展","提示词工程"],"domainKey":"Claude Code","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"anthropic","source":"github","weight":0.7},{"name":"anthropic-claude","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"trending","url":"https://github.com/davila7/claude-code-templates","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.33747854281899675,"copyOk":true},{"repo":"pocket-id/pocket-id","owner":"pocket-id","name":"pocket-id","desc":"The most user-friendly OpenID Connect Certified™ and OAuth 2.0 provider that lets users sign in to your applications with passkeys.","summaryCn":"Passkey就是你的数字万能钥匙，再也不用记各种密码","reasonCn":"pocket-id是一款开源的身份认证服务，支持OpenID Connect和OAuth 2.0协议，让开发者能把Passkey登录能力集成到自己的应用中。它解决了传统密码体系的痛点——密码泄露、重复注册、管理混乱。部署简单，Docker一键启动，可作为独立的认证提供者嵌入各种Web应用。","stars":9257,"starGrowth":8,"createdAt":"2024-08-11T18:57:32Z","silentRounds":0,"language":"Go","topics":["self-hosted"],"aiDims":["安全工具","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["身份认证","Passkey","OAuth"],"domainKey":"身份认证","pushedAt":"2026-09-23T20:43:29Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/pocket-id/pocket-id","ts":"2026-09-23T20:43:29Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.026592632602354976,"copyOk":false},{"repo":"JetBrains/kotlinconf-app","owner":"JetBrains","name":"kotlinconf-app","desc":"The official KotlinConf application","summaryCn":"开会党必备！JetBrains官方Kotlin桌面会议应用","reasonCn":"这款应用是JetBrains为KotlinConf大会精心开发的官方桌面客户端，采用Kotlin Multiplatform跨平台技术构建，完美支持Windows、macOS和Linux三大操作系统。应用整合了会议日程管理、演讲者资料查询、会场导航定位以及参会者社交互动等核心功能，界面设计现代简洁，操作流程直观顺畅。项目代码结构严谨规范，是Kotlin桌面开发领域的优秀参考案例，充分展示了JetBrains在桌面端的技术积累与创新能力。","stars":3562,"starGrowth":0,"createdAt":"2017-10-26T11:37:25Z","silentRounds":22,"language":"Kotlin","topics":["desktop"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kotlin","会议助手","桌面应用","日程管理"],"domainKey":"Kotlin","pushedAt":"2026-09-23T12:17:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/JetBrains/kotlinconf-app","ts":"2026-09-23T12:17:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"K-Dense-AI/scientific-agent-skills","owner":"K-Dense-AI","name":"scientific-agent-skills","desc":"Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 170,000+ scientists worldwide. 161 ready-to-use validated skills plus 100+ scientific databases covering biology, chemistry, medicine, and drug discovery. Compatible with Cursor, Claude Code, Codex, Pi, Antigravity, and the open Agent Skills standard.","summaryCn":"给AI装上千种科研技能，让它秒变全能科学家助手","reasonCn":"这是GitHub上最火的AI科学家技能库，拥有158项内置技能，覆盖生物学、化学、数学等上百个专业数据库。它让普通AI助手瞬间具备执行复杂科研任务的能力，无论是文献挖掘、数据模拟还是公式推导，都能通过工具调用自动完成，极大提升科研效率。","stars":46135,"starGrowth":188,"createdAt":"2025-10-19T20:54:15Z","silentRounds":0,"language":"Python","topics":["bioinformatics"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["Agent技能库","科研自动化","生物信息学","化学计算"],"domainKey":"Agent技能库","pushedAt":"2026-09-21T09:27:50Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"bioinformatics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/K-Dense-AI/scientific-agent-skills","ts":"2026-09-21T09:27:50Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.37466999024601705,"copyOk":false},{"repo":"pixijs/pixijs","owner":"pixijs","name":"pixijs","desc":"The HTML5 Creation Engine: Create beautiful digital content with the fastest, most flexible 2D WebGL renderer.","summaryCn":"用代码画2D动画的神器，做小游戏网页特效都超顺手","reasonCn":"基于WebGL的高性能2D渲染引擎，专为游戏开发和创意动画打造。支持精灵、粒子系统、动画、物理效果等丰富功能，代码简洁优雅，跨平台兼容性好。全球有大量游戏开发者在用，适合想快速做出炫酷2D视觉效果的项目，上手友好，社区教程丰富。","stars":48207,"starGrowth":8,"createdAt":"2013-01-21T22:40:50Z","silentRounds":0,"language":"TypeScript","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["2D渲染引擎","WebGL","游戏开发库","图形渲染"],"domainKey":"2D渲染引擎","pushedAt":"2026-09-23T20:02:06Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/pixijs/pixijs","ts":"2026-09-23T20:02:06Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024497853008899124,"copyOk":false},{"repo":"google-ai-edge/mediapipe","owner":"google-ai-edge","name":"mediapipe","desc":"Cross-platform, customizable ML solutions for live and streaming media.","summaryCn":"Google开源的多平台AI框架，手眼追踪面部检测全覆盖","reasonCn":"MediaPipe是Google AI团队开发的跨平台机器学习框架，专注于实时媒体处理。支持手部追踪、面部网格、姿态估计、物体检测等多种AI功能，可在手机、电脑、Web和嵌入式设备上运行。项目star超过36000，是AI应用开发领域最热门的项目之一。提供Python、C++、Swift、Kotlin等多语言绑定，文档完善，示例丰富，几乎涵盖了实时AI视觉应用的所有常见需求。","stars":37047,"starGrowth":19,"createdAt":"2019-06-13T19:16:41Z","silentRounds":0,"language":"C++","topics":["ml"],"aiDims":["多模态Agent","AI基础设施","AI应用"],"zone":"AI","funScore":0,"domainTags":["计算机视觉","实时媒体","多平台"],"domainKey":"计算机视觉","pushedAt":"2026-09-23T21:39:33Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/google-ai-edge/mediapipe","ts":"2026-09-23T21:39:33Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05853858612033362,"copyOk":false},{"repo":"sherlock-project/sherlock","owner":"sherlock-project","name":"sherlock","desc":"Hunt down social media accounts by username across social networks","summaryCn":"这工具能一键查出某人的所有社媒账号，查号神器。","reasonCn":"Sherlock 通过并发请求各大社交平台，根据用户输入的用户名精准匹配并返回对应的账号链接。项目支持高度自定义的请求头、代理设置以及多线程并发，能显著提升信息搜集效率。它完全开源，运行在本地，无需注册任何服务，且对隐私非常友好，适合用于市场调研、社交媒体管理或寻找失联好友。","stars":92492,"starGrowth":68,"createdAt":"2018-12-24T14:30:48Z","silentRounds":0,"language":"Python","topics":["osint","reconnaissance","linux","cli","sherlock","python3","redteam","tools","information-gathering","hacktoberfest","python","cybersecurity","cti","forensics","infosec","pentesting"],"aiDims":["非AI-实用","安全工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["OSINT侦查","用户名枚举","社交媒体取证"],"domainKey":"OSINT侦查","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"osint","source":"github","weight":0.7},{"name":"reconnaissance","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"sherlock","source":"github","weight":0.7},{"name":"python3","source":"github","weight":0.7},{"name":"redteam","source":"github","weight":0.7},{"name":"tools","source":"github","weight":0.7},{"name":"information-gathering","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"cti","source":"github","weight":0.7},{"name":"forensics","source":"github","weight":0.7},{"name":"infosec","source":"github","weight":0.7},{"name":"pentesting","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/sherlock-project/sherlock","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.20620559615966788,"copyOk":false},{"repo":"twpayne/chezmoi","owner":"twpayne","name":"chezmoi","desc":"Manage your dotfiles across multiple diverse machines, securely.","summaryCn":"安全同步你所有配置文件的瑞士军刀，换电脑再也不用手动拷贝","reasonCn":"技术上，chezmoi把工作目录当成一个加密的Git仓库，支持模板变量（比如不同机器用不同的主机名）、加密敏感信息（密码API密钥自动加密存储）、幂等部署（不管跑多少次结果都一样）。它用Go写成，单个二进制文件，没有外部依赖，Mac/Linux/Windows都支持。","stars":21688,"starGrowth":5,"createdAt":"2018-11-12T16:45:43Z","silentRounds":0,"language":"Go","topics":["security"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["配置文件管理","多机同步","敏感信息加密","模板引擎","单二进制部署"],"domainKey":"配置文件管理","pushedAt":"2026-09-20T19:33:46Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/twpayne/chezmoi","ts":"2026-09-20T19:33:46Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01569162670601254,"copyOk":true},{"repo":"ever-co/ever-gauzy","owner":"ever-co","name":"ever-gauzy","desc":"Ever® Gauzy™ - Open Business Management Platform (ERP/CRM/HRM/ATS/PM) - https://gauzy.co","summaryCn":"开源版ERP加CRM， freelancers和中小企业省钱办公司必备","reasonCn":"这是一款全功能的开源商业管理平台，集成了ERP、CRM、HRM、招聘和项目管理。专为自由职业者和中小企业打造，提供时间追踪、 invoicing、库存管理等功能，支持自定义和自托管，避免被SaaS服务商绑定，数据完全自主可控，技术栈基于TypeScript和Angular，社区活跃且持续迭代中。","stars":7841,"starGrowth":8,"createdAt":"2019-06-02T09:47:13Z","silentRounds":0,"language":"TypeScript","topics":["hr","salary","accounting","expenses","billing","productivity","time-tracker","team-management","invoices","invoicing","bookkeeping","payroll","crm","erp","timetracking","time-tracking","crm-platform","human-resources","issue-tracker","project-management"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["商业管理","ERP","CRM","HRM"],"domainKey":"商业管理","pushedAt":"2026-09-23T18:57:29Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"hr","source":"github","weight":0.7},{"name":"salary","source":"github","weight":0.7},{"name":"accounting","source":"github","weight":0.7},{"name":"expenses","source":"github","weight":0.7},{"name":"billing","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"time-tracker","source":"github","weight":0.7},{"name":"team-management","source":"github","weight":0.7},{"name":"invoices","source":"github","weight":0.7},{"name":"invoicing","source":"github","weight":0.7},{"name":"bookkeeping","source":"github","weight":0.7},{"name":"payroll","source":"github","weight":0.7},{"name":"crm","source":"github","weight":0.7},{"name":"erp","source":"github","weight":0.7},{"name":"timetracking","source":"github","weight":0.7},{"name":"time-tracking","source":"github","weight":0.7},{"name":"crm-platform","source":"github","weight":0.7},{"name":"human-resources","source":"github","weight":0.7},{"name":"issue-tracker","source":"github","weight":0.7},{"name":"project-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/ever-co/ever-gauzy","ts":"2026-09-23T18:57:29Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.027060834077286062,"copyOk":false},{"repo":"heygen-com/hyperframes","owner":"heygen-com","name":"hyperframes","desc":"Write HTML. Render video. Built for agents.","summaryCn":"用HTML写代码就能生成AI视频，专为Agent打造的新范式","reasonCn":"HeyGen团队推出的开源项目，允许用纯HTML描述视频内容并渲染成片，专门为AI Agent设计。将传统视频制作流程抽象为可代码化的片段，支持动态内容插入、多语言配音、真人播报等能力。这是AI生成视频领域的一个创新尝试，让Agent能够程序化地创建个性化视频内容。","stars":52390,"starGrowth":293,"createdAt":"2026-03-10T01:51:06Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI Agent","创意工具","AI应用"],"zone":"AI","funScore":1,"domainTags":["AI视频生成","Agent","HTML渲染","TTS配音"],"domainKey":"Agent","pushedAt":"2026-09-23T21:48:53Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/heygen-com/hyperframes","ts":"2026-09-23T21:48:53Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.380133613284978,"copyOk":false},{"repo":"earthtojake/text-to-cad","owner":"earthtojake","name":"text-to-cad","desc":"A library of agent skills for CAD, CAE and CAM","summaryCn":"AI 画图一键变 CAD，机械设计从此告别手绘","reasonCn":"在技术实现上，text-to-cad 利用了现有的图像生成模型（如 Stable Diffusion）来理解你的自然语言描述，然后通过一系列后处理算法，将这些描述转化为符合 CAD 标准的矢量数据。它支持多种 CAD 格式，比如 STEP、IGES 和 DWG，这意味着生成的模型可以直接导入 SolidWorks、AutoCAD ","stars":16277,"starGrowth":37,"createdAt":"2026-04-22T01:09:45Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":1,"domainTags":["CAD生成","自然语言处理","机械设计"],"domainKey":"CAD生成","pushedAt":"2026-09-23T05:55:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/earthtojake/text-to-cad","ts":"2026-09-23T05:55:11Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.14138332616575536,"copyOk":true},{"repo":"expo/expo","owner":"expo","name":"expo","desc":"An open-source framework for making universal native apps with React. Expo runs on Android, iOS, and the web.","summaryCn":"用 React 写跨平台 App，一次开发三大平台都能跑","reasonCn":"Expo 是基于 React 的跨平台移动应用开发框架，用 JavaScript/TypeScript 编写一套代码，同时生成 Android、iOS 和 Web 应用。提供丰富的内置组件和工具链，开发体验极佳。支持 OTA 热更新，无需重新发布即可推送代码。社区活跃，star 超过 5 万，是 React Native 生态中最受欢迎的解决方案之一，适合快速构建生产级移动应用。","stars":52386,"starGrowth":6,"createdAt":"2016-08-15T17:14:25Z","silentRounds":0,"language":"TypeScript","topics":["frontend","web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["跨平台开发","移动应用框架","React"],"domainKey":"跨平台开发","pushedAt":"2026-09-23T21:51:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/expo/expo","ts":"2026-09-23T21:51:25Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01834360325277746,"copyOk":false},{"repo":"SeldonIO/alibi","owner":"SeldonIO","name":"alibi","desc":"","summaryCn":"用于机器学习模型检查与解释的高质量开源Python库","reasonCn":"Alibi 是一个专注于可解释性人工智能（XAI）的 Python 库。它实现了多种黑盒与白盒模型的解释算法，适用于分类和回归任务。用户可以通过该库生成局部或全局解释，理解模型决策依据。它还包含异常检测和对抗样本检测等辅助功能。对于需要满足合规或审计要求，希望洞察“黑盒”模型内部逻辑的数据科学家和ML工程师，Alibi 提供了标准化的评估工具和可视化支持。","stars":2646,"starGrowth":0,"createdAt":"2019-02-26T10:10:56Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["可解释性","开发者工具","机器学习"],"zone":"工具","funScore":0.2,"domainTags":["模型解释","黑盒方法","白盒方法","异常检测"],"domainKey":"模型解释","pushedAt":"2026-09-21T19:19:33.440Z","tags":[{"name":"可解释性","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"机器学习","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.44,"source":"tier","url":"https://github.com/SeldonIO/alibi","ts":"2026-09-21T19:19:33.440Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"vadimdemedes/ink","owner":"vadimdemedes","name":"ink","desc":"🌈 React for interactive command-line apps","summaryCn":"用 React 语法搭命令行界面，告别枯燥终端开发","reasonCn":"技术上，它的核心是把 React 的声明式范式移植到了终端环境。每个 CLI 组件都是一个 React 组件，支持完整的状态管理和生命周期。内置了按钮、复选框、进度条、图表等丰富 UI 组件，还支持 SVG 图标和 ANSI 颜色。和传统的 commander 或 inquirer 库相比，它的优势在于开发体验和交","stars":39926,"starGrowth":9,"createdAt":"2017-06-12T06:12:28Z","silentRounds":0,"language":"TypeScript","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["CLI","React","UI-Components","Command-Line"],"domainKey":"CLI","pushedAt":"2026-09-21T05:22:19Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/vadimdemedes/ink","ts":"2026-09-21T05:22:19Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02767625106446927,"copyOk":false},{"repo":"hesreallyhim/awesome-claude-code","owner":"hesreallyhim","name":"awesome-claude-code","desc":"A hand-picked collection of the finest of resources for the most awesome of agents, Claude Code, the undisputed champion of coding companions, from the unstoppable team at Anthropic PBC. A delectable showcase of top tier skills, ambidextrous agents, scintillating status lines, top notch developer tooling, and also we have plugins","summaryCn":"Claude Code最强资源合集，提升编程效率的宝藏指南","reasonCn":"这是一个精心整理的Claude Code相关资源集合，涵盖官方文档、社区教程、扩展插件和最佳实践。帮助用户快速掌握Anthropic推出的编程助手工具，实现代码生成、重构和调试自动化。包含大量实际案例和工作流示例，适合希望提升开发效率的工程师。作为awesome系列项目，其内容质量经过筛选，节省了大量信息搜集时间。","stars":54445,"starGrowth":34,"createdAt":"2025-04-19T20:55:59Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI Agent","代码助手"],"zone":"资源","funScore":0,"domainTags":["AI编程助手","资源合集","提示词工程","MCP协议"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T21:09:27Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/hesreallyhim/awesome-claude-code","ts":"2026-09-23T21:09:27Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10387345027091562,"copyOk":true},{"repo":"ComposioHQ/composio","owner":"ComposioHQ","name":"composio","desc":"Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.","summaryCn":"让AI Agent一键连接上千种工具的超级胶水平台","reasonCn":"Composio是AI Agent的工具集成平台，支持1000多种工具快速接入，提供工具搜索、身份验证、上下文管理和安全沙箱。基于TypeScript开发，兼容LangChain、OpenAI等主流框架。GitHub近3万星标，是目前构建AI Agent最实用的基础设施之一。","stars":30288,"starGrowth":17,"createdAt":"2024-02-23T13:58:27Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools","ai","llm","agents"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","工具集成","API 管理","TypeScript"],"domainKey":"AI Agent","pushedAt":"2026-09-23T20:05:51Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/ComposioHQ/composio","ts":"2026-09-23T20:05:51Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0526838351822504,"copyOk":false},{"repo":"is-a-dev/register","owner":"is-a-dev","name":"register","desc":"Grab your own sweet-looking '.is-a.dev' subdomain.","summaryCn":"免费获取个性子域名，轻松打造你的专属网络身份标识","reasonCn":"is-a.dev是一个提供免费子域名的服务，你可以获得形如yourname.is-a.dev的个人域名。无需购买域名，无需配置DNS服务器，只需简单几步就能拥有一个个性化的网络地址。适合开发者、博主、学生等需要个人品牌展示的人群。","stars":11364,"starGrowth":14,"createdAt":"2020-10-04T20:05:28Z","silentRounds":0,"language":"JavaScript","topics":["subdomain","developer","foss","website","github-pages","dev","free","free-for-dev","free-for-developers","cloudflare","dns","domain","free-domain","oss"],"aiDims":["非AI-好玩","效率工具","开发者工具"],"zone":"工具","funScore":0.7,"domainTags":["子域名","个人品牌","DNS"],"domainKey":"子域名","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"subdomain","source":"github","weight":0.7},{"name":"developer","source":"github","weight":0.7},{"name":"foss","source":"github","weight":0.7},{"name":"website","source":"github","weight":0.7},{"name":"github-pages","source":"github","weight":0.7},{"name":"dev","source":"github","weight":0.7},{"name":"free","source":"github","weight":0.7},{"name":"free-for-dev","source":"github","weight":0.7},{"name":"free-for-developers","source":"github","weight":0.7},{"name":"cloudflare","source":"github","weight":0.7},{"name":"dns","source":"github","weight":0.7},{"name":"domain","source":"github","weight":0.7},{"name":"free-domain","source":"github","weight":0.7},{"name":"oss","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"trending","url":"https://github.com/is-a-dev/register","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.045695881731784585,"copyOk":false},{"repo":"organicmaps/organicmaps","owner":"organicmaps","name":"organicmaps","desc":"🍃 Organic Maps is a free Android & iOS offline maps app for more than 6M travelers, tourists, hikers, and cyclists. It uses crowd-sourced OpenStreetMap data and is developed with love by the community. No ads, no tracking, no data collection, no crapware. Please donate to support the development!","summaryCn":"离线地图神器，出国旅行没网也能精准导航不迷路","reasonCn":"技术核心是用C++编写的高性能地图渲染引擎，结合OpenStreetMap开源数据构建。它不依赖任何商业地图API，完全自主渲染，所以能做到零广告、零追踪、零数据上传。支持路线规划、兴趣点搜索、海拔剖面图等功能，界面简洁无干扰。","stars":15469,"starGrowth":5,"createdAt":"2020-12-27T19:02:26Z","silentRounds":0,"language":"C++","topics":["mobile","privacy"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["离线地图","移动导航","OpenStreetMap","隐私保护"],"domainKey":"离线地图","pushedAt":"2026-09-23T21:17:58Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/organicmaps/organicmaps","ts":"2026-09-23T21:17:58Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01596968129808003,"copyOk":false},{"repo":"evildmp/diataxis-documentation-framework","owner":"evildmp","name":"diataxis-documentation-framework","desc":"A systematic approach to creating better documentation.","summaryCn":"一本帮技术团队告别文档混乱的实战框架，1200星验证有效","reasonCn":"系统化的技术文档创作方法论框架，将文档分为教程、操作指南、参考文档和说明四大类型，帮助技术团队建立清晰、一致的文档体系。Python实现，star数1216，是技术文档领域最具影响力的开源框架之一，适合需要规范文档写作的团队。","stars":1239,"starGrowth":2,"createdAt":"2020-03-25T15:29:08Z","silentRounds":0,"language":"Python","topics":["documentation"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"资源","funScore":0.7,"domainTags":["技术文档","文档写作","方法论"],"domainKey":"技术文档","pushedAt":"2026-09-20T21:26:50Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"documentation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/evildmp/diataxis-documentation-framework","ts":"2026-09-20T21:26:50Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Leonxlnx/taste-skill","owner":"Leonxlnx","name":"taste-skill","desc":"Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop","summaryCn":"给AI装上品味筛选器，专治各种生成‘工业废土’味道的无聊代码和废话","reasonCn":"这是一个提示词工程工具包，专门针对AI生成内容‘千篇一律’、‘油腻俗套’的问题。它内置了经过精心设计的系统提示，引导LLM在生成代码、文案或设计时剔除平庸表达，保留更有创意、更精准的风格，让输出结果更有‘人味’而非机器味。","stars":89305,"starGrowth":211,"createdAt":"2026-02-19T21:44:05Z","silentRounds":0,"language":"JavaScript","topics":["design","frontend","ai"],"aiDims":["提示工程","AI应用","AI写作"],"zone":"AI","funScore":0.5,"domainTags":["提示词工程","代码风格","AI质量"],"domainKey":"提示词工程","pushedAt":"2026-09-23T07:26:15Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/Leonxlnx/taste-skill","ts":"2026-09-23T07:26:15Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3685056827725211,"copyOk":false},{"repo":"public-apis/public-apis","owner":"public-apis","name":"public-apis","desc":"A collective list of free APIs","summaryCn":"全网最全免费API大汇总，开发者的宝藏清单","reasonCn":"这是一个由社区维护的免费API集合，涵盖天气、AI、金融、健身等各个领域。开发者不再需要逐个搜索接口，打开就能找到大量开箱即用的接口。非常适合做原型开发、填坑Demo或构建AI应用时快速调用第三方服务，省时省力。","stars":482341,"starGrowth":226,"createdAt":"2016-03-20T23:49:42Z","silentRounds":0,"language":"Python","topics":["api"],"aiDims":["开发者工具","AI应用","学习资源"],"zone":"资源","funScore":0,"domainTags":["API集合","开发者工具","Web服务"],"domainKey":"API集合","pushedAt":"2026-09-23T19:38:11Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/public-apis/public-apis","ts":"2026-09-23T19:38:11Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3014056445543713,"copyOk":false},{"repo":"koalaman/shellcheck","owner":"koalaman","name":"shellcheck","desc":"ShellCheck, a static analysis tool for shell scripts","summaryCn":"写Shell脚本的避坑神器，帮你提前发现潜在问题","reasonCn":"ShellCheck 是一个专注于 Shell 脚本的静态分析工具，能够检测脚本中的语法错误、潜在 bug 和安全风险。它支持 Bash、Dash、KornShell 等主流 shell 方言，通过词法分析和抽象语法树进行深度检查。内置了超过两百条检查规则，涵盖未引用变量、语法混淆、危险命令使用等多种常见问题。支持在主流编辑器中实时提示，配合 CI/CD 流程可以实现自动化脚本审查，大幅提升运维脚本和自动化任务的可靠性与可维护性。","stars":40070,"starGrowth":8,"createdAt":"2012-11-17T03:15:11Z","silentRounds":0,"language":"Haskell","topics":["developer-tools"],"aiDims":["开发者工具","效率工具","安全工具"],"zone":"工具","funScore":0,"domainTags":["Shell脚本","静态分析","语法检查","运维安全","代码审查"],"domainKey":"Shell脚本","pushedAt":"2026-09-21T06:43:16Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Haskell","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/koalaman/shellcheck","ts":"2026-09-21T06:43:16Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024598951834289992,"copyOk":false},{"repo":"x64dbg/x64dbg","owner":"x64dbg","name":"x64dbg","desc":"An open-source user mode debugger for Windows. Optimized for reverse engineering and malware analysis.","summaryCn":"Windows平台最火的开源逆向调试神器","reasonCn":"这是一个专为Windows设计的开源用户态调试器，主要用于逆向工程、恶意软件分析和软件调试。它提供了完整的调试功能，包括断点、单步执行、内存查看、反汇编窗口等。界面类似经典工具OllyDbg，但功能更强大且持续更新，是安全研究者和逆向工程师的必备工具。","stars":49584,"starGrowth":10,"createdAt":"2015-04-11T20:48:23Z","silentRounds":0,"language":"C++","topics":["security"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["调试器","逆向工程","安全"],"domainKey":"调试器","pushedAt":"2026-09-23T10:46:25Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/x64dbg/x64dbg","ts":"2026-09-23T10:46:25Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030605033881897385,"copyOk":false},{"repo":"louis-e/arnis","owner":"louis-e","name":"arnis","desc":"Generate any location from the real world in Minecraft with a high level of detail.","summaryCn":"用真实地理数据一键生成 Minecraft 世界。","reasonCn":"这款开源工具能把真实的城市地形和自然景观，精准地转化为 Minecraft 的游戏世界。它利用 OpenStreetMap 的地理数据和海拔信息，自动生成高精度的建筑与地形，让你轻松打造属于自己的家乡或宏伟的自然景观，无需复杂的绘图技巧。","stars":18042,"starGrowth":4,"createdAt":"2022-09-10T13:42:29Z","silentRounds":0,"language":"Rust","topics":["maps","procedural-generation"],"aiDims":["游戏","创意工具","可视化"],"zone":"创意","funScore":0.7,"domainTags":["Minecraft","地理数据可视化","OpenStreetMap","体素生成","Rust","游戏模组"],"domainKey":"Minecraft","pushedAt":"2026-09-21T15:38:59Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"maps","source":"github","weight":0.7},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/louis-e/arnis","ts":"2026-09-21T15:38:59Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"web-infra-dev/midscene","owner":"web-infra-dev","name":"midscene","desc":"AI-powered, vision-driven UI automation for every platform.","summaryCn":"AI 视觉驱动的 GUI 自动化测试工具，跨平台运行。","reasonCn":"Midscene 是一款基于 AI 视觉能力的 GUI Agent 测试套件，它通过“看屏幕、做操作、验结果”的闭环逻辑，帮助开发者用自然语言或代码编写自动化测试。它内置测试报告，能直观展示操作轨迹与断言结果，支持在 Web、移动端和桌面应用间复用同一套 Agent 逻辑。","stars":14992,"starGrowth":14,"createdAt":"2024-07-23T04:03:50Z","silentRounds":0,"language":"TypeScript","topics":["testing"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["视觉驱动","UI自动化","测试","大模型","RPA"],"domainKey":"视觉驱动","pushedAt":"2026-09-23T10:55:04Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/web-infra-dev/midscene","ts":"2026-09-23T10:55:04Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.044801494130202776,"copyOk":true},{"repo":"google-gemini/gemini-cli","owner":"google-gemini","name":"gemini-cli","desc":"An open-source AI agent that brings the power of Gemini directly into your terminal.","summaryCn":"想象一下，你平时写代码或者管理服务器时，如果能有个懂技术的AI助手就坐在","reasonCn":"技术架构上，它采用了现代化的TypeScript开发，完美利用了Gemini强大的多模态处理能力和超长上下文窗口。与传统的命令行工具相比，它不仅能理解你的字面指令，还能结合当前项目的目录结构和文件内容进行深度推理。它通过精妙的提示词工程和函数调用机制，安全地将自然语言转化为终","stars":107130,"starGrowth":9,"createdAt":"2025-04-17T17:04:31Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI Agent","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["Gemini模型","命令行工具","AI代理","代码理解"],"domainKey":"Gemini模型","pushedAt":"2026-09-23T20:48:28Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/google-gemini/gemini-cli","ts":"2026-09-23T20:48:28Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.027252030243629235,"copyOk":false},{"repo":"liangleslie/nea_sg_weather","owner":"liangleslie","name":"nea_sg_weather","desc":"Home Assistant integration for NEA Weather forecasts","summaryCn":"给 Home Assistant 加上新加坡天气插件的集成包","reasonCn":"在技术实现上，它利用 Python 编写，通过 Home Assistant 的集成 API 将 NEA 的 XML 数据解析并结构化，支持多种气象要素展示。它采用了异步请求处理机制，不会阻塞 Home Assistant 的主进程，保证了系统的稳定性。配置过程也相对简单，只需在 YAML 配置文件中添加几行代码即可启用","stars":50,"starGrowth":0,"createdAt":"2021-11-06T02:55:18Z","silentRounds":11,"language":"Python","topics":["weather"],"aiDims":["非AI-实用","学习资源"],"zone":"工具","funScore":0,"domainTags":["新加坡天气","Home Assistant集成","气象数据解析","智能家居传感器","NEA数据源"],"domainKey":"新加坡天气","pushedAt":"2026-09-22T15:03:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"weather","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/liangleslie/nea_sg_weather","ts":"2026-09-22T15:03:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AChep/keyguard-app","owner":"AChep","name":"keyguard-app","desc":"A password manager that supports Bitwarden platform and KeePass (KDBX) files. It autofills your logins, supports passkeys, works offline, and runs a Watchtower that finds leaked and reused passwords and other issues.","summaryCn":"这款密码管理器能自动填表登录，还能检查你的密码有没有泄露","reasonCn":"支持Bitwarden平台和KeePass格式，基于Kotlin开发专为Android设计。提供自动填充服务可集成到各类App实现一键登录。内置Watchtower安全审计能扫描泄露密码和重复密码。支持passkeys标准离线运行无需联网数据本地加密存储。界面简洁无订阅费用，开源项目可自由使用，是移动端密码管理的优秀选择。","stars":3294,"starGrowth":6,"createdAt":"2023-07-23T06:12:36Z","silentRounds":0,"language":"Kotlin","topics":["desktop"],"aiDims":["安全工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Android密码管理","Bitwarden兼容","KeePass支持","Passkeys","Watchtower审计"],"domainKey":"Android密码管理","pushedAt":"2026-09-23T17:15:13Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/AChep/keyguard-app","ts":"2026-09-23T17:15:13Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.023464480874316938,"copyOk":false},{"repo":"getsentry/sentry","owner":"getsentry","name":"sentry","desc":"Developer-first error tracking and performance monitoring","summaryCn":"开源界的Bug雷达，自动捕获报错并定位根因，让代码上线更省心","reasonCn":"Sentry是领先的错误追踪平台，支持Python、JS、Go等主流语言。它能自动捕获应用运行时的异常和性能瓶颈，提供完整的堆栈跟踪。企业版功能丰富，开源版自托管也强大，是运维和开发的必备监控利器，极大降低排查成本。","stars":44825,"starGrowth":4,"createdAt":"2010-08-30T22:06:41Z","silentRounds":0,"language":"Python","topics":["crash-reporting","crash-reports","error-monitoring","monitoring","devops","csp-report","django","error-logging","sentry","python","monitor","apm","hacktoberfest","tag-production","fair-source"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["错误追踪","性能监控","开发运维"],"domainKey":"错误追踪","pushedAt":"2026-09-23T21:50:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"crash-reporting","source":"github","weight":0.7},{"name":"crash-reports","source":"github","weight":0.7},{"name":"error-monitoring","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"csp-report","source":"github","weight":0.7},{"name":"django","source":"github","weight":0.7},{"name":"error-logging","source":"github","weight":0.7},{"name":"sentry","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"monitor","source":"github","weight":0.7},{"name":"apm","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"tag-production","source":"github","weight":0.7},{"name":"fair-source","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/getsentry/sentry","ts":"2026-09-23T21:50:57Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"esengine/DeepSeek-Reasonix","owner":"esengine","name":"DeepSeek-Reasonix","desc":"DeepSeek-native AI coding agent for your terminal. Engineered around prefix-cache stability — leave it running.","summaryCn":"AI编程界的持久管家，边用越聪明，帮你干完一整天活都不停歇","reasonCn":"DeepSeek-Reasonix是专为DeepSeek模型设计的终端AI编程代理，核心创新在于前缀缓存优化——项目跑起来后，它会自动缓存已处理过的代码上下文。处理大型项目时越用越快，可长时间稳定运行，自动完成编码、调试、文档编写等复杂任务链，是终端用户的AI结对程序员。","stars":35680,"starGrowth":12,"createdAt":"2026-04-21T08:27:02Z","silentRounds":0,"language":"Go","topics":["ai-agent","llm"],"aiDims":["AI Agent","AI基础设施","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["AI编程助手","终端Agent","前缀缓存","DeepSeek生态","代码上下文管理"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T21:48:53Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/esengine/DeepSeek-Reasonix","ts":"2026-09-23T21:48:53Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.044410762331838556,"copyOk":true},{"repo":"ruvnet/RuView","owner":"ruvnet","name":"RuView","desc":"π RuView turns commodity WiFi signals into real-time spatial intelligence, vital sign monitoring, and presence detection — all without a single pixel of video.","summaryCn":"不用摄像头，用WiFi信号就能感知房间里有人没","reasonCn":"技术原理很有意思。WiFi信号在空间里传播时会碰到人体并发生反射，人体即使静止不动，呼吸也会引起胸腔微小起伏，这些变化会微妙地改变信号的相位和强度。RuView通过采集多个天线的CSI（信道状态信息）数据，用信号处理算法提取这些微变化，再用机器学习模型分类识别——是有人在动、","stars":94773,"starGrowth":72,"createdAt":"2025-06-07T04:32:30Z","silentRounds":0,"language":"Rust","topics":["esp32","home-automation"],"aiDims":["AI应用","可视化"],"zone":"工具","funScore":0.6,"domainTags":["WiFi感知","CSI分析","生命体征监测","存在检测","信号处理","隐私感知"],"domainKey":"WiFi感知","pushedAt":"2026-09-23T18:30:48Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"esp32","source":"github","weight":0.7},{"name":"home-automation","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/ruvnet/RuView","ts":"2026-09-23T18:30:48Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.21827913013199962,"copyOk":false},{"repo":"matplotlib/matplotlib","owner":"matplotlib","name":"matplotlib","desc":"matplotlib: plotting with Python","summaryCn":"Python里画图最强大的开源库，论文图表和数据分析全靠它","reasonCn":"matplotlib的设计哲学是灵活和精确。它提供了两种接口：一种是类似MATLAB的简洁API，几行代码就能画出图表；另一种是面向对象的API，让你对图表的每个细节进行精确控制——坐标轴样式、颜色映射、字体大小、标注位置，你想改什么都能改。","stars":23252,"starGrowth":6,"createdAt":"2011-02-19T03:17:12Z","silentRounds":0,"language":"Python","topics":["data-visualization","data-science"],"aiDims":["非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["数据可视化","Python","科学计算"],"domainKey":"数据可视化","pushedAt":"2026-09-23T16:59:07Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/matplotlib/matplotlib","ts":"2026-09-23T16:59:07Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018774126956820916,"copyOk":true},{"repo":"bmorcelli/Launcher","owner":"bmorcelli","name":"Launcher","desc":"","summaryCn":"ESP32设备的多应用启动器，开机选程序，支持在线刷写新固件","reasonCn":"这是一个为ESP32系列开发板设计的应用启动器，相当于给硬件设备装了一个引导菜单。它支持M5Stack、Lilygo等众多主流开发板，启动后可以选择安装或运行不同的固件程序。通过内置的OTA功能，可以直接从M5Burner或GitHub链接远程安装新二进制文件，省去插线刷机的麻烦。安装方式多样，可用官方刷写器、M5Burner或esptool.py。适合玩ESP32硬件生态的开发者，让设备管理更便捷。","stars":2146,"starGrowth":34,"createdAt":"2024-02-10T22:39:03Z","silentRounds":0,"language":"C++","topics":[],"aiDims":["其他"],"zone":"工具","funScore":0.5,"domainTags":["ESP32","嵌入式固件","设备启动器","OTA更新"],"domainKey":"ESP32","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"其他","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/bmorcelli/Launcher","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.14953028890959927,"copyOk":true},{"repo":"Pumpkin-MC/Pumpkin","owner":"Pumpkin-MC","name":"Pumpkin","desc":"Empowering everyone to host fast and efficient Minecraft servers","summaryCn":"用Rust重写的神级MC服务器，性能炸裂让卡顿成为历史","reasonCn":"基于Rust从零开发的高性能Minecraft服务器，相比Paper/Spigot有显著性能提升。采用多线程架构充分利用多核CPU，内存管理更高效，TPS稳定性更好。支持完整Minecraft协议和主流插件系统，能承载大量玩家同时在线，是自建游戏服务器的优质选择。","stars":11323,"starGrowth":8,"createdAt":"2024-07-28T21:07:38Z","silentRounds":0,"language":"Rust","topics":["minecraft","minecraft-server","rust","game-server","minecraft-protocol","gamedev","networking","docker","server","minecraft-bedrock-edition"],"aiDims":["游戏"],"zone":"创意","funScore":0.5,"domainTags":["Minecraft","Rust","游戏服务器","高性能"],"domainKey":"Minecraft","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"minecraft","source":"github","weight":0.7},{"name":"minecraft-server","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"game-server","source":"github","weight":0.7},{"name":"minecraft-protocol","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"networking","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"server","source":"github","weight":0.7},{"name":"minecraft-bedrock-edition","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/Pumpkin-MC/Pumpkin","ts":"2026-09-21T04:10:58.868Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.026119579616709354,"copyOk":true},{"repo":"livekit/agents","owner":"livekit","name":"agents","desc":"A framework for building realtime voice AI agents 🤖🎙️📹","summaryCn":"用代码给你的AI装上个嘴，让智能体开口就能聊、张嘴就能听。","reasonCn":"LiveKit Agents是一个基于Python的实时语音AI框架，核心在于它能让你轻松构建具备通话能力的智能体。它深度集成了LiveKit平台，支持双向低延迟音频流，让AI不仅能听懂人类说话，还能实时回应，实现真正的语音对话体验。","stars":14312,"starGrowth":10,"createdAt":"2023-10-19T23:00:55Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["AI Agent","多模态Agent","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["实时语音AI","智能体框架","WebRTC"],"domainKey":"实时语音AI","pushedAt":"2026-09-23T19:09:07Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.96,"source":"trending","url":"https://github.com/livekit/agents","ts":"2026-09-23T19:09:07Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03209614309670207,"copyOk":false},{"repo":"ZhuLinsen/daily_stock_analysis","owner":"ZhuLinsen","name":"daily_stock_analysis","desc":"LLM 驱动的多市场股票智能分析系统：多源行情、实时新闻、决策看板与自动推送，支持零成本定时运行。 LLM-powered multi-market stock analysis system with multi-source market data, real-time news, decision dashboard, automated notifications, and cost-free scheduled runs.","summaryCn":"LLM全自动扒股票，新闻数据看板定时推送到你手机","reasonCn":"这是一个由LLM驱动的开源股票分析系统，能自动抓取多市场数据、新闻，生成决策看板并定时推送。它支持零成本定时运行，适合个人投资者或量化初学者。项目热度极高，六万多的Star证明了大家对AI炒股辅助工具的强烈需求，是目前最接地气的大模型金融应用之一。","stars":65505,"starGrowth":51,"createdAt":"2026-01-10T06:43:20Z","silentRounds":0,"language":"Python","topics":["ai-agent","llm"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["股票分析","LLM应用","金融Agent","RAG检索","数据可视化"],"domainKey":"股票分析","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/ZhuLinsen/daily_stock_analysis","ts":"2026-09-22T04:07:30.369Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.18640283947790245,"copyOk":true},{"repo":"zauberzeug/nicegui","owner":"zauberzeug","name":"nicegui","desc":"Create web-based user interfaces with Python. The nice way.","summaryCn":"用Python轻松构建交互式Web界面，告别繁琐的前端开发","reasonCn":"NiceGUI是基于FastAPI的Python库，仅用Python代码就能创建现代Web界面。它提供丰富的组件如按钮、图表、表格等，支持实时通信和数据绑定。相比传统前端开发，它消除了HTML/CSS/JavaScript的学习成本，特别适合数据科学家、AI工程师快速搭建展示原型或内部工具，可集成Vue.js等框架实现复杂交互。","stars":16231,"starGrowth":6,"createdAt":"2021-05-07T13:55:05Z","silentRounds":0,"language":"Python","topics":["frontend","robotics"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["Web界面","Python","前端框架","UI开发"],"domainKey":"Web界面","pushedAt":"2026-09-23T14:56:40Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/zauberzeug/nicegui","ts":"2026-09-23T14:56:40Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01910898897172078,"copyOk":true},{"repo":"urfave/cli","owner":"urfave","name":"cli","desc":"A declarative, simple, fast, and fun package for building command line tools in Go","summaryCn":"Go语言中最流行的命令行工具构建库，写法优雅上手快。","reasonCn":"urfave/cli是Go生态中最受欢迎的命令行框架，采用声明式编程风格，代码简洁易读。支持子命令、标志参数、自动帮助生成等功能。Docker、GitHub CLI、kubectl等知名工具都基于它构建。对于需要开发CLI工具的Go开发者来说，这是首选方案。","stars":24246,"starGrowth":4,"createdAt":"2013-07-13T19:32:06Z","silentRounds":0,"language":"Go","topics":["command-line"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行框架","Go语言开发","CLI工具","开发库","声明式编程"],"domainKey":"命令行框架","pushedAt":"2026-09-22T02:33:15Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/urfave/cli","ts":"2026-09-22T02:33:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"excaliburjs/Excalibur","owner":"excaliburjs","name":"Excalibur","desc":"🎮 Your friendly TypeScript 2D game engine for the web 🗡️","summaryCn":"用TypeScript写2D游戏，浏览器里就能跑，轻松做出好玩的小游戏","reasonCn":"Excalibur是基于TypeScript的轻量级2D游戏引擎，专为Web平台打造。内置Sprite渲染、物理模拟、碰撞检测、输入管理和动画系统，支持ECS架构和插件扩展。相比Phaser等老牌引擎，它在类型安全和现代化架构上更胜一筹，适合快速开发HTML5网页游戏和交互演示。","stars":2346,"starGrowth":1,"createdAt":"2013-01-05T17:05:56Z","silentRounds":3,"language":"TypeScript","topics":["excalibur","excaliburjs","game-development","game-engine","game-framework","gamedev","games","html5-canvas","typescript"],"aiDims":["游戏","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["TypeScript","2D游戏引擎","Web游戏开发","HTML5"],"domainKey":"TypeScript","pushedAt":"2026-09-20T19:32:16Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"excalibur","source":"github","weight":0.7},{"name":"excaliburjs","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"game-framework","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"games","source":"github","weight":0.7},{"name":"html5-canvas","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"bigbro","url":"https://github.com/excaliburjs/Excalibur","ts":"2026-09-20T19:32:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"topoteretes/cognee","owner":"topoteretes","name":"cognee","desc":"Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.","summaryCn":"给AI装个永久大脑，实现跨会话长时记忆的智能方案","reasonCn":"专门为AI Agent设计的开源记忆平台，解决大模型“健忘”痛点。利用知识图谱加向量数据库的双层架构，实现跨会话的长期记忆存储。Python开发，几行代码即可接入，让智能体拥有持续学习和回顾的能力，是构建企业级Agent的关键基础设施。","stars":30918,"starGrowth":22,"createdAt":"2023-08-16T16:16:33Z","silentRounds":0,"language":"Python","topics":["vector-db","ai"],"aiDims":["AI Agent","RAG与知识","向量数据库"],"zone":"AI","funScore":0.6,"domainTags":["AI记忆","知识图谱","向量数据库","Agent"],"domainKey":"知识图谱","pushedAt":"2026-09-23T21:37:33Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/topoteretes/cognee","ts":"2026-09-23T21:37:33Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06813467882786726,"copyOk":false},{"repo":"microsoft/agent-framework","owner":"microsoft","name":"agent-framework","desc":"A framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and .NET.","summaryCn":"构建、部署AI代理和多代理工作流程的工具","reasonCn":"微软的agent-framework是一个强大的工具，可以帮助开发人员快速地构建和管理复杂的AI系统。它不仅支持Python编程，还提供了对多种AI模型（如TensorFlow、PyTorch等）的支持，并且可以在不同的云平台（如Azure、AWS等）上部署。","stars":13732,"starGrowth":22,"createdAt":"2025-04-28T19:40:42Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["Python","Agent"],"zone":"AI","funScore":0,"domainTags":["AI代理","多智能体","工作流编排"],"domainKey":"AI代理","pushedAt":"2026-09-23T19:23:26Z","tags":[{"name":"Python","source":"llm","weight":1},{"name":"Agent","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7}],"aiScore":0.85,"source":"search","url":"https://github.com/microsoft/agent-framework","ts":"2026-09-23T19:23:26Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.07080629187299738,"copyOk":false},{"repo":"responsively-org/responsively-app","owner":"responsively-org","name":"responsively-app","desc":"A modified web browser that helps in responsive web development. A web developer's must have dev-tool.","summaryCn":"让网页多端同步预览，前端开发效率直接起飞的神仙工具","reasonCn":"这是一个专为响应式网页开发打造的浏览器工具，可同时预览多个设备尺寸并实时同步交互操作，无需反复切换浏览器窗口或开发者工具。支持Chrome扩展和独立桌面应用两种模式，兼容主流前端框架，适合追求高效工作流的前端工程师。","stars":25195,"starGrowth":2,"createdAt":"2019-08-10T02:48:18Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["响应式开发","多端预览","浏览器工具","前端工具"],"domainKey":"响应式开发","pushedAt":"2026-09-20T15:05:44Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/responsively-org/responsively-app","ts":"2026-09-20T15:05:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"radzenhq/radzen-blazor","owner":"radzenhq","name":"radzen-blazor","desc":"Radzen Blazor is the most sophisticated free UI component library for Blazor, featuring 145+ native components including DataGrid, Scheduler, Charts, and advanced theming with full support for Material Design and Fluent UI. If Radzen Blazor is useful to you, a ⭐ on GitHub helps other Blazor developers find it.","summaryCn":"Blazor的免费UI组件库，145个组件让Web界面开发效率翻倍","reasonCn":"组件库包含145个以上原生Blazor组件，覆盖了表格、图表、表单、调度器、导航等常用场景。DataGrid组件支持排序筛选分页等高级功能，Charts组件支持多种图表类型，Scheduler组件可以搭建日历和任务管理界面。主题系统基于Material Design，支持完整定制，让你的应用既有专业外观又不","stars":4354,"starGrowth":2,"createdAt":"2019-02-07T10:46:19Z","silentRounds":0,"language":"C#","topics":["data-visualization"],"aiDims":["非AI-实用","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["UI组件库","数据可视化","Blazor"],"domainKey":"UI组件库","pushedAt":"2026-09-23T20:51:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/radzenhq/radzen-blazor","ts":"2026-09-23T20:51:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/terminal","owner":"microsoft","name":"terminal","desc":"The new Windows Terminal and the original Windows console host, all in the same place!","summaryCn":"微软开源的现代终端神器，多标签GPU加速渲染超流畅","reasonCn":"Windows Terminal是微软推出的现代化命令行终端，支持多标签页、GPU硬件加速渲染、自定义配色和字体、以及高度可配置的JSON设置文件。它可以同时运行PowerShell、CMD、WSL Linux和Azure Cloud Shell，界面比传统cmd.exe强大数倍。","stars":104963,"starGrowth":2,"createdAt":"2017-08-11T18:38:22Z","silentRounds":0,"language":"C++","topics":["command-line"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["终端模拟器","命令行","Windows"],"domainKey":"终端模拟器","pushedAt":"2026-09-23T20:38:14Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/microsoft/terminal","ts":"2026-09-23T20:38:14Z","category":"tool","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"bjarneo/cliamp","owner":"bjarneo","name":"cliamp","desc":"cliamp - Terminal music player inspired by winamp","summaryCn":"命令行里跑的音乐播放器，复古Winamp风。","reasonCn":"这款Go语言编写的终端音乐播放器，完美复刻了经典Winamp的视觉风格。它支持播放本地音频文件，拥有现代化的图形界面和流畅的交互体验。项目虽然不涉及AI技术，但凭借其独特的复古设计和实用功能，成为了程序员和极客群体中备受关注的实用工具。","stars":4311,"starGrowth":30,"createdAt":"2026-02-24T07:18:59Z","language":"Go","topics":["music"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["终端音乐播放器","复古UI","Winamp","命令行"],"domainKey":"终端音乐播放器","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/bjarneo/cliamp","ts":"2026-09-23T21:53:43.262Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.13305219206680582,"copyOk":false},{"repo":"agno-agi/agno","owner":"agno-agi","name":"agno","desc":"Build, run, and manage agent platforms.","summaryCn":"Python写的AI智能体开发框架，几行代码就能造自己的AI助手","reasonCn":"技术上它基于原生Python构建，不依赖第三方重型框架，开箱即用。支持多轮对话、长期记忆存储、工具调用和任务规划。可以集成OpenAI、Anthropic或本地LLM，也能接LangChain和LlamaIndex的生态组件，灵活度很高。 安装很简单：pip install agno，然后写几行Python代码就能启动对话系","stars":42303,"starGrowth":12,"createdAt":"2022-05-04T15:23:02Z","silentRounds":0,"language":"Python","topics":["developer-tools","ai","agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent框架","多模型集成","记忆与工具调用"],"domainKey":"AI Agent框架","pushedAt":"2026-09-23T20:23:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/agno-agi/agno","ts":"2026-09-23T20:23:17Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03685100347493085,"copyOk":false},{"repo":"ellisonleao/magictools","owner":"ellisonleao","name":"magictools","desc":":video_game: :pencil: A list of Game Development resources to make magic happen.","summaryCn":"游戏开发的瑞士军刀，一站式收齐所有魔法资源","reasonCn":"这是一个精心整理的大型游戏开发资源目录，覆盖2D/3D美术素材、音频音效、程序库、教程文档等方方面面。由开发者Ellison Leao持续维护，包含数十个精选子列表，从引擎选择到发布平台全部涵盖。适合独立游戏开发者、游戏设计学生以及想转行进游戏行业的程序员。GitHub星标超过17000，是游戏开发领域最受欢迎的资源汇总之一，虽然不直接涉及AI技术，但对于想做游戏AI或者游戏项目开发的创意人群非常实用。","stars":17347,"starGrowth":5,"createdAt":"2014-07-17T21:27:07Z","silentRounds":0,"language":"Markdown","topics":["design","game-development"],"aiDims":["非AI-好玩","游戏","学习资源"],"zone":"资源","funScore":0.6,"domainTags":["游戏开发资源","美术素材","音频资源","开发工具","教程指南"],"domainKey":"游戏开发资源","pushedAt":"2026-09-23T14:38:21Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"Markdown","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ellisonleao/magictools","ts":"2026-09-23T14:38:21Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.015864702830460597,"copyOk":false},{"repo":"sergix44/xbackbone","owner":"sergix44","name":"xbackbone","desc":"A lightweight file manager with full ShareX support and more","summaryCn":"轻量级文件管理器，完美集成ShareX截图上传神器","reasonCn":"这个项目的架构非常简洁，基于PHP开发，配合SQLite或MySQL数据库即可运行。界面清爽现代，支持拖拽上传、自动生成分享链接、设置密码保护、限制文件大小等功能。特别值得一提的是它对ShareX的完整支持——安装后只需在ShareX中配置目标URL，截图后一键上传即可获得分享链接，整个流程","stars":1168,"starGrowth":0,"createdAt":"2018-04-28T12:16:46Z","silentRounds":7,"language":"PHP","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["文件管理","ShareX集成","文件分享"],"domainKey":"文件管理","pushedAt":"2026-09-18T12:12:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/sergix44/xbackbone","ts":"2026-09-18T12:12:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"iii-hq/iii","owner":"iii-hq","name":"iii","desc":"Effortlessly compose, extend, and observe every service in real-time for the first time ever.","summaryCn":"一把瑞士军刀，让所有服务实时连线，改代码就像搭积木一样简单","reasonCn":"iii是一个用Rust编写的服务编排与可观测性工具，支持实时热插拔、动态扩展和可视化调试。开发者可以快速搭建微服务架构，通过声明式配置灵活组合各个服务组件，同时内置实时监控面板，方便排查问题。相比传统方案，它消除了服务间依赖的复杂性，让开发调试像搭积木一样简单直观。","stars":18798,"starGrowth":4,"createdAt":"2025-01-02T17:45:02Z","silentRounds":0,"language":"Rust","topics":["agents"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["服务编排","微服务开发","可观测性","热重载","开发工具"],"domainKey":"服务编排","pushedAt":"2026-09-23T15:45:24Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/iii-hq/iii","ts":"2026-09-23T15:45:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"danielmiessler/Fabric","owner":"danielmiessler","name":"Fabric","desc":"Fabric is an open-source framework for augmenting humans using AI. It provides a modular system for solving specific problems using a crowdsourced set of AI prompts that can be used anywhere.","summaryCn":"把AI提示词像乐高一样拼装，让普通程序员秒变超级AI特工","reasonCn":"Fabric 是一个用 Go 语言写的开源框架，主打通过模块化方式增强人类能力。它核心是一个庞大的、由社区贡献的提示词库，你可以像搭积木一样把不同的提示词组合成工作流。支持浏览器插件、VS Code 扩展等多种客户端，非常适合需要频繁调用 AI 解决具体问题的开发者。","stars":44047,"starGrowth":7,"createdAt":"2024-01-03T23:18:31Z","silentRounds":0,"language":"Go","topics":["ai"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"工具","funScore":0,"domainTags":["提示词工程","模块化","AI工作流"],"domainKey":"提示词工程","pushedAt":"2026-09-23T17:16:35Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/danielmiessler/Fabric","ts":"2026-09-23T17:16:35Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.021476763457216157,"copyOk":true},{"repo":"marimo-team/marimo","owner":"marimo-team","name":"marimo","desc":"A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.","summaryCn":"纯Python构建的交互式笔记本，可部署为脚本或Web应用。","reasonCn":"marimo 是一款响应式的 Python 笔记本工具，它将单元格和 UI 元素关联起来，用户交互时自动更新相关代码，保证代码与输出的一致性。其笔记以纯 Python 代码形式保存，具备 SQL 一等公民支持，既可作为脚本直接运行，也能部署为 Web 应用，解决了传统笔记本环境部署难、代码可复用性差的问题。","stars":22880,"starGrowth":13,"createdAt":"2023-08-14T18:56:20Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","代码助手","可视化"],"zone":"工具","funScore":0,"domainTags":["交互式笔记本","响应式编程","数据科学","Python开发","版本控制"],"domainKey":"交互式笔记本","pushedAt":"2026-09-23T21:44:17Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/marimo-team/marimo","ts":"2026-09-23T21:44:17Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04070454545454545,"copyOk":true},{"repo":"alibaba/yalantinglibs","owner":"alibaba","name":"yalantinglibs","desc":"","summaryCn":"阿里开源的 C++20 高性能实用工具集。","reasonCn":"yalantinglibs 是阿里巴巴开源的一套现代 C++20 库集合，涵盖了序列化、网络通信、日志记录等多个领域。它致力于为开发者提供简单易用且高性能的底层组件，帮助快速构建复杂的 C++ 应用程序。项目包含了 struct_pack 结构化序列化、coro_rpc 协程 RPC 以及 easylog 日志库等核心模块，是 C++ 工程师的得力助手。","stars":2180,"starGrowth":0,"silentRounds":0,"language":"C++","topics":[],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.26,"domainTags":["C++20","结构序列化","协程网络","日志库"],"domainKey":"C++20","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/alibaba/yalantinglibs","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alexhallam/tv","owner":"alexhallam","name":"tv","desc":"","summaryCn":"Tidy Viewer (tv) 通过在终端实现优雅的列样式排版，从根本上改善了处理 CS","reasonCn":"为了处理大型数据集，项目引入了先进的流式打印机制，能够处理超过 5MB 甚至更大的文件，而不会耗尽你的系统内存。这一特性意味着即便在内存有限的设备上，你也能流畅浏览大型表格数据，无需等待漫长的全量加载过程。","stars":2164,"starGrowth":0,"createdAt":"2021-04-30T13:18:16Z","silentRounds":1,"language":"Rust","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["数据处理","CLI工具","CSV分析","终端显示"],"domainKey":"数据处理","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/alexhallam/tv","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DWilliames/paddy-sketch-plugin","owner":"DWilliames","name":"paddy-sketch-plugin","desc":"","summaryCn":"该插件通过识别特定命名规则的图层组，自动计算并应用内边距，从而让设计稿","reasonCn":"项目核心在于其简单直观的操作逻辑，使用者只需在图层名称中加上特定的填充值，插件便会自动重绘布局，无需通过复杂的面板进行参数配置。这种基于命名即代码的思想，曾一度在 UI 设计圈内广为流传，成为了提升设计系统一致性与组件复用率的高效工具。","stars":2146,"starGrowth":0,"createdAt":"2018-02-03T01:05:47Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["UI设计插件","Sketch自动化","自动布局"],"domainKey":"UI设计插件","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/DWilliames/paddy-sketch-plugin","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NanmiCoder/cc-haha","owner":"NanmiCoder","name":"cc-haha","desc":"Local-first cross-platform desktop workspace for Claude Code / agents: multi-agent, Git worktrees, code diffs, skill marketplace, multi-model, Computer Use, task-aware desktop pets, with WeChat, Feishu, DingTalk, Telegram, WhatsApp and H5 access.","summaryCn":"Claude Code的本地桌面伴侣，带宠物和插件市场的超级工作台","reasonCn":"基于TypeScript开发的跨平台桌面应用，专为Claude Code和AI Agent打造。支持多智能体协作、Git worktrees代码差异对比、技能市场扩展以及Computer Use功能。内置多模型支持，还能挂一个任务感知的桌面宠物，更有微信和飞书深度集成，让AI助手真正融入日常办公流。","stars":14678,"starGrowth":10,"createdAt":"2026-03-31T17:30:14Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI Agent桌面客户端","Computer Use","多模型聊天","桌面宠物","微信飞书集成"],"domainKey":"AI Agent桌面客户端","pushedAt":"2026-09-23T17:04:33Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/NanmiCoder/cc-haha","ts":"2026-09-23T17:04:33Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.038452650224826274,"copyOk":false},{"repo":"Specy/asm-editor","owner":"Specy","name":"asm-editor","desc":"An app to write, run and learn M68K, MIPS, RISC-V, X86, Z80 assembly","summaryCn":"学习M68K、MIPS等汇编语言的实用工具，提升编程技能","reasonCn":"这款应用程序允许用户学习并编写、运行和调试M68K、MIPS等古老计算机架构的汇编代码，是学习现代编程语言前了解这些古老技术的好帮手。它不仅提供实时错误提示和性能分析，还允许用户探索各种编程模式和算法，非常适合对计算机历史感兴趣的开发者。","stars":326,"starGrowth":0,"createdAt":"2022-04-04T20:32:40Z","silentRounds":1,"language":"TypeScript","topics":["learning"],"aiDims":["AI应用"],"zone":"工具","funScore":0.6,"domainTags":["汇编语言学习","CPU模拟器","计算机教学"],"domainKey":"汇编语言学习","pushedAt":"2026-09-23T16:31:04Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":1,"source":"search","url":"https://github.com/Specy/asm-editor","ts":"2026-09-23T16:31:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ankit/stylebot","owner":"ankit","name":"stylebot","desc":"Change the appearance of the web instantly","summaryCn":"浏览器一键换装神器，想改网页长啥样就改啥样，解放你的眼睛","reasonCn":"Stylebot是一款基于TypeScript开发的浏览器扩展，核心原理是通过Content Script向当前网页动态注入自定义CSS样式，无需修改源文件即可实现实时预览与一键保存。支持所有常见CSS属性，包括颜色、字体、间距、隐藏元素等，改动会持久化存储并按域名自动应用，无需每次重复设置。","stars":1641,"starGrowth":1,"createdAt":"2010-05-01T18:03:02Z","silentRounds":5,"language":"TypeScript","topics":["design"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["浏览器扩展","CSS样式","网页定制"],"domainKey":"浏览器扩展","pushedAt":"2026-09-23T21:39:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ankit/stylebot","ts":"2026-09-23T21:39:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wdhdev/free-for-life","owner":"wdhdev","name":"free-for-life","desc":"A huge list of great stuff you can get for free!","summaryCn":"收集各种免费神仙资源，设计开发一条龙全都有","reasonCn":"这个项目是一个超级大清单，汇总了网上能找到的各种免费资源，包括设计素材、开发工具、软件服务、在线平台等等。从UI图标、字体、图片库，到云服务器免费额度、AI工具免费额度，应有尽有。适合做设计、搞开发、创业的朋友，能省下一大笔开支，是实用的宝藏清单。","stars":1736,"starGrowth":1,"createdAt":"2023-01-28T10:22:02Z","silentRounds":3,"language":"HTML","topics":["design"],"aiDims":["非AI-实用","学习资源","创意工具"],"zone":"资源","funScore":0,"domainTags":["免费资源","设计","开发工具","清单","薅羊毛"],"domainKey":"免费资源","pushedAt":"2026-09-23T20:56:39Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/wdhdev/free-for-life","ts":"2026-09-23T20:56:39Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TwiN/gatus","owner":"TwiN","name":"gatus","desc":"Automated developer-oriented status page with alerting and incident support","summaryCn":"自动化运维状态页面，监控你的服务并自动告警通知","reasonCn":"技术上，Gatus采用Go语言编写，性能优秀，资源占用低。它支持多种健康检查类型：HTTP请求验证、DNS查询、TCP连接、SMTP验证等。配置通过YAML文件管理，支持环境变量和模板变量。告警系统支持Webhook、Slack、Discord、Telegram、邮件等多种通知渠道。","stars":12132,"starGrowth":3,"createdAt":"2019-09-04T02:35:40Z","silentRounds":0,"language":"Go","topics":["devops","self-hosted"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["运维监控","状态页面","DevOps","健康检查"],"domainKey":"运维监控","pushedAt":"2026-09-22T01:08:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/TwiN/gatus","ts":"2026-09-22T01:08:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"danny-avila/LibreChat","owner":"danny-avila","name":"LibreChat","desc":"Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active","summaryCn":"增强版的ChatGPT克隆，具有丰富的AI功能和特性","reasonCn":"LibreChat是一款基于ChatGPT的克隆程序，它集成了多种AI功能，包括Agent、MCP（多客户端协议）、Skills、DeepSeek等，旨在提供更丰富和智能的交流体验。","stars":44672,"starGrowth":79,"createdAt":"2023-02-12T01:06:52Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["ChatGPT Clone","AI Agent"],"zone":"AI","funScore":0,"domainTags":["AI聊天","多模型集成","ChatGPT克隆","Agent"],"domainKey":"Agent","pushedAt":"2026-09-23T21:28:45Z","tags":[{"name":"ChatGPT Clone","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/danny-avila/LibreChat","ts":"2026-09-23T21:28:45Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.24230533667621776,"copyOk":false},{"repo":"Mintplex-Labs/anything-llm","owner":"Mintplex-Labs","name":"anything-llm","desc":"Stop renting your intelligence. Own it with AnythingLLM. Everything you need for a powerful local-first agent experience","summaryCn":"本地运行的全能 AI 办公套件，支持文档问答与智能体自动化。","reasonCn":"AnythingLLM 是一款面向本地部署的全能 AI 应用，旨在解决用户在处理文档和自动化工作流时的繁琐配置问题。它支持超个性化配置和多用户模式，无需繁琐的安装步骤即可直接运行，让 AI 能力触手可及。项目特别集成了 RAG 技术，能精准理解本地文档内容，同时提供 AI Agent 能力，帮助用户自动化处理复杂任务，无论是个人笔记整理还是团队协作都能轻松应对。","stars":66338,"starGrowth":24,"createdAt":"2023-06-04T02:29:14Z","silentRounds":0,"language":"JavaScript","topics":["vector-db","llm","rag"],"aiDims":["AI应用","AI Agent","RAG与知识"],"zone":"AI","funScore":0.6,"domainTags":["本地LLM","RAG","多模型聚合","向量数据库"],"domainKey":"RAG","pushedAt":"2026-09-22T19:06:03Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Mintplex-Labs/anything-llm","ts":"2026-09-22T19:06:03Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07308535077934215,"copyOk":true},{"repo":"vcmi/vcmi","owner":"vcmi","name":"vcmi","desc":"Open-source engine for Heroes of Might and Magic III","summaryCn":"经典英雄无敌3的免费开源重制版，电脑手机都能玩","reasonCn":"VCMI是《魔法门之英雄无敌III》的跨平台开源引擎重实现，基于Qt框架开发，支持Windows、macOS、Linux甚至Android/iOS。保留了原版所有功能，包括战役、地图编辑器和多人对战，同时修复了旧版BUG并添加了现代UI改进。","stars":5875,"starGrowth":1,"createdAt":"2014-04-06T13:42:23Z","silentRounds":2,"language":"C++","topics":["game"],"aiDims":["游戏"],"zone":"创意","funScore":0.6,"domainTags":["英雄无敌3","游戏引擎重写","跨平台游戏","开源游戏移植","怀旧策略游戏"],"domainKey":"英雄无敌3","pushedAt":"2026-09-23T20:44:16Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/vcmi/vcmi","ts":"2026-09-23T20:44:16Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"webdriverio/webdriverio","owner":"webdriverio","name":"webdriverio","desc":"Next-gen browser and mobile automation test framework for Node.js","summaryCn":"用代码操控浏览器自动干活，测试网页、爬取数据全靠它","reasonCn":"这是一个基于Node.js的浏览器自动化测试框架，支持Chrome、Firefox、Safari等主流浏览器以及移动端设备。采用TypeScript编写，类型提示友好，内置WebdriverIO命令链式API，可与Cucumber、Jest、Mocha等测试框架无缝集成，适合构建稳定的自动化测试套件。","stars":9840,"starGrowth":1,"createdAt":"2011-08-30T19:02:35Z","silentRounds":2,"language":"TypeScript","topics":["automation"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["浏览器自动化","WebDriver协议","端到端测试","Node.js测试框架","跨浏览器测试"],"domainKey":"浏览器自动化","pushedAt":"2026-09-23T20:00:07Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/webdriverio/webdriverio","ts":"2026-09-23T20:00:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kubescape/kubescape","owner":"kubescape","name":"kubescape","desc":"Kubescape is an open-source Kubernetes security platform for your IDE, CI/CD pipelines, and clusters. It includes risk analysis, security, compliance, and misconfiguration scanning, saving Kubernetes users and administrators precious time, effort, and resources.","summaryCn":"K8s集群安全扫描神器，一键检测配置漏洞和合规风险","reasonCn":"Kubescape是Kubernetes安全扫描开源平台，支持IDE、CI/CD和集群多场景。内置CIS、MITRE ATT&CK、NIST等多个合规框架，能自动检测配置错误、权限过高、镜像漏洞等安全问题。提供丰富的集成能力，可与GitHub Actions、Jenkins等流水线无缝对接，实现安全左移。","stars":11743,"starGrowth":3,"createdAt":"2021-08-12T10:39:29Z","silentRounds":0,"language":"Go","topics":["devops","kubernetes"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes安全","合规扫描","DevSecOps"],"domainKey":"Kubernetes安全","pushedAt":"2026-09-23T16:16:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/kubescape/kubescape","ts":"2026-09-23T16:16:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dramaclaw/dramaclaw","owner":"dramaclaw","name":"dramaclaw","desc":"A general-purpose AIGC video engine: script to finished film in one pipeline — dramas, ads, product videos, otome games, and more. | 通用 AIGC 视频引擎 —— 从剧本到成片一条流水线，漫剧、广告、电商、乙游皆可","summaryCn":"用AI代理打造影视工业化全流程创作系统，打破大厂垄断。","reasonCn":"DramaClaw 是一个面向影视与剧本创作的AI代理项目，旨在构建从剧本到成片的全流程自动化生产系统。它试图通过软件工具链将创意生产工业化，让创作者摆脱对大公司的依赖，自主掌控生产流程。项目采用Python开发，提供完整的剧本、分镜和资产管理功能，适合独立创作者和小型团队使用。","stars":6272,"starGrowth":13,"createdAt":"2026-03-27T04:34:30Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","创意工具","多模态Agent"],"zone":"AI","funScore":0.7,"domainTags":["AIGC视频生成","多智能体协作","短剧制作","脚本转视频","内容自动化"],"domainKey":"AIGC视频生成","pushedAt":"2026-09-23T10:33:05Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/dramaclaw/dramaclaw","ts":"2026-09-23T10:33:05Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.05426173469387755,"copyOk":true},{"repo":"CursorTouch/Windows-MCP","owner":"CursorTouch","name":"Windows-MCP","desc":"MCP Server for Computer Use in Windows","summaryCn":"让AI直接操控Windows电脑，告别手动操作","reasonCn":"这是一个为Windows系统打造的MCP服务器，允许AI Agent直接访问和操作Windows桌面。它支持点击、输入、截图识别等能力，让大模型不再只是聊天，而是能真正充当你的电脑助手，执行自动化任务，是AI应用落地的前沿探索项目。","stars":7090,"starGrowth":18,"createdAt":"2025-05-13T02:27:53Z","silentRounds":0,"language":"Python","topics":["desktop"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":1,"domainTags":["Windows自动化","MCP","AI Agent","桌面操作","Agent框架"],"domainKey":"AI Agent","pushedAt":"2026-09-23T16:34:29Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/CursorTouch/Windows-MCP","ts":"2026-09-23T16:34:29Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0616163610719323,"copyOk":true},{"repo":"hundredrabbits/Orca","owner":"hundredrabbits","name":"Orca","desc":"","summaryCn":"Orca 彻底打破了传统的音乐编辑逻辑，它将每一个字母视为一个操作指令。","reasonCn":"由于其轻量化设计，该工具几乎可以在任何设备上运行。除了在 Windows、Linux 和 OSX 的桌面端表现优异，它甚至还支持在终端运行，甚至可以部署在 Monome Norns 这种小巧的硬件上。这种跨平台的特性使得它能够深度融入各种工作流，从专业的录音棚环境到便携的移动演出配置，都能看到","stars":5097,"starGrowth":1,"createdAt":"2018-02-14T19:42:19Z","silentRounds":2,"language":"JavaScript","topics":["livecoding","osc","udp","midi","programming-language","esoteric-language"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"创意","funScore":0.8,"domainTags":["实时音序器","MIDI","创意编程","实时编码"],"domainKey":"实时音序器","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"livecoding","source":"github","weight":0.7},{"name":"osc","source":"github","weight":0.7},{"name":"udp","source":"github","weight":0.7},{"name":"midi","source":"github","weight":0.7},{"name":"programming-language","source":"github","weight":0.7},{"name":"esoteric-language","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/hundredrabbits/Orca","ts":"2026-09-16T22:54:48.334Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"clockworklabs/SpacetimeDB","owner":"clockworklabs","name":"SpacetimeDB","desc":"Development at the speed of light","summaryCn":"把后端代码写进数据库，前端直接调用，告别繁琐 API 层","reasonCn":"Rust 写的实时数据库，能在数据库内部直接跑业务逻辑代码。前端不用写 API 接口，直接通过 WebSocket 调用数据库里的模块，实现超低延迟的数据访问。支持多语言客户端绑定、事务处理和自动推送到前端，比传统前后端分离架构开发速度快得多。","stars":25230,"starGrowth":0,"createdAt":"2023-06-17T07:28:29Z","silentRounds":2,"language":"Rust","topics":["game-development","web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.7,"domainTags":["实时数据库","后端开发","WebSocket","WASM","游戏后端"],"domainKey":"实时数据库","pushedAt":"2026-09-23T21:38:29Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"web-framework","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/clockworklabs/SpacetimeDB","ts":"2026-09-23T21:38:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ImranR98/Obtainium","owner":"ImranR98","name":"Obtainium","desc":"Get Android app updates straight from the source.","summaryCn":"直接从源头获取Android应用更新，不用依赖应用商店","reasonCn":"一个Android应用更新工具，能从GitHub、F-Droid、SourceForge等来源直接获取应用更新，无需依赖Google Play商店。支持手动或自动检查更新，能验证签名和GPG签名确保来源可靠。界面简洁易用，特别适合关注安全和隐私的Android用户。","stars":19884,"starGrowth":24,"createdAt":"2022-08-11T00:16:36Z","silentRounds":0,"language":"Dart","topics":["security","automation","privacy"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["Android","APK更新","F-Droid","应用分发","隐私安全"],"domainKey":"Android","pushedAt":"2026-09-13T22:38:31Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/ImranR98/Obtainium","ts":"2026-09-13T22:38:31Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0756210018105009,"copyOk":false},{"repo":"run-llama/llama_index","owner":"run-llama","name":"llama_index","desc":"LlamaIndex is the leading document agent and OCR platform","summaryCn":"让大模型读你的私有文档，秒变专属知识助手","reasonCn":"LlamaIndex的核心能力在于它提供了一套完整的数据处理管道。你可以把PDF、Word文档、网页链接、数据库甚至API返回的数据喂给它，它会自动把内容切成小块，生成嵌入向量，存入向量数据库，然后构建可检索的知识索引。","stars":52283,"starGrowth":11,"createdAt":"2022-11-02T04:24:54Z","silentRounds":0,"language":"Python","topics":["rag","llm","agents","vector-db"],"aiDims":["RAG与知识","AI Agent","文档理解"],"zone":"AI","funScore":0.5,"domainTags":["RAG","私有数据","知识库","LLM","向量检索"],"domainKey":"RAG","pushedAt":"2026-09-23T15:00:46Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/run-llama/llama_index","ts":"2026-09-23T15:00:46Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.033631180307174414,"copyOk":false},{"repo":"lancedb/lancedb","owner":"lancedb","name":"lancedb","desc":"Developer-friendly OSS embedded retrieval library for multimodal AI. Search More; Manage Less.","summaryCn":"嵌入式的多模态向量数据库，文本图片音频都能语义搜索","reasonCn":"LanceDB是专为多模态AI设计的嵌入式向量数据库，支持文本、图片、音频、视频的语义搜索。基于Apache Arrow和Lance格式，查询性能极高。无需单独部署数据库服务，直接嵌入到你的应用中。Python和Rust双语言支持，几行代码就能实现向量检索。","stars":11504,"starGrowth":10,"createdAt":"2023-02-28T01:15:17Z","silentRounds":0,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI搜索"],"zone":"AI","funScore":0,"domainTags":["向量数据库","多模态检索","嵌入式数据库","RAG","Lance格式"],"domainKey":"向量数据库","pushedAt":"2026-09-23T21:42:00Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/lancedb/lancedb","ts":"2026-09-23T21:42:00Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.032607788595271206,"copyOk":false},{"repo":"drawdb-io/drawdb","owner":"drawdb-io","name":"drawdb","desc":"Free, simple, and intuitive online database diagram editor and SQL generator.","summaryCn":"画数据库关系图的在线神器，拖拖拽拽就能出图","reasonCn":"这是一个基于浏览器的数据库设计工具，支持拖拽生成ER图，并能反向生成SQL语句。虽然不直接涉及AI技术，但作为开发者的实用效率工具，能极大简化数据库建模流程，适合前端和后端工程师日常使用，且具备较高的社区热度。","stars":39683,"starGrowth":32,"createdAt":"2023-07-16T17:09:20Z","silentRounds":0,"language":"JavaScript","topics":["database","editor"],"aiDims":["可视化","非AI-实用","创意工具"],"zone":"工具","funScore":0,"domainTags":["数据库设计","ER图","SQL生成"],"domainKey":"数据库设计","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"editor","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/drawdb-io/drawdb","ts":"2026-09-23T07:35:20.006Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09841917193760552,"copyOk":true},{"repo":"sqlmapproject/sqlmap","owner":"sqlmapproject","name":"sqlmap","desc":"Automatic SQL injection and database takeover tool","summaryCn":"全球最知名的SQL注入神器，黑客必备、安全测试绕不过的硬核工具","reasonCn":"sqlmap是全球最流行的自动化SQL注入工具，支持检测并利用各类数据库的注入漏洞。它兼容MySQL、PostgreSQL、Oracle等主流数据库，能自动提取数据、执行命令甚至接管数据库服务器。作为安全测试和渗透测试的必备神器，它在CTF比赛和Web安全审计中出镜率极高，是安全从业者必须掌握的经典开源项目。","stars":38493,"starGrowth":7,"createdAt":"2012-06-26T09:52:15Z","silentRounds":0,"language":"Python","topics":["database"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["SQL注入","数据库安全","渗透测试","自动化"],"domainKey":"SQL注入","pushedAt":"2026-09-23T07:25:25Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/sqlmapproject/sqlmap","ts":"2026-09-23T07:25:25Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0215455537370431,"copyOk":false},{"repo":"vikejs/vike","owner":"vikejs","name":"vike","desc":"(Replaces Next.js/Nuxt) 🔨 Build mission-critical applications with stability and development freedom.","summaryCn":"告别Next.js选择困难，这个框架让你专注业务不折腾","reasonCn":"Vike是一个基于Vite的Web框架，可以替代Next.js和Nuxt。它支持React、Vue、Svelte等主流框架，提供服务端渲染、静态生成、API路由等完整能力。架构上保持框架无关性，代码风格更简洁灵活，适合追求开发体验和长期可维护性的团队。","stars":5830,"starGrowth":1,"createdAt":"2021-01-28T10:27:12Z","silentRounds":3,"language":"TypeScript","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web框架","前端开发","SSR"],"domainKey":"Web框架","pushedAt":"2026-09-23T18:55:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/vikejs/vike","ts":"2026-09-23T18:55:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"coder/coder","owner":"coder","name":"coder","desc":"Secure environments for developers and their agents","summaryCn":"给AI打工仔配专属工位，代码跑在云端不踩坑","reasonCn":"Coder是一款安全的云端开发环境平台，为开发者和AI Agent提供隔离、可复用的远程工作空间。基于Docker容器技术，支持VS Code、JetBrains等编辑器远程访问，可部署在Kubernetes集群上。内置认证、权限管理、审计日志等企业级安全特性，让AI Agent能在沙箱环境里安全跑代码，保护本地机器不被污染。","stars":16623,"starGrowth":290,"createdAt":"2021-12-22T06:08:52Z","silentRounds":0,"language":"Go","topics":["agents"],"aiDims":["AI Agent","开发者工具"],"zone":"工具","funScore":0,"domainTags":["云端开发环境","Docker","Kubernetes"],"domainKey":"云端开发环境","pushedAt":"2026-09-23T21:45:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/coder/coder","ts":"2026-09-23T21:45:46Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.3523371232629489,"copyOk":false},{"repo":"OuterTune/OuterTune","owner":"OuterTune","name":"OuterTune","desc":"","summaryCn":"双模音乐播放器，本地与网络音源一键切换。","reasonCn":"这是一款基于Material Design 3设计风格的Android音乐播放器。它不仅能播放你手机里的本地音频文件，还能作为YouTube Music的客户端使用，实现了本地与在线流媒体的无缝切换。项目由InnerTune超级分叉而来，旨在提供更强大的功能体验。","stars":5399,"starGrowth":2,"createdAt":"2024-05-13T13:55:37Z","silentRounds":0,"language":"Kotlin","topics":["android","jetpack-compose","material-ui","music-player"],"aiDims":["非AI-好玩"],"zone":"创意","funScore":0.5,"domainTags":["音乐播放","流媒体","Material Design","Android"],"domainKey":"音乐播放","pushedAt":"2026-09-16T18:07:39.903Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"android","source":"github","weight":0.7},{"name":"jetpack-compose","source":"github","weight":0.7},{"name":"material-ui","source":"github","weight":0.7},{"name":"music-player","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/OuterTune/OuterTune","ts":"2026-09-16T18:07:39.903Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"xmanrui/dsh-im","owner":"xmanrui","name":"dsh-im","desc":"通过扫码或机器人凭据把IM机器人接入DeepSeek Harness（支持飞书、微信、钉钉、企业微信、QQ、Slack、Telegram、Discord和WhatsApp）。 Connect IM bots to DeepSeek Harness via QR code or credentials (9 channels).","summaryCn":"把DeepSeek大模型塞进你常用的聊天软件里，实现多端互通。","reasonCn":"该项目通过提供扫码或凭证方式，将IM机器人接入DeepSeek Harness，支持飞书、微信、钉钉等9个主流通讯平台。它解决了单一渠道接入的痛点，让DeepSeek的智能体能力在办公协作场景中无处不在，极大提升了跨平台交互的便捷性。","stars":1446,"starGrowth":13,"createdAt":"2026-08-14T20:40:06Z","silentRounds":0,"language":"JavaScript","topics":["chatbot"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["DeepSeek","IM集成","即时通讯","多端适配"],"domainKey":"DeepSeek","pushedAt":"2026-09-23T18:10:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/xmanrui/dsh-im","ts":"2026-09-23T18:10:46Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0989564315352697,"copyOk":false},{"repo":"AlkaidLab/foundation-sunshine","owner":"AlkaidLab","name":"foundation-sunshine","desc":"","summaryCn":"这个项目是LizardByte Sunshine的增强分支，专门解决Windows平台下游戏串流","reasonCn":"除了画面，它在音频传输上也非常讲究。项目支持7.1.4绕声（共12声道）的传输，甚至包含了Opus DRED丢包恢复技术。对于拥有家庭影院系统的用户来说，这意味着远程连接时也能听到临场感极强的音效，而不是简单的立体声压缩。","stars":6943,"starGrowth":7,"createdAt":"2023-02-06T03:27:06Z","silentRounds":0,"language":"C++","topics":["cloudgaming","moonlight","remote-desktop","stream","sunshine","game-streaming","rdp","amf","hdr","moonlight-streaming","remote","remote-desktop-protocol","self-hosted","streaming","virtual-display","dolby-vision","vivid"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["游戏串流","音视频传输","HDR支持","家庭影院"],"domainKey":"游戏串流","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cloudgaming","source":"github","weight":0.7},{"name":"moonlight","source":"github","weight":0.7},{"name":"remote-desktop","source":"github","weight":0.7},{"name":"stream","source":"github","weight":0.7},{"name":"sunshine","source":"github","weight":0.7},{"name":"game-streaming","source":"github","weight":0.7},{"name":"rdp","source":"github","weight":0.7},{"name":"amf","source":"github","weight":0.7},{"name":"hdr","source":"github","weight":0.7},{"name":"moonlight-streaming","source":"github","weight":0.7},{"name":"remote","source":"github","weight":0.7},{"name":"remote-desktop-protocol","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"streaming","source":"github","weight":0.7},{"name":"virtual-display","source":"github","weight":0.7},{"name":"dolby-vision","source":"github","weight":0.7},{"name":"vivid","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/AlkaidLab/foundation-sunshine","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.024024629122857556,"copyOk":false},{"repo":"taiga-family/taiga-ui","owner":"taiga-family","name":"taiga-ui","desc":"Angular components library for awesome people","summaryCn":"Angular开发者的颜值担当，一套能让页面瞬间变好看的现代化组件库","reasonCn":"这是一个专为Angular框架设计的现代化UI组件库，包含大量开箱即用的移动端和桌面端组件。支持暗黑模式、主题定制、响应式布局，内置表单验证、数据表格、图表等常用组件，帮助开发者快速构建美观的Web应用界面，大幅提升前端开发效率。","stars":4057,"starGrowth":0,"createdAt":"2020-09-25T16:09:18Z","silentRounds":2,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Angular组件库","UI组件","前端开发"],"domainKey":"Angular组件库","pushedAt":"2026-09-23T19:26:11Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/taiga-family/taiga-ui","ts":"2026-09-23T19:26:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"HKUDS/LightRAG","owner":"HKUDS","name":"LightRAG","desc":"[EMNLP2025] \"LightRAG: Simple and Fast Retrieval-Augmented Generation\"","summaryCn":"把知识图谱和向量检索强强联合，让AI回答问题不再断章取义","reasonCn":"香港大学团队提出的新型RAG框架，创新地将知识图谱与向量检索相结合。它不只是简单地把文档切片存入向量库，而是先提取文档中的实体和关系构建知识图谱，再用图遍历和向量检索双路径同时查找信息。这种方式让大模型获得更结构化、更全面的知识，尤其适合复杂推理和多跳问答场景，38685的star数证明了这个方向被广泛认可。","stars":39816,"starGrowth":12,"createdAt":"2024-10-02T11:57:54Z","silentRounds":0,"language":"Python","topics":["rag","llm"],"aiDims":["RAG与知识","大语言模型","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["检索增强生成","知识图谱","向量检索"],"domainKey":"知识图谱","pushedAt":"2026-09-23T18:36:45Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/HKUDS/LightRAG","ts":"2026-09-23T18:36:45Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03690415913200723,"copyOk":false},{"repo":"YouMind-OpenLab/awesome-gpt-image-2","owner":"YouMind-OpenLab","name":"awesome-gpt-image-2","desc":"🚀 World's largest GPT Image 2 prompt library, updated daily — 2000+ curated prompts with preview images, 16 languages. OpenAI's next-gen image model with pixel-perfect text rendering, cross-image consistency, and commercial-grade illustration. Free & open source.","summaryCn":"全球最大的GPT Image 2 prompt库，提供丰富的图像生成提示。","reasonCn":"这个库包含了超过2000个精心挑选的图像生成提示，涵盖了各种风格和场景。","stars":9934,"starGrowth":1,"createdAt":"2026-04-16T13:57:42Z","silentRounds":2,"language":"TypeScript","topics":["awesome"],"aiDims":["Awesome"],"zone":"资源","funScore":0,"domainTags":["提示词库","GPT Image 2","AI绘画素材"],"domainKey":"提示词库","pushedAt":"2026-09-23T15:44:33Z","tags":[{"name":"Awesome","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9762,"source":"search","url":"https://github.com/YouMind-OpenLab/awesome-gpt-image-2","ts":"2026-09-23T15:44:33Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PipedreamHQ/pipedream","owner":"PipedreamHQ","name":"pipedream","desc":"Connect APIs, remarkably fast.  Free for developers.","summaryCn":"不用写代码就能把各种API串联起来，自动化你的工作流。","reasonCn":"技术上，它提供了一个可视化的工作流编辑器，每个节点代表一个服务（比如发送邮件、查询数据库、调用AI模型），你只需要把节点连起来，设置好参数，就能让整个流程自动跑起来。它内置了几百种预建集成，覆盖常见的SaaS工具，包括OpenAI、Anthropic等AI服务。","stars":11702,"starGrowth":0,"createdAt":"2020-01-22T00:28:31Z","silentRounds":2,"language":"JavaScript","topics":["automation"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["自动化","API集成","无代码"],"domainKey":"自动化","pushedAt":"2026-09-23T19:38:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/PipedreamHQ/pipedream","ts":"2026-09-23T19:38:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chef/chef","owner":"chef","name":"chef","desc":"Chef Infra, a powerful automation platform that transforms infrastructure into code automating how infrastructure is configured, deployed and managed across any environment, at any scale","summaryCn":"老牌基础设施自动化神器，用代码管理服务器，适合运维工程师。","reasonCn":"Chef是经典的服务器配置管理工具，使用Ruby DSL将基础设施即代码落地。支持Chef Server协调、Inspec合规检查及Habitat容器打包。它不依赖AI，是DevOps领域稳定可靠的工业级方案，适合大型传统架构的精细化管控与自动化部署。","stars":8242,"starGrowth":0,"createdAt":"2009-01-15T20:40:48Z","silentRounds":3,"language":"Ruby","topics":["devops","automation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["基础设施即代码","配置管理","DevOps"],"domainKey":"基础设施即代码","pushedAt":"2026-09-23T19:50:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/chef/chef","ts":"2026-09-23T19:50:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lobehub/lobe-icons","owner":"lobehub","name":"lobe-icons","desc":"🥨 Lobe Icons - Brings AI/LLM brand logos to your React & React Native apps — static SVG/PNG/WebP, no dependencies.","summaryCn":"汇集了AI和LLM大厂的精美图标，React项目开箱即用","reasonCn":"lobe-icons 是一个专注于 AI/LLM 品牌的图标库，包含 OpenAI、Anthropic、Google、DeepSeek 等数十家公司的官方品牌图标。以静态 SVG、PNG、WebP 格式提供，零依赖，可直接嵌入 React 和 React Native 项目，让开发者轻松在 UI 中展示 AI 品牌标识，省去到处找图或手绘的麻烦。","stars":2528,"starGrowth":1,"createdAt":"2024-01-30T03:13:18Z","silentRounds":2,"language":"TypeScript","topics":["design"],"aiDims":["非AI-实用","开发者工具","创意工具"],"zone":"工具","funScore":0,"domainTags":["图标库","React组件","SVG图标"],"domainKey":"图标库","pushedAt":"2026-09-23T17:21:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/lobehub/lobe-icons","ts":"2026-09-23T17:21:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langbot-app/LangBot","owner":"langbot-app","name":"LangBot","desc":"Production-grade platform for building agentic IM bots - 生产级多平台智能机器人开发平台/ Agent、知识库编排、插件系统 / Bots for Discord / Slack / LINE / Telegram / WeChat(企业微信, 企微智能机器人, 公众号) / 飞书 / 钉钉 / QQ / Matrix e.g. Integrated with ChatGPT(GPT), DeepSeek, Dify, n8n, Langflow, Coze, Claude, Gemini, GLM, Ollama, SiliconFlow, Moonshot, openclaw / hermes agent, deerflow","summaryCn":"一个平台帮你快速搭建支持微信飞书钉钉的多平台AI聊天机器人","reasonCn":"技术上，它采用插件化架构，核心层负责消息协议适配，插件层负责业务逻辑。支持主流大模型接入，包括OpenAI、Claude、国内的各种大模型。内置RAG知识库模块，可以上传文档让机器人基于你的知识回答。插件系统很强大，社区已经开发了很多现成的插件，从天气查询到代码生成都有。","stars":17951,"starGrowth":7,"createdAt":"2022-12-07T08:48:02Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["聊天机器人","多平台","RAG"],"domainKey":"RAG","pushedAt":"2026-09-23T20:22:01Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/langbot-app/LangBot","ts":"2026-09-23T20:22:01Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.022169851261768146,"copyOk":true},{"repo":"deepset-ai/haystack","owner":"deepset-ai","name":"haystack","desc":"Open-source AI orchestration framework for building context-engineered, production-ready LLM applications. Design modular pipelines and agent workflows with explicit control over retrieval, routing, memory, and generation. Built for scalable agents, RAG, multimodal applications, semantic search, and conversational systems.","summaryCn":"生产级AI应用开发框架，搭RAG系统和智能Agent的瑞士军刀，社区超活跃","reasonCn":"技术上，它采用Pipeline架构，把文档加载、预处理、分块、嵌入、检索、重排序、生成等环节组装成可复用的工作流。支持多种嵌入模型和LLM后端，包括OpenAI、Anthropic、本地模型和Hugging Face模型。内置多种检索策略，支持向量检索和关键词检索的组合。","stars":26580,"starGrowth":5,"createdAt":"2019-11-14T09:05:28Z","silentRounds":0,"language":"Python","topics":["rag","llm","agents"],"aiDims":["AI基础设施","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["RAG框架","LLM编排","Agent工作流","向量检索","生产级AI"],"domainKey":"RAG框架","pushedAt":"2026-09-23T17:06:11Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/deepset-ai/haystack","ts":"2026-09-23T17:06:11Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.01556433408577878,"copyOk":false},{"repo":"plotly/dash","owner":"plotly","name":"dash","desc":"Data Apps & Dashboards for Python. No JavaScript Required.","summaryCn":"用Python画数据面板，不用懂JS也能做出炫酷的交互式仪表板","reasonCn":"Plotly Dash是专为Python开发者打造的数据应用框架，让你仅用Python就能构建交互式Web仪表板。内置丰富的Plotly图表组件，支持实时数据刷新和动态交互，无需配置前端技术栈即可快速搭建从数据监控到商业智能的完整应用。","stars":24424,"starGrowth":1,"createdAt":"2015-04-10T01:53:08Z","silentRounds":1,"language":"Python","topics":["data-visualization","finance"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据可视化","仪表板","Python","交互式图表"],"domainKey":"数据可视化","pushedAt":"2026-09-23T16:56:59Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/plotly/dash","ts":"2026-09-23T16:56:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ansible/awx","owner":"ansible","name":"awx","desc":"AWX provides a web-based user interface, REST API, and task engine built on top of Ansible. It is one of the upstream projects for Red Hat Ansible Automation Platform.","summaryCn":"自动化运维界的“指挥官”，让成千上万台服务器听话照办，不用人肉敲命令。","reasonCn":"AWX是Ansible的开源Web控制台，提供可视化任务编排、库存管理和REST API。它把枯燥的命令行自动化变成了拖拉拽式的图形操作，支持定时任务、审批流程和团队协作，是企业级IT自动化的核心基础设施，虽然不直接涉及AI，但却是现代DevOps的必备利器。","stars":15567,"starGrowth":0,"createdAt":"2017-05-17T15:50:14Z","silentRounds":2,"language":"Python","topics":["automation"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["运维自动化","任务编排","IT基础设施"],"domainKey":"运维自动化","pushedAt":"2026-09-23T18:27:21Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ansible/awx","ts":"2026-09-23T18:27:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"getarcaneapp/arcane","owner":"getarcaneapp","name":"arcane","desc":"Modern Docker Management, Designed for Everyone","summaryCn":"可视化管Docker容器，比命令行更直观的好帮手","reasonCn":"Arcane是一款基于Go开发的现代化Docker可视化面板，让Docker管理告别晦涩的命令行。它提供直观的界面来管理容器、镜像、网络和卷，支持一键启停、实时日志查看和资源监控。界面简洁美观，操作流畅，非技术用户也能快速上手。项目采用跨平台架构，支持Windows、macOS和Linux，是日常开发和运维的实用利器。","stars":7520,"starGrowth":32,"createdAt":"2025-04-19T00:43:29Z","silentRounds":0,"language":"Go","topics":["docker","docker-management","self-hosted","sveltekit","typescript","go","compose","docker-compose","container-management","containers","web-ui"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Docker管理","容器可视化","运维面板","跨平台桌面应用"],"domainKey":"Docker管理","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"docker-management","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"sveltekit","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"compose","source":"github","weight":0.7},{"name":"docker-compose","source":"github","weight":0.7},{"name":"container-management","source":"github","weight":0.7},{"name":"containers","source":"github","weight":0.7},{"name":"web-ui","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/getarcaneapp/arcane","ts":"2026-09-18T10:23:42.801Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.10876595744680852,"copyOk":true},{"repo":"github/github-mcp-server","owner":"github","name":"github-mcp-server","desc":"GitHub's official MCP Server","summaryCn":"给AI装上GitHub大脑的官方神器，让大模型直接读写你的仓库代码。","reasonCn":"这是GitHub官方推出的Model Context Protocol服务器，让AI Agent能通过标准协议直接读写GitHub数据。支持仓库搜索、Issue管理、PR创建等操作，帮助开发者构建智能工作流。内置了丰富的工具链，方便AI理解和操作代码库。","stars":33153,"starGrowth":20,"language":"Go","topics":["github","mcp","mcp-server"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["Model Context Protocol","AI Agent","GitHub API","代码分析"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"github","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/github/github-mcp-server","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06180979096914306,"copyOk":false},{"repo":"traefik/traefik","owner":"traefik","name":"traefik","desc":"The Cloud Native Application Proxy","summaryCn":"让服务器自动搞流量分发，不用写代码的神器","reasonCn":"Traefik 是一款现代化的云原生反向代理和负载均衡器，专为微服务架构设计。它通过自动发现服务实例、动态更新配置和内嵌健康检查，实现了零停机部署。相比 Nginx，Traefik 支持自动 HTTPS 证书申请和配置热更新，无需手动重启服务。","stars":64930,"starGrowth":11,"createdAt":"2015-09-13T19:04:02Z","silentRounds":0,"language":"Go","topics":["docker","kubernetes"],"aiDims":["非AI-实用","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["反向代理","负载均衡","云原生"],"domainKey":"反向代理","pushedAt":"2026-09-23T14:40:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/traefik/traefik","ts":"2026-09-23T14:40:06Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03350823964269213,"copyOk":false},{"repo":"BuilderIO/builder","owner":"BuilderIO","name":"builder","desc":"Visual Development for React, Vue, Svelte, Qwik, and more","summaryCn":"拖拽组件就能搭网页的开源平台，设计师和开发者都能用","reasonCn":"Builder.io是一个可视化CMS和页面构建器，支持React、Vue、Svelte等主流框架。它让非技术人员通过拖拽就能搭建页面，同时也为开发者提供完整的代码导出能力。核心亮点包括可视化编辑、内容管理、A/B测试和个性化推荐，深度集成到现有开发工作流中，无需从零写代码。","stars":8845,"starGrowth":1,"createdAt":"2019-01-30T23:06:02Z","silentRounds":2,"language":"TypeScript","topics":["frontend"],"aiDims":["可视化","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["可视化CMS","页面构建器","前端开发","A/B测试"],"domainKey":"可视化CMS","pushedAt":"2026-09-23T16:07:25Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/BuilderIO/builder","ts":"2026-09-23T16:07:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"podman-desktop/podman-desktop","owner":"podman-desktop","name":"podman-desktop","desc":"Podman Desktop is the best free and open source tool to work with Containers and Kubernetes for developers. Get an intuitive and user-friendly interface to effortlessly build, manage, and deploy containers and Kubernetes — all from your desktop.","summaryCn":"零命令行搞定容器和K8s，桌面端运维新体验","reasonCn":"Podman Desktop 是用 TypeScript 开发的开源桌面应用，提供可视化界面管理容器和 Kubernetes 集群。无需记忆复杂命令，即可构建镜像、运行容器、管理 Pod 和集群，内置安全扫描功能，适合新手入门也满足开发者日常需求，是 Docker Desktop 的有力替代品。","stars":8024,"starGrowth":2,"createdAt":"2022-03-03T18:45:51Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["容器管理","Kubernetes","桌面应用","DevOps"],"domainKey":"容器管理","pushedAt":"2026-09-23T14:03:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/podman-desktop/podman-desktop","ts":"2026-09-23T14:03:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SimplifyJobs/Summer2027-Internships","owner":"SimplifyJobs","name":"Summer2027-Internships","desc":"Summer 2027 software engineering, data science, AI, quant, product management, and hardware internship postings. Updated daily by Simplify and Pitt CSC.","summaryCn":"GitHub上最火的实习岗位汇总库，每周更新，找工作必备神器","reasonCn":"由Simplify和匹兹堡大学CSC联合维护，每日更新2027年暑期实习信息，涵盖软件工程、数据科学、AI、量化金融、产品经理和硬件等方向。已获4.6万星，是求职季获取一手招聘信息的高价值聚合资源，虽非AI工具但极具实用价值。","stars":47601,"starGrowth":22,"createdAt":"2020-05-23T20:53:18Z","silentRounds":0,"language":"Python","topics":["data-science"],"aiDims":["学习资源"],"zone":"资源","funScore":0,"domainTags":["实习求职","岗位汇总","数据聚合"],"domainKey":"实习求职","pushedAt":"2026-09-23T21:31:50Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/SimplifyJobs/Summer2027-Internships","ts":"2026-09-23T21:31:50Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06738652549316192,"copyOk":false},{"repo":"Canner/WrenAI","owner":"Canner","name":"WrenAI","desc":"GenBI (Generative BI) for AI agents, an open-source, governed text-to-SQL through an open context layer that turns natural-language questions into trusted dashboards, charts, and SQL across 20+ data sources, such as BigQuery, Snowflake, PostgreSQL, ClickHouse, Amazon Redshift, Databricks and more.","summaryCn":"把自然语言直接变图表，小白也能当数据分析师","reasonCn":"WrenAI是一款生成式BI工具，通过开放语义层解决企业数据孤岛问题。它利用RAG技术理解业务语境，将自然语言问题转化为可信任的SQL查询、图表及仪表板。支持二十多种主流数据源，内置安全治理机制，让AI Agent能精准调用数据工具，无需代码基础即可快速完成复杂数据分析任务。","stars":17722,"starGrowth":11,"createdAt":"2024-03-13T06:18:20Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI应用","工具调用","RAG与知识"],"zone":"AI","funScore":0.5,"domainTags":["生成式BI","Text-to-SQL","语义层","数据分析","RAG"],"domainKey":"RAG","pushedAt":"2026-09-23T12:46:51Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Canner/WrenAI","ts":"2026-09-23T12:46:51Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.034862092314637175,"copyOk":true},{"repo":"chaitanyagiri/munder-difflin","owner":"chaitanyagiri","name":"munder-difflin","desc":"local multi-agent harness","summaryCn":"本地跑多智能体协作的测试神器，模拟真实开发团队干活。","reasonCn":"这是一个本地运行的多智能体编排框架，基于TypeScript开发。它模拟多个AI Agent协同工作，通过函数调用和工具使用来完成任务，非常适合测试AI Agent的协作能力。项目支持自定义Agent行为和工具定义，提供了一个沙盒环境来观察Agent如何决策。","stars":7835,"starGrowth":27,"createdAt":"2026-05-31T15:25:05Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["多Agent系统","AI测试","沙盒环境","Agent框架"],"domainKey":"多Agent系统","pushedAt":"2026-09-23T07:26:06Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/chaitanyagiri/munder-difflin","ts":"2026-09-23T07:26:06Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.10960587109125718,"copyOk":false},{"repo":"microsoft/Data-Science-For-Beginners","owner":"microsoft","name":"Data-Science-For-Beginners","desc":"10 Weeks, 20 Lessons, Data Science for All!","summaryCn":"微软出品的数据科学入门课程，20节课带你从零学会数据分析","reasonCn":"微软官方推出的数据科学入门系列教程，采用Jupyter Notebook形式，每节课都配有可运行的代码和数据集。内容涵盖数据处理、可视化、机器学习等核心环节，循序渐进设计，零基础也能跟上。项目完全免费，代码直接可用，适合想系统学习数据科学的同学。","stars":37256,"starGrowth":22,"createdAt":"2021-03-03T16:30:15Z","silentRounds":0,"language":"Jupyter Notebook","topics":["data-visualization","data-science"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["数据科学教程","Jupyter Notebook教学","机器学习入门","数据可视化教学"],"domainKey":"数据科学教程","pushedAt":"2026-09-13T21:52:54Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/microsoft/Data-Science-For-Beginners","ts":"2026-09-13T21:52:54Z","category":"learning","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.0677715267339489,"copyOk":true},{"repo":"bbc/sqs-consumer","owner":"bbc","name":"sqs-consumer","desc":"Build Amazon Simple Queue Service (SQS) based applications without the boilerplate","summaryCn":"让处理AWS队列消息不再繁琐的实用TypeScript库","reasonCn":"专为TypeScript开发者打造，简化了Amazon SQS消息队列的消费逻辑。它封装了复杂的长轮询和错误处理，让开发者只需关注业务逻辑。支持自动批处理、并发控制和自定义配置，极大提升了后端开发效率，是构建稳定消息驱动架构的理想选择。","stars":1908,"starGrowth":0,"createdAt":"2014-12-18T14:01:33Z","silentRounds":22,"language":"TypeScript","topics":["backend"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["AWS SQS","消息队列","TypeScript","后端开发"],"domainKey":"AWS SQS","pushedAt":"2026-09-22T01:42:56Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/bbc/sqs-consumer","ts":"2026-09-22T01:42:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"webstudio-is/webstudio","owner":"webstudio-is","name":"webstudio","desc":"Open source website builder and Webflow alternative. Webstudio is an advanced visual builder that connects to any headless CMS, supports all CSS properties, and can be hosted anywhere, including with us.","summaryCn":"不用写代码就能拖拽出专业级网站的建站神器","reasonCn":"这是开源版Webflow，纯可视化拖拽建站，支持所有CSS属性和任意Headless CMS接入，可部署到任何托管平台。无需编程基础，设计师或运营也能做出专业网站。代码完全开源免费，支持导出干净HTML和React/Vue组件，灵活可控，省去大量外包开发成本。","stars":8978,"starGrowth":3,"createdAt":"2022-03-13T04:19:01Z","silentRounds":0,"language":"TypeScript","topics":["design","frontend"],"aiDims":["非AI-实用","创意工具","效率工具"],"zone":"工具","funScore":1,"domainTags":["网站构建","可视化编辑","无代码"],"domainKey":"网站构建","pushedAt":"2026-09-23T11:03:37Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/webstudio-is/webstudio","ts":"2026-09-23T11:03:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"browser-use/browser-harness","owner":"browser-use","name":"browser-harness","desc":"Browser Harness | Self-healing harness that enables LLMs to complete any task.","summaryCn":"让大模型自己操作浏览器完成任何任务的智能体","reasonCn":"这是一个能让LLM像人一样操作浏览器的工具库，自动处理页面加载、点击、输入等动作，甚至能自我修复定位失败的问题。支持Playwright和Selenium，可完成网页填写、数据抓取、自动化测试等任务，是构建浏览器自动化Agent的核心组件。","stars":18028,"starGrowth":62,"createdAt":"2026-04-17T01:56:15Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["浏览器自动化","LLM","Playwright"],"domainKey":"浏览器自动化","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/browser-use/browser-harness","ts":"2026-09-21T04:10:58.868Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.23558074106944749,"copyOk":false},{"repo":"perspective-dev/perspective","owner":"perspective-dev","name":"perspective","desc":"A data visualization and analytics component, especially well-suited for large and/or streaming datasets.","summaryCn":"高性能数据可视化组件，专门处理大规模和实时流数据。","reasonCn":"perspective是一个基于WebAssembly的高性能数据可视化库，专为大数据集和实时流设计。支持HTML、JS、Python、R等多语言绑定，单文件即可运行。相比D3.js等库，它能轻松处理百万级数据点，渲染流畅，适合数据大屏和实时监控场景。","stars":11244,"starGrowth":3,"createdAt":"2017-11-02T16:27:54Z","silentRounds":0,"language":"Rust","topics":["data-visualization"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["数据可视化","WebAssembly","大规模数据","实时数据流"],"domainKey":"数据可视化","pushedAt":"2026-09-21T17:46:11Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/perspective-dev/perspective","ts":"2026-09-21T17:46:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yamadashy/repomix","owner":"yamadashy","name":"repomix","desc":"📦 Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file. Perfect for when you need to feed your codebase to Large Language Models (LLMs) or other AI tools like Claude, ChatGPT, DeepSeek, Perplexity, Gemini, Gemma, Llama, Grok, and more.","summaryCn":"把整个代码库塞进一个文件，让AI秒变你的专属程序员","reasonCn":"用一行命令把你的整个项目目录打包成一个结构化的文本文件，可以直接喂给ChatGPT、Claude等大模型进行分析、调试或重构建议。支持TypeScript、JavaScript、Python等主流语言，还能自动过滤node_modules等无关目录，输出高度精简却信息完整的代码上下文。","stars":28466,"starGrowth":18,"createdAt":"2024-07-13T07:11:32Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools","llm","chatbot"],"aiDims":["代码助手","工具调用","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["LLM","代码上下文","AI编程","Tree-sitter"],"domainKey":"LLM","pushedAt":"2026-09-23T10:15:11Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/yamadashy/repomix","ts":"2026-09-23T10:15:11Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.055896999929740745,"copyOk":false},{"repo":"Snailclimb/JavaGuide","owner":"Snailclimb","name":"JavaGuide","desc":"Java 面试 & 后端通用面试指南，覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发","summaryCn":"一份超全面的Java后端面试宝典，从基础到分布式全覆盖","reasonCn":"这个仓库是Java后端学习者的宝藏，内容覆盖Java基础、集合框架、并发编程、JVM原理、数据库、分布式、高并发、系统设计等多个领域。虽然标题提到AI应用开发，但主体仍是Java面试指南，适合准备大厂后端岗位的开发者系统复习。","stars":158810,"starGrowth":30,"createdAt":"2018-05-07T13:27:00Z","silentRounds":0,"language":"JavaScript","topics":["ai"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["Java","面试指南","后端开发","计算机科学","编程"],"domainKey":"Java","pushedAt":"2026-09-22T09:58:35Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"trending","url":"https://github.com/Snailclimb/JavaGuide","ts":"2026-09-22T09:58:35Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09056671494238398,"copyOk":true},{"repo":"beekeeper-studio/beekeeper-studio","owner":"beekeeper-studio","name":"beekeeper-studio","desc":"Modern and easy to use SQL client for MySQL, Postgres, SQLite, SQL Server, and more. Linux, MacOS, and Windows.","summaryCn":"颜值超高的跨平台数据库管理工具，连小白也能轻松上手","reasonCn":"基于Electron构建的开源SQL客户端，支持MySQL、PostgreSQL、SQLite、SQL Server等主流数据库。界面简洁现代，支持实时查询执行、数据可视化编辑、智能代码补全、查询历史管理等功能，跨平台兼容Linux/macOS/Windows，上手成本极低。","stars":23644,"starGrowth":6,"createdAt":"2019-07-23T18:12:33Z","silentRounds":0,"language":"TypeScript","topics":["database"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["SQL 客户端","数据库管理","跨平台"],"domainKey":"SQL 客户端","pushedAt":"2026-09-23T03:59:53Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/beekeeper-studio/beekeeper-studio","ts":"2026-09-23T03:59:53Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018761292505498222,"copyOk":false},{"repo":"coleam00/Archon","owner":"coleam00","name":"Archon","desc":"The first open-source harness builder for AI coding. Make AI coding deterministic and repeatable.","summaryCn":"让AI写代码不再靠运气，每一次都稳定复现","reasonCn":"Archon是一个开源的AI编码工具集，专为让AI辅助编程变得可预测、可重复而设计。它提供标准化工作流、执行追踪、自动验证等能力，支持Claude、GPT等多种模型接入。与Cursor、Copilot等工具不同，Archon强调工程化思维，适合需要将AI编码纳入团队协作流程的开发者，降低AI不确定性的风险。","stars":23531,"starGrowth":9,"createdAt":"2025-02-07T21:04:12Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools","automation"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI编码","工作流","模型管理"],"domainKey":"AI编码","pushedAt":"2026-09-23T21:51:15Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/coleam00/Archon","ts":"2026-09-23T21:51:15Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.028147422548977945,"copyOk":false},{"repo":"cactus-compute/needle","owner":"cactus-compute","name":"needle","desc":"Automation foundation model for tiny devices: 2-bit, 8-29 MB, tool calls, structured extraction and embeddings on phones, wearables, smart homes, robots, cars and microcontrollers.","summaryCn":"14MB的轻量级AI模型，让手机和穿戴设备也能跑大模型。","reasonCn":"Needle采用了极其激进的模型压缩技术，在保持较高精度的同时，将模型体积压缩到了极致。它针对ARM架构进行了深度优化，利用了现代移动处理器的向量指令集，使得推理速度非常可观。它甚至不需要联网，所有计算都在本地完成，保护隐私又省流量。","stars":12206,"starGrowth":144,"createdAt":"2026-02-24T04:50:47Z","silentRounds":0,"language":"Python","topics":["cactus","llm","on-device-ai","edge-ai","foundation-model","function-calling","tinyml","tool-calling"],"aiDims":["模型部署","AI基础设施","实用工具"],"zone":"AI","funScore":1,"domainTags":["模型压缩","边缘AI","移动推理"],"domainKey":"模型压缩","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"实用工具","source":"llm","weight":1},{"name":"cactus","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"on-device-ai","source":"github","weight":0.7},{"name":"edge-ai","source":"github","weight":0.7},{"name":"foundation-model","source":"github","weight":0.7},{"name":"function-calling","source":"github","weight":0.7},{"name":"tinyml","source":"github","weight":0.7},{"name":"tool-calling","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/cactus-compute/needle","ts":"2026-09-21T04:10:58.868Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.40247091594297885,"copyOk":false},{"repo":"ktorio/ktor","owner":"ktorio","name":"ktor","desc":"Framework for quickly creating connected applications in Kotlin with minimal effort","summaryCn":"JetBrains亲儿子，Kotlin异步框架，搭后端服务快准狠","reasonCn":"JetBrains官方出品的异步框架，基于Kotlin协程实现，支持JVM、JavaScript和Native多平台。采用声明式DSL定义路由，中间件链式调用，内置WebSocket和多种序列化方案。相比Spring Boot更轻量灵活，代码类型安全，适合微服务和API后端开发。","stars":14531,"starGrowth":0,"createdAt":"2015-08-03T16:49:36Z","silentRounds":2,"language":"Kotlin","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Kotlin","异步框架","Web服务"],"domainKey":"Kotlin","pushedAt":"2026-09-23T14:34:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ktorio/ktor","ts":"2026-09-23T14:34:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dgraph-io/badger","owner":"dgraph-io","name":"badger","desc":"Fast key-value DB in Go.","summaryCn":"Go语言开发的超快键值数据库，轻量级存储神器","reasonCn":"Badger是用Go语言开发的高性能键值数据库，专为现代云原生应用设计。它采用LSM树存储架构，支持ACID事务、范围查询和复杂数据序列化。与LevelDB和RocksDB等传统数据库相比，Badger完全用Go编写，在Go生态中更易于集成，性能表现优异，IO延迟极低。特别适合需要嵌入式数据库的微服务架构场景，单二进制文件部署，零依赖运行。","stars":15764,"starGrowth":1,"createdAt":"2017-01-26T05:09:49Z","silentRounds":7,"language":"Go","topics":["database"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["键值存储","LSM树","嵌入式数据库","高性能存储"],"domainKey":"键值存储","pushedAt":"2026-09-21T20:08:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/dgraph-io/badger","ts":"2026-09-21T20:08:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Wisser/Jailer","owner":"Wisser","name":"Jailer","desc":"Database Subsetting and Relational Data Browsing Tool.","summaryCn":"从海量数据里挑出你需要的那一小撮，还能看清表格之间啥关系","reasonCn":"Jailer是一款数据库子集化工具，专为从大型生产数据库中提取安全、可用的数据子集而设计。它能智能分析表间关系，自动遵循外键约束，生成完整且一致的小数据集。支持多种主流关系型数据库，让开发者在本地环境即可进行真实数据的测试和开发，无需接触敏感生产数据。","stars":3202,"starGrowth":1,"createdAt":"2011-06-20T13:28:18Z","silentRounds":4,"language":"Java","topics":["frontend","testing"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["数据库子集","数据提取","测试数据管理"],"domainKey":"数据库子集","pushedAt":"2026-09-23T12:25:28Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"testing","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/Wisser/Jailer","ts":"2026-09-23T12:25:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"IBM/mcp-context-forge","owner":"IBM","name":"mcp-context-forge","desc":"An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimizes Agent & Tool calling, and supports plugins.","summaryCn":"想象一下，你的AI Agent就像一个需要不断学习新知识的职员，而外部的各种AP","reasonCn":"它的核心技术在于将MCP协议与多种接口标准进行了深度整合。它具备智能的API发现能力，自动帮你匹配正确的工具，同时还内置了关键的“安全护栏”，比如请求限制、流量控制、以及自动化的API响应缓存。这种设计不仅提高了调用的响应速度，还极大地保障了智能体在生产环境运行时的健壮性","stars":4515,"starGrowth":3,"createdAt":"2025-05-08T08:16:59Z","silentRounds":0,"language":"Python","topics":["devops"],"aiDims":["工具调用","AI基础设施","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["AI网关","Agent框架","API治理","多协议整合"],"domainKey":"AI网关","pushedAt":"2026-09-23T16:08:33Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/IBM/mcp-context-forge","ts":"2026-09-23T16:08:33Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"virgiliojr94/book-to-skill","owner":"virgiliojr94","name":"book-to-skill","desc":"Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work.","summaryCn":"把PDF技术书秒变Claude编程技能，边工作边学习超方便","reasonCn":"该项目将技术书籍PDF自动转换为Claude Code的skill格式，提取书中的核心概念、代码示例和最佳实践，构建可查询的知识库。用户可以在编写代码时直接调用这些skill，无需反复翻阅纸质书或PDF，实现边写边学的高效工作流。","stars":31991,"starGrowth":136,"createdAt":"2026-05-01T02:09:39Z","silentRounds":0,"language":"Python","topics":["llm","rag"],"aiDims":["工具调用","RAG与知识","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["PDF解析","知识库","Claude Code","技能包","RAG"],"domainKey":"RAG","pushedAt":"2026-09-22T17:45:44Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/virgiliojr94/book-to-skill","ts":"2026-09-22T17:45:44Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.37530430433559436,"copyOk":false},{"repo":"cryptomator/cryptomator","owner":"cryptomator","name":"cryptomator","desc":"Cryptomator for Windows, macOS, and Linux: Secure client-side encryption for your cloud storage, ensuring privacy and control over your data.","summaryCn":"给你的云端文件加个物理级保险柜，数据只有你能看","reasonCn":"Cryptomator采用客户端加密架构，在文件上传云端前在本地进行AES-256加密，解密则发生在设备本地。它通过虚拟文件系统技术，让用户像操作普通文件夹一样操作加密数据，无需关心复杂的加密细节。项目支持Windows、macOS和Linux多平台，并提供了浏览器插件以实现跨平台访问。","stars":16176,"starGrowth":2,"createdAt":"2014-02-02T00:43:47Z","silentRounds":0,"language":"Java","topics":["security","cryptography","privacy"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["客户端加密","云存储隐私","AES-256","虚拟文件系统"],"domainKey":"客户端加密","pushedAt":"2026-09-23T10:38:51Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"cryptography","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/cryptomator/cryptomator","ts":"2026-09-23T10:38:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/struts","owner":"apache","name":"struts","desc":"Apache Struts is a free, open-source, MVC framework for creating elegant, modern Java web applications","summaryCn":"老牌Java Web开发框架，虽已过气但历史地位不可撼动","reasonCn":"Apache Struts是经典的Java MVC Web框架，曾统治企业级Web开发多年。虽然近年热度下降，被Spring Boot取代，但它仍是许多遗留系统的基石，对于维护老项目或理解Java Web演进历程仍有重要价值，适合Java后端开发者了解历史与技术债务。","stars":1369,"starGrowth":0,"createdAt":"2014-01-15T08:00:08Z","silentRounds":22,"language":"Java","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Java Web框架","MVC架构","遗留系统维护","Struts 2","Servlet规范"],"domainKey":"Java Web框架","pushedAt":"2026-09-23T13:33:54Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/apache/struts","ts":"2026-09-23T13:33:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"simonrw/rust-fitsio","owner":"simonrw","name":"rust-fitsio","desc":"FFI wrapper around cfitsio in Rust","summaryCn":"第一段：天文观测数据大多数以 FITS（Flexible Image Transport System）格","reasonCn":"如果需要写入，只需调用 `f.create_image(&header, &data)` 即可。项目同时提供了 `ndarray` 的特化实现，方便进行矩阵运算。\n第四段：该库特别适合以下人群：使用 Rust 开发天文数据分析 pipeline 的科研人员；想在高性能计算环境中统一语言栈的团队；以及对内存安全有严格要求的","stars":46,"starGrowth":0,"createdAt":"2016-05-20T00:50:11Z","silentRounds":10,"language":"C","topics":["astronomy"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["天文数据","FITS格式","Rust FFI","CFITSIO","科学计算"],"domainKey":"天文数据","pushedAt":"2026-09-21T20:30:36Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"astronomy","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/simonrw/rust-fitsio","ts":"2026-09-21T20:30:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"paperless-ngx/paperless-ngx","owner":"paperless-ngx","name":"paperless-ngx","desc":"A community-supported supercharged document management system: scan, index and archive all your documents","summaryCn":"你有没有这种经历——家里或办公室堆了一大堆纸质文件，发票、合同、收据，想","reasonCn":"技术层面，它基于Python和Django构建，核心是一个OCR引擎，使用Tesseract和OCRmyPDF来识别文档内容。识别后的文字会被存入PostgreSQL数据库，支持全文检索。系统还会自动提取元数据如日期、发送者、接收者等信息，并为每份文档生成唯一的缩略图和标签。","stars":45920,"starGrowth":75,"createdAt":"2022-02-12T21:56:52Z","silentRounds":0,"language":"Python","topics":["ml","llm","ai"],"aiDims":["文档理解","AI应用","效率工具"],"zone":"工具","funScore":0,"domainTags":["文档管理","OCR","自托管","企业归档"],"domainKey":"文档管理","pushedAt":"2026-09-23T21:33:59Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/paperless-ngx/paperless-ngx","ts":"2026-09-23T21:33:59Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2298998257839721,"copyOk":false},{"repo":"huggingface/diffusers","owner":"huggingface","name":"diffusers","desc":"🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.","summaryCn":"如果你经常在网上看到各种惊艳的AI绘画、AI生成的酷炫视频，或者听过AI合成","reasonCn":"技术架构上，它采用PyTorch构建，对主流的Stable Diffusion、Flux、SVD等各种尖端扩散模型提供了完美支持。它把模型的噪声预测器、调度器、变分自编码器等核心组件进行了模块化解耦，开发者可以像搭积木一样自由组合不同的采样算法和管道。","stars":34581,"starGrowth":14,"createdAt":"2022-05-30T16:04:02Z","silentRounds":0,"language":"Python","topics":["deep-learning"],"aiDims":["模型与训练","创意工具","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["扩散模型","AI生成","图像生成","PyTorch"],"domainKey":"图像生成","pushedAt":"2026-09-23T16:50:32Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/huggingface/diffusers","ts":"2026-09-23T16:50:32Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.043214539776177674,"copyOk":false},{"repo":"mingrammer/go-web-framework-stars","owner":"mingrammer","name":"go-web-framework-stars","desc":":star: Web frameworks for Go, most starred on GitHub","summaryCn":"Go Web框架全面盘点，帮你快速找到最适合项目的HTTP框架","reasonCn":"这是一个Go语言Web框架的精选合集，按GitHub星标数排序收录了最热门的项目。涵盖从轻量级HTTP路由到完整Web框架的各类选择，每个项目都有star数量展示。它本身不是框架，而是帮你快速了解Go生态中有哪些可用选项，适合刚开始接触Go Web开发的工程师做技术选型参考。","stars":3843,"starGrowth":0,"createdAt":"2017-09-05T00:08:59Z","silentRounds":22,"language":"Go","topics":["web-framework"],"aiDims":["开发者工具","效率工具","教程"],"zone":"资源","funScore":0,"domainTags":["Go语言","Web框架","技术选型","开源清单","HTTP框架"],"domainKey":"Go语言","pushedAt":"2026-09-20T03:07:01Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/mingrammer/go-web-framework-stars","ts":"2026-09-20T03:07:01Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"clash-verge-rev/clash-verge-rev","owner":"clash-verge-rev","name":"clash-verge-rev","desc":"A modern GUI client based on Tauri, designed to run in Windows, macOS and Linux for tailored proxy experience","summaryCn":"一套可视化的网络代理管理神器，一键切换全球节点让上网更自由","reasonCn":"Clash Verge Rev是基于Tauri框架构建的现代化代理客户端，支持Windows、macOS和Linux三大平台。它提供直观的图形界面，帮助用户轻松管理代理规则、节点切换和流量监控。相比命令行工具，它的优势在于开箱即用、配置可视化，并且底层集成Clash核心，支持多种代理协议。用户可以直接拖拽节点排序、设置规则分组，还可以导入订阅链接自动更新节点列表，适合需要经常翻墙或管理多节点的用户。","stars":146829,"starGrowth":395,"language":"Rust","topics":["clash","clash-meta","clash-verge","linux","mac","tauri-app","windows","mihomo"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["代理客户端","网络代理","跨平台","翻墙工具"],"domainKey":"代理客户端","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"clash","source":"github","weight":0.7},{"name":"clash-meta","source":"github","weight":0.7},{"name":"clash-verge","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"mac","source":"github","weight":0.7},{"name":"tauri-app","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"mihomo","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.25,"source":"trending","url":"https://github.com/clash-verge-rev/clash-verge-rev","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3080706127536113,"copyOk":false},{"repo":"mainmatter/100-exercises-to-learn-rust","owner":"mainmatter","name":"100-exercises-to-learn-rust","desc":"A self-paced course to learn Rust, one exercise at a time.","summaryCn":"如果你对Rust语言一向感兴趣，但苦于找不到入门台阶，Mainmatter公司推出的","reasonCn":"这套课程的设计核心在于“动手”。它没有长篇大论的理论推导，而是让学习者在解决第1题到第100题的过程中，逐步感知Rust的所有权系统、模式匹配、运算符等核心特性。从“了解什么都没有”到“能够开始编写自己的程序”，这100个练习构成了一个完整的闭环。","stars":9596,"starGrowth":3,"createdAt":"2024-05-12T20:20:12Z","silentRounds":0,"language":"Rust","topics":["exercises","learning-by-doing","rust"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["Rust","编程练习","语言学习","题库"],"domainKey":"Rust","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"exercises","source":"github","weight":0.7},{"name":"learning-by-doing","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.38,"source":"trending","url":"https://github.com/mainmatter/100-exercises-to-learn-rust","ts":"2026-09-22T04:07:30.369Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ChatLab/ChatLab","owner":"ChatLab","name":"ChatLab","desc":"Local-first chat history analyzer with AI. | 本地优先的 AI 聊天记录分析工具","summaryCn":"把你和朋友的聊天记录喂给AI，自动生成精彩摘要和趣味分析","reasonCn":"ChatLab是一款本地优先的AI聊天记录分析工具，基于TypeScript开发，支持导入多种聊天应用格式，如WhatsApp、Telegram等。核心能力包括情感趋势分析、高频话题提取、对话风格识别和关键信息总结。采用本地优先架构，所有数据处理均在本地完成，确保隐私安全，无需上传云端。提供可视化分析报告，帮助用户快速理解聊天历史中的重点内容。","stars":7416,"starGrowth":11,"createdAt":"2025-11-26T05:54:26Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","data-visualization"],"aiDims":["AI应用","工具调用","AI Agent"],"zone":"AI","funScore":0.6,"domainTags":["聊天记录","数据分析","隐私计算"],"domainKey":"聊天记录","pushedAt":"2026-09-23T11:00:26Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/ChatLab/ChatLab","ts":"2026-09-23T11:00:26Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.04493980582524272,"copyOk":false},{"repo":"newton-physics/newton","owner":"newton-physics","name":"newton","desc":"An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.","summaryCn":"让机器人物理模拟像看电影一样丝滑的GPU引擎","reasonCn":"Newton物理引擎是个硬核家伙，专门帮搞机器人的人解决“虚拟世界里的物理规律”问题。在现实里，机器人要移动得符合物理定律，在电脑里跑仿真也一样。但传统的CPU算物理太慢，稍微复杂点就卡顿。Newton直接用GPU来干这个脏活累活，速度飞快，而且精度很高，能处理刚体碰撞、软体形变","stars":5677,"starGrowth":7,"createdAt":"2025-04-22T04:12:07Z","silentRounds":0,"language":"Python","topics":["robotics"],"aiDims":["游戏","创意工具","非AI-实用"],"zone":"工具","funScore":0.7,"domainTags":["物理仿真","机器人","GPU计算"],"domainKey":"物理仿真","pushedAt":"2026-09-23T18:13:55Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/newton-physics/newton","ts":"2026-09-23T18:13:55Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02469913686806412,"copyOk":false},{"repo":"HakanSeven12/OpenCADStudio","owner":"HakanSeven12","name":"OpenCADStudio","desc":"A CAD application built with Rust — 2D/3D drawing, DWG/DXF support, and GPU-accelerated rendering","summaryCn":"用Rust写的开源CAD软件，2D/3D绘图+GPU加速很硬核","reasonCn":"一个用Rust从头构建的开源CAD应用，支持2D/3D绘图、DWG/DXF文件格式读写和GPU加速渲染。跨平台运行，填补了开源CAD领域的空白。对于需要CAD工具但预算有限的个人开发者、学生和小型工作室来说，这是一个极具吸引力的免费替代方案。","stars":2213,"starGrowth":33,"createdAt":"2026-03-22T11:53:50Z","silentRounds":0,"language":"Rust","topics":["cad","dwg","dxf","rust","dwg-compatible","dwg-editor","dwg-files","dwg-parser","dwg-reader","dwg-support","dwg-viewer","dxf-editor","dxf-files","dxf-parser","dxf-reader","dxf-viewer","dxf-writer","dxf-compatible","dxf-support","dwg-writer"],"aiDims":["非AI-实用","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["CAD软件","2D绘图","3D建模","DWG格式","GPU渲染"],"domainKey":"CAD软件","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cad","source":"github","weight":0.7},{"name":"dwg","source":"github","weight":0.7},{"name":"dxf","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"dwg-compatible","source":"github","weight":0.7},{"name":"dwg-editor","source":"github","weight":0.7},{"name":"dwg-files","source":"github","weight":0.7},{"name":"dwg-parser","source":"github","weight":0.7},{"name":"dwg-reader","source":"github","weight":0.7},{"name":"dwg-support","source":"github","weight":0.7},{"name":"dwg-viewer","source":"github","weight":0.7},{"name":"dxf-editor","source":"github","weight":0.7},{"name":"dxf-files","source":"github","weight":0.7},{"name":"dxf-parser","source":"github","weight":0.7},{"name":"dxf-reader","source":"github","weight":0.7},{"name":"dxf-viewer","source":"github","weight":0.7},{"name":"dxf-writer","source":"github","weight":0.7},{"name":"dxf-compatible","source":"github","weight":0.7},{"name":"dxf-support","source":"github","weight":0.7},{"name":"dwg-writer","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.48,"source":"trending","url":"https://github.com/HakanSeven12/OpenCADStudio","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.1724827835517397,"copyOk":true},{"repo":"jacomyal/sigma.js","owner":"jacomyal","name":"sigma.js","desc":"A JavaScript library aimed at visualizing graphs of thousands of nodes and edges","summaryCn":"画大规模社交网络图的神器，几千个节点也能丝滑流畅","reasonCn":"技术上，它的核心优势在于渲染性能。项目底层使用Canvas进行2D渲染，同时也支持WebGL加速，这让它能够处理普通图表库无法承载的数据量。布局算法方面，它内置了力导向布局等多种算法，可以根据数据特性选择最合适的展示方式。","stars":12170,"starGrowth":1,"createdAt":"2012-03-11T15:17:40Z","silentRounds":3,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["图可视化","网络图","力导向布局","Canvas渲染"],"domainKey":"图可视化","pushedAt":"2026-09-16T09:47:20Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/jacomyal/sigma.js","ts":"2026-09-16T09:47:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"trailofbits/algo","owner":"trailofbits","name":"algo","desc":"Set up a personal VPN in the cloud","summaryCn":"在云服务器上一键搭建个人VPN，再也不用担心公共WiFi泄露隐私了","reasonCn":"由安全研究公司Trail of Bits开发，支持AWS、DigitalOcean、GCP等主流云服务商，一条命令就能部署个人VPN。支持WireGuard、OpenVPN等协议，内置DNS泄漏防护和IPv6支持，界面简洁但功能扎实，非常适合注重隐私的用户群体。","stars":30390,"starGrowth":0,"createdAt":"2016-05-15T03:42:48Z","silentRounds":2,"language":"Python","topics":["security"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["个人VPN","WireGuard","安全工具","云部署","隐私保护"],"domainKey":"个人VPN","pushedAt":"2026-09-23T08:45:21Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/trailofbits/algo","ts":"2026-09-23T08:45:21Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"onyx-dot-app/onyx","owner":"onyx-dot-app","name":"onyx","desc":"Open Source AI Platform - AI Chat with advanced features that works with every LLM","summaryCn":"开源AI聊天平台，支持各种大模型还能自动检索知识库回答问题","reasonCn":"Onyx是一款开源AI聊天平台，支持接入OpenAI、Claude、本地模型等多种大语言模型。核心亮点是内置RAG检索增强功能，用户可上传文档构建私有知识库，AI回答时会先检索相关知识再输出答案。界面美观现代，支持多模型并行对比，可作为个人AI助手或团队协作的知识管理工具，开箱即用。","stars":32207,"starGrowth":12,"createdAt":"2023-04-27T06:04:01Z","silentRounds":0,"language":"Python","topics":["ai","llm","rag"],"aiDims":["AI应用","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["RAG检索增强","多模型聊天平台","私有知识库","本地部署AI","企业级AI助手"],"domainKey":"RAG检索增强","pushedAt":"2026-09-23T21:45:04Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/onyx-dot-app/onyx","ts":"2026-09-23T21:45:04Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03711776942900612,"copyOk":false},{"repo":"skypilot-org/skypilot","owner":"skypilot-org","name":"skypilot","desc":"The AI Compute Platform for frontier teams. SkyPilot turns fragmented AI compute into one AI supercomputer, so frontier AI teams build custom intelligence faster.","summaryCn":"一键管理全球云算力，让大模型训练像用本地电脑一样方便。","reasonCn":"SkyPilot 是一款跨云 AI 计算平台，能自动调度和管理 AWS、GCP、Azure 等多种云服务，实现成本优化和资源自动扩展。它将碎片化的云资源整合成统一超算集群，支持大模型训练和部署，显著降低 AI 团队的技术门槛和运营成本。","stars":10641,"starGrowth":8,"createdAt":"2021-08-11T23:32:15Z","silentRounds":0,"language":"Python","topics":["ml","deep-learning"],"aiDims":["AI基础设施","模型部署","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["多云管理","GPU调度","分布式训练","Spot实例优化","AI基础设施"],"domainKey":"多云管理","pushedAt":"2026-09-23T21:28:05Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/skypilot-org/skypilot","ts":"2026-09-23T21:28:05Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.026255427121511136,"copyOk":false},{"repo":"javascript-obfuscator/javascript-obfuscator","owner":"javascript-obfuscator","name":"javascript-obfuscator","desc":"A powerful obfuscator for JavaScript and Node.js","summaryCn":"给JavaScript代码穿隐身衣，让黑客看了都直摇头","reasonCn":"javascript-obfuscator 是一款专业的 JavaScript/Node.js 代码混淆工具，通过字符串数组加密、控制流扁平化、死代码注入等多种混淆策略，有效保护源代码知识产权，防止代码被轻易逆向工程，在开源社区拥有超过 1.6 万 star，被广泛应用于前端项目保护场景。","stars":16270,"starGrowth":1,"createdAt":"2016-05-09T08:16:53Z","silentRounds":1,"language":"TypeScript","topics":["security","cryptography"],"aiDims":["开发者工具","安全工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["代码混淆","安全","AST"],"domainKey":"代码混淆","pushedAt":"2026-09-20T20:23:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"cryptography","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/javascript-obfuscator/javascript-obfuscator","ts":"2026-09-20T20:23:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"denisidoro/navi","owner":"denisidoro","name":"navi","desc":"An interactive cheatsheet tool for the command-line","summaryCn":"一个超酷的可交互式命令行查表工具，再也不怕忘命令了","reasonCn":"这是一个互动式的 Cheat Sheet 工具，能帮你快速回忆各种命令行参数。支持 Bash、Fish、Zsh 等主流 Shell，内置大量常见工具的配置片段。通过 AI 对话风格搜索，输入自然语言即可找到对应的命令示例，极大提升日常开发效率。","stars":17611,"starGrowth":5,"createdAt":"2019-09-20T13:38:20Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行工具","效率工具","Shell","Cheatsheet"],"domainKey":"命令行工具","pushedAt":"2026-09-20T10:07:15Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/denisidoro/navi","ts":"2026-09-20T10:07:15Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01585174038952927,"copyOk":true},{"repo":"ninenines/cowboy","owner":"ninenines","name":"cowboy","desc":"Small, fast, modern HTTP server for Erlang/OTP.","summaryCn":"Erlang生态里最流行的轻量级HTTP服务器，高并发性能极强","reasonCn":"基于Erlang/OTP构建的轻量级HTTP服务器，专为高并发场景设计。支持WebSocket、HTTP/2等现代协议，能在普通硬件上轻松支撑百万级并发连接。与Phoenix框架配合使用可快速搭建后端服务，是erlang生态中最成熟的web框架之一，部署简单且性能稳定。","stars":7521,"starGrowth":0,"createdAt":"2011-03-09T19:55:52Z","silentRounds":22,"language":"Erlang","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Erlang","HTTP服务器","WebSocket","高并发","OTP框架"],"domainKey":"Erlang","pushedAt":"2026-09-16T09:02:05Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Erlang","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/ninenines/cowboy","ts":"2026-09-16T09:02:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tursodatabase/libsql","owner":"tursodatabase","name":"libsql","desc":"libSQL is a fork of SQLite that is both Open Source, and Open Contributions.","summaryCn":"SQLite开源升级版，解锁远程连接和分布式能力","reasonCn":"libSQL是SQLite的开源分支，完整保留了SQLite的全部特性，同时新增了LibSQL扩展支持远程数据库连接、分布式复制以及Fleet分布式架构。开发者可以自由使用、修改和贡献代码。相比商业数据库，它更轻量、部署简单，适合从单机应用到云端部署的各种场景，是数据库开发者的理想选择。","stars":17226,"starGrowth":0,"createdAt":"2022-09-29T18:56:16Z","silentRounds":3,"language":"C","topics":["database"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["SQLite","数据库","分布式","开源"],"domainKey":"SQLite","pushedAt":"2026-09-16T08:12:49Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/tursodatabase/libsql","ts":"2026-09-16T08:12:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"androoAGI/starnet","owner":"androoAGI","name":"starnet","desc":"A living pixel-art station where real AI agents do real work. Local-first desktop agent harness - bring your own key, watch your crew actually run.","summaryCn":"在复古像素风空间里，看着一群AI小人儿在屏幕里忙前忙后地为你打工。","reasonCn":"这是一个本地运行的桌面AI代理控制台，将复杂的后端任务可视化为生动有趣的像素小镇。你只需提供API Key，便能观摩你的数字crew在屏幕中执行任务、交互协作，将枯燥的代码工程转化为充满生活气息的经营游戏。既是功能强大的生产力工具，也是让人忍不住想一直盯着看的精美数字工艺品。","stars":103,"starGrowth":2,"createdAt":"2026-06-14T23:11:55Z","silentRounds":0,"language":"JavaScript","topics":["pixel-art"],"aiDims":["非AI-好玩","AI Agent"],"zone":"AI","funScore":1,"domainTags":["AI Agent","像素艺术","可视化","本地AI"],"domainKey":"AI Agent","pushedAt":"2026-09-23T17:52:14Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/androoAGI/starnet","ts":"2026-09-23T17:52:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cookiecutter/cookiecutter-django","owner":"cookiecutter","name":"cookiecutter-django","desc":"Cookiecutter Django is a framework for jumpstarting production-ready Django projects quickly.","summaryCn":"很多人刚接触Django时，会遇到这样一个痛点：跟着教程跑通了Hello World，但","reasonCn":"这个项目就是来解决这个问题的。它本质上是一个高度封装的Cookiecutter模板——你只需运行一条命令，回答几个简单问题（比如项目名称、邮箱），它就能自动生成一个结构完整的Django工程。内置了PostgreSQL数据库、Redis缓存、Celery异步任务队列、Docker Compose容器编排、GitHub Ac","stars":13609,"starGrowth":0,"createdAt":"2013-08-14T17:43:31Z","silentRounds":5,"language":"Python","topics":["docker"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Django","项目脚手架","Python","后端开发"],"domainKey":"Django","pushedAt":"2026-09-23T05:13:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/cookiecutter/cookiecutter-django","ts":"2026-09-23T05:13:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"reactive-resume/reactive-resume","owner":"reactive-resume","name":"reactive-resume","desc":"A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!","summaryCn":"一个隐私至上、可自定义的开源简历生成器，帮你轻松搞定专业简历。","reasonCn":"Reactive‑Resume 用 TypeScript 打造，可离线部署，支持 AI 辅助填充、模板切换与数据导出，全部本地运行保证信息不泄露，并通过可插件化架构让用户自由定制布局和样式，是兼顾安全与美观的现代简历编辑工具，并支持多语言导入导出。","stars":43331,"starGrowth":78,"createdAt":"2020-03-25T14:08:22Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI写作","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["简历生成","隐私保护","开源工具","文档编辑"],"domainKey":"简历生成","pushedAt":"2026-09-22T22:12:49Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/reactive-resume/reactive-resume","ts":"2026-09-22T22:12:49Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2394002907848884,"copyOk":false},{"repo":"openspug/spug","owner":"openspug","name":"spug","desc":"开源运维平台：面向中小型企业设计的轻量级无Agent的自动化运维平台，整合了主机管理、主机批量执行、主机在线终端、文件在线上传下载、应用发布部署、在线任务计划、配置中心、监控、报警等一系列功能。","summaryCn":"国产开源运维神器，无需Agent，主机管理部署报警一站式搞定。","reasonCn":"Spug是一款面向中小企业的轻量级自动化运维平台，无需在被控主机安装Agent。它整合了主机管理、批量执行、在线终端、文件传输、应用发布、计划任务和监控报警等功能，极大降低了DevOps门槛，让运维工作变得简单直观。","stars":11079,"starGrowth":3,"createdAt":"2018-01-31T15:52:53Z","silentRounds":0,"language":"JavaScript","topics":["devops"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["运维","DevOps","自动化部署","Web终端","主机管理"],"domainKey":"运维","pushedAt":"2026-09-21T03:51:28Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/openspug/spug","ts":"2026-09-21T03:51:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"EverMind-AI/EverOS","owner":"EverMind-AI","name":"EverOS","desc":"One portable memory layer for every AI agent: local-first, Markdown-native, user-owned, and self-evolving across apps, tools, and workflows.","summaryCn":"给所有AI助手装上统一的个人记忆大脑，让你的数据永远只属于你自己","reasonCn":"在底层架构上，它主打本地优先和Markdown原生存储。所有的记忆片段和知识库都以纯文本的Markdown格式保存在本地磁盘，这意味着你可以用任何文本编辑器直接查看和修改它们。通过内置的检索增强生成与智能演化机制，系统能够随着你日常使用的深入，自动整理、归纳和更新这些记忆，让","stars":13143,"starGrowth":10,"createdAt":"2025-10-28T15:09:32Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI Agent","RAG与知识","效率工具"],"zone":"AI","funScore":0.8,"domainTags":["RAG","本地记忆","Markdown","AI Agent"],"domainKey":"RAG","pushedAt":"2026-09-23T17:03:13Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/EverMind-AI/EverOS","ts":"2026-09-23T17:03:13Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03873910066194932,"copyOk":true},{"repo":"opentofu/opentofu","owner":"opentofu","name":"opentofu","desc":"OpenTofu lets you declaratively manage your cloud infrastructure.","summaryCn":"云管平替版 Terraform，一招搞定所有云资源部署。","reasonCn":"作为 Terraform 的开源替代品，OpenTofu 提供声明式基础设施即代码管理能力，支持 AWS、Azure、GCP 等主流云平台。它允许开发者用统一语法编写基础设施配置，实现版本控制和自动化部署，解决了多云环境管理碎片化难题。","stars":30270,"starGrowth":13,"language":"Go","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["基础设施即代码","云管理","Terraform替代","DevOps"],"domainKey":"基础设施即代码","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/opentofu/opentofu","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04028840436075322,"copyOk":false},{"repo":"magwo/elevatorsaga","owner":"magwo","name":"elevatorsaga","desc":"","summaryCn":"项目通过将抽象的电梯调度问题转化为可视化的编程挑战，让玩家直观地感受到","reasonCn":"在核心玩法上，玩家需要利用 JavaScript 语言编写处理函数，通过定义电梯在各个楼层的动作指令，优化等待时间并解决突发的乘客需求。这种设计完美契合了程序员对逻辑控制的掌控欲，同时又能通过流畅的动画演示，让你一眼看出自己的代码策略是否科学，是学习复杂逻辑控制与算法设计","stars":2556,"starGrowth":1,"createdAt":"2014-08-15T01:28:03Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0.68,"domainTags":["编程游戏","逻辑模拟","算法挑战"],"domainKey":"编程游戏","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/magwo/elevatorsaga","ts":"2026-09-21T23:17:36.060Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"feature-sliced/documentation","owner":"feature-sliced","name":"documentation","desc":"Architectural methodology for frontend projects","summaryCn":"前端架构领域超实用的方法论文档，帮你的项目井井有条","reasonCn":"这是Feature-Sliced Design架构模式的核心文档库，详细讲解了如何组织前端代码、划分模块、定义层级关系。它是一套专门解决大型前端项目代码混乱的架构方法论，通过标准化目录结构和命名规范，让团队开发更高效、代码更易维护，特别适合中大型Web应用开发场景。","stars":2368,"starGrowth":1,"createdAt":"2021-01-10T19:37:29Z","silentRounds":10,"language":"MDX","topics":["frontend"],"aiDims":["开发者工具","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["前端架构","代码组织","规范设计","方法论"],"domainKey":"前端架构","pushedAt":"2026-09-23T00:39:40Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"MDX","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/feature-sliced/documentation","ts":"2026-09-23T00:39:40Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lapce/lapce","owner":"lapce","name":"lapce","desc":"Lightning-fast and Powerful Code Editor written in Rust","summaryCn":"Rust编写的高性能代码编辑器，启动速度碾压VS Code体验丝滑","reasonCn":"Lapce是用Rust编写的高性能现代代码编辑器，号称速度最快的编辑器之一。基于Druid框架实现GPU加速渲染，支持智能补全、多光标编辑、内置终端、Git集成等完整开发功能。采用Lua语言编写插件系统，启动速度远超Electron应用。","stars":38864,"starGrowth":1,"createdAt":"2018-02-06T08:41:06Z","silentRounds":2,"language":"Rust","topics":["developer-tools"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["代码编辑器","Rust","开发者工具","GPU渲染"],"domainKey":"代码编辑器","pushedAt":"2026-09-23T00:51:31Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/lapce/lapce","ts":"2026-09-23T00:51:31Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"triggerdotdev/trigger.dev","owner":"triggerdotdev","name":"trigger.dev","desc":"Trigger.dev – build and deploy fully‑managed AI agents and workflows","summaryCn":"让大模型工作流像搭积木一样简单，自动跑后台任务","reasonCn":"核心技术亮点在于它的工作流引擎架构。每个Agent可以被定义为独立的函数，支持输入输出类型约束，确保数据流清晰可控。平台提供了内置的工具调用机制，Agent可以随时发起HTTP请求、查询数据库、调用外部API。相比LangChain这类框架，Trigger.dev更专注于部署和运维层面，让你在开发","stars":16373,"starGrowth":10,"createdAt":"2022-11-30T14:59:07Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI Agent","工作流编排","服务端自动化","任务调度","TypeScript"],"domainKey":"AI Agent","pushedAt":"2026-09-23T16:43:57Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/triggerdotdev/trigger.dev","ts":"2026-09-23T16:43:57Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03183228485921945,"copyOk":true},{"repo":"botpress/botpress","owner":"botpress","name":"botpress","desc":"The open-source hub to build & deploy GPT/LLM Agents ⚡️","summaryCn":"大模型时代的开源超级大脑，轻松构建专属智能体工场。","reasonCn":"技术层面，它采用了前后端分离架构，核心使用TypeScript编写。平台内置了完善的自然语言处理管道、向量数据库对接模块以及强大的状态机工作流引擎。开发者可以通过图形化界面或者代码灵活定义智能体的行为逻辑，让大模型在安全可控的框架内执行复杂的多步推理任务。","stars":14922,"starGrowth":4,"createdAt":"2016-11-16T21:57:59Z","silentRounds":0,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","AI应用","大语言模型"],"zone":"AI","funScore":0,"domainTags":["智能体","工作流引擎","RAG","向量数据库","可视化开发"],"domainKey":"RAG","pushedAt":"2026-09-23T19:30:54Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/botpress/botpress","ts":"2026-09-23T19:30:54Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"10miaomiao/bilimiao2","owner":"10miaomiao","name":"bilimiao2","desc":"","summaryCn":"B站安卓第三方客户端，主打时光机回溯历史热门，可屏蔽关键词与UP主。","reasonCn":"这是一个用安卓原生开发的B站第三方APP，前身是「哔哩哔哩时光机」。它核心保留查看各分区历史时间段热门视频的能力，并新增列表屏蔽关键字和UP主屏蔽功能，让用户能定制更干净的浏览体验，适合喜欢挖掘B站历史内容或需要精简信息流的安卓用户。","stars":2935,"starGrowth":0,"createdAt":"2018-07-24T14:20:33Z","silentRounds":1,"language":"Kotlin","topics":[],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["B站客户端","时光机功能","安卓原生开发","Jetpack Compose"],"domainKey":"B站客户端","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/10miaomiao/bilimiao2","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"heroiclabs/nakama","owner":"heroiclabs","name":"nakama","desc":"Scalable open-source game backend server: multiplayer, matchmaking, leaderboards, chat, and social features for games.","summaryCn":"轻量级游戏后端服务，轻松搞定多人对战和社交功能","reasonCn":"Nakama是一个强大的开源游戏后端服务器，专为多人游戏设计。它内置了匹配系统、排行榜、聊天和社交关系链等功能，开发者无需从零搭建基础设施，就能快速实现复杂的多人交互逻辑。支持多种语言SDK，稳定性经过大量商业游戏验证，是中小型游戏团队的降本增效神器。","stars":13377,"starGrowth":5,"createdAt":"2017-01-13T17:15:12Z","silentRounds":0,"language":"Go","topics":["game"],"aiDims":["非AI-实用","游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["游戏后端","多人游戏","实时通信","游戏服务器"],"domainKey":"游戏后端","pushedAt":"2026-09-22T15:16:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/heroiclabs/nakama","ts":"2026-09-22T15:16:46Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.016121327651939896,"copyOk":false},{"repo":"liam-hq/liam","owner":"liam-hq","name":"liam","desc":"Automatically generates beautiful and easy-to-read ER diagrams from your database.","summaryCn":"数据库表格关系一键生成高清ER图，告别手绘烦恼","reasonCn":"这个工具专门针对数据库实体关系图，能自动读取数据库Schema并生成美观、可读性强的ER图。支持多种格式导出，适合开发文档编写、系统架构梳理和团队协作。虽然和AI技术无关，但作为开发者常用的实用效率工具，能极大提升数据库设计的可视化体验，让复杂的数据结构一目了然，节省大量时间。","stars":5118,"starGrowth":4,"createdAt":"2024-08-07T07:11:45Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["ER-Diagram","Database","Visualization","CLI"],"domainKey":"ER-Diagram","pushedAt":"2026-09-22T17:00:58Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/liam-hq/liam","ts":"2026-09-22T17:00:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Automattic/harper","owner":"Automattic","name":"harper","desc":"Offline, privacy-first grammar checker. Fast, open-source, Rust-powered","summaryCn":"不用联网就能帮你捉文字虫子的隐私安全语法检查器","reasonCn":"基于Rust开发的离线语法检查工具，专为开发者和写作者设计。不依赖云端API，所有处理都在本地完成，确保文本隐私安全。能检测拼写错误、语法问题和标点符号使用不当。支持多种编程语言和Markdown文档，可集成到VS Code、Neovim等编辑器中。采用LLM辅助判断，比传统规则匹配更智能准确。","stars":15839,"starGrowth":30,"createdAt":"2023-10-22T02:58:35Z","silentRounds":0,"language":"Rust","topics":["developer-tools"],"aiDims":["AI写作","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["语法检查","离线","隐私保护","Rust","LLM辅助"],"domainKey":"语法检查","pushedAt":"2026-09-23T21:08:27Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/Automattic/harper","ts":"2026-09-23T21:08:27Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.09568217690510764,"copyOk":false},{"repo":"flexprice/flexprice","owner":"flexprice","name":"flexprice","desc":"Usage-based pricing and billing for developers 🔓 Cloud or self-hosted ⚙️ No-code UI 💰 Realtime usage metering 🎟 Credits & top-ups 🔑 Control feature access","summaryCn":"给开发者量身定制的订阅计费神器，自动搞定用量监控和账单生成。","reasonCn":"这是一个专为开发者设计的B2B SaaS计费平台，支持按用量计费、无代码UI和实时监控。它解决了传统订阅模式在动态计费场景下的痛点，让开发者能专注于产品创新而非财务系统搭建。项目由Go语言编写，提供自托管和云端部署两种方式，通过插件系统灵活扩展。","stars":6930,"starGrowth":157,"language":"Go","topics":["billing","clickhouse","events","go","kafka","postgresql","pricing","usage-based-pricing","stripe","stripe-subscriptions","subscription-management","temporal","hacktoberfest"],"aiDims":["开发者工具","AI基础设施","效率工具"],"zone":"工具","funScore":0,"domainTags":["计费系统","B2B SaaS","开源","Go"],"domainKey":"计费系统","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"billing","source":"github","weight":0.7},{"name":"clickhouse","source":"github","weight":0.7},{"name":"events","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"kafka","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"pricing","source":"github","weight":0.7},{"name":"usage-based-pricing","source":"github","weight":0.7},{"name":"stripe","source":"github","weight":0.7},{"name":"stripe-subscriptions","source":"github","weight":0.7},{"name":"subscription-management","source":"github","weight":0.7},{"name":"temporal","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/flexprice/flexprice","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.36796536796536794,"copyOk":false},{"repo":"openpaperwork/paperwork","owner":"openpaperwork","name":"paperwork","desc":"","summaryCn":"基于Python的本地大语言模型Web交互界面，用于便捷聊天调试。","reasonCn":"Paperwork是一个基于Python开发的Web工具，主要用途是为用户提供一个简洁的前端界面来与本地部署的大语言模型进行交互。它通常作为中间层，连接浏览器与LLM API，使得开发者无需编写复杂的前端代码即可快速测试模型响应。结构轻量，适合用于企业内部的知识助手原型开发或个人使用。","stars":2433,"starGrowth":0,"createdAt":"2011-06-26T13:47:56Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["AI聊天界面","本地LLM","Web应用","Python"],"domainKey":"AI聊天界面","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/openpaperwork/paperwork","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"keepassxreboot/keepassxc","owner":"keepassxreboot","name":"keepassxc","desc":"KeePassXC is a cross-platform community-driven port of the Windows application “KeePass Password Safe”.","summaryCn":"本地加密的密码保险箱，全平台通用告别忘密码的烦恼","reasonCn":"KeePassXC是KeePass的跨平台社区分支，采用AES-256和ChaCha20加密算法保护数据库，支持Argon2密钥派生。提供密码生成器、自动类型、TOTP双因素认证和浏览器扩展一键填充。原生支持Linux、macOS、Windows三大平台，数据本地存储不依赖云端，隐私可控性强。","stars":28908,"starGrowth":11,"createdAt":"2016-02-28T15:52:40Z","silentRounds":0,"language":"C++","topics":["security","privacy"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["密码管理","本地存储","加密工具"],"domainKey":"密码管理","pushedAt":"2026-09-23T06:20:27Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/keepassxreboot/keepassxc","ts":"2026-09-23T06:20:27Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03414155251141553,"copyOk":false},{"repo":"timescale/timescaledb","owner":"timescale","name":"timescaledb","desc":"A time-series database for high-performance real-time analytics packaged as a Postgres extension","summaryCn":"PostgreSQL的时序数据库插件，让时间序列数据查询快如闪电","reasonCn":"基于PostgreSQL构建的时序数据库扩展，专门为时间序列数据的存储和查询性能做了深度优化。内置自动分区、连续聚合、降采样等核心特性，能轻松应对物联网设备、实时监控、金融交易等场景的海量时间数据查询需求，查询性能远超普通关系型数据库，是处理时间序列数据的理想选择。","stars":23564,"starGrowth":5,"createdAt":"2017-03-07T20:03:41Z","silentRounds":0,"language":"C","topics":["database","iot"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["时序数据库","PostgreSQL扩展","实时分析"],"domainKey":"时序数据库","pushedAt":"2026-09-23T17:20:39Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"iot","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/timescale/timescaledb","ts":"2026-09-23T17:20:39Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01563656425055169,"copyOk":true},{"repo":"KnockOutEZ/wigolo","owner":"KnockOutEZ","name":"wigolo","desc":"The go-to web for your AI coding agent — local-first search, fetch, crawl & research over MCP. No API keys, no cloud, $0/query. Public beta.","summaryCn":"不用云端API，本地就能帮AI写代码时上网查资料的神器","reasonCn":"Wigolo是基于MCP协议的本地优先搜索工具，让AI coding agent无需调用任何云端API，就能在本地完成网页搜索、抓取和研究任务。支持零成本查询，保护隐私且不依赖网络，特别适合敏感场景下的自动化编程流程，目前star已超四千。","stars":5390,"starGrowth":17,"createdAt":"2026-04-12T15:04:11Z","silentRounds":0,"language":"TypeScript","topics":["rag","ai-agent"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["MCP协议","本地网页搜索","AI编程助手"],"domainKey":"MCP协议","pushedAt":"2026-09-23T09:57:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/KnockOutEZ/wigolo","ts":"2026-09-23T09:57:11Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0725543599257885,"copyOk":false},{"repo":"google/adk-python","owner":"google","name":"adk-python","desc":"An open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.","summaryCn":"现在大家都想做自己的AI Agent也就是智能体，但市面上的框架要么太重不好魔","reasonCn":"在底层架构上，它采用了高度模块化的设计，把智能体的感知、思考、工具调用和记忆模块拆分得清清楚楚。你可以非常方便地接入各种大语言模型，编写自定义的Python函数作为AI的双手，甚至构建复杂的智能体网络让它们互相开会协作。","stars":21602,"starGrowth":7,"createdAt":"2025-04-01T20:44:40Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["AI Agent","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI Agent","智能体框架","Google AI"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:14:47Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/google/adk-python","ts":"2026-09-23T21:14:47Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.02197213220998056,"copyOk":false},{"repo":"CorsixTH/CorsixTH","owner":"CorsixTH","name":"CorsixTH","desc":"Open source clone of Theme Hospital","summaryCn":"怀旧经营策略游戏，复刻经典医院运营模拟体验","reasonCn":"这是一个开源复刻的经典模拟经营游戏，完美重现了Theme Hospital的玩法与风格。玩家可以扮演院长管理医院，雇佣医护人员，研究疾病，打造高效医疗帝国。全程充满黑色幽默，运行在Lua脚本引擎上，支持跨平台游玩，是老游戏新玩法的绝佳选择。","stars":4566,"starGrowth":1,"createdAt":"2013-08-25T18:16:56Z","silentRounds":3,"language":"Lua","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["模拟经营游戏","Theme Hospital复刻","开源游戏","Lua游戏引擎","怀旧游戏"],"domainKey":"模拟经营游戏","pushedAt":"2026-09-22T21:30:35Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Lua","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/CorsixTH/CorsixTH","ts":"2026-09-22T21:30:35Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Kinnara/ModernWpf","owner":"Kinnara","name":"ModernWpf","desc":"Modern styles and controls for your WPF applications","summaryCn":"让Win10老旧软件穿上现代科技风，颜值瞬间翻倍。","reasonCn":"专为WPF应用打造的一套现代化UI控件库，将Windows传统桌面应用的界面风格升级到类Win10/Win11的现代设计语言。它提供了统一的视觉主题、圆角控件和流畅动画，能让开发者在C#项目中快速实现高颜值界面，无需手写复杂样式代码，特别适合追求用户体验的工具类软件。","stars":4957,"starGrowth":0,"createdAt":"2019-10-18T15:08:32Z","silentRounds":7,"language":"C#","topics":["desktop"],"aiDims":["非AI-好玩","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WPF","UI设计","C#","桌面应用"],"domainKey":"WPF","pushedAt":"2026-09-21T05:35:18Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Kinnara/ModernWpf","ts":"2026-09-21T05:35:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"scalatra/scalatra","owner":"scalatra","name":"scalatra","desc":"Tiny Scala high-performance, async web framework, inspired by Sinatra","summaryCn":"Scala里的小清新Web框架，语法简洁灵活上手快","reasonCn":"Scalatra是一个轻量级Scala Web框架，灵感来自Ruby的Sinatra。它支持异步非阻塞架构，可与Akka Actor系统无缝集成，适合构建高性能REST API。相比重型框架，Scalatra代码量更少、启动更快，是Scala生态中追求简洁开发体验的优选方案。","stars":2648,"starGrowth":0,"createdAt":"2009-05-31T09:20:32Z","silentRounds":22,"language":"Scala","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Scala","Web框架","轻量级","REST API"],"domainKey":"Scala","pushedAt":"2026-09-22T23:30:58Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Scala","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/scalatra/scalatra","ts":"2026-09-22T23:30:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GetStream/Vision-Agents","owner":"GetStream","name":"Vision-Agents","desc":"Open Vision Agents by Stream. Build voice and vision agents quickly with any model or video provider. Uses Stream's edge network for ultra-low latency.","summaryCn":"几行代码就能打造具备超低延迟视听能力的AI智能体框架","reasonCn":"它最牛的地方在于把复杂的流媒体传输和多模态大模型完美结合。依托底层的全球边缘节点网络，它把音视频处理的延迟压到了极致。架构上设计得非常模块化，你可以自由挑选喜欢的大语言模型、语音识别模型以及视频流提供商，像搭积木一样拼出自己的专属Agent。","stars":8140,"starGrowth":2,"createdAt":"2025-08-11T18:02:04Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["多模态Agent","AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["多模态智能体","实时语音","视频理解","低延迟流媒体","AI框架"],"domainKey":"多模态智能体","pushedAt":"2026-09-23T21:19:25Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/GetStream/Vision-Agents","ts":"2026-09-23T21:19:25Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"callstack/agent-device","owner":"callstack","name":"agent-device","desc":"CLI to control iOS and Android devices for AI agents","summaryCn":"让AI助手直接操控你的手机平板，点击滑动打字全自动","reasonCn":"这个项目让AI Agent能直接控制iOS和Android设备，通过命令行发送点击、滑动、输入等操作。适合需要AI自动操作手机应用的场景，比如自动化测试、重复任务执行、远程设备管理，是AI Agent生态里连接数字世界和物理设备的桥梁。","stars":4731,"starGrowth":22,"createdAt":"2026-01-30T18:41:54Z","silentRounds":0,"language":"TypeScript","topics":["mobile","automation"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["AI Agent","移动设备控制","自动化测试","CLI工具","移动端API"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:26:23Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/callstack/agent-device","ts":"2026-09-23T21:26:23Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.09594064679771719,"copyOk":false},{"repo":"skeletonlabs/skeleton","owner":"skeletonlabs","name":"skeleton","desc":"Skeleton is an adaptive design system powered by Tailwind CSS.","summaryCn":"用TailwindCSS打造的设计系统，快速搭建美观的响应式网页界面","reasonCn":"技术上，它不是一个独立的UI框架，而是深度整合Tailwind CSS，通过配置tailwind.config.js来管理颜色、间距、字体等设计令牌。它支持响应式布局和深色模式，组件设计注重无障碍访问。和Ant Design或Material UI不同，它不依赖React或Vue，纯CSS方案，几乎零运行时开销。","stars":6061,"starGrowth":0,"createdAt":"2022-02-23T16:04:46Z","silentRounds":4,"language":"TypeScript","topics":["design"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["设计系统","Tailwind CSS","UI组件","响应式设计","无障碍访问"],"domainKey":"设计系统","pushedAt":"2026-09-22T23:18:31Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/skeletonlabs/skeleton","ts":"2026-09-22T23:18:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"civitai/civitai","owner":"civitai","name":"civitai","desc":"","summaryCn":"该项目解决的是模型碎片化的问题，让分散在各个角落的优秀AI绘画权重能够在","reasonCn":"在技术架构上，该项目采用了高性能的Web框架，能够支撑海量模型的日常检索与下载请求。开发者可以通过学习其源码，了解大规模模型资产分发系统是如何构建的，包括如何处理文件的版本控制、权限管理以及社区互动机制。","stars":7256,"starGrowth":3,"createdAt":"2022-10-11T20:03:20Z","silentRounds":0,"language":"TypeScript","topics":["ai","social-network","stable-diffusion"],"aiDims":["模型部署","模型与训练"],"zone":"AI","funScore":0,"domainTags":["AI绘画","模型托管","社区分享","生成艺术"],"domainKey":"AI绘画","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"social-network","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/civitai/civitai","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"playframework/playframework","owner":"playframework","name":"playframework","desc":"The Community Maintained High Velocity Web Framework For Java and Scala.","summaryCn":"Java/Scala开发者的后端神器，优雅高效地构建高性能Web应用","reasonCn":"一款成熟的企业级Web框架，基于Java和Scala，强调开发效率和运行性能。拥有清晰的文档、非阻塞I/O和高生产力特性，适合构建大型、可扩展的Web应用。虽然与AI无直接关联，但作为部署AI模型后端的基础设施，它是不错的选择。","stars":12617,"starGrowth":0,"createdAt":"2011-09-07T09:24:08Z","silentRounds":23,"language":"Scala","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Web 框架","Java","Scala","异步框架"],"domainKey":"Web 框架","pushedAt":"2026-09-22T21:46:39Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Scala","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/playframework/playframework","ts":"2026-09-22T21:46:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"basicmachines-co/basic-memory","owner":"basicmachines-co","name":"basic-memory","desc":"AI conversations that actually remember. Never re-explain your project to your AI again. Join our Discord: https://discord.gg/tyvKNccgqN","summaryCn":"你有没有过这种经历：刚给AI交代完项目背景，换了个窗口再问，它就全忘了，","reasonCn":"技术核心在于一套精巧的记忆管理系统，它通过智能的向量检索技术，将对话记录转化为AI能读懂的“经验值”。当它检测到当前对话与过往任务相关时，会自动调取历史记忆，这意味着它不需要重新学习就能理解你的意图，甚至能基于之前的反馈进行动态迭代。","stars":4023,"starGrowth":3,"createdAt":"2024-12-02T22:40:43Z","silentRounds":0,"language":"Python","topics":["productivity","knowledge-management"],"aiDims":["AI Agent","效率工具"],"zone":"AI","funScore":0.5,"domainTags":["AI记忆系统","向量检索","RAG","对话管理"],"domainKey":"RAG","pushedAt":"2026-09-23T14:21:21Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/basicmachines-co/basic-memory","ts":"2026-09-23T14:21:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HelixDB/helix-db","owner":"HelixDB","name":"helix-db","desc":"HelixDB is an OLTP graph database with native vector and full-text search built in Rust on Object Storage.","summaryCn":"存图和向量的一站式数据库，让AI既懂关系又懂相似性","reasonCn":"Rust 打造的高性能 OLTP 图-向量混合数据库，基于对象存储构建，专为检索增强生成 (RAG) 设计。它将向量相似度搜索与图关系查询结合，解决传统方案需多种存储拼凑的痛点，让 AI 应用同时具备语义理解和逻辑推理能力。","stars":6083,"starGrowth":81,"createdAt":"2024-11-23T11:02:14Z","silentRounds":0,"language":"Rust","topics":["vector-db","rag"],"aiDims":["向量数据库","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["图数据库","向量数据库","RAG"],"domainKey":"图数据库","pushedAt":"2026-09-23T21:26:32Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/HelixDB/helix-db","ts":"2026-09-23T21:26:32Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.2829473943777741,"copyOk":false},{"repo":"apache/shiro","owner":"apache","name":"shiro","desc":"Apache Shiro is a powerful and easy-to-use Java security framework that performs authentication, authorization, cryptography, and session management","summaryCn":"Java应用的安全小秘书，帮你搞定登录认证和权限管控，代码简洁好上手","reasonCn":"Apache Shiro是Java生态中最轻量的安全框架，核心功能包括认证、授权、加密和会话管理。它不依赖Spring等重量级容器， standalone使用也完全没问题。内置Realm机制，对接数据库、LDAP、OAuth等认证源很方便，权限粒度能细到方法级别，代码风格简洁，适合中小型项目快速落地安全功能。","stars":4463,"starGrowth":0,"createdAt":"2009-08-29T08:00:17Z","silentRounds":6,"language":"Java","topics":["web-framework"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Java安全","认证授权","权限管理"],"domainKey":"Java安全","pushedAt":"2026-09-22T21:48:09Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/apache/shiro","ts":"2026-09-22T21:48:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"martenbjork/github-xp","owner":"martenbjork","name":"github-xp","desc":"","summaryCn":"在浏览器扩展的历史中，GitHub XP是一个将互联网服务与经典操作系统视觉风格","reasonCn":"该项目的兴起迅速引发了开发社区的共鸣，许多人将其作为个性化的“桌宠”或整活工具。从技术角度看，它需要深入修改DOM结构并覆盖GitHub原有的CSS样式，同时模拟原生操作系统的交互逻辑。这种对现有Web应用的深度视觉重构在当时并不多见，展示了前端开发者在CSS布局与DOM操作上的创意","stars":2652,"starGrowth":0,"createdAt":"2018-06-04T08:32:16Z","silentRounds":2,"language":"","topics":[],"aiDims":["非AI-好玩"],"zone":"创意","funScore":0.5599999999999999,"domainTags":["Chrome扩展","Windows模拟","GitHub集成"],"domainKey":"Chrome扩展","pushedAt":"2026-09-21T19:19:33.440Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1}],"aiScore":0.38,"source":"tier","url":"https://github.com/martenbjork/github-xp","ts":"2026-09-21T19:19:33.440Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wxWidgets/wxWidgets","owner":"wxWidgets","name":"wxWidgets","desc":"Cross-Platform C++ GUI Library","summaryCn":"一套代码三端通吃，C++跨平台GUI库，桌面开发者的神器级存在","reasonCn":"wxWidgets是历史最悠久的C++跨平台GUI框架之一，用原生控件渲染而非自绘，确保应用在Windows、macOS、Linux上都保持系统原生外观。支持多后端（MSW、Carbon、GTK），提供完整的事件驱动模型、布局管理器、文档视图架构，性能优异且稳定。","stars":7267,"starGrowth":0,"createdAt":"2011-05-18T05:54:03Z","silentRounds":3,"language":"C++","topics":["desktop"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["跨平台GUI","C++框架","原生渲染"],"domainKey":"跨平台GUI","pushedAt":"2026-09-22T21:25:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/wxWidgets/wxWidgets","ts":"2026-09-22T21:25:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"1Panel-dev/MaxKB","owner":"1Panel-dev","name":"MaxKB","desc":"🔥 MaxKB is an open-source platform for building enterprise-grade agents.  强大易用的开源企业级智能体平台。","summaryCn":"小白也能搭的企业级智能体平台，知识库问答开箱即用","reasonCn":"MaxKB是基于RAG技术的开源智能体平台，无需代码即可搭建企业知识库。支持多种文档解析与向量检索，集成主流大模型。它让没有技术背景的业务人员也能快速构建专属AI助手，实现精准问答与知识沉淀，大幅降低AI应用落地门槛，实用价值极高。","stars":22864,"starGrowth":9,"createdAt":"2023-09-14T02:05:12Z","silentRounds":0,"language":"Python","topics":["rag","llm","chatbot"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["RAG","智能体平台","知识库问答","文档解析","私有化部署"],"domainKey":"RAG","pushedAt":"2026-09-23T11:02:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/1Panel-dev/MaxKB","ts":"2026-09-23T11:02:11Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.028180895731280617,"copyOk":false},{"repo":"vercel/ai","owner":"vercel","name":"ai","desc":"The AI Toolkit for TypeScript. From the creators of Next.js, the AI SDK is a free open-source library for building AI-powered applications and agents","summaryCn":"TypeScript驱动的人工智能SDK","reasonCn":"开发者可以使用该SDK创建各种智能应用，包括聊天机器人、图像识别系统等。它支持多种编程语言，如Python、JavaScript等，并且具有高度的可定制性，可以根据需求调整API接口和功能。此外，它还提供了一些预训练模型，帮助开发者更快地实现想法。","stars":26901,"starGrowth":14,"createdAt":"2023-05-23T15:04:08Z","silentRounds":0,"language":"TypeScript","topics":["llm"],"aiDims":["AI基础设施","AI Agent"],"zone":"AI","funScore":0,"domainTags":["AI SDK","TypeScript","LLM","Agent开发","Vercel"],"domainKey":"AI SDK","pushedAt":"2026-09-23T21:51:41Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/vercel/ai","ts":"2026-09-23T21:51:41Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.043561280249804844,"copyOk":false},{"repo":"HerbertHe/iptv-sources","owner":"HerbertHe","name":"iptv-sources","desc":"","summaryCn":"喜欢在电脑或电视盒子上看直播的朋友经常会遇到源地址失效的问题。","reasonCn":"在部署和使用方面，项目提供了极高的灵活性。你可以选择通过 GitHub Pages 免费托管生成静态直播源页面，也可以使用 Docker 容器或 Node.js 环境在本地或云端服务器上自行搭建。通过配置环境变量，用户还能灵活关闭或设置 GitHub 原始内容代理、启用直播源有效性检查器，甚至添加自","stars":8988,"starGrowth":1,"createdAt":"2023-11-25T06:22:20Z","silentRounds":2,"language":"TypeScript","topics":[],"aiDims":["效率工具"],"zone":"工具","funScore":0.2,"domainTags":["IPTV直播源","自动更新","Docker部署","GitHub Pages"],"domainKey":"IPTV直播源","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/HerbertHe/iptv-sources","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fosrl/pangolin","owner":"fosrl","name":"pangolin","desc":"Modern networking and security platform providing secure access and connectivity to apps, infrastructure, and AI workloads. Connect and protect your users.","summaryCn":"基于WireGuard的安全VPN和反向代理，让本地服务安全暴露在公网","reasonCn":"Pangolin是一款基于WireGuard构建的身份感知VPN和反向代理工具，专为远程访问本地服务而设计。它提供端到端加密通信，支持多种身份验证方式，可以将内网服务安全地暴露到公网。对于需要远程访问NAS、摄像头或开发环境的用户来说非常实用。不过项目相对小众，社区规模和知名度都比较有限。","stars":22878,"starGrowth":11,"createdAt":"2024-09-27T02:16:47Z","silentRounds":0,"language":"TypeScript","topics":["reverse-proxy","wireguard","single-sign-on","self-hosted","oidc","vpn","zero-trust","zero-trust-network-access","ztna","pam","remote-access","ssh","tunneling","nat-traversal","sase","sso","wireguard-vpn","pangolin","ai-gateway","gateway"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["WireGuard","反向代理","内网穿透","网络安全"],"domainKey":"WireGuard","pushedAt":"2026-09-23T19:33:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"reverse-proxy","source":"github","weight":0.7},{"name":"wireguard","source":"github","weight":0.7},{"name":"single-sign-on","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"oidc","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"zero-trust","source":"github","weight":0.7},{"name":"zero-trust-network-access","source":"github","weight":0.7},{"name":"ztna","source":"github","weight":0.7},{"name":"pam","source":"github","weight":0.7},{"name":"remote-access","source":"github","weight":0.7},{"name":"ssh","source":"github","weight":0.7},{"name":"tunneling","source":"github","weight":0.7},{"name":"nat-traversal","source":"github","weight":0.7},{"name":"sase","source":"github","weight":0.7},{"name":"sso","source":"github","weight":0.7},{"name":"wireguard-vpn","source":"github","weight":0.7},{"name":"pangolin","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"gateway","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/fosrl/pangolin","ts":"2026-09-23T19:33:08Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0344424337791765,"copyOk":false},{"repo":"argoproj/argo-cd","owner":"argoproj","name":"argo-cd","desc":"Declarative Continuous Deployment for Kubernetes","summaryCn":"Kubernetes集群的自动化部署管家，Git驱动持续交付","reasonCn":"Argo CD是CNCF孵化的Kubernetes声明式持续部署工具，通过监听Git仓库的变化自动同步应用状态到集群，实现GitOps工作流。支持多集群管理、应用级健康检查、回滚机制和审批工作流，是目前Kubernetes生态中最主流的持续交付解决方案之一。","stars":24223,"starGrowth":11,"createdAt":"2018-02-09T11:12:01Z","silentRounds":0,"language":"Go","topics":["devops","docker"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","持续交付","GitOps","DevOps"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T19:24:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/argoproj/argo-cd","ts":"2026-09-23T19:24:52Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03436234157618792,"copyOk":false},{"repo":"teableio/teable","owner":"teableio","name":"teable","desc":"✨ AI Spreadsheet for Business","summaryCn":"AI加持的智能电子表格，把Notion和Excel的优点融合成了神器","reasonCn":"Teable是一个基于Airtable理念开发的开源智能表格，主打多人协作与AI增强功能。它支持多种字段类型和视图，并内置AI Agent能力，可自动处理数据、生成内容。对于喜欢用表格管理项目、知识库或进行自动化工作的用户来说，是极佳的生产力工具。","stars":21813,"starGrowth":2,"createdAt":"2022-11-01T07:02:49Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","database"],"aiDims":["AI Agent","AI应用","效率工具"],"zone":"AI","funScore":0,"domainTags":["智能表格","AI代理","数据库协作"],"domainKey":"智能表格","pushedAt":"2026-09-23T14:31:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/teableio/teable","ts":"2026-09-23T14:31:32Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Blaizzy/mlx-audio","owner":"Blaizzy","name":"mlx-audio","desc":"A text-to-speech (TTS), speech-to-text (STT) and speech-to-speech (STS) library built on Apple's MLX framework, providing efficient speech analysis on Apple Silicon.","summaryCn":"苹果芯片上的全能语音套件，文本转语音和语音转文字都在行。","reasonCn":"这是一个基于Apple MLX框架构建的语音AI库，支持文本转语音、语音转文本以及语音转语音三种核心功能。它利用Apple Silicon芯片的硬件加速特性，提供高效的语音处理能力。项目由Blaizzy维护，在GitHub上获得了超过7.7千个Star，显示出极高的社区关注度和实用价值。作为一款多模态工具，它打破了纯文本AI的局限，让开发者能轻松在本地实现各种语音交互场景。","stars":7931,"starGrowth":7,"createdAt":"2024-11-27T21:14:34Z","silentRounds":0,"language":"Python","topics":["transformers"],"aiDims":["多模态Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["语音合成","语音识别","Apple MLX"],"domainKey":"语音识别","pushedAt":"2026-09-23T18:59:14Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Blaizzy/mlx-audio","ts":"2026-09-23T18:59:14Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02364783759929391,"copyOk":true},{"repo":"wrenfold/wrenfold","owner":"wrenfold","name":"wrenfold","desc":"Toolkit for generating code from symbolic math expressions.","summaryCn":"只写符号数学，自动生成可编译的C++代码，省时省力的神器","reasonCn":"wrenfold 是用 C++ 实现的符号计算库，能读取 LaTeX 或自定义公式，在编译期完成求导、展开、简化等操作，并生成可直接编译的 C++ 源码。它利用模板元编程和向量化指令，支持多线程加速，适合物理仿真、机器学习前向传播等高性能计算。项目附带 CMake 构建示例，几行代码即可把数学表达式转为高效实现，省去手写繁琐代码的时间。","stars":77,"starGrowth":0,"createdAt":"2021-10-30T22:37:31Z","silentRounds":9,"language":"C++","topics":["mathematics"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["符号计算","代码生成","C++库","数学表达式","性能优化"],"domainKey":"符号计算","pushedAt":"2026-09-19T06:55:59Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/wrenfold/wrenfold","ts":"2026-09-19T06:55:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tuteng/Best-websites-a-programmer-should-visit-zh","owner":"tuteng","name":"Best-websites-a-programmer-should-visit-zh","desc":"","summaryCn":"这款开源项目本质上是一份经过人工筛选的网站导航地图。","reasonCn":"除了情绪疏导和求助，该清单还涵盖了学习路径的优化建议。文档中推荐了 Learn Anything 这样的工具，旨在帮助用户规划学习任意知识时的最佳路线，避免了盲目刷课或阅读无效资料的陷阱。对于新手而言，这里特别设置了针对性的代码训练章节，引导初学者从基础起步。","stars":13719,"starGrowth":0,"createdAt":"2017-06-18T09:27:20Z","silentRounds":5,"language":"","topics":[],"aiDims":["学习资源","非AI-实用","开发者工具"],"zone":"资源","funScore":0,"domainTags":["程序员","网站导航","Awesome List","学习资源","开发者社区"],"domainKey":"程序员","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/tuteng/Best-websites-a-programmer-should-visit-zh","ts":"2026-09-14T23:04:28.868Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"doccano/doccano","owner":"doccano","name":"doccano","desc":"","summaryCn":"doccano 是一款开源的文本标注工具，旨在帮助用户快速构建用于机器学习的标","reasonCn":"该项目最突出的功能在于协作标注。多人可同时进入同一项目，实时看到彼此的标注进度和结果，极大提高团队协作效率。支持多语言标注，能满足多语种数据需求；移动端支持让用户能够在手机或平板上随时随地完成标注任务。","stars":10772,"starGrowth":0,"createdAt":"2018-05-09T06:10:05Z","silentRounds":5,"language":"Python","topics":["natural-language-processing","machine-learning","annotation-tool","python","datasets","dataset","data-labeling","text-annotation","nuxtjs","vue","vuejs","nuxt"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文本标注","数据集构建","协作工具","机器学习","NLP"],"domainKey":"文本标注","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"natural-language-processing","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"annotation-tool","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"datasets","source":"github","weight":0.7},{"name":"dataset","source":"github","weight":0.7},{"name":"data-labeling","source":"github","weight":0.7},{"name":"text-annotation","source":"github","weight":0.7},{"name":"nuxtjs","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"vuejs","source":"github","weight":0.7},{"name":"nuxt","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/doccano/doccano","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"labring/FastGPT","owner":"labring","name":"FastGPT","desc":"FastGPT is a knowledge-based platform built on the LLMs, offers a comprehensive suite of out-of-the-box capabilities such as data processing, RAG retrieval, and visual AI workflow orchestration, letting you easily develop and deploy complex question-answering systems without the need for extensive setup or configuration.","summaryCn":"不用写代码就能搭出懂业务知识的智能客服系统","reasonCn":"FastGPT是基于大模型的开源RAG知识库平台，提供可视化工作流编排、多种数据处理方式和智能检索能力。它让开发者或业务人员能零代码快速搭建专属知识库应用，支持对接主流大模型。相比直接调API，它能帮你把文档变成会问答的智能助手，上手简单且功能强大，是企业和个人构建私有知识库的首选工具之一。","stars":29723,"starGrowth":8,"createdAt":"2023-02-23T16:53:25Z","silentRounds":0,"language":"TypeScript","topics":["rag","llm"],"aiDims":["RAG与知识","AI应用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["RAG知识库","AI工作流编排","智能客服","LLM应用开发"],"domainKey":"RAG知识库","pushedAt":"2026-09-23T10:12:44Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/labring/FastGPT","ts":"2026-09-23T10:12:44Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.024807455505837232,"copyOk":false},{"repo":"googleworkspace/cli","owner":"googleworkspace","name":"cli","desc":"Google Workspace CLI — one command-line tool for Drive, Gmail, Calendar, Sheets, Docs, Chat, Admin, and more. Dynamically built from Google Discovery Service. Includes AI agent skills.","summaryCn":"命令行里的Google全能助手，AI加持更高效","reasonCn":"这是Google官方推出的命令行工具，集成Drive、Gmail、Calendar、Sheets等所有Workspace服务。更令人兴奋的是，它内置了AI agent skills，允许用户通过自然语言指令完成跨服务任务，比如“把上周邮件里的重要文件整理到云端”。这是效率工具与AI Agent结合的典型代表。","stars":31106,"starGrowth":9,"createdAt":"2026-03-02T19:46:06Z","silentRounds":0,"language":"Rust","topics":["ai-agent"],"aiDims":["AI Agent","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Google Workspace","CLI工具","自动化"],"domainKey":"Google Workspace","pushedAt":"2026-09-23T20:10:28Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/googleworkspace/cli","ts":"2026-09-23T20:10:28Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03344159969137787,"copyOk":false},{"repo":"flutter-webrtc/flutter-webrtc","owner":"flutter-webrtc","name":"flutter-webrtc","desc":"WebRTC plugin for Flutter Mobile/Desktop/Web","summaryCn":"Flutter开发者的实时音视频通信神器，轻松实现视频通话和屏幕共享","reasonCn":"flutter-webrtc是Flutter框架的WebRTC插件，支持移动端、桌面端和Web端。它让开发者能在Flutter应用中快速集成实时音视频通话、屏幕共享、数据传输等功能。基于C++底层实现，性能优秀，跨平台兼容性好，是构建社交、在线教育、远程协作等实时通信应用的实用工具。","stars":4491,"starGrowth":0,"createdAt":"2018-03-07T02:26:03Z","silentRounds":7,"language":"C++","topics":["desktop"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["实时音视频","Flutter插件","WebRTC","跨平台开发"],"domainKey":"实时音视频","pushedAt":"2026-09-22T16:52:31Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/flutter-webrtc/flutter-webrtc","ts":"2026-09-22T16:52:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"silexlabs/Silex","owner":"silexlabs","name":"Silex","desc":"Silex is an online tool for visually creating static sites with dynamic data. With the free/libre spirit of internet, together.","summaryCn":"拖拽就能搭出动态网站，不用写代码也能让数据活起来","reasonCn":"一款在线可视化建站工具，通过拖拽方式创建静态网站并接入动态数据，支持MySQL等数据库，内置多种模板和组件。技术栈基于TypeScript，可导出纯静态站点部署到任意服务器或CDN。适合内容创作者快速搭建产品展示页或小型数据驱动型网站。","stars":2977,"starGrowth":0,"createdAt":"2013-09-29T10:22:10Z","silentRounds":3,"language":"TypeScript","topics":["design"],"aiDims":["创意工具","可视化"],"zone":"工具","funScore":0,"domainTags":["可视化建站","静态网站","低代码"],"domainKey":"可视化建站","pushedAt":"2026-09-22T16:27:57Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/silexlabs/Silex","ts":"2026-09-22T16:27:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"addmix/godot_aerodynamic_physics","owner":"addmix","name":"godot_aerodynamic_physics","desc":"A Godot plugin that adds nodes for realistic Aerodynamic simulation and control.","summaryCn":"给 Godot 加逼真气动节点，让飞行更真实","reasonCn":"Godot 开发者常需手写空气动力学模型。此插件提供升力、阻力、侧向力等节点，拖拽配置即可得到逼真飞行物理。兼容 Godot 4，插件代码开源，支持自定义方程，帮助游戏制作者快速实现飞行器、飞机等动态效果，省去繁琐调试。插件内部实现了基于线性方程的升阻模型，并可通过参数调节来模拟不同气象条件。你可以在场景树里直接拖拽 Aerodynamic Body 节点，调整速度、角度、质量等属性，插件会实时更新力学效果。该项目已集成单元测试，持续维护，兼容多平台。","stars":204,"starGrowth":0,"createdAt":"2022-04-21T10:39:45Z","silentRounds":12,"language":"GDScript","topics":["physics"],"aiDims":["游戏","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["游戏开发插件","空气动力学","物理模拟","Godot引擎","飞行模拟器"],"domainKey":"游戏开发插件","pushedAt":"2026-09-19T20:03:02Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"physics","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.43,"source":"search","url":"https://github.com/addmix/godot_aerodynamic_physics","ts":"2026-09-19T20:03:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DayuanJiang/next-ai-draw-io","owner":"DayuanJiang","name":"next-ai-draw-io","desc":"A next.js web application that integrates AI capabilities with draw.io diagrams. This app allows you to create, modify, and enhance diagrams through natural language commands and AI-assisted visualization.","summaryCn":"AI帮你画图，输入文字就能生成专业流程图，告别手动拖拽的时代","reasonCn":"这个项目把draw.io图表工具和AI大语言模型结合，用户只需输入自然语言描述，比如画一个微服务架构图，AI就能自动帮你生成专业流程图。支持创建、修改、增强各种图表，涵盖流程图、UML、网络拓扑等，让不懂绘图的人也能快速产出高质量可视化内容。","stars":36002,"starGrowth":11,"createdAt":"2025-03-23T15:03:48Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI应用","可视化","创意工具"],"zone":"AI","funScore":0.5,"domainTags":["AI绘图","流程图生成","draw.io","架构图可视化"],"domainKey":"AI绘图","pushedAt":"2026-09-23T03:36:39Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/DayuanJiang/next-ai-draw-io","ts":"2026-09-23T03:36:39Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0339166157435698,"copyOk":false},{"repo":"ToolJet/ToolJet","owner":"ToolJet","name":"ToolJet","desc":"Open-source foundation of ToolJet AI - the enterprise app generation platform for internal tools, dashboards, business applications, workflows and AI agents. Build visually, from a prompt, or from Claude Code, Codex and Cursor over MCP 🚀","summaryCn":"零代码搭建企业AI应用和内部工具的开源平台","reasonCn":"开源低代码平台，支持拖拽搭建内部工具和仪表板，连接多种数据源和API，支持JavaScript和Python自定义逻辑，可集成AI Agent构建智能工作流。企业无需大量开发人力，即可快速部署业务应用，支持自托管保障数据安全，是AI时代的企业级应用开发利器。","stars":40981,"starGrowth":8,"createdAt":"2021-03-30T08:51:34Z","language":"JavaScript","topics":["self-hosted"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"工具","funScore":0,"domainTags":["低代码","内部工具开发","AI应用集成","自托管"],"domainKey":"低代码","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/ToolJet/ToolJet","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02458563724652888,"copyOk":true},{"repo":"activist-org/activist","owner":"activist-org","name":"activist","desc":"An open-source activism platform","summaryCn":"开源请愿与活动平台，帮普通人发起社会运动","reasonCn":"技术架构方面，activist 采用 TypeScript 后端，提供了一套完整的 API 接口。它支持多种活动类型，包括请愿签名、线上/线下聚会、社交媒体传播等。与 GitHub 的深度集成是它的一大亮点——创建活动后会自动生成对应的 issue，方便团队协作和进度跟踪。","stars":748,"starGrowth":0,"createdAt":"2021-01-22T21:17:28Z","silentRounds":8,"language":"TypeScript","topics":["backend","frontend","learning"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["社会活动","开源平台","请愿签名"],"domainKey":"社会活动","pushedAt":"2026-09-22T17:12:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"learning","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/activist-org/activist","ts":"2026-09-22T17:12:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chen08209/FlClash","owner":"chen08209","name":"FlClash","desc":"A multi-platform proxy client based on ClashMeta,simple and easy to use, open-source and ad-free.","summaryCn":"免费开源的多平台代理管理客户端，支持ClashMeta核心","reasonCn":"FlClash是一款跨平台代理客户端，基于ClashMeta核心，支持V2Ray、Trojan、Hysteria等多种协议。界面简洁，支持系统代理、规则路由和订阅管理。无需翻墙知识即可配置，适合需要管理多个代理配置的用户，提供稳定高效的网络代理体验。","stars":53059,"starGrowth":97,"createdAt":"2023-08-15T07:47:22Z","silentRounds":0,"language":"Dart","topics":["clash","clash-meta","flutter","hysteria","multi-platform","proxy","v2ray","vless","vpn"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["代理客户端","ClashMeta","网络工具","跨平台"],"domainKey":"代理客户端","pushedAt":"2026-08-17T09:34:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"clash","source":"github","weight":0.7},{"name":"clash-meta","source":"github","weight":0.7},{"name":"flutter","source":"github","weight":0.7},{"name":"hysteria","source":"github","weight":0.7},{"name":"multi-platform","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"v2ray","source":"github","weight":0.7},{"name":"vless","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.58,"source":"bigbro","url":"https://github.com/chen08209/FlClash","ts":"2026-08-17T09:34:23Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.29648446069469836,"copyOk":true},{"repo":"algerkong/AlgerMusicPlayer","owner":"algerkong","name":"AlgerMusicPlayer","desc":"一个第三方音乐播放器、本地服务、桌面歌词、音乐下载、远程控制","summaryCn":"把电脑里的曲库变成随身入口，桌面还能挂歌词","reasonCn":"它用前端框架搭建播放界面，并运行一个本地服务来读取歌曲元数据、处理播放队列和音源请求。除播放本机文件外，也能接入第三方音乐源，把搜索结果整理成可下载音频，同时在桌面显示同步歌词。项目还开放远程控制能力，让手机或平板变成遥控器，实现切歌、暂停、音量调节操作，适合喜欢折腾私有播放器的用户。","stars":16783,"starGrowth":9,"createdAt":"2023-10-27T15:52:55Z","silentRounds":0,"language":"Vue","topics":["music"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["本地音乐服务","桌面歌词","远程控制","音频播放"],"domainKey":"本地音乐服务","pushedAt":"2026-09-19T04:15:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/algerkong/AlgerMusicPlayer","ts":"2026-09-19T04:15:17Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.028608770779955908,"copyOk":true},{"repo":"screenpipe/screenpipe","owner":"screenpipe","name":"screenpipe","desc":"YC (S26) | Record your screen 24/7 and plug into your agents. Local, private, secure. Connect to OpenClaw, Hermes agent and 100+ apps","summaryCn":"24小时本地屏幕录制神器AI自动理解你屏幕上的所有内容还能搜索","reasonCn":"YC孵化的本地屏幕录制AI分析工具，24小时不间断录制屏幕内容并接入AI Agent。支持100+应用集成，本地处理保护隐私。Rust编写性能优异，支持OpenClaw等Agent框架。自动标签搜索让屏幕内容变得可检索，堪称数字生活的第二大脑。","stars":21668,"starGrowth":9,"createdAt":"2024-06-19T13:23:56Z","silentRounds":0,"language":"Rust","topics":["agents","ml"],"aiDims":["多模态Agent","AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["屏幕录制","AI分析","本地隐私"],"domainKey":"屏幕录制","pushedAt":"2026-09-23T20:51:58Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/screenpipe/screenpipe","ts":"2026-09-23T20:51:58Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.028246077164482186,"copyOk":true},{"repo":"microsoft/flint-chart","owner":"microsoft","name":"flint-chart","desc":"🪄 Flint is a visualization language that lets AI agents reliably create expressive, good-looking charts from simple, human-editable chart specs.","summaryCn":"让AI像画图一样简单，生成好看图表的神器","reasonCn":"Flint 是微软推出的可视化语言，旨在让 AI 智能体可靠地通过简单的规格说明创建出美观的图表。它将复杂的图表渲染过程抽象为人类可读的描述，并针对 AI 的输出特点进行了优化。项目目前基于 TypeScript 开发，获得了 4000+ 的 Star，表明了其在数据可视化领域的潜力。","stars":4268,"starGrowth":4,"createdAt":"2026-05-13T00:00:28Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["AI Agent","工具调用","可视化"],"zone":"AI","funScore":0.8,"domainTags":["数据可视化","AI生成","图表"],"domainKey":"数据可视化","pushedAt":"2026-09-23T15:57:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/microsoft/flint-chart","ts":"2026-09-23T15:57:32Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"tldr-pages/tldr","owner":"tldr-pages","name":"tldr","desc":"Collaborative cheatsheets for console commands 📚.","summaryCn":"忘掉冗长的man手册，用简洁示例快速记住命令行技巧","reasonCn":"tldr的设计哲学非常清晰：少即是多。它不追求覆盖所有参数和选项，而是精选每个命令最实用的几个场景，用代码示例直观展示。比如你想知道怎么用curl，它不会把curl所有选项都列出来，而是给出最常见的几个用法示例：下载文件、提交表单、设置headers等。","stars":63718,"starGrowth":9,"createdAt":"2013-12-08T07:34:43Z","silentRounds":0,"language":"Markdown","topics":["command-line","documentation","cheatsheet"],"aiDims":["开发者工具","非AI-实用","学习资源"],"zone":"资源","funScore":0,"domainTags":["命令行","备忘单","文档","开发效率"],"domainKey":"命令行","pushedAt":"2026-09-23T18:20:48Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"documentation","source":"github","weight":0.7},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"Markdown","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/tldr-pages/tldr","ts":"2026-09-23T18:20:48Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.027423742113688438,"copyOk":true},{"repo":"ajeetdsouza/zoxide","owner":"ajeetdsouza","name":"zoxide","desc":"A smarter cd command. Supports all major shells.","summaryCn":"比cd smarter的智能导航工具，记住你常去的目录","reasonCn":"zoxide是一个智能cd替代工具，用Rust编写，通过学习你的目录访问习惯自动推荐目标路径。它不依赖模糊匹配，而是根据访问频率和距离打分排序，输入关键字就能跳到最常去的目录。支持Zsh、Bash、Fish等主流shell，只需一行命令即可完成安装配置。对开发者日常效率提升明显。","stars":39654,"starGrowth":9,"createdAt":"2020-03-05T13:11:40Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["命令行","终端导航","效率工具"],"domainKey":"命令行","pushedAt":"2026-09-21T19:30:17Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/ajeetdsouza/zoxide","ts":"2026-09-21T19:30:17Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02768088969586927,"copyOk":false},{"repo":"huxingyi/autoremesher","owner":"huxingyi","name":"autoremesher","desc":"","summaryCn":"这是一款能将高模网格自动转换为干净四边形拓扑的跨平台实用工具。","reasonCn":"在三维建模和游戏开发中，高多边形模型往往结构杂乱，不便进行后续的动画绑定与细节雕刻。这款工具能够自动把复杂的高模转化为整洁干净的四边形拓扑结构，极大减轻了技术美术和建模师手动拓扑的负担。它基于多个开源了几何处理能力的专业库开发，支持跨平台编译运行，为三维图形工作流提供了一个高效实用的基础处理环节，让繁琐的手工布线工作变得自动化。","stars":3495,"starGrowth":3,"createdAt":"2020-06-17T21:48:08Z","silentRounds":0,"language":"C++","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["网格重构","三维建模","多边形转换","C++"],"domainKey":"网格重构","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/huxingyi/autoremesher","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"shiwenwen/hope-agent","owner":"shiwenwen","name":"hope-agent","desc":"🦭 会记忆、能持续推进目标、会动态编排多 Agent 的跨端桌面 AI 助手，也可服务化常驻 NAS / 云端 | A cross-device desktop AI agent with memory, autonomous goals, dynamic workflows, and headless deployment","summaryCn":"一个有记忆且能自主帮你在电脑上干活的 AI 助手","reasonCn":"技术架构上，使用 Rust 语言确保了极高的运行效率和内存安全，这对于常驻后台的 Agent 至关重要。它最核心的竞争力在于‘动态编排’，即它不会死板地执行脚本，而是根据任务进度动态决定下一步调用哪个子 Agent。同时，它内置了记忆系统，能够从历史交互中学习用户的偏好，真正实现个","stars":1622,"starGrowth":5,"createdAt":"2026-03-13T15:57:46Z","silentRounds":0,"language":"Rust","topics":["chatbot"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI代理","记忆系统","自动化"],"domainKey":"AI代理","pushedAt":"2026-09-23T21:06:35Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/shiwenwen/hope-agent","ts":"2026-09-23T21:06:35Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.029097410604192353,"copyOk":true},{"repo":"huggingface/datasets","owner":"huggingface","name":"datasets","desc":"🤗 The largest hub of ready-to-use datasets for AI models with fast, easy-to-use and efficient data manipulation tools","summaryCn":"给AI喂数据的超级仓库，海量数据集随便挑","reasonCn":"这是一个专门处理AI数据集的Python库，支持从Hugging Face Hub加载数千个预置数据集，也支持本地CSV、JSON、Parquet等多种格式。内置流式处理功能，可以逐个样本读取而无需全部载入内存，大幅简化数据预处理流程，还能直接对接TF Dataset和PyTorch DataLoader。","stars":22004,"starGrowth":5,"createdAt":"2020-03-26T09:23:22Z","silentRounds":0,"language":"Python","topics":["ml","llm"],"aiDims":["AI基础设施","模型与训练","RAG与知识"],"zone":"资源","funScore":0,"domainTags":["Hugging Face","数据集","机器学习"],"domainKey":"Hugging Face","pushedAt":"2026-09-23T13:45:27Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/huggingface/datasets","ts":"2026-09-23T13:45:27Z","category":"learning","momentum":["daily"],"fromOfficial":true,"heatScore":0.015681694237411378,"copyOk":true},{"repo":"0xType/0xProto","owner":"0xType","name":"0xProto","desc":"","summaryCn":"一款降低代码误读的编程字体，让相近字符一眼区分。","reasonCn":"它面向需要长时间写代码的人，把字符辨别作为第一优先。字体文件安装后就能替换编辑器默认显示，减少易混符号带来的阅读摩擦。它把清晰区分相似字母作为设计目标，用来降低编码中的歧义。项目文案还用一句玩笑明确提醒：最终代码可读性仍要使用者自己负责。","stars":2902,"starGrowth":4,"createdAt":"2023-05-07T01:35:24Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["编程字体","源代码可读性","相似字母区分","字体安装"],"domainKey":"编程字体","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.41,"source":"search","url":"https://github.com/0xType/0xProto","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Meloong-Git/PCL","owner":"Meloong-Git","name":"PCL","desc":"","summaryCn":"知名Minecraft启动器开源代码库，供开发者学习其模块化设计。","reasonCn":"这是著名Minecraft启动器Plain Craft Launcher（PCL）的源代码仓库。PCL是国内玩家使用率极高的启动器，以界面流畅、功能丰富著称。这个仓库把启动器大部分代码都公开了，包括UI库、动画、下载和启动游戏的模块，但注意代码更新有滞后，只在官方发布新版本后同步。适合想学习VB.NET或研究启动器原理的开发者。","stars":7305,"starGrowth":8,"createdAt":"2022-07-26T15:43:26Z","silentRounds":0,"language":"Visual Basic .NET","topics":["minecraft","launcher","minecraft-launcher"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Minecraft","启动器","游戏启动","模块化设计"],"domainKey":"Minecraft","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"minecraft","source":"github","weight":0.7},{"name":"launcher","source":"github","weight":0.7},{"name":"minecraft-launcher","source":"github","weight":0.7},{"name":"Visual Basic .NET","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/Meloong-Git/PCL","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.027285420944558523,"copyOk":true},{"repo":"fastfetch-cli/fastfetch","owner":"fastfetch-cli","name":"fastfetch","desc":"A maintained, feature-rich and performance oriented, neofetch like system information tool.","summaryCn":"终端系统信息展示神器，比neofetch更快更酷。","reasonCn":"fastfetch是一个系统信息展示工具，类似neofetch但速度更快、功能更丰富。支持几乎所有主流Linux发行版、macOS和Windows，能够显示CPU、内存、磁盘、GPU等硬件信息，以及系统版本、主题、终端等软件信息，颜值很高。","stars":24730,"starGrowth":9,"createdAt":"2021-02-18T21:25:19Z","silentRounds":0,"language":"C","topics":["command-line"],"aiDims":["非AI-好玩","开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["系统信息","终端工具","Linux","硬件检测"],"domainKey":"系统信息","pushedAt":"2026-09-23T15:09:34Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/fastfetch-cli/fastfetch","ts":"2026-09-23T15:09:34Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02809179134654266,"copyOk":false},{"repo":"Yeachan-Heo/oh-my-claudecode","owner":"Yeachan-Heo","name":"oh-my-claudecode","desc":"Teams-first Multi-agent orchestration for Claude Code","summaryCn":"让Claude Code化身团队指挥官，多智能体协作写代码超爽","reasonCn":"这是一个基于Claude Code的多智能体编排框架，专为团队开发场景设计。它打破了单人使用AI辅助编程的局限，允许多个Claude实例分工协作，分别负责规划、编码、测试等不同任务，通过自动调度实现高效配合，显著提升复杂项目的开发效率与代码质量。","stars":39302,"starGrowth":10,"createdAt":"2026-01-09T03:36:29Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["多智能体","Claude Code","Agent编排","自动化"],"domainKey":"多智能体","pushedAt":"2026-09-23T16:42:30Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Yeachan-Heo/oh-my-claudecode","ts":"2026-09-23T16:42:30Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03691598391939341,"copyOk":false},{"repo":"langchain-ai/open-swe","owner":"langchain-ai","name":"open-swe","desc":"An Open-Source Asynchronous Coding Agent","summaryCn":"让AI自己读代码写代码修bug的异步编程助手","reasonCn":"Open-SWE是LangChain团队推出的开源异步编码Agent，能让AI自主读取代码仓库、理解需求、编写和修复代码。它支持多步骤异步任务编排，集成代码分析和自动生成能力，可独立完成从问题诊断到代码提交的完整流程，是开发者提效的神器。","stars":10753,"starGrowth":11,"createdAt":"2025-05-21T21:44:24Z","silentRounds":0,"language":"Python","topics":["agents"],"aiDims":["AI Agent","代码助手","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["AI编码","Agent","代码生成","LangChain"],"domainKey":"Agent","pushedAt":"2026-09-23T21:49:19Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/langchain-ai/open-swe","ts":"2026-09-23T21:49:19Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.03606891100158095,"copyOk":false},{"repo":"microsoft/microsoft-ui-xaml","owner":"microsoft","name":"microsoft-ui-xaml","desc":"WinUI: a modern UI framework with a rich set of controls and styles to build dynamic and high-performing Windows applications.","summaryCn":"微软官方Windows UI框架，做出炫酷桌面应用的好帮手","reasonCn":"WinUI 3是微软推出的现代化UI框架，基于XAML和Fluent Design设计语言，专为Windows 10/11打造。提供丰富的控件库和样式资源，支持动态主题切换、动画效果、自定义控件等高级特性。底层采用WinAppSDK架构，摆脱了旧版对Windows SDK的强依赖，开发体验更轻量灵活，构建的高性能桌面应用能完美融合系统原生风格。","stars":8462,"starGrowth":5,"createdAt":"2018-07-26T18:45:29Z","silentRounds":0,"language":"C++","topics":["desktop"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Windows桌面开发","UI框架","XAML","Fluent Design"],"domainKey":"Windows桌面开发","pushedAt":"2026-09-23T20:59:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/microsoft/microsoft-ui-xaml","ts":"2026-09-23T20:59:54Z","category":"tool","momentum":["daily"],"fromOfficial":true,"heatScore":0.01677263058378634,"copyOk":false},{"repo":"tiny-craft/tiny-rdm","owner":"tiny-craft","name":"tiny-rdm","desc":"","summaryCn":"Tiny RDM的核心竞争力在于其基于Webview2的技术架构，这一设计使其彻底摆脱","reasonCn":"在功能层面，Tiny RDM不仅仅是简单的数据查看器，它提供了针对Lists、Hashes、Strings、Sets、Sorted Sets以及Streams这六种核心Redis数据类型的可视化CRUD操作。这意味着用户无需记忆复杂的Redis命令行指令，就能通过直观的图形界面完成数据的增删改查。","stars":13120,"starGrowth":2,"createdAt":"2023-06-27T07:06:25Z","silentRounds":0,"language":"Vue","topics":["icon-pack","naive-ui","redis","redis-client","vue","wails","wails-app","web-view","go","golang","redis-gui","redis-manager","redis-desktop","redis-web"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Redis GUI客户端","数据库管理工具","Webview2桌面应用","跨平台数据库工具","连接管理"],"domainKey":"Redis GUI客户端","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"icon-pack","source":"github","weight":0.7},{"name":"naive-ui","source":"github","weight":0.7},{"name":"redis","source":"github","weight":0.7},{"name":"redis-client","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"wails","source":"github","weight":0.7},{"name":"wails-app","source":"github","weight":0.7},{"name":"web-view","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"redis-gui","source":"github","weight":0.7},{"name":"redis-manager","source":"github","weight":0.7},{"name":"redis-desktop","source":"github","weight":0.7},{"name":"redis-web","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/tiny-craft/tiny-rdm","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"harness/harness","owner":"harness","name":"harness","desc":"Harness Open Source is an end-to-end developer platform with Source Control Management, CI/CD Pipelines, Hosted Developer Environments, and Artifact Registries.","summaryCn":"一站式开发平台，代码托管到部署全流程搞定","reasonCn":"Harness是一款自托管的端到端开发者平台，涵盖源码管理、CI/CD流水线、托管开发环境和制品注册表四大核心能力。相比GitHub Actions或GitLab CI，它最大的亮点是能把完整开发工具链部署在私有环境内，企业合规场景特别香。","stars":38420,"starGrowth":8,"createdAt":"2014-02-07T07:54:44Z","silentRounds":0,"language":"Go","topics":["docker","git"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["CI/CD","源码管理","DevOps平台"],"domainKey":"CI/CD","pushedAt":"2026-09-21T16:39:26Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/harness/harness","ts":"2026-09-21T16:39:26Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02462467464862051,"copyOk":false},{"repo":"brndnmtthws/thetagang","owner":"brndnmtthws","name":"thetagang","desc":"","summaryCn":"盈透证券期权自动化交易机器人，支持轮动策略与动态对冲组合管理。","reasonCn":"这是一个针对盈透证券的自动化交易工具，最初实现了经典的期权轮动策略。它已扩展为综合组合管理方案，支持直接股票再平衡、现金管理及VIX看涨期权对冲。通过制度感知再平衡和交易时段门控，它能自动执行复杂配置，帮助投资者以系统化方式管理风险并优化收益。","stars":2730,"starGrowth":0,"createdAt":"2020-12-03T20:17:09Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["IBKR接口","期权策略","组合再平衡","现金管理","VIX对冲"],"domainKey":"IBKR接口","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/brndnmtthws/thetagang","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"noisefactorllc/noisemaker","owner":"noisefactorllc","name":"noisemaker","desc":"A friendly and extendable shader art engine.","summaryCn":"写 Shader 代码就能生成迷幻画作的编程玩具。","reasonCn":"在技术层面，Noisemaker 的核心是 GLSL 着色器语言。它提供了一个友好的抽象层，隐藏了复杂的图形 API 细节。它的强大之处在于强大的“噪声”算法支持，能够生成平滑且自然的随机纹理，配合时间变量，可以做出极具动感的动画效果。","stars":121,"starGrowth":0,"createdAt":"2017-03-25T23:09:15Z","silentRounds":17,"language":"JavaScript","topics":["generative-art","creative-coding"],"aiDims":["创意工具","游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["生成艺术","GLSL","创意编程"],"domainKey":"生成艺术","pushedAt":"2026-09-23T18:34:52Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"generative-art","source":"github","weight":0.7},{"name":"creative-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/noisefactorllc/noisemaker","ts":"2026-09-23T18:34:52Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"risesoft-y9/Digital-Infrastructure","owner":"risesoft-y9","name":"Digital-Infrastructure","desc":"数字底座是一款面向大型政府、企业数字化转型，基于身份认证、组织架构、岗位职务、应用系统、资源角色、数据目录、安全控制等功能构建的统一且安全的管理支撑平台。数字底座基于三员管理模式，具备微服务、多租户、容器化和国产化，支持用户利用代码生成器快速构建自己的业务应用，同时可关联诸多成熟且好用的内部生态应用。","summaryCn":"面向政府和企业的数字化底座平台，国产微服务架构","reasonCn":"一套面向政府企业数字化转型的统一管理平台，基于微服务、多租户、容器化架构，支持身份认证、组织架构、资源角色等核心功能。提供代码生成器快速构建业务应用，支持国产化环境部署，适合中大型组织搭建信息化系统。","stars":2613,"starGrowth":0,"createdAt":"2023-06-26T03:13:33Z","silentRounds":22,"language":"Java","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["政务数字化","微服务架构","多租户管理","代码生成器","身份认证"],"domainKey":"政务数字化","pushedAt":"2026-09-22T07:14:24Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/risesoft-y9/Digital-Infrastructure","ts":"2026-09-22T07:14:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mastra-ai/mastra","owner":"mastra-ai","name":"mastra","desc":"Mastra is the modern TypeScript framework for AI-powered applications and agents.","summaryCn":"TypeScript开发者的AI应用脚手架，搭个聊天机器人几分钟搞定","reasonCn":"mastra是专为TypeScript开发者打造的AI应用开发框架，深度集成Next.js生态，提供完整的agent构建、评估和MCP协议支持。它让开发者能快速搭建智能客服、自动化工作流等AI应用，同时内置了代码级调试和性能评估工具，是目前TypeScript生态中最全能的AI开发底座之一。","stars":28267,"starGrowth":15,"createdAt":"2024-08-06T20:44:31Z","silentRounds":0,"language":"TypeScript","topics":["agents","ai","chatbots","evals","javascript","llm","mcp","nextjs","nodejs","reactjs","tts","typescript","workflows"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI开发框架","TypeScript","Agent框架","MCP协议"],"domainKey":"AI开发框架","pushedAt":"2026-09-23T21:50:25Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"chatbots","source":"github","weight":0.7},{"name":"evals","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"reactjs","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"workflows","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"bigbro","url":"https://github.com/mastra-ai/mastra","ts":"2026-09-23T21:50:25Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.046591962358934445,"copyOk":false},{"repo":"rasbt/LLMs-from-scratch","owner":"rasbt","name":"LLMs-from-scratch","desc":"Implement a ChatGPT-like LLM in PyTorch from scratch, step by step","summaryCn":"从零手写一个ChatGPT，彻底搞懂大模型到底咋运行","reasonCn":"这本书配套代码手把手教你用纯Python和PyTorch，从最基础的字符级RNN开始，一步步构建出Transformer架构，并训练出能对话的迷你大模型。作者Daniel Ruszcz的讲解非常细致，每一个数学公式都配有直观的代码实现，让原本晦涩难懂的理论变得触手可及。","stars":105403,"starGrowth":19,"createdAt":"2023-07-23T18:15:57Z","silentRounds":0,"language":"Jupyter Notebook","topics":["ml","llm","ai","transformers"],"aiDims":["教程","模型与训练","大语言模型"],"zone":"资源","funScore":0.5,"domainTags":["PyTorch","Transformer","深度学习","教学代码","LLM"],"domainKey":"PyTorch","pushedAt":"2026-09-22T13:58:26Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/rasbt/LLMs-from-scratch","ts":"2026-09-22T13:58:26Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05754078157168201,"copyOk":true},{"repo":"tModLoader/tModLoader","owner":"tModLoader","name":"tModLoader","desc":"A mod to make and play Terraria mods. Supports Terraria 1.4 (and earlier) installations ","summaryCn":"给泰拉瑞亚装的超级模组神器，让你自己设计装备地图和怪物","reasonCn":"这是泰拉瑞亚游戏的官方模组加载器，支持C#语言编写各类MOD。你可以自己制作新武器、新 Boss、新生物群落，甚至完全改变游戏机制。提供图形化模组管理界面，一键安装卸载社区作品。社区生态成熟，Steam创意工坊直接对接，下载即玩。","stars":5665,"starGrowth":4,"createdAt":"2015-07-21T05:30:07Z","silentRounds":0,"language":"C#","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["泰拉瑞亚","游戏模组","沙盒游戏","C#开发"],"domainKey":"泰拉瑞亚","pushedAt":"2026-09-23T21:25:19Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/tModLoader/tModLoader","ts":"2026-09-23T21:25:19Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Bitterbot-AI/bitterbot-desktop","owner":"Bitterbot-AI","name":"bitterbot-desktop","desc":"A local-first AI agent with persistent memory, emotional intelligence, and a peer-to-peer skills economy.","summaryCn":"如果说一般的AI是毫无感情的翻译机，那么Bitterbot就是一个性格鲜明的数字伙","reasonCn":"技术架构上，它采用了本地优先（Local-first）的原则，优先保障用户隐私，所有对话数据都存储在本地。它利用先进的Agent框架构建，集成了情感分析模型和持久化记忆模块，同时还包含一个P2P网络层，让你的AI可以“学习”并输出特定的技能，这种架构在目前的AI应用中显得非常前卫。","stars":2460,"starGrowth":1,"createdAt":"2026-03-28T15:15:01Z","silentRounds":15,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","非AI-好玩"],"zone":"AI","funScore":1,"domainTags":["AI 伴侣","本地 Agent","情感计算","P2P"],"domainKey":"AI 伴侣","pushedAt":"2026-09-23T21:39:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Bitterbot-AI/bitterbot-desktop","ts":"2026-09-23T21:39:45Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"algorithm-visualizer/algorithm-visualizer","owner":"algorithm-visualizer","name":"algorithm-visualizer","desc":"","summaryCn":"交互式在线算法可视化工具，支持多种编程语言展示算法运行过程。","reasonCn":"这是一个让算法“活”起来的在线平台。它不仅能展示算法结果，还能让你看到代码一步步运行的动画过程。内置了丰富的教程、文章和视频，无论是学生、老师还是开发者，都能用它来直观理解排序、搜索等复杂算法背后的逻辑。","stars":48738,"starGrowth":1,"createdAt":"2016-05-15T00:27:45Z","silentRounds":2,"language":"JavaScript","topics":["algorithm","data-structure","visualization","animation"],"aiDims":["可视化","学习资源","教程"],"zone":"工具","funScore":1,"domainTags":["算法可视化","交互式教学","编程学习","数据结构"],"domainKey":"算法可视化","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"algorithm","source":"github","weight":0.7},{"name":"data-structure","source":"github","weight":0.7},{"name":"visualization","source":"github","weight":0.7},{"name":"animation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/algorithm-visualizer/algorithm-visualizer","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"PixelGuys/Cubyz","owner":"PixelGuys","name":"Cubyz","desc":"Voxel sandbox game with a large render distance, procedurally generated content and some cool graphical effects.","summaryCn":"这是一款体素沙盒游戏，能生成超远视距的方块世界","reasonCn":"用Zig语言开发的体素沙盒游戏，支持超大渲染距离和程序化地形生成，自带一些酷炫的图形效果。你可以像MC一样自由建造和探索，程序化地形生成让每次冒险都独一无二。Zig语言赋予接近C的性能优势，即使地图范围巨大也能保持流畅，同时配备了水面反射和光影效果等现代图形渲染技术。","stars":3668,"starGrowth":1,"createdAt":"2019-02-04T10:16:37Z","silentRounds":4,"language":"Zig","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["体素游戏","沙盒","Zig语言","程序化地形","超远视距"],"domainKey":"体素游戏","pushedAt":"2026-09-20T07:55:13Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Zig","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/PixelGuys/Cubyz","ts":"2026-09-20T07:55:13Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dailydotdev/daily","owner":"dailydotdev","name":"daily","desc":"daily.dev is the personalized developer news feed and community. Get the best tech content from 2,000+ sources in your browser new tab or on mobile. Free and open source.","summaryCn":"聚合2000+技术源的智能资讯流，新标签页就能看到最新资讯","reasonCn":"Daily是daily.dev开源版的核心项目，一个为开发者定制的个性化新闻聚合平台。它聚合了2000多个技术来源的内容，通过浏览器新标签页或移动端APP呈现，支持自定义内容源和过滤规则。技术上采用可扩展的插件架构，可以轻松添加新的RSS源或API，通过算法推荐用户感兴趣的技术内容，让开发者不用到处翻就能掌握最新动态。","stars":20071,"starGrowth":0,"createdAt":"2017-11-10T07:30:51Z","silentRounds":3,"language":"JavaScript","topics":["developer-tools"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["开发者资讯","浏览器扩展","技术新闻","内容聚合"],"domainKey":"开发者资讯","pushedAt":"2026-09-14T11:13:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/dailydotdev/daily","ts":"2026-09-14T11:13:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hpcaitech/ColossalAI","owner":"hpcaitech","name":"ColossalAI","desc":"Making large AI models cheaper, faster and more accessible","summaryCn":"让训练百亿参数大模型变得便宜又快，是AI模型训练界的“降本增效”专家。","reasonCn":"这是由高性能计算团队打造的深度学习加速系统，通过多维并行策略极大地降低了大模型的训练门槛。它支持单机多卡到超大规模算力集群的无缝扩展，不仅提升了训练速度，更通过显存优化技术让普通显卡也能跑起超大模型，是目前构建AI基础设施与模型训练领域的标杆级工具，助力企业与研究者以极低成本实现模型从零到一的快速迭代与高效部署。","stars":41440,"starGrowth":0,"createdAt":"2021-10-28T16:19:44Z","silentRounds":1,"language":"Python","topics":["ai"],"aiDims":["模型与训练","AI基础设施","大语言模型"],"zone":"AI","funScore":0,"domainTags":["大模型训练","分布式训练","PyTorch","显存优化"],"domainKey":"大模型训练","pushedAt":"2026-09-23T11:22:44Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/hpcaitech/ColossalAI","ts":"2026-09-23T11:22:44Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"astral-sh/ruff","owner":"astral-sh","name":"ruff","desc":"An extremely fast Python linter and code formatter, written in Rust.","summaryCn":"用Rust写的Python代码检查工具，比Pylint快100倍","reasonCn":"Ruff的核心是一个用Rust编写的解析器，它利用Rust的高性能特性，实现了对Python代码的极速解析和检查。它不仅能做传统的Lint和格式化，还支持代码安全扫描和类型检查。它的规则库非常丰富，兼容了绝大多数的配置文件，迁移成本几乎为零。","stars":49750,"starGrowth":13,"language":"Rust","topics":["linter","pep8","python","python3","rust","rustpython","static-analysis","static-code-analysis","style-guide","styleguide","ruff"],"aiDims":["开发者工具","效率工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["Python Linter","代码格式化","Rust","静态分析","开发工具"],"domainKey":"Python Linter","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"linter","source":"github","weight":0.7},{"name":"pep8","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"python3","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"rustpython","source":"github","weight":0.7},{"name":"static-analysis","source":"github","weight":0.7},{"name":"static-code-analysis","source":"github","weight":0.7},{"name":"style-guide","source":"github","weight":0.7},{"name":"styleguide","source":"github","weight":0.7},{"name":"ruff","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/astral-sh/ruff","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03978391959798995,"copyOk":true},{"repo":"simplex-chat/simplex-chat","owner":"simplex-chat","name":"simplex-chat","desc":"SimpleX - the first messaging network operating without user identifiers of any kind - 100% private by design! iOS, Android and desktop apps 📱!","summaryCn":"无用户ID的极致隐私聊天软件，彻底告别追踪","reasonCn":"SimpleX是首款不依赖任何用户标识符的 messaging 网络，从设计源头就实现了100%隐私保护。它支持iOS、Android和桌面端，没有手机号、没有邮箱、没有用户名，彻底杜绝了基于身份数据的追踪和元数据泄露，是隐私爱好者的终极选择。","stars":19465,"starGrowth":3,"createdAt":"2019-12-21T09:50:09Z","silentRounds":0,"language":"Haskell","topics":["security","privacy"],"aiDims":["安全工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":1,"domainTags":["隐私通讯","去中心化","即时通讯"],"domainKey":"隐私通讯","pushedAt":"2026-09-23T21:15:55Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"Haskell","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/simplex-chat/simplex-chat","ts":"2026-09-23T21:15:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rzashakeri/beautify-github-profile","owner":"rzashakeri","name":"beautify-github-profile","desc":"","summaryCn":"一个帮助开发者轻松美化GitHub个人主页README的创意开源工具集","reasonCn":"这个项目专门为想要提升GitHub个人主页颜值的开发者打造，提供从徽章装饰到智能生成器的一整套解决方案。用户无需掌握复杂设计技能，就能通过内置的widgets和icons快速组装出专业级的个人展示页面，让访客一眼记住你的技术栈和成果。","stars":12527,"starGrowth":0,"createdAt":"2022-05-13T13:39:42Z","silentRounds":1,"language":"","topics":["github","github-profile","github-profile-readme","profile","profile-readme","readme","readme-profile","beautifygithubprofile","githubprofile","awsome","awsome-profile","github-api","markdown"],"aiDims":["非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["GitHub主页美化","README生成器","开发者个人品牌","徽章装饰组件","个人展示页面"],"domainKey":"GitHub主页美化","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"github","source":"github","weight":0.7},{"name":"github-profile","source":"github","weight":0.7},{"name":"github-profile-readme","source":"github","weight":0.7},{"name":"profile","source":"github","weight":0.7},{"name":"profile-readme","source":"github","weight":0.7},{"name":"readme","source":"github","weight":0.7},{"name":"readme-profile","source":"github","weight":0.7},{"name":"beautifygithubprofile","source":"github","weight":0.7},{"name":"githubprofile","source":"github","weight":0.7},{"name":"awsome","source":"github","weight":0.7},{"name":"awsome-profile","source":"github","weight":0.7},{"name":"github-api","source":"github","weight":0.7},{"name":"markdown","source":"github","weight":0.7}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/rzashakeri/beautify-github-profile","ts":"2026-09-13T22:08:09.499Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hubotio/hubot","owner":"hubotio","name":"hubot","desc":"","summaryCn":"Hubot 的 codebase 目前已完全转换为 ESM 现代模块规范，这标志着该框架彻底","reasonCn":"作为一款模拟 GitHub 内部 Campfire 机器人的框架，它的核心优势在于其极简的脚本化设计。你可以像添加插件一样，快速为机器人注入如搜索天气、查询库存或执行服务器命令等功能。这种模块化的设计逻辑非常适合那些希望在办公协作平台中引入自动化能力的团队，既能通过命令提高沟通","stars":16796,"starGrowth":0,"createdAt":"2011-08-27T08:30:59Z","silentRounds":5,"language":"JavaScript","topics":["hubot","chat","bot"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["聊天机器人框架","办公自动化","脚本插件系统","团队协作工具","ESM模块化"],"domainKey":"聊天机器人框架","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"hubot","source":"github","weight":0.7},{"name":"chat","source":"github","weight":0.7},{"name":"bot","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/hubotio/hubot","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gastownhall/beads","owner":"gastownhall","name":"beads","desc":"Beads - A memory upgrade for your coding agent","summaryCn":"给AI编程助手装上一个持久记忆，让它不再忘东忘西记不住上下文","reasonCn":"beads是为AI编码代理设计的记忆层组件，通过向量数据库持久化存储对话历史和项目上下文。它解决了AI助手在多轮对话中丢失上下文的核心问题，让编码代理能够记住之前的交互和项目状态，实现更连贯的智能辅助编程体验。","stars":27366,"starGrowth":21,"createdAt":"2025-10-12T03:09:46Z","silentRounds":0,"language":"Go","topics":["agents"],"aiDims":["AI Agent","向量数据库","RAG与知识"],"zone":"AI","funScore":1,"domainTags":["RAG","向量数据库","AI记忆","编程Agent"],"domainKey":"RAG","pushedAt":"2026-09-23T12:46:48Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/gastownhall/beads","ts":"2026-09-23T12:46:48Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.07836255207191406,"copyOk":false},{"repo":"alibaba/skill-up","owner":"alibaba","name":"skill-up","desc":"An evaluation and evolution tool for Agent Skills.","summaryCn":"给AI Agent的技能装个“进化轮”，让它们越用越强。","reasonCn":"这是一个专为AI智能体设计的技能评估与进化框架，旨在解决智能体在实际应用中技能退化或无法适应新任务的核心痛点。通过自动化测试与反馈机制，它能持续评估智能体调用工具的能力，并自动生成针对性的训练数据，实现从“静态技能”到“动态进化”的转变。","stars":1051,"starGrowth":14,"language":"Go","topics":["agent-skills","alibaba","ai","ai-agents","skills"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["Agent评估","技能进化","自动反馈","训练数据生成"],"domainKey":"Agent评估","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"alibaba","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/alibaba/skill-up","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.08196194100856327,"copyOk":false},{"repo":"Ruben2776/PicView","owner":"Ruben2776","name":"PicView","desc":"Fast, free and customizable picture viewer","summaryCn":"国产开源看图神器，比系统自带流畅十倍，支持格式多到离谱","reasonCn":"这款开源看图工具基于C#开发，专为Windows系统设计，主打极速加载与高自由度自定义。相比系统自带看图，它原生支持RAW、HEIC、SVG等丰富格式，且内存占用极低。界面干净无广告，支持插件扩展与主题换肤，无论是日常浏览照片还是专业图像处理，都能提供流畅且个性化的视觉体验。","stars":3589,"starGrowth":2,"createdAt":"2017-01-23T08:52:18Z","silentRounds":0,"language":"C#","topics":["desktop"],"aiDims":["非AI-实用","可视化","非AI-好玩"],"zone":"工具","funScore":0.6,"domainTags":["图片查看","Windows应用","图像格式"],"domainKey":"图片查看","pushedAt":"2026-09-23T21:19:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/Ruben2776/PicView","ts":"2026-09-23T21:19:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"composer/composer","owner":"composer","name":"composer","desc":"","summaryCn":"Packagist.org 平台汇聚了海量的 PHP 组件，Composer 正是通过与该仓库的无","reasonCn":"该项目的核心亮点在于其出色的版本语义控制，它能够精确地管理项目需要加载的外部代码版本，避免了由于组件更新导致的意外报错。同时，它支持通过简单的配置文件来定义项目结构，当你切换开发设备或部署至服务器时，仅需运行安装指令，系统便会自动补齐所有缺失的依赖，确保本地环","stars":29524,"starGrowth":0,"createdAt":"2011-06-08T08:53:00Z","silentRounds":3,"language":"PHP","topics":["php","composer","dependency-manager","package-manager","packages"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["PHP 依赖管理","包管理器","版本控制","命令行工具"],"domainKey":"PHP 依赖管理","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"php","source":"github","weight":0.7},{"name":"composer","source":"github","weight":0.7},{"name":"dependency-manager","source":"github","weight":0.7},{"name":"package-manager","source":"github","weight":0.7},{"name":"packages","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/composer/composer","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opentiny/tiny-vue","owner":"opentiny","name":"tiny-vue","desc":"TinyVue is an enterprise-class UI component library of OpenTiny community, support both Vue.js 2 and Vue.js 3, as well as PC and mobile.","summaryCn":"腾讯开源的企业级Vue组件库，兼容Vue2和Vue3双版本","reasonCn":"TinyVue是腾讯开源的企业级前端UI组件库，同时支持Vue 2和Vue 3，覆盖PC和移动端场景。提供超过100个高质量组件，内置国际化、主题定制、无障碍访问等企业级特性，适合中大型项目的快速搭建，是Element UI / Ant Design Vue之外的优质替代方案。","stars":2271,"starGrowth":0,"createdAt":"2022-11-26T07:32:44Z","silentRounds":22,"language":"Less","topics":["frontend"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["Vue组件库","前端框架","企业级UI","双版本兼容"],"domainKey":"Vue组件库","pushedAt":"2026-09-17T06:34:59Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Less","source":"language","weight":0.5}],"aiScore":0.48,"source":"search","url":"https://github.com/opentiny/tiny-vue","ts":"2026-09-17T06:34:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"gitkraken/vscode-gitlens","owner":"gitkraken","name":"vscode-gitlens","desc":"Supercharge Git inside VS Code and unlock untapped knowledge within each repository — Visualize code authorship at a glance via Git blame annotations and CodeLens, seamlessly navigate and explore Git repositories, gain valuable insights via rich visualizations and powerful comparison commands, and so much more","summaryCn":"开发过程中最头疼的莫过于看到一段诡异的代码，却不知道是谁写的、为什么写","reasonCn":"它不仅是简单的Blame工具，还内置了极其强大的可视化对比系统。你可以通过CodeLens快速查看函数调用历史，利用交互式的Git历史面板查看整个项目的演进路径。它甚至能帮你管理复杂的分支合并冲突，让你在合并代码时像玩游戏一样直观地看到每一个变更。","stars":9929,"starGrowth":1,"createdAt":"2016-08-08T14:50:30Z","silentRounds":7,"language":"TypeScript","topics":["git"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Git","代码分析","可视化","VSCode插件"],"domainKey":"Git","pushedAt":"2026-09-23T21:24:44Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/gitkraken/vscode-gitlens","ts":"2026-09-23T21:24:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cobusgreyling/loop-engineering","owner":"cobusgreyling","name":"loop-engineering","desc":"Practical patterns, starters & CLI tools for loop engineering with AI coding agents. Design systems that prompt and orchestrate agents (inspired by Addy Osmani and Boris Cherny). Includes loop-audit, loop-init, loop-cost.","summaryCn":"让AI编程Agent形成闭环的智能自动化框架","reasonCn":"这个项目聚焦于AI编程Agent的循环工程（Loop Engineering），提供了设计、编排和审计AI编码代理的实用模式和工具链。它帮助开发者构建能自我反思、自我修复的AI编程系统，让AI不再是一次性问答，而是形成持续迭代的智能工作流。包含CLI工具loop-audit等组件，支持多Agent协作场景。","stars":11284,"starGrowth":7,"createdAt":"2026-06-09T06:28:20Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.8,"domainTags":["AI Agent","循环工程","自动化","编程助手","工作流"],"domainKey":"AI Agent","pushedAt":"2026-09-23T08:01:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/cobusgreyling/loop-engineering","ts":"2026-09-23T08:01:40Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.027433250620347394,"copyOk":false},{"repo":"alteryx/featuretools","owner":"alteryx","name":"featuretools","desc":"","summaryCn":"自动合成机器学习特征的 Python 库，极大简化数据预处理流程。","reasonCn":"Featuretools 是一款用于自动化特征工程的 Python 库，它通过深度特征合成算法，能够自动挖掘原始数据中的潜在关系并生成新特征。这解决了机器学习中特征工程耗时且依赖人工经验的痛点，支持 NLP 等扩展原语，让数据科学家能更专注于模型调优而非繁琐的数据处理工作。","stars":7684,"starGrowth":1,"createdAt":"2017-09-08T22:15:17Z","silentRounds":2,"language":"Python","topics":["data-science"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["特征工程","机器学习","数据科学"],"domainKey":"特征工程","pushedAt":"2026-09-18T16:08:19.621Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/alteryx/featuretools","ts":"2026-09-18T16:08:19.621Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rorkai/App-Store-Connect-CLI","owner":"rorkai","name":"App-Store-Connect-CLI","desc":"Fast, scriptable CLI for the App Store Connect API. Automate TestFlight, builds, submissions, signing, analytics, screenshots, subscriptions, and more","summaryCn":"iOS打包提交一条龙，命令行自动化搞定App Store全套流程","reasonCn":"从技术角度，它封装了Apple官方的App Store Connect API，提供了一个干净利落的命令行接口。所有操作都通过JSON配置文件来驱动，完全没有交互式提示，这意味着你可以把它无缝集成到CI/CD流水线里。支持的功能覆盖了iOS发版的全流程：构建管理、TestFlight分发、版本号管理、签名证","stars":7332,"starGrowth":5,"createdAt":"2026-01-20T11:17:23Z","silentRounds":0,"language":"Go","topics":["devops","developer-tools"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["iOS开发","App Store","CI/CD","自动化"],"domainKey":"iOS开发","pushedAt":"2026-09-23T15:30:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/rorkai/App-Store-Connect-CLI","ts":"2026-09-23T15:30:34Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.020454991816693944,"copyOk":true},{"repo":"gruntwork-io/terragrunt","owner":"gruntwork-io","name":"terragrunt","desc":"Terragrunt is a flexible orchestration tool that allows Infrastructure as Code written in OpenTofu/Terraform to scale.","summaryCn":"给Terraform套上自动化马甲，让几十台服务器的配置不再抓瞎","reasonCn":"这款Go语言开发的编排工具，专为解决大规模基础设施管理痛点而生。它通过自动提取公共配置、管理远程状态和依赖关系，让你只需编写一份代码，就能在多个环境批量部署。虽无AI功能，但能大幅减少重复劳动，是DevOps工程师的提效利器。","stars":9847,"starGrowth":2,"createdAt":"2016-05-23T22:17:48Z","silentRounds":0,"language":"Go","topics":["devops","developer-tools"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["DevOps","基础设施即代码","Terraform","云原生"],"domainKey":"DevOps","pushedAt":"2026-09-23T20:54:20Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/gruntwork-io/terragrunt","ts":"2026-09-23T20:54:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Writeopia/Writeopia","owner":"Writeopia","name":"Writeopia","desc":"Write like nobody’s watching. Because they aren't.","summaryCn":"一款笔记应用程序，支持多种笔记形式和云同步功能","reasonCn":"这个项目提供了一个简单易用的笔记应用程序，支持Markdown、HTML等多种格式的笔记输入和编辑，同时支持云同步功能，方便用户在不同设备间访问和共享他们的笔记。","stars":279,"starGrowth":0,"createdAt":"2023-04-14T15:06:45Z","silentRounds":1,"language":"Kotlin","topics":["note-taking"],"aiDims":["笔记","创意工具"],"zone":"工具","funScore":0,"domainTags":["笔记应用","Markdown编辑","云同步","写作"],"domainKey":"笔记应用","pushedAt":"2026-09-22T20:47:30Z","tags":[{"name":"笔记","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":1,"source":"search","url":"https://github.com/Writeopia/Writeopia","ts":"2026-09-22T20:47:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Hmbown/Codewhale","owner":"Hmbown","name":"Codewhale","desc":"Open-source coding agent for your terminal, built in Rust and on a journey of continuous community improvement. Issues and PRs welcome.","summaryCn":"在终端里写代码的AI特工，让你的命令行变身超级大脑","reasonCn":"这是一个基于Rust编写的开源AI编程助手，专为命令行环境打造。它不仅能理解复杂的代码指令，还能直接调用终端命令、执行Git操作和运行测试，将自然语言需求转化为具体的代码实现。项目由社区持续驱动，目前拥有超过4万颗星，标志着它在开发者工具领域的极高人气。","stars":41031,"starGrowth":6,"createdAt":"2026-01-19T18:21:01Z","silentRounds":0,"language":"Rust","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","开发工具"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","终端智能体","Rust开发工具","代码自动化","工具调用Agent"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T21:50:33Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Hmbown/Codewhale","ts":"2026-09-23T21:50:33Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.022126431234919936,"copyOk":false},{"repo":"space-wizards/space-station-14","owner":"space-wizards","name":"space-station-14","desc":"A multiplayer game about paranoia and chaos on a space station. Remake of the cult-classic Space Station 13.","summaryCn":"太空版狼人杀沙盒游戏，在空间站里制造混乱，多人联机超好玩。","reasonCn":"Space Station 14是经典游戏Space Station 13的重制版，一款多人在线的沙盒模拟游戏。玩家在空间站中扮演不同角色，有的负责维修、医疗，有的则忙着策划阴谋。游戏强调玩家互动与混乱美学，没有固定主线，剧情由玩家共同演绎，充满了黑话、搞笑事件和突发危机，是开源游戏开发社区的杰作。","stars":3829,"starGrowth":1,"createdAt":"2017-05-22T01:57:51Z","silentRounds":2,"language":"C#","topics":["game"],"aiDims":["非AI-好玩","游戏","创意工具"],"zone":"创意","funScore":1,"domainTags":["太空沙盒","多人联机","角色扮演"],"domainKey":"太空沙盒","pushedAt":"2026-09-23T21:35:57Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/space-wizards/space-station-14","ts":"2026-09-23T21:35:57Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"naver/egjs-flicking","owner":"naver","name":"egjs-flicking","desc":"🎠 ♻️ Everyday 30 million people experience. It's reliable, flexible and extendable carousel.","summaryCn":"Naver开源的稳定轮播组件，支持多种框架","reasonCn":"来自Naver的开源轮播组件，每天被3000万人使用，特点是稳定可靠、灵活可扩展。支持移动端和桌面端，兼容React、Vue、Angular等主流框架，采用虚拟化技术优化性能。内置丰富的动画效果和手势导航，API设计清晰，易于集成到现有项目中。","stars":2923,"starGrowth":0,"createdAt":"2017-04-28T02:32:33Z","silentRounds":22,"language":"TypeScript","topics":["mobile","desktop"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["轮播组件","Carousel","虚拟化渲染","多框架适配","手势导航"],"domainKey":"轮播组件","pushedAt":"2026-09-16T07:44:27Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/naver/egjs-flicking","ts":"2026-09-16T07:44:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"openebs/openebs","owner":"openebs","name":"openebs","desc":"A popular & widely deployed Open Source Container Native Storage platform for Stateful Persistent Applications on Kubernetes.","summaryCn":"Kubernetes上给有状态应用持久化存储的神器，让数据不丢不烂","reasonCn":"OpenEBS是专为Kubernetes设计的容器原生存储平台，支持动态卷供给、数据复制和高可用。提供本地磁盘、分布式块存储、Replicated存储等多种存储类型，兼容主流存储后端，让有状态应用无需管理底层存储即可获得持久化支持。","stars":9816,"starGrowth":1,"createdAt":"2016-08-01T09:33:32Z","silentRounds":3,"language":"","topics":["devops","kubernetes"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes存储","容器存储","DevOps","持久化卷"],"domainKey":"Kubernetes存储","pushedAt":"2026-09-23T21:15:14Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7}],"aiScore":0.3,"source":"search","url":"https://github.com/openebs/openebs","ts":"2026-09-23T21:15:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nocobase/nocobase","owner":"nocobase","name":"nocobase","desc":"NocoBase is an open-source AI + no-code platform for building business systems fast. Instead of generating everything from scratch, AI works on top of production-proven infrastructure and a WYSIWYG no-code interface, so you get both speed and reliability.","summaryCn":"用AI帮你搭业务系统，零代码就能快速上手的开源神器","reasonCn":"NocoBase是一个开源的AI+无代码平台，让开发者无需从零写代码，就能通过可视化界面快速搭建业务系统。它提供WYSIWYG拖拽式配置能力，内置自动化工作流引擎、多数据源支持和丰富的插件生态。相比传统开发，它能将业务系统的搭建时间从数周缩短到数小时，且数据可完全私有化部署，特别适合中小企业和独立开发者快速验证想法。","stars":24319,"starGrowth":7,"createdAt":"2020-10-24T07:27:48Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI应用","AI基础设施","AI Agent"],"zone":"工具","funScore":0,"domainTags":["无代码平台","业务系统","工作流引擎","可视化配置","自托管部署"],"domainKey":"无代码平台","pushedAt":"2026-09-23T17:40:35Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/nocobase/nocobase","ts":"2026-09-23T17:40:35Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02186352234878079,"copyOk":false},{"repo":"ggml-org/ggml","owner":"ggml-org","name":"ggml","desc":"Tensor library for machine learning","summaryCn":"让普通笔记本电脑也能流畅运行大模型的纯C++高性能张量计算库","reasonCn":"在技术层面，它完全用纯C++手写，没有任何沉重的第三方框架依赖，代码精炼到极致。它支持各种精妙的量化算法（比如把32位浮点数压缩成4位或5位），在大幅减少内存占用的同时，几乎不损失大模型的聪明才智。同时它对各大主流硬件平台的CPU指令集做了极致优化，哪怕没有独立显卡，单","stars":15393,"starGrowth":8,"createdAt":"2022-09-18T17:07:19Z","silentRounds":0,"language":"C++","topics":["ml"],"aiDims":["推理引擎","AI基础设施","机器学习"],"zone":"AI","funScore":0,"domainTags":["机器学习","C++库","张量计算","量化算法","底层优化"],"domainKey":"机器学习","pushedAt":"2026-09-23T17:56:18Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"机器学习","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/ggml-org/ggml","ts":"2026-09-23T17:56:18Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02555915026310661,"copyOk":true},{"repo":"adrienverge/yamllint","owner":"adrienverge","name":"yamllint","desc":"","summaryCn":"它给 YAML 文件做体检，专挑语法之外的小毛病和格式问题。","reasonCn":"它面向 YAML 配置和清单文件，在提交前帮开发者发现容易漏掉的格式问题。除了判断能不能正常解析，它还会指出重复键、过长行、多余尾随空格以及缩进不一致，适合把团队规范写进自动检查流程。对运维、CI 或前端项目来说，它能减少因一个小逗号或空格导致发布失败的风险。","stars":3465,"starGrowth":1,"createdAt":"2016-01-11T08:26:36Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["YAML","静态检查","Python","代码质量","配置格式"],"domainKey":"YAML","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.428,"source":"search","url":"https://github.com/adrienverge/yamllint","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bitwarden/clients","owner":"bitwarden","name":"clients","desc":"Bitwarden client apps (web, browser extension, desktop, and cli).","summaryCn":"开源免费密码管理器，全平台覆盖一键自动填充守护账号安全","reasonCn":"Bitwarden是开源免费的密码管理器，提供Web、浏览器扩展、桌面和命令行全平台客户端。核心是让用户只需记住一个主密码即可安全存储所有账号密码，支持自动生成高强度随机密码和一键自动填充。采用端到端加密架构，服务器无法读取用户密码数据。","stars":13819,"starGrowth":0,"createdAt":"2016-03-09T23:14:01Z","silentRounds":2,"language":"TypeScript","topics":["desktop"],"aiDims":["安全工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["密码管理","端到端加密","跨平台","客户端"],"domainKey":"密码管理","pushedAt":"2026-09-23T21:42:51Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/bitwarden/clients","ts":"2026-09-23T21:42:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"electron/fiddle","owner":"electron","name":"fiddle","desc":":electron: 🚀 The easiest way to get started with Electron","summaryCn":"写Electron桌面应用最快的起飞机场，三秒跑起来给你看","reasonCn":"Electron官方出品的快速实验工具，专为测试和分享Electron代码片段设计。内置代码编辑器、应用预览窗口和输出日志三栏布局，支持实时修改代码并即时看到效果。可一键分享代码到Gist，也支持导出为完整项目。它消除了从零搭建Electron脚手架的时间成本，是Electron开发者日常调试和演示的利器。","stars":7771,"starGrowth":0,"createdAt":"2018-06-07T10:52:57Z","silentRounds":3,"language":"TypeScript","topics":["desktop"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Electron开发","桌面应用调试","代码示例","开发者工具","跨平台框架"],"domainKey":"Electron开发","pushedAt":"2026-09-23T21:31:18Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/electron/fiddle","ts":"2026-09-23T21:31:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"chroma-core/chroma","owner":"chroma-core","name":"chroma","desc":"Search infrastructure for AI","summaryCn":"给AI装上记忆系统，海量知识一秒就能精准检索到","reasonCn":"Chroma是开源嵌入式向量数据库，专为AI应用设计，可将文本自动转为向量存储，支持语义搜索和元数据过滤，与LangChain、LlamaIndex无缝集成，几行代码即可实现向量化、存储和检索全流程，是搭建RAG应用的理想基础设施。","stars":29356,"starGrowth":3,"createdAt":"2022-10-05T17:58:44Z","silentRounds":0,"language":"Rust","topics":["database","agents"],"aiDims":["向量数据库","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["向量数据库","语义搜索","RAG"],"domainKey":"向量数据库","pushedAt":"2026-09-23T21:45:26Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/chroma-core/chroma","ts":"2026-09-23T21:45:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/peft","owner":"huggingface","name":"peft","desc":"🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.","summaryCn":"想象一下，你有一本厚如字典的百科全书（大语言模型），现在想教它学会你的","reasonCn":"技术上，PEFT核心在于参数高效性。它实现了LoRA、Prefix Tuning等尖端方法，将训练参数量从数千亿压缩到仅占原模型的0.1%左右。这不仅大幅降低了显存消耗，还减少了灾难性遗忘问题，让模型在学会新知识的同时，依然保持原有的通用能力。","stars":21715,"starGrowth":5,"createdAt":"2022-11-25T03:51:09Z","silentRounds":0,"language":"Python","topics":["transformers"],"aiDims":["微调","模型与训练"],"zone":"AI","funScore":0,"domainTags":["参数高效微调","LoRA","大语言模型","模型训练"],"domainKey":"大语言模型","pushedAt":"2026-09-23T11:19:03Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/huggingface/peft","ts":"2026-09-23T11:19:03Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.015690766751093713,"copyOk":false},{"repo":"howie6879/weekly","owner":"howie6879","name":"weekly","desc":"老胡的信息技术周刊❤️记录我本周看到的有价值的信息，针对优秀项目、软件、教程资料、网站等。","summaryCn":"每周精选信息技术干货，开发者的宝藏资讯订阅源","reasonCn":"老胡整理的信息技术周刊，每周精选有价值的开源项目、软件工具、教程资料和优质网站，覆盖后端开发、AI、运维等多个方向。通过Markdown格式发布在GitHub上，内容经过筛选和整理，比刷社交媒体更高效地获取技术前沿信息，适合需要持续学习的开发者。","stars":1036,"starGrowth":0,"createdAt":"2021-09-17T15:15:26Z","silentRounds":9,"language":"Python","topics":["backend"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["技术周刊","开源项目","资讯聚合"],"domainKey":"技术周刊","pushedAt":"2026-09-23T21:25:35Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/howie6879/weekly","ts":"2026-09-23T21:25:35Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"infracost/infracost","owner":"infracost","name":"infracost","desc":"Cloud cost intelligence for engineers, AI coding agents, and CI/CD 💰📉 Shift FinOps Left!","summaryCn":"算云服务账单的神器，再也不用被AWS账单吓哭","reasonCn":"这个工具能帮你提前估算在AWS、Azure、GCP等主流云厂商上部署基础设施要花多少钱。支持Terraform、CloudFormation等主流IaC工具，集成到CI/CD流水线中，在代码合并前就能发现成本异常。还能生成详细的成本差异报告，让财务透明化。特别适合DevOps团队做FinOps实践，避免账单超预期。","stars":12530,"starGrowth":2,"createdAt":"2020-06-24T09:11:56Z","silentRounds":0,"language":"Go","topics":["devops"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Terraform","云成本","CI/CD","基础设施"],"domainKey":"Terraform","pushedAt":"2026-09-21T18:37:04Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/infracost/infracost","ts":"2026-09-21T18:37:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"clap-rs/clap","owner":"clap-rs","name":"clap","desc":"A full featured, fast Command Line Argument Parser for Rust","summaryCn":"Rust生态最流行的命令行参数解析库，写CLI工具的必备","reasonCn":"Clap的核心理念是类型安全和表达力并重。它支持从简单的positional参数到复杂的子命令嵌套、可选参数、标志位、环境变量注入等各种用法。通过宏或builder API定义参数后，Clap自动生成帮助信息、版本信息和错误提示，让开发者专注于业务逻辑而非参数解析的繁琐细节。","stars":16723,"starGrowth":2,"createdAt":"2015-02-25T13:36:33Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["命令行","参数解析","Rust库"],"domainKey":"命令行","pushedAt":"2026-09-21T18:47:13Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/clap-rs/clap","ts":"2026-09-21T18:47:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"inkeep/open-knowledge","owner":"inkeep","name":"open-knowledge","desc":"Beautiful, AI-native markdown IDE and LLM wiki","summaryCn":"很多时候我们写的文档都成了死数据，存进电脑就再也没人看。","reasonCn":"其架构亮点在于“AI原生设计”。它不是在编辑器里简单嵌入聊天窗口，而是让LLM直接参与文档的解析、索引与生成。它能轻松处理海量Markdown文件，通过高效的语义检索，让你在写作时随时可以引用过去写下的想法，实现了真正的知识沉淀与再利用。","stars":4289,"starGrowth":13,"createdAt":"2026-06-03T11:12:40Z","silentRounds":0,"language":"TypeScript","topics":["knowledge-management"],"aiDims":["AI应用","AI写作","文档理解"],"zone":"工具","funScore":0,"domainTags":["知识管理","Markdown编辑器","AI写作辅助","语义检索"],"domainKey":"知识管理","pushedAt":"2026-09-23T20:55:47Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/inkeep/open-knowledge","ts":"2026-09-23T20:55:47Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.05771163441361622,"copyOk":false},{"repo":"MervinPraison/PraisonAI","owner":"MervinPraison","name":"PraisonAI","desc":"PraisonAI 🦞 — Hire a 24/7 AI Workforce. Stop writing boilerplate and start shipping autonomous self-improving agents that research, plan, code, and execute tasks. Deployed in 5 lines of code with built-in memory, RAG, and support for 100+ LLMs.","summaryCn":"让AI当你的全职员工，24小时帮你查资料写代码干杂活","reasonCn":"PraisonAI是一个开源多Agent协作框架，用Python构建，支持多个AI代理互相通信和分工协作。内置丰富的工具调用能力，可对接浏览器、代码编辑器、文件系统和各类API。核心亮点是自改进机制，代理能根据执行结果自动优化策略，只需5行代码即可部署完整AI工作流。","stars":9082,"starGrowth":1,"createdAt":"2024-03-19T16:45:25Z","silentRounds":2,"language":"Python","topics":["agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["多Agent协作","AI工作流","自动化"],"domainKey":"多Agent协作","pushedAt":"2026-09-23T17:22:30Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/MervinPraison/PraisonAI","ts":"2026-09-23T17:22:30Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ToFuProject/tofu","owner":"ToFuProject","name":"tofu","desc":"Project for an open-source python library for synthetic diagnostics and tomography for Fusion devices","summaryCn":"核聚变装置诊断神器，给等离子体拍CT，物理学家专属","reasonCn":"面向核聚变装置的开源Python诊断库，专注于合成数据处理层析成像技术。提供从信号采集、噪声过滤到三维重建的完整算法链路，支持多种物理模型与反问题求解。虽然应用场景垂直且小众，但为核物理研究提供了标准化工具，降低了底层代码重复造轮子的成本，是科研人员探索等离子体动力学结构的重要辅助手段，具备极高的专业价值。","stars":88,"starGrowth":0,"createdAt":"2016-09-13T15:56:56Z","silentRounds":14,"language":"Python","topics":["physics"],"aiDims":["学习资源"],"zone":"工具","funScore":0,"domainTags":["核聚变","等离子体诊断","科学计算","层析成像"],"domainKey":"核聚变","pushedAt":"2026-09-23T20:53:24Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"physics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/ToFuProject/tofu","ts":"2026-09-23T20:53:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dotnet/winforms","owner":"dotnet","name":"winforms","desc":"Windows Forms is a .NET UI framework for building Windows desktop applications.","summaryCn":"微软官方老牌Windows桌面开发框架，拖拽控件就能做软件","reasonCn":"这是微软官方的经典Windows窗体应用开发框架，支持拖拽式UI设计。C#开发者用它可快速构建传统桌面客户端，集成Visual Studio设计器，兼容.NET全版本。虽非AI项目，但对开发者极度实用，适合构建企业内网工具、数据录入系统，生态成熟稳定。","stars":4859,"starGrowth":1,"createdAt":"2018-10-19T01:54:02Z","silentRounds":8,"language":"C#","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Windows开发","桌面应用","C#","GUI框架"],"domainKey":"Windows开发","pushedAt":"2026-09-23T18:09:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/dotnet/winforms","ts":"2026-09-23T18:09:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apioo/fusio","owner":"apioo","name":"fusio","desc":"Self-Hosted API Management for Builders","summaryCn":"自托管的API管理平台，轻松搭建和维护自己的API网关。","reasonCn":"Fusio是一个自托管的API管理工具，基于PHP开发，提供完整的API网关能力。支持API路由、认证、限流、监控等功能，拥有可视化管理界面。对于需要自建API管理的中后端团队来说，是一个轻量级且功能全面的替代方案，无需依赖第三方云服务。","stars":2118,"starGrowth":2,"createdAt":"2015-07-29T16:34:52Z","silentRounds":0,"language":"PHP","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["API网关","PHP","自托管","API管理"],"domainKey":"API网关","pushedAt":"2026-09-21T17:57:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/apioo/fusio","ts":"2026-09-21T17:57:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dataease/SQLBot","owner":"dataease","name":"SQLBot","desc":"🔥 基于大模型和 RAG 的智能问数系统，对话式数据分析神器。Text-to-SQL Generation via LLMs using RAG.","summaryCn":"会聊天的数据分析师，问一句就能看懂数据报表。","reasonCn":"这是一个基于大模型和RAG技术的智能问数系统，支持自然语言转SQL查询。用户只需用口语描述问题，系统就能自动生成查询语句并返回分析结果。内置丰富的可视化组件，让数据分析变得像聊天一样简单，无需掌握复杂的SQL语法。","stars":6847,"starGrowth":6,"createdAt":"2025-04-21T05:33:53Z","silentRounds":0,"language":"JavaScript","topics":["rag"],"aiDims":["AI应用","AI搜索","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["数据分析","RAG","Text-to-SQL","自然语言查询"],"domainKey":"RAG","pushedAt":"2026-09-23T04:45:01Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/dataease/SQLBot","ts":"2026-09-23T04:45:01Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.020628888564334744,"copyOk":false},{"repo":"supermemoryai/supermemory","owner":"supermemoryai","name":"supermemory","desc":"Memory and context engine + app that is extremely fast, scalable, and can be run fully locally. The Memory API for the AI era.","summaryCn":"在使用大模型进行对话或生成内容时，常常会遇到上下文短、信息遗忘的问题。","reasonCn":"在使用大模型进行对话或生成内容时，常常会遇到上下文短、信息遗忘的问题。Supermemory 的目标是为 AI 提供持久且高速的记忆层，让模型能够随时查找之前的交互、文档或代码片段，像人类的大脑一样进行信息检索。\n技术核心是基于 Hierarchical Navigable Small World（HNSW）算法实","stars":30823,"starGrowth":114,"createdAt":"2024-02-27T20:10:04Z","silentRounds":0,"language":"TypeScript","topics":["cloudflare-pages","cloudflare-workers","drizzle-orm","tailwindcss","typescript","cloudflare-kv","postgres","remix","vite","agent-memory","ai-memory","memory"],"aiDims":["AI应用","RAG与知识","工具调用"],"zone":"AI","funScore":0,"domainTags":["向量检索","RAG","本地部署"],"domainKey":"RAG","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"cloudflare-pages","source":"github","weight":0.7},{"name":"cloudflare-workers","source":"github","weight":0.7},{"name":"drizzle-orm","source":"github","weight":0.7},{"name":"tailwindcss","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"cloudflare-kv","source":"github","weight":0.7},{"name":"postgres","source":"github","weight":0.7},{"name":"remix","source":"github","weight":0.7},{"name":"vite","source":"github","weight":0.7},{"name":"agent-memory","source":"github","weight":0.7},{"name":"ai-memory","source":"github","weight":0.7},{"name":"memory","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/supermemoryai/supermemory","ts":"2026-09-21T04:10:58.868Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.31109561042078965,"copyOk":false},{"repo":"microsoft/winget-cli","owner":"microsoft","name":"winget-cli","desc":"WinGet is the Windows Package Manager. This project includes a CLI (Command Line Interface), PowerShell modules, and a COM (Component Object Model) API (Application Programming Interface).","summaryCn":"Windows自带的应用商店命令行版，一行命令装软件","reasonCn":"技术上，WinGet基于JSON清单文件管理软件源，支持多种软件分发方式，包括MSIX、EXE、MSI安装包。它内置了版本解析、依赖管理和自动更新检测功能。作为CLI工具，它可以集成到PowerShell脚本、CI/CD流水线中，实现开发环境的自动化部署。","stars":26445,"starGrowth":0,"createdAt":"2019-07-16T22:16:49Z","silentRounds":2,"language":"C++","topics":["command-line"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["包管理","命令行","Windows"],"domainKey":"包管理","pushedAt":"2026-09-23T20:14:03Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/microsoft/winget-cli","ts":"2026-09-23T20:14:03Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"NetHack/NetHack","owner":"NetHack","name":"NetHack","desc":"Official NetHack Git Repository","summaryCn":"地牢探险鼻祖级roguelike，每局都是独一无二的冒险体验","reasonCn":"游戏采用回合制战斗系统，玩家控制角色在多层地牢中探索、战斗、收集装备，目标是获取神器并活着离开。永久死亡是最大特色——角色一旦死亡就彻底结束，这迫使玩家步步为营，珍惜每一个决策。游戏内置数十种职业和种族组合，每种都有不同的起始装备和技能倾向，重玩价值极高。","stars":3918,"starGrowth":0,"createdAt":"2015-12-08T02:33:01Z","silentRounds":3,"language":"C","topics":["game"],"aiDims":["游戏"],"zone":"创意","funScore":1,"domainTags":["Roguelike","地牢探险","回合制游戏"],"domainKey":"Roguelike","pushedAt":"2026-09-23T20:37:38Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/NetHack/NetHack","ts":"2026-09-23T20:37:38Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"apache/ofbiz-framework","owner":"apache","name":"ofbiz-framework","desc":"Apache OFBiz is an open source product for the automation of enterprise processes. It includes framework components and business applications for ERP, CRM, E-Business/E-Commerce, Supply Chain Management and Manufacturing Resource Planning. OFBiz provides a foundation and starting point for reliable, secure and scalable enterprise solutions.","summaryCn":"Apache老牌企业框架，ERP、CRM、电商一站式搞定","reasonCn":"Apache顶级项目OFBiz，专注企业业务流程自动化。提供完整的ERP、CRM和电商解决方案，涵盖供应链、财务、人力资源等核心模块。纯Java技术栈，灵活可扩展的业务模型和组件引擎。十年以上持续迭代，被全球众多企业用于核心系统搭建。","stars":1125,"starGrowth":1,"createdAt":"2017-02-13T08:00:07Z","silentRounds":23,"language":"Java","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["ERP","CRM","电商","企业流程自动化"],"domainKey":"ERP","pushedAt":"2026-09-23T17:53:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/apache/ofbiz-framework","ts":"2026-09-23T17:53:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Lumiwealth/lumibot","owner":"Lumiwealth","name":"lumibot","desc":"Backtestable AI trading agents and Python algorithmic trading strategies for stocks, options, crypto, futures, forex, SEC filings, FRED macro data, and real brokers.","summaryCn":"你是否曾幻想过让代码代替自己盯着大盘，在该买入时自动下单，在该止损时果","reasonCn":"核心亮点在于其架构的灵活性，它将策略逻辑与执行逻辑解耦。你可以利用它强大的回测引擎，用历史数据测试你的AI模型或量化策略是否真的赚钱。此外，它能处理各种类型的数据流，比如通过爬虫获取SEC文件，或者接入宏观经济指标，让你的AI决策不仅仅依赖于价格走势，还能感知市场的大","stars":2081,"starGrowth":2,"createdAt":"2020-09-10T10:00:16Z","silentRounds":0,"language":"Python","topics":["finance"],"aiDims":["AI Agent","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["量化交易","回测引擎","金融数据","Python"],"domainKey":"量化交易","pushedAt":"2026-09-23T21:46:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Lumiwealth/lumibot","ts":"2026-09-23T21:46:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/graphrag","owner":"microsoft","name":"graphrag","desc":"A modular graph-based Retrieval-Augmented Generation (RAG) system","summaryCn":"微软开源的图RAG系统，用知识图谱让AI回答问题更有深度","reasonCn":"核心技术亮点在于它的图谱构建 pipeline。它首先用 LLM 对文档进行实体抽取和关系提取，构建知识图谱，然后利用图算法（如 Leiden 社区发现）对图谱进行聚类，生成不同粒度的摘要。查询时，系统会在多个层级上进行检索和聚合，既保留细节又兼顾全局。","stars":36071,"starGrowth":10,"createdAt":"2024-03-27T17:57:52Z","silentRounds":0,"language":"Python","topics":["rag","llm"],"aiDims":["RAG与知识","知识图谱","AI应用"],"zone":"AI","funScore":0.8,"domainTags":["RAG","知识图谱","检索增强生成","LLM","图算法"],"domainKey":"RAG","pushedAt":"2026-09-23T19:29:14Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/microsoft/graphrag","ts":"2026-09-23T19:29:14Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.030831693049818414,"copyOk":true},{"repo":"angular/angular","owner":"angular","name":"angular","desc":"Deliver web apps with confidence 🚀","summaryCn":"谷歌亲儿子前端框架，构建大型Web应用的工业级标准","reasonCn":"由Google维护的最主流前端框架之一，生态成熟、工具链完善。适合构建复杂的企业级单页应用，提供声明式模板、依赖注入等强大特性。虽然不直接涉及AI，但它是开发AI应用前端界面的绝佳选择，稳定性和性能备受全球开发者信赖。","stars":101022,"starGrowth":0,"createdAt":"2014-09-18T16:12:01Z","silentRounds":2,"language":"TypeScript","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Web框架","TypeScript","企业级应用","前端开发"],"domainKey":"Web框架","pushedAt":"2026-09-23T20:03:16Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/angular/angular","ts":"2026-09-23T20:03:16Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"amir1376/ab-download-manager","owner":"amir1376","name":"ab-download-manager","desc":"A Download Manager that speeds up your downloads","summaryCn":"多线程下载加速器，让文件下载速度直接翻倍，支持断点续传和任务管理","reasonCn":"AB Download Manager是一款基于Kotlin开发的跨平台下载管理器，利用多线程技术将文件分段并行下载，速度可达单线程的数倍。支持HTTP、HTTPS、FTP协议，具备断点续传、批量下载、任务队列管理和下载速度限制等实用功能，界面简洁易用。","stars":18036,"starGrowth":5,"createdAt":"2024-07-07T03:56:25Z","silentRounds":0,"language":"Kotlin","topics":["desktop"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["下载管理器","多线程","Kotlin"],"domainKey":"下载管理器","pushedAt":"2026-09-23T09:58:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/amir1376/ab-download-manager","ts":"2026-09-23T09:58:51Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01583166999334664,"copyOk":true},{"repo":"mikefarah/yq","owner":"mikefarah","name":"yq","desc":"yq is a portable command-line YAML, JSON, XML, CSV, TOML, HCL and properties processor","summaryCn":"命令行处理YAML JSON的神器，比jq功能更强大","reasonCn":"yq是个强大的命令行数据处理工具，堪称jq的多格式版。能处理YAML、JSON、XML、CSV、TOML、HCL等各种格式，支持读取查询修改等操作。语法与jq兼容，Go语言编写，单个二进制文件即可运行，特别适合CI/CD和配置文件管理场景。","stars":15998,"starGrowth":8,"createdAt":"2015-09-26T22:14:44Z","silentRounds":0,"language":"Go","topics":["yaml-processor","yaml","cli","golang","splat","devops-tools","portable","bash","xml","json","csv","properties","toml","hcl","terraform"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行工具","YAML处理","JSON处理"],"domainKey":"命令行工具","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"yaml-processor","source":"github","weight":0.7},{"name":"yaml","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"splat","source":"github","weight":0.7},{"name":"devops-tools","source":"github","weight":0.7},{"name":"portable","source":"github","weight":0.7},{"name":"bash","source":"github","weight":0.7},{"name":"xml","source":"github","weight":0.7},{"name":"json","source":"github","weight":0.7},{"name":"csv","source":"github","weight":0.7},{"name":"properties","source":"github","weight":0.7},{"name":"toml","source":"github","weight":0.7},{"name":"hcl","source":"github","weight":0.7},{"name":"terraform","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/mikefarah/yq","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02550018752344043,"copyOk":true},{"repo":"jamiepine/voicebox","owner":"jamiepine","name":"voicebox","desc":"The open-source AI voice studio. Clone, dictate, create.","summaryCn":"本地优先的AI语音工作室，克隆声音、语音输入、给Agent配声。","reasonCn":"Voicebox 是一个完全本地运行的开源 AI 语音工作室，旨在替代 ElevenLabs 与 WisprFlow 等云服务。它仅需几秒参考音频即可克隆声音，支持 23 种语言及 7 大 TTS 引擎合成语音。用户可通过全局热键在任意输入框语音输入，并通过 MCP 协议为 AI 智能体赋予专属声音，实现语音交互闭环。","stars":55466,"starGrowth":124,"createdAt":"2026-01-25T12:27:03Z","silentRounds":0,"language":"TypeScript","topics":["ai","voice-clone","qwen3-tts","voice-ai","whisper","qwen3-tts-ui","cuda","mlx"],"aiDims":["创意工具","非AI-好玩","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["语音合成","声音克隆","TTS","语音转换","AI语音"],"domainKey":"语音合成","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"voice-clone","source":"github","weight":0.7},{"name":"qwen3-tts","source":"github","weight":0.7},{"name":"voice-ai","source":"github","weight":0.7},{"name":"whisper","source":"github","weight":0.7},{"name":"qwen3-tts-ui","source":"github","weight":0.7},{"name":"cuda","source":"github","weight":0.7},{"name":"mlx","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.96,"source":"trending","url":"https://github.com/jamiepine/voicebox","ts":"2026-09-19T04:31:19.347Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3680481736559333,"copyOk":true},{"repo":"godotengine/godot","owner":"godotengine","name":"godot","desc":"Godot Engine – Multi-platform 2D and 3D game engine","summaryCn":"轻量免费且全能的开源游戏引擎，让你用双手轻松创造出心中的奇妙世界","reasonCn":"它的核心亮点在于独特的节点与场景设计哲学，万物皆可组合，逻辑清晰得像搭积木一样。它内置了性能强劲的2D和3D渲染器、物理引擎以及动画编辑器，并且专门为它量身定制了一种叫GDScript的脚本语言，语法极其类似Python，对新手极其友好。","stars":117612,"starGrowth":19,"createdAt":"2014-01-04T16:05:36Z","silentRounds":0,"language":"C++","topics":["game-development","godot"],"aiDims":["游戏","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","2D游戏","3D游戏","GDScript"],"domainKey":"游戏引擎","pushedAt":"2026-09-23T17:49:44Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"godot","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/godotengine/godot","ts":"2026-09-23T17:49:44Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05748464442403836,"copyOk":true},{"repo":"binaryjs/binaryjs","owner":"binaryjs","name":"binaryjs","desc":"","summaryCn":"在这个实时数据交互日益频繁的时代，如何在浏览器与服务器之间高效传输底层","reasonCn":"对于想要上手的开发者而言，可以通过获取其服务器端与客户端的相关代码或通过 Git 直接集成来开始使用。在实际开发中，它免去了手动编写复杂二进制封包解包逻辑的烦恼，使 StreamID 不再局限于数字，同时优化了套接字关闭时的异常处理机制，避免在写入已关闭的底层套接字时抛出异常","stars":2189,"starGrowth":0,"silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["WebSocket","Node.js","流处理","二进制传输"],"domainKey":"WebSocket","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/binaryjs/binaryjs","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"faiface/beep","owner":"faiface","name":"beep","desc":"","summaryCn":"Go语言音频库，简化多格式播放编码流式处理，现已归档迁移至新仓库。","reasonCn":"这是一个让Go应用获得声音能力的轻量级库。它抽象出类似文件读取的音频流接口，让开发者能轻松处理WAV、MP3等多种格式。虽然功能强大且API简洁，但该仓库已停止维护，新开发建议直接使用其继任项目gopxl/Beep，以确保获得持续支持和代码更新。","stars":2220,"starGrowth":0,"createdAt":"2017-07-14T16:29:59Z","silentRounds":1,"language":"Go","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["音频解码","WAV编码","Go语言","音频流式处理"],"domainKey":"音频解码","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/faiface/beep","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AaronLasseigne/active_interaction","owner":"AaronLasseigne","name":"active_interaction","desc":"","summaryCn":"ActiveInteraction 的核心目标是管理 Rails 应用中的特定业务逻辑，它实现了","reasonCn":"该项目的一个显著特点是其对输入验证的重视。文档指出，它通过验证输入是否符合预期来帮助编写更安全的代码。这种设计思想借鉴了 Rails 的命名规范，如果 ActiveModel 处理的是“名词”（数据实体），那么 ActiveInteraction 处理的则是“动词”（业务操作）。","stars":2147,"starGrowth":0,"createdAt":"2013-06-26T22:19:51Z","silentRounds":1,"language":"Ruby","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Rails","业务逻辑","服务对象","输入验证","过滤"],"domainKey":"Rails","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/AaronLasseigne/active_interaction","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ShareX/ShareX","owner":"ShareX","name":"ShareX","desc":"ShareX is a free and open-source application that enables users to capture or record any area of their screen with a single keystroke. It also supports uploading images, text, and various file types to a wide range of destinations.","summaryCn":"对于经常需要分享截图、录屏或做教程的程序员、设计师、技术博主来说，Shar","reasonCn":"它的核心优势在于高度的可定制化。你可以通过配置文件添加自定义的上传任务、截图编辑功能、快捷键设置。它内置了强大的截图编辑器，支持添加马赛克、箭头、文字、模糊等标注，做完直接上传。它的任务队列系统也非常强大，支持批量截图和批量处理，比如你可以设置一个任务，每隔 1","stars":39696,"starGrowth":15,"createdAt":"2013-10-08T23:32:10Z","silentRounds":0,"language":"C#","topics":["productivity"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["截图工具","录屏","效率软件"],"domainKey":"截图工具","pushedAt":"2026-09-23T07:23:18Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ShareX/ShareX","ts":"2026-09-23T07:23:18Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04613361547762999,"copyOk":false},{"repo":"MemPalace/mempalace","owner":"MemPalace","name":"mempalace","desc":"The best-benchmarked open-source AI memory system. And it's free.","summaryCn":"让AI记住你每次对话的神仙系统，越用越聪明的那种","reasonCn":"开源AI记忆系统，让大模型具备持久化记忆能力，支持多轮对话记忆存储和检索。基于向量数据库实现语义搜索，AI能记住用户偏好和历史对话内容，大幅提升交互体验。部署简单，文档完善，是目前AI记忆方向最热门开源项目之一，非常适合做个性化AI应用。","stars":59226,"starGrowth":18,"createdAt":"2026-04-05T01:12:07Z","silentRounds":0,"language":"Python","topics":["llm","ai"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI记忆","向量数据库","对话历史"],"domainKey":"向量数据库","pushedAt":"2026-09-23T06:48:41Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/MemPalace/mempalace","ts":"2026-09-23T06:48:41Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06589411407152264,"copyOk":false},{"repo":"CarpenterLee/JCFInternals","owner":"CarpenterLee","name":"JCFInternals","desc":"","summaryCn":"深入解析Java集合框架内部结构，通过源码与图解详解常用容器实现原理。","reasonCn":"这是一个专注于Java集合框架（JCF）内部机制的学习资料库。许多资料集中在C++ STL，但Java集合框架有其独特的设计哲学。该项目通过结合生动的图解和核心源代码，深入分析了List、Set、Map、Stack、Queue等典型容器的数据结构与算法实现。它不浮于表面，而是直接切入ArrayList、HashMap等类的底层逻辑，帮助读者从代码层面彻底理解Java集合的运行机制。适合想要突破Java基础瓶颈、深入理解集合框架设计思想的中级及以上开发者。","stars":2222,"starGrowth":0,"silentRounds":0,"language":"","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["Java集合","数据结构","源码分析"],"domainKey":"Java集合","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1}],"aiScore":0.32,"source":"search","url":"https://github.com/CarpenterLee/JCFInternals","ts":"2026-09-22T22:51:50.275Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"SuperMonster003/Ant-Forest","owner":"SuperMonster003","name":"Ant-Forest","desc":"","summaryCn":"基于Auto.js的安卓端蚂蚁森林能量自动收取脚本，解放双手。","reasonCn":"项目的部署过程相对硬核，需要具备基本的安卓文件管理能力。用户首先需要在手机上安装 AutoJs6，然后通过项目提供的部署工具或直接解压数据包，将脚本放置到 AutoJs6 的默认工作目录中，例如中文系统的 ./脚本/ 或英文系统的 ./Scripts/ 路径。","stars":2851,"starGrowth":0,"createdAt":"2019-03-19T01:36:10Z","silentRounds":3,"language":"JavaScript","topics":[],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["Auto.js","自动收取","安卓脚本","蚂蚁森林"],"domainKey":"Auto.js","pushedAt":"2026-09-20T15:45:52.959Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/SuperMonster003/Ant-Forest","ts":"2026-09-20T15:45:52.959Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"89luca89/distrobox","owner":"89luca89","name":"distrobox","desc":"Use any linux distribution inside your terminal. Enable both backward and forward compatibility with software and freedom to use whatever distribution you’re more comfortable with. Mirror available at: https://gitlab.com/89luca89/distrobox","summaryCn":"不用装双系统，在终端里随时切换Linux发行版，开发环境零折腾","reasonCn":"Distrobox利用Linux容器技术，让你在不重装系统、不碰双系统的情况下，在任意发行版中工作。一条命令就能创建容器，安装各种开发环境，且容器与宿主机无缝共享文件和命令。项目Star数超过1.2万，社区活跃，是开发者的环境管理神器。","stars":12997,"starGrowth":3,"createdAt":"2021-11-18T15:41:03Z","silentRounds":0,"language":"Go","topics":["command-line","docker"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Linux容器","开发环境管理","发行版切换"],"domainKey":"Linux容器","pushedAt":"2026-09-17T15:05:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/89luca89/distrobox","ts":"2026-09-17T15:05:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Annyfee/agent-craft","owner":"Annyfee","name":"agent-craft","desc":"AI Agent 教学仓库 | 系统化 LangChain、RAG、LangGraph、MCP 全栈实战代码 | 万字博客详解 | 开源可运行示例 | 从零构建智能体","summaryCn":"手把手教你从零开发智能体的全栈实战仓库，把复杂技术拆成可运行的代码。","reasonCn":"这是一个聚焦AI智能体开发的系统化实战教程，深度整合了LangChain与LangGraph等前沿框架。仓库通过通俗易懂的万字博客和丰富的可运行代码示例，带你攻克RAG检索增强和MCP工具调用等核心技术难题，非常适合想要快速掌握智能体架构的开发者。","stars":490,"starGrowth":0,"createdAt":"2025-10-23T09:29:06Z","silentRounds":4,"language":"Python","topics":["tutorial"],"aiDims":["AI Agent","教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["AI教程","LangChain","代码示例"],"domainKey":"AI教程","pushedAt":"2026-09-23T08:50:29Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Annyfee/agent-craft","ts":"2026-09-23T08:50:29Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"samber/cc-skills-golang","owner":"samber","name":"cc-skills-golang","desc":"🧑‍🎨 A collection of Golang agentic skills that works","summaryCn":"Golang版AI技能库，让AI像人类一样干活","reasonCn":"这是一个集合了Golang编写AI智能体技能的开源项目，旨在帮助开发者快速构建具备特定专业能力的AI助手。项目通过模块化的技能定义，实现了工具调用、文件操作、HTTP请求等常见任务的标准化封装，支持多种大语言模型接入。","stars":3316,"starGrowth":9,"silentRounds":0,"language":"Go","topics":["agent","ai","antigravity","claude","claude-code","code","codex","coding","copilot","cursor","gemini","opencode","plugin","skills","vibe-coding","openclaw","gemini-cli-extension","agent-skills","skillsmp"],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","技能包","Golang","工具调用","模块化"],"domainKey":"AI Agent","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"coding","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"plugin","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"vibe-coding","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"gemini-cli-extension","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"skillsmp","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/samber/cc-skills-golang","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03514234016887817,"copyOk":false},{"repo":"yairm210/Unciv","owner":"yairm210","name":"Unciv","desc":"Open-source Android/Desktop remake of Civ V","summaryCn":"免费开源的文明5重制版，手机电脑都能玩，策略游戏迷必入","reasonCn":"这是一款完全免费的开源游戏，复刻了文明5的核心玩法。你可以在Android手机或电脑上体验完整的回合制战略游戏，从远古时代一路发展到未来。支持单机、网络多人对战，还能导入原版MOD。游戏采用像素艺术风格，AI难度可调，适合各层次玩家。无需购买原版即可体验经典文明策略乐趣，社区持续更新维护，是移动设备上最完善的4X策略游戏之一。","stars":11302,"starGrowth":2,"createdAt":"2017-11-21T21:58:12Z","silentRounds":0,"language":"Kotlin","topics":["game"],"aiDims":["游戏"],"zone":"创意","funScore":1,"domainTags":["策略游戏","4X","开源游戏","移动游戏"],"domainKey":"策略游戏","pushedAt":"2026-09-23T14:24:37Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/yairm210/Unciv","ts":"2026-09-23T14:24:37Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fastlane/fastlane","owner":"fastlane","name":"fastlane","desc":"🚀 The easiest way to automate building and releasing your iOS and Android apps","summaryCn":"一键自动化打包iOS安卓App，上架应用商店不再头疼","reasonCn":"fastlane是一套完整的移动应用发布自动化工具链，覆盖从代码构建、测试、签名到上架应用商店的全流程。支持iOS的xcodebuild和App Store Connect API，以及Android的Gradle和Play Store API。内置超过100个lane（自动化任务），可自定义扩展。配合CI/CD系统如GitHub Actions和Jenkins使用，能大幅减少人工操作，避免上架时因配置错误导致的反复折腾。","stars":42147,"starGrowth":3,"createdAt":"2014-12-02T17:00:38Z","silentRounds":0,"language":"Ruby","topics":["automation","mobile"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["移动应用发布","自动化构建","持续集成"],"domainKey":"移动应用发布","pushedAt":"2026-09-23T11:15:34Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"mobile","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/fastlane/fastlane","ts":"2026-09-23T11:15:34Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"cncf/tag-security","owner":"cncf","name":"tag-security","desc":"","summaryCn":"在由微服务、Kubernetes 和快速迭代代码构成的庞大云原生生态中，传统的单体","reasonCn":"传统的防范方式往往将隔离视为第一要务，但这会导致开发与部署效率的急剧下降，而该安全团队强调必须彻底转变研发思路，从日常开发者本身的舒适程度和操作习惯出发，将安全防护无缝编织进开发流程的最前端。这要求将安全要求变得极简，让遵守安全规范的成本大幅低于为了规避防火墙","stars":2262,"starGrowth":0,"silentRounds":0,"language":"HTML","topics":[],"aiDims":["安全工具","学习资源"],"zone":"工具","funScore":0.2,"domainTags":["云原生安全","开源软件治理","云防护","技术规范"],"domainKey":"云原生安全","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/cncf/tag-security","ts":"2026-09-22T21:05:02.747Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yeahhe365/AMC-WebUI","owner":"yeahhe365","name":"AMC-WebUI","desc":"面向 Gemini 的 Local-First AI 工作流 WebUI，集成多模态聊天、Canvas、文件处理、实时搜索、代码执行与高级推理。","summaryCn":"面向Gemini的AI工作流WebUI，集成多模态聊天和实时交互。","reasonCn":"该软件提供了一个友好的用户界面，使非技术用户也能方便地进行AI对话和任务执行。","stars":913,"starGrowth":1,"createdAt":"2025-06-24T10:44:40Z","silentRounds":1,"language":"TypeScript","topics":["chatbot"],"aiDims":["Chatbot"],"zone":"AI","funScore":0,"domainTags":["AI聊天","Gemini","多模态交互","自托管","工作流"],"domainKey":"AI聊天","pushedAt":"2026-09-23T15:20:30Z","tags":[{"name":"Chatbot","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/yeahhe365/AMC-WebUI","ts":"2026-09-23T15:20:30Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AdguardTeam/AdGuardHome","owner":"AdguardTeam","name":"AdGuardHome","desc":"Network-wide ads & trackers blocking DNS server","summaryCn":"给全屋网络装个强力过滤器，广告追踪全拦截。","reasonCn":"技术上，它基于 DNS 53 端口工作，通过劫持域名解析来实现拦截。它内置了一个庞大的广告域名列表，并允许用户通过 Web 界面实时添加自定义规则。它不像传统的 Hosts 文件那样需要手动维护，也不需要你修改每台设备的网络设置，只需要在路由器上设置一下 DNS 地址即可。","stars":37030,"starGrowth":10,"createdAt":"2016-07-06T10:31:47Z","silentRounds":0,"language":"TypeScript","topics":["privacy"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["网络广告拦截","DNS服务","自托管软件","隐私保护","路由器插件"],"domainKey":"网络广告拦截","pushedAt":"2026-09-23T15:40:19Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/AdguardTeam/AdGuardHome","ts":"2026-09-23T15:40:19Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030810153929246555,"copyOk":false},{"repo":"crossplane/crossplane","owner":"crossplane","name":"crossplane","desc":"The Cloud Native Control Plane","summaryCn":"用Kubernetes管云资源，一个平台搞定AWS、阿里云等多云管理","reasonCn":"Crossplane是基于Kubernetes的云原生控制平面，让你用熟悉的K8s资源声明式地管理AWS、GCP、Azure、阿里云等多云资源。它通过Provider插件机制扩展，支持数据库、存储、网络等各类云服务的自动配置，适合多云架构和DevOps团队。","stars":12098,"starGrowth":2,"createdAt":"2018-09-08T00:10:35Z","silentRounds":0,"language":"Go","topics":["kubernetes"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","多云管理","云原生","IaC"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T11:39:48Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/crossplane/crossplane","ts":"2026-09-23T11:39:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"5-stones/react-native-readium","owner":"5-stones","name":"react-native-readium","desc":"📚 📖   React Native ebook reader for iOS, Android, & Web via Readium","summaryCn":"跨平台电子书阅读器，Android和iOS都能用","reasonCn":"从技术角度看，这个项目深度集成了Readium SDK，这是一个在数字出版界非常成熟的开放标准。这意味着它不仅仅是个简单的阅读器，而是支持复杂的数字出版物特性，比如翻页动画、进度保存、书签管理以及EPUB3里的交互元素。","stars":173,"starGrowth":0,"createdAt":"2022-03-11T16:53:04Z","silentRounds":11,"language":"HTML","topics":["ebook"],"aiDims":["非AI-实用","工具调用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["电子书","Readium","React Native","EPUB","阅读器"],"domainKey":"电子书","pushedAt":"2026-09-21T17:54:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/5-stones/react-native-readium","ts":"2026-09-21T17:54:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vercel/next.js","owner":"vercel","name":"next.js","desc":"The React Framework","summaryCn":"扩展 React 功能，集 Rust 工具链的全栈开发框架。","reasonCn":"Next.js 是一款强大的全栈 Web 应用开发框架，它通过扩展 React 的最新特性，并结合基于 Rust 的 JavaScript 工具链，极大地提升了开发效率和构建速度。开发者可以利用它快速构建功能完备的网站，无论是前端交互还是后端逻辑都能一并处理。","stars":142406,"starGrowth":13,"createdAt":"2016-10-05T23:32:51Z","silentRounds":0,"language":"JavaScript","topics":["react","server-rendering","universal","node","components","browser","nextjs","vercel","static-site-generator","hybrid","ssg","static","blog","compiler"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["React","全栈开发","SSR","前端框架"],"domainKey":"React","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"react","source":"github","weight":0.7},{"name":"server-rendering","source":"github","weight":0.7},{"name":"universal","source":"github","weight":0.7},{"name":"node","source":"github","weight":0.7},{"name":"components","source":"github","weight":0.7},{"name":"browser","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"vercel","source":"github","weight":0.7},{"name":"static-site-generator","source":"github","weight":0.7},{"name":"hybrid","source":"github","weight":0.7},{"name":"ssg","source":"github","weight":0.7},{"name":"static","source":"github","weight":0.7},{"name":"blog","source":"github","weight":0.7},{"name":"compiler","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/vercel/next.js","ts":"2026-09-18T10:23:42.801Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.039273864865244444,"copyOk":true},{"repo":"ArcadeData/arcadedb","owner":"ArcadeData","name":"arcadedb","desc":"ArcadeDB Multi-Model Database, one DBMS that supports SQL, Cypher, Gremlin, HTTP/JSON, MongoDB and Redis. ArcadeDB is a conceptual fork of OrientDB, the first Multi-Model DBMS. ArcadeDB supports Vector Embeddings.","summaryCn":"把向量库、图数据库和SQL混在一起的神奇数据库，AI好用的瑞士军刀","reasonCn":"ArcadeDB 是一款革命性的多模型数据库，它颠覆了传统数据库只能选一种模型的限制。它不仅能用 SQL、Cypher 和 Gremlin 查询，还直接支持 HTTP/JSON 和 MongoDB、Redis 协议。最关键的是，它天生支持向量检索和图计算，完美解决了 AI 应用中“数据关联”和“语义搜索”的痛点。作为 OrientDB 的概念分支，它继承了多模型数据库的架构优势，专为 AI 场景优化。","stars":1159,"starGrowth":0,"createdAt":"2021-08-16T16:01:29Z","silentRounds":1,"language":"Java","topics":["vector-db"],"aiDims":["向量数据库","知识图谱","多模态Agent"],"zone":"工具","funScore":0.6,"domainTags":["多模型数据库","向量检索","图数据库"],"domainKey":"多模型数据库","pushedAt":"2026-09-23T21:43:08Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/ArcadeData/arcadedb","ts":"2026-09-23T21:43:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rullerzhou-afk/clawd-on-desk","owner":"rullerzhou-afk","name":"clawd-on-desk","desc":"A pixel desktop pet that watches Claude Code, Codex, Cursor & other AI coding agents — so you don't have to.","summaryCn":"AI 写代码时陪你犯傻的像素小宠物，太有爱了。","reasonCn":"它的萌点在于与 AI 编程工具的联动。当你使用 Cursor、Claude Code 或 Codex 等工具写代码时，Clawd 会根据你的光标移动和代码输入做出各种有趣的反应。比如你写代码写得快，它可能会跑过来给你点赞；如果你停下来思考，它可能会坐在旁边陪着你。","stars":6270,"starGrowth":5,"createdAt":"2026-03-18T09:04:20Z","silentRounds":0,"language":"JavaScript","topics":["pixel-art"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["像素宠物","AI编程","桌面助手","前端动画"],"domainKey":"像素宠物","pushedAt":"2026-09-23T14:51:58Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/rullerzhou-afk/clawd-on-desk","ts":"2026-09-23T14:51:58Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.020870813397129182,"copyOk":false},{"repo":"pocketbase/pocketbase","owner":"pocketbase","name":"pocketbase","desc":"Open Source realtime backend in 1 file","summaryCn":"一个Go文件搞定后端，自带数据库和认证，极简开发者福音","reasonCn":"这是一个单文件开源的实时后端服务，集成了数据库、认证、存储和管理面板。用Go写成，部署极其简单，只要一个二进制文件就能跑起来。对于独立开发者或需要快速搭建原型的小团队，它能省去大量配置数据库和API的时间，是构建轻量级应用的利器。","stars":61125,"starGrowth":3,"createdAt":"2022-07-05T06:06:21Z","silentRounds":0,"language":"Go","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.7,"domainTags":["实时后端","SQLite","认证系统","单文件部署"],"domainKey":"实时后端","pushedAt":"2026-09-21T09:46:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/pocketbase/pocketbase","ts":"2026-09-21T09:46:23Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aalhour/C-Sharp-Algorithms","owner":"aalhour","name":"C-Sharp-Algorithms","desc":":books: :chart_with_upwards_trend: Plug-and-play class-library project of standard Data Structures and Algorithms in C#","summaryCn":"一本C#程序员面试前偷师算法数据结构的随身小抄","reasonCn":"项目把数组链表栈队列优先队列堆哈希表图树等常用结构，以及排序查找遍历递归分治等算法封装成可直接引用的类库，并附示例与测试，能帮读者对照伪代码理解复杂度边界条件和接口设计，也能直观演示时空复杂度与常见错误，适合入门补基础、老手查细节，很稳。","stars":6163,"starGrowth":2,"createdAt":"2015-05-14T11:56:48Z","silentRounds":0,"language":"C#","topics":["algorithms"],"aiDims":["学习资源","开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["Algorithms","Data-Structures","C#","Reference"],"domainKey":"Algorithms","pushedAt":"2026-09-17T12:44:35Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"algorithms","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/aalhour/C-Sharp-Algorithms","ts":"2026-09-17T12:44:35Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"robotframework/robotframework","owner":"robotframework","name":"robotframework","desc":"Generic automation framework for acceptance testing and RPA","summaryCn":"通用的自动化测试框架，用自然语言写测试脚本。","reasonCn":"Robot Framework 是一个基于 Python 的通用自动化测试框架，以可读性强的表格语法著称。广泛用于软件测试和 RPA（机器人流程自动化）。它允许非程序员通过简单的关键词驱动方式编写测试用例，适合企业级自动化场景，稳定性极高。","stars":11906,"starGrowth":0,"createdAt":"2014-06-27T11:10:38Z","silentRounds":2,"language":"Python","topics":["automation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["自动化测试","验收测试","RPA","关键字驱动测试","Python测试框架"],"domainKey":"自动化测试","pushedAt":"2026-09-21T15:19:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/robotframework/robotframework","ts":"2026-09-21T15:19:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/optimum","owner":"huggingface","name":"optimum","desc":"🚀 Accelerate inference and training of 🤗 Transformers, Diffusers, TIMM and Sentence Transformers with easy to use hardware optimization tools","summaryCn":"当你训练好一个模型准备上线，发现它在服务器上跑得慢如蜗牛，这就是Optimu","reasonCn":"技术上，它通过图优化、算子融合以及针对特定指令集的量化技术来提升效率。比如，它可以将模型转换为ONNX格式，进而利用Intel的OpenVINO进行极致优化，或者在NVIDIA显卡上使用TensorRT加速。它不仅加速推理，在分布式训练的某些场景下也有显著的吞吐量提升。","stars":3495,"starGrowth":0,"createdAt":"2021-07-20T12:36:40Z","silentRounds":3,"language":"Python","topics":["transformers"],"aiDims":["模型部署","推理引擎","模型与训练"],"zone":"AI","funScore":0,"domainTags":["模型加速","推理优化","量化","ML部署"],"domainKey":"模型加速","pushedAt":"2026-09-23T14:16:56Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/huggingface/optimum","ts":"2026-09-23T14:16:56Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"continuedev/continue","owner":"continuedev","name":"continue","desc":"open-source coding agent","summaryCn":"VS Code里的AI编程副驾驶，免费开源比Copilot更自由","reasonCn":"开源AI编码助手插件，深度集成VS Code和JetBrains系列IDE，支持GPT-4、Claude、本地Ollama等多种后端。可自动补全代码、解释复杂逻辑、生成测试用例，还能读取整个代码库上下文进行智能回答，完全免费且数据不出本地。","stars":35995,"starGrowth":15,"createdAt":"2023-05-24T03:39:39Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools","ai"],"aiDims":["代码助手","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","IDE插件","代码补全","代码库上下文"],"domainKey":"AI编程助手","pushedAt":"2026-09-22T08:09:07Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/continuedev/continue","ts":"2026-09-22T08:09:07Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04625017363522711,"copyOk":true},{"repo":"dail8859/NotepadNext","owner":"dail8859","name":"NotepadNext","desc":"A cross-platform, reimplementation of Notepad++","summaryCn":"Notepad++的现代化重制版，跨平台支持更自由","reasonCn":"基于Qt框架重新实现的跨平台文本编辑器，完美继承Notepad++的快捷键和操作习惯。支持插件系统，可以在Windows、macOS、Linux上使用，解决了Notepad++仅支持Windows的痛点，适合习惯Notepad++的用户迁移到macOS或Linux。","stars":14583,"starGrowth":7,"createdAt":"2019-11-27T16:06:10Z","silentRounds":0,"language":"C++","topics":["desktop","editor"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文本编辑器","跨平台","Qt框架"],"domainKey":"文本编辑器","pushedAt":"2026-09-22T23:44:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"editor","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/dail8859/NotepadNext","ts":"2026-09-22T23:44:26Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.022440032915038058,"copyOk":true},{"repo":"Blazemeter/taurus","owner":"Blazemeter","name":"taurus","desc":"","summaryCn":"性能测试往往意味着要学习JMeter的XML配置、Gatling的Scala DSL或者Locust的","reasonCn":"安装非常简便，通过pip即可完成安装。上手也是几行YAML搞定：定义测试执行的次数、并发用户数、请求地址，保存为test.yml，然后运行bzt test.yml就好。命令执行后，控制台会输出汇总统计——吞吐量、响应时间、错误率等关键指标一目了然，测试过程中生成的所有产物文件也会自动整理到","stars":2113,"starGrowth":1,"createdAt":"2015-02-17T22:05:08Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["其他"],"zone":"工具","funScore":0.2,"domainTags":["性能测试","负载测试","自动化测试","JMeter","Gatling"],"domainKey":"性能测试","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"其他","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/Blazemeter/taurus","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"privacy-protection-tools/anti-AD","owner":"privacy-protection-tools","name":"anti-AD","desc":"","summaryCn":"中文区高命中率广告过滤列表，支持多种网络组件，屏蔽广告与保护隐私。","reasonCn":"该项目的核心亮点在于覆盖范围广泛且针对性强，不仅能屏蔽普通网页广告，还能针对电视盒子广告和 App 内置广告进行拦截，同时屏蔽涉及个人隐私信息的日志收集和大数据统计站点，防止隐私数据被偷偷上传。项目支持 AdGuardHome、dnsmasq、Surge、Pi-Hole、Clash、mihomo、smartdns、","stars":10772,"starGrowth":4,"createdAt":"2015-10-22T08:59:28Z","silentRounds":0,"language":"","topics":["dnsmasq","easylist","php","neohosts","adblock","adguardhome","pi-hole","surge"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["广告过滤","隐私保护","Hosts","网络工具","去广告"],"domainKey":"广告过滤","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"dnsmasq","source":"github","weight":0.7},{"name":"easylist","source":"github","weight":0.7},{"name":"php","source":"github","weight":0.7},{"name":"neohosts","source":"github","weight":0.7},{"name":"adblock","source":"github","weight":0.7},{"name":"adguardhome","source":"github","weight":0.7},{"name":"pi-hole","source":"github","weight":0.7},{"name":"surge","source":"github","weight":0.7}],"aiScore":0.5,"source":"tier","url":"https://github.com/privacy-protection-tools/anti-AD","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"codeofaxel/Kiln","owner":"codeofaxel","name":"Kiln","desc":"Describe it or draw it — Kiln makes it real. The open-source MCP server for 3D printing: AI agents (Claude, Codex, Cursor, or any MCP client) design, generate, slice & print on Bambu Lab, Creality, Prusa, Elegoo, Klipper/Moonraker, OctoPrint, Duet/RepRapFirmware & any Marlin printer.","summaryCn":"动动嘴或画个草图就能直接3D打印，把AI智能体变成你专属的造物工厂。","reasonCn":"技术层面上，它巧妙地利用了Anthropic推出的Model Context Protocol（模型上下文协议，简称MCP）。通过这个协议，AI智能体不再局限于动嘴皮子聊天，而是获得了调用本地专业软件和硬件的双手。它整合了开源建模工具、自动切片引擎以及各类3D打印机的控制接口，形成了一条从创意到物","stars":67,"starGrowth":1,"createdAt":"2026-02-10T19:01:32Z","silentRounds":3,"language":"Python","topics":["3d-printing"],"aiDims":["多模态Agent","工具调用","创意工具"],"zone":"AI","funScore":0.8,"domainTags":["3D打印","MCP","AI智能体","CAD自动建模"],"domainKey":"3D打印","pushedAt":"2026-09-23T19:34:53Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/codeofaxel/Kiln","ts":"2026-09-23T19:34:53Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"OpenCut-app/OpenCut","owner":"OpenCut-app","name":"OpenCut","desc":"The open-source CapCut alternative","summaryCn":"完全开源的剪映替代品，视频剪辑也能自由掌控","reasonCn":"OpenCut是一款基于Web技术的开源视频编辑器，使用TypeScript开发，支持多平台运行。它提供了丰富的视频剪辑功能，包括轨道编辑、转场特效、字幕添加等，采用类似CapCut的现代化界面设计，让用户可以轻松上手进行视频创作。","stars":90476,"starGrowth":146,"silentRounds":0,"language":"TypeScript","topics":["editor"],"aiDims":["创意工具","非AI-实用","视频编辑"],"zone":"工具","funScore":0,"domainTags":["视频剪辑","Electron","TypeScript","多媒体处理"],"domainKey":"视频剪辑","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"视频编辑","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/OpenCut-app/OpenCut","ts":"2026-09-23T07:35:20.006Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30484106282329015,"copyOk":false},{"repo":"smkent/monoscad","owner":"smkent","name":"monoscad","desc":"🏗️ 🖨️ Monorepository of OpenSCAD models for 3D printing","summaryCn":"给桌面打印找能改数的小物件，这里像脚本货架逛一逛","reasonCn":"仓库汇集了一批参数化三维打印模型，读者可拿到代码图纸，按需调整尺寸、孔位、壁厚和外形，适合做桌面零件与小型机构。若愿意学脚本化建模，它能替代固定图纸，把设计修改与导出切片的流程串起来。特别适合打印机玩家把想法变成可试错原型，也方便比较不同结构参数对强度和耗材的影响。","stars":231,"starGrowth":0,"createdAt":"2023-10-07T04:14:17Z","silentRounds":8,"language":"OpenSCAD","topics":["3d-printing"],"aiDims":["非AI-实用","学习资源"],"zone":"创意","funScore":0.6,"domainTags":["参数化建模","OpenSCAD","3D打印模型","桌面制造","脚本化设计"],"domainKey":"参数化建模","pushedAt":"2026-09-11T06:48:20Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"OpenSCAD","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/smkent/monoscad","ts":"2026-09-11T06:48:20Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fatedier/frp","owner":"fatedier","name":"frp","desc":"A fast reverse proxy to help you expose a local server behind a NAT or firewall to the internet.","summaryCn":"把内网服务暴露到公网的瑞士军刀，远程访问神器","reasonCn":"frp是一款高性能反向代理工具，支持TCP、UDP、HTTP等多种协议穿透NAT和防火墙。它能在家里电脑上搭建隧道，让你从 anywhere 访问NAS、本地Web服务或开发应用。配置灵活，支持加密和权限控制，是私有服务和远程办公的必备利器。","stars":109565,"starGrowth":20,"createdAt":"2015-12-21T15:24:59Z","silentRounds":0,"language":"Go","topics":["proxy","reverse-proxy","tunnel","nat","go","firewall","frp","expose","http-proxy","p2p"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["反向代理","内网穿透","NAT","网络工具","Go"],"domainKey":"反向代理","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"proxy","source":"github","weight":0.7},{"name":"reverse-proxy","source":"github","weight":0.7},{"name":"tunnel","source":"github","weight":0.7},{"name":"nat","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"firewall","source":"github","weight":0.7},{"name":"frp","source":"github","weight":0.7},{"name":"expose","source":"github","weight":0.7},{"name":"http-proxy","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/fatedier/frp","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06054762013416693,"copyOk":false},{"repo":"leeoniya/uPlot","owner":"leeoniya","name":"uPlot","desc":"📈 A small, fast chart for time series, lines, areas, ohlc & bars","summaryCn":"轻量极速的 JS 图表库，专注时间序列、线图和蜡烛图。","reasonCn":"uPlot 体积仅约 10KB（gzip），采用 Canvas 高效渲染，支持上万点毫秒级刷新。内置交互功能如缩放、悬停提示，API 简洁，只需几行配置即可绘制折线、面积、OHLC、柱状等多种图形，兼容 React、Vue 等主流框架，适合仪表盘、监控面板和金融图表等高频数据展示场景。","stars":10511,"starGrowth":3,"createdAt":"2019-09-27T18:45:21Z","silentRounds":0,"language":"JavaScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["图表库","数据可视化","Canvas渲染"],"domainKey":"图表库","pushedAt":"2026-09-22T07:47:46Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/leeoniya/uPlot","ts":"2026-09-22T07:47:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jo-inc/camofox-browser","owner":"jo-inc","name":"camofox-browser","desc":"Stealth headless browser for AI agents — bypass Cloudflare, bot detection, and anti-scraping. Drop-in Puppeteer/Playwright replacement.","summaryCn":"AI Agent的隐身神器，过Cloudflare如探囊取物","reasonCn":"专为AI Agent打造的无头浏览器，内置指纹随机化和反检测对抗机制，能自动绕过Cloudflare、CAPTCHA等反爬策略。与Puppeteer/Playwright完全兼容，替换配置即可使用，让Agent在无感状态下稳定抓取网页数据，无需手动处理验证码。","stars":11158,"starGrowth":8,"createdAt":"2026-01-26T07:31:13Z","silentRounds":0,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"工具","funScore":0.6,"domainTags":["反爬虫","浏览器自动化","AI Agent","指纹混淆"],"domainKey":"反爬虫","pushedAt":"2026-09-22T23:39:29Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/jo-inc/camofox-browser","ts":"2026-09-22T23:39:29Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.031381107725398816,"copyOk":true},{"repo":"recommenders-team/recommenders","owner":"recommenders-team","name":"recommenders","desc":"Best Practices on Recommendation Systems","summaryCn":"微软出品的推荐系统最佳实践库，从协同过滤到大模型推荐全覆盖","reasonCn":"技术亮点在于覆盖面广，从经典的协同过滤、矩阵分解，到深度的序列推荐、多任务学习，再到最新的大语言模型推荐，都有完整实现。每个算法都有详细注释和可复现的实验代码，不是那种只讲理论的教程项目。它还提供了评估框架，方便你对比不同算法的效果。","stars":21913,"starGrowth":3,"createdAt":"2018-09-19T10:06:07Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["模型与训练","AI基础设施","大语言模型"],"zone":"工具","funScore":0,"domainTags":["推荐系统","协同过滤","序列推荐","算法评估框架"],"domainKey":"推荐系统","pushedAt":"2026-09-23T20:09:46Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/recommenders-team/recommenders","ts":"2026-09-23T20:09:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rgeometry/rgeometry","owner":"rgeometry","name":"rgeometry","desc":"Computational Geometry library written in Rust","summaryCn":"用Rust写的计算几何瑞士军刀，专治各种空间算法排版难题","reasonCn":"技术上，rgeometry选择用Rust编写并非偶然。计算几何涉及大量的坐标运算和数据结构操作，使用C++虽然性能好，但容易出现内存泄漏或非法访问等问题。Rust的所有权机制确保了代码的安全性和高效性，让你在不牺牲性能的前提下，避免了不少底层bug。","stars":183,"starGrowth":0,"createdAt":"2021-04-29T11:32:42Z","silentRounds":3,"language":"Rust","topics":["algorithms"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["计算几何","空间算法","Rust库","凸包计算","碰撞检测"],"domainKey":"计算几何","pushedAt":"2026-09-21T09:31:31Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"algorithms","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/rgeometry/rgeometry","ts":"2026-09-21T09:31:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"varletjs/varlet","owner":"varletjs","name":"varlet","desc":"A Vue3 component library based on Material Design 2 and 3, supporting mobile and desktop.","summaryCn":"Vue3的Material Design组件库，手机端电脑端全搞定","reasonCn":"基于Vue3和TypeScript构建的Material Design 2/3组件库，提供60+高质量组件，支持主题定制、按需加载和国际化。覆盖表单、导航、数据展示等常用场景，响应式布局适配移动端和桌面端，是开发跨平台Vue应用的得力工具。","stars":5333,"starGrowth":0,"createdAt":"2020-10-30T13:32:43Z","silentRounds":23,"language":"Vue","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Vue3","UI组件库","Material Design","移动端"],"domainKey":"Vue3","pushedAt":"2026-09-21T08:30:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/varletjs/varlet","ts":"2026-09-21T08:30:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pranshuparmar/witr","owner":"pranshuparmar","name":"witr","desc":"Why is this running? Trace any process, port, container, or file back to what started it - CLI + TUI.","summaryCn":"系统里的「福尔摩斯」，一键揪出是谁在偷偷占用你的内存和端口","reasonCn":"witr 是个 CLI+TUI 双模工具，能追踪任意进程、端口、容器或文件是由什么发起的。基于 Linux 的 fanotify/audit 机制，实时记录元数据并可视化关联关系，帮 DevOps 快速定位资源被谁占用了，排查系统瓶颈和异常进程非常高效。","stars":22461,"starGrowth":17,"createdAt":"2025-12-20T17:26:59Z","silentRounds":0,"language":"Go","topics":["cli","devops","freebsd","golang","linux","macos","observability","sysadmin","troubleshooting","tui","windows","containers","docker","incident-response","process-management","systemd","go","kubernetes","monitoring","terminal"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.7,"domainTags":["进程溯源","fanotify","TUI","系统诊断"],"domainKey":"进程溯源","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"cli","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"freebsd","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"observability","source":"github","weight":0.7},{"name":"sysadmin","source":"github","weight":0.7},{"name":"troubleshooting","source":"github","weight":0.7},{"name":"tui","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"containers","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"incident-response","source":"github","weight":0.7},{"name":"process-management","source":"github","weight":0.7},{"name":"systemd","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/pranshuparmar/witr","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0639247228529451,"copyOk":false},{"repo":"ChromeDevTools/chrome-devtools-mcp","owner":"ChromeDevTools","name":"chrome-devtools-mcp","desc":"Chrome DevTools for coding agents","summaryCn":"给 AI 赋予浏览器操作能力，让它真干活的神器","reasonCn":"该项目通过 MCP 协议为 AI Agent 提供了 Chrome DevTools 的完整接口，允许编程助手直接控制浏览器、读取 DOM 元素、执行网络请求并解析页面结构。它实现了浏览器与 AI Agent 的双向通信，解决了 Agent 无法自主交互网页的问题，让智能体从“读屏助手”进化为“真干活”的自动化工具。","stars":52485,"starGrowth":66,"createdAt":"2025-09-11T10:39:55Z","silentRounds":0,"language":"TypeScript","topics":["mcp-server","puppeteer","chrome-devtools","browser","chrome","debugging","devtools","mcp"],"aiDims":["AI Agent","工具调用","开发工具"],"zone":"AI","funScore":0.6,"domainTags":["浏览器自动化","Agent开发","MCP协议","AI工具","Web开发"],"domainKey":"浏览器自动化","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"mcp-server","source":"github","weight":0.7},{"name":"puppeteer","source":"github","weight":0.7},{"name":"chrome-devtools","source":"github","weight":0.7},{"name":"browser","source":"github","weight":0.7},{"name":"chrome","source":"github","weight":0.7},{"name":"debugging","source":"github","weight":0.7},{"name":"devtools","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/ChromeDevTools/chrome-devtools-mcp","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2017725064304087,"copyOk":false},{"repo":"google/artemis","owner":"google","name":"artemis","desc":"ARTEMIS turns natural-language instructions into reliable Android automation. It automates end-to-end workflows, captures logs, and integrates seamlessly with AI coding assistants such as Antigravity, Codex, and Claude Code.  It also achieves 99%+ success rate on AndroidWorld Benchmark.  Created by Google's Pixel-Test-Engineering (PTE) Fusion team.","summaryCn":"把自然语言指令变成安卓自动化操作，AI直接帮你点点点","reasonCn":"技术上，它构建了一个从指令理解到动作执行的完整闭环。核心在于将大模型的生成能力与安卓的自动化能力结合起来，不仅能执行简单的点击，还能处理复杂的逻辑判断和异常处理。它内置了强大的日志系统和截图功能，能实时监控运行状态，一旦遇到问题能自动修复或上报，保证了自动化流","stars":9245,"starGrowth":309,"createdAt":"2026-08-13T17:59:37Z","silentRounds":0,"language":"Python","topics":["testing"],"aiDims":["AI Agent","工具调用","开发工具"],"zone":"AI","funScore":1,"domainTags":["Android","自动化","RPA","LLM","指令执行"],"domainKey":"Android","pushedAt":"2026-09-12T03:00:58Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/google/artemis","ts":"2026-09-12T03:00:58Z","category":"ai","momentum":["rising","daily"],"fromOfficial":true,"heatScore":0.6004056246619794,"copyOk":true},{"repo":"typeorm/typeorm","owner":"typeorm","name":"typeorm","desc":"TypeScript & JavaScript ORM for Node.js — supports PostgreSQL, MySQL, MariaDB, SQLite, SQL Server, Oracle, and more.","summaryCn":"用写类的方式操作数据库，告别手写SQL的烦恼","reasonCn":"TypeORM是Node.js生态中最成熟的TypeScript/JavaScript ORM框架之一，36626颗星证明其社区认可度极高。它支持PostgreSQL、MySQL、SQLite等多种数据库，可以用面向对象的方式定义数据模型，自动生成表结构。具备迁移管理、事务支持、查询构建器、实体关系映射等核心能力，适合中大型项目使用。","stars":36655,"starGrowth":0,"createdAt":"2016-02-29T07:41:14Z","silentRounds":3,"language":"TypeScript","topics":["database"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["Node.js","TypeScript","ORM","数据库迁移","关系映射"],"domainKey":"Node.js","pushedAt":"2026-09-21T08:49:54Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/typeorm/typeorm","ts":"2026-09-21T08:49:54Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ikas-mc/ContextMenuForWindows11","owner":"ikas-mc","name":"ContextMenuForWindows11","desc":"","summaryCn":"Windows 11系统为了追求简洁，简化了原本丰富的右键菜单，但这给许多追求操","reasonCn":"核心亮点在于其极高的兼容性与灵活性，项目提供了多种分发方式，涵盖了微软商店版本、GitHub Release包以及Scoop安装，确保用户能够根据自己的网络环境和使用习惯选择最适合的安装方案。项目版本迭代迅速，v3.8版本后甚至允许用户同时运行不同渠道的安装包，这种设计对于同时测试新","stars":2889,"starGrowth":0,"createdAt":"2021-07-14T06:32:41Z","silentRounds":3,"language":"C#","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Windows定制","右键菜单","系统美化"],"domainKey":"Windows定制","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/ikas-mc/ContextMenuForWindows11","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"majd/ipatool","owner":"majd","name":"ipatool","desc":"Command-line tool that allows you to search for iOS, iPadOS, tvOS, visionOS, and macOS apps on the App Store, and download .ipa or macOS .pkg app packages.","summaryCn":"命令行下载App Store应用的神器，免费版也能轻松获得","reasonCn":"技术上，ipatool利用了Apple的私有API实现账户认证和应用下载。你只需要一个Apple ID，工具会自动完成登录、购买（免费应用）和下载流程。它支持搜索应用、查看应用详情、下载指定版本，还能切换不同地区的App Store。","stars":11360,"starGrowth":17,"createdAt":"2021-05-22T08:27:21Z","silentRounds":0,"language":"Go","topics":["security","command-line"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["iOS应用下载","命令行工具","App Store"],"domainKey":"iOS应用下载","pushedAt":"2026-09-19T17:13:29Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/majd/ipatool","ts":"2026-09-19T17:13:29Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.055489436619718314,"copyOk":false},{"repo":"rush86999/atom","owner":"rush86999","name":"atom","desc":"Atom Agent, Open-Source Governed AI Agent Platform for Self-Hosted Automation","summaryCn":"一个开源的自动化工具平台，用于管理、控制和自动化各种自动化任务。","reasonCn":"它提供了一个灵活的环境，让开发者可以创建和管理自己的自动化任务，无需编写复杂的代码。","stars":901,"starGrowth":0,"createdAt":"2023-01-22T15:55:04Z","silentRounds":3,"language":"Python","topics":["finance"],"aiDims":["AI Agent","开源Governed AI Agent Platform"],"zone":"工具","funScore":0,"domainTags":["自动化平台","工作流引擎","自托管","任务调度"],"domainKey":"自动化平台","pushedAt":"2026-09-23T16:01:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开源Governed AI Agent Platform","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/rush86999/atom","ts":"2026-09-23T16:01:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"michaelwilhelmsen/humla","owner":"michaelwilhelmsen","name":"humla","desc":"Open-source AI meeting notes for Mac. Records mic + system audio with no bot, transcribes on-device or via OpenAI / Deepgram / Groq, identifies speakers offline, and writes summaries that fuse your notes with the transcript. Ask your notes and get cited answers. Tauri 2 + Rust + Swift.","summaryCn":"想象一下你正在开会，没时间做笔记，开完会还得手忙脚乱整理录音，这款工具","reasonCn":"技术上，Humla采取了隐私优先的设计架构。它使用Rust编写，确保了在Mac设备上的高性能表现。它最酷的地方在于支持完全离线处理，通过本地的音频处理模型进行说话人识别和转录，这意味着你的会议内容不会上传到任何服务器。","stars":289,"starGrowth":0,"createdAt":"2026-04-27T21:08:53Z","silentRounds":1,"language":"Rust","topics":["note-taking"],"aiDims":["AI应用","效率工具"],"zone":"AI","funScore":0.5,"domainTags":["会议记录","语音转录","本地AI","隐私保护"],"domainKey":"会议记录","pushedAt":"2026-09-23T21:44:10Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/michaelwilhelmsen/humla","ts":"2026-09-23T21:44:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"git-learning-game/oh-my-git","owner":"git-learning-game","name":"oh-my-git","desc":"","summaryCn":"Oh My Git! 是一款专为版本控制系统 Git 设计的教育型独立游戏，它打破了传","reasonCn":"该游戏的最大亮点在于其跨平台的易上手性。不同于需要配置复杂环境的编程沙盒，开发者为 Linux、macOS 和 Windows 三大主流操作系统提供了直接在 itch.io 上可下载的二进制可执行文件。用户无需安装任何依赖库或配置开发环境，下载后双击即可启动游戏，这种“零门槛”的体验极大地降","stars":2889,"starGrowth":0,"createdAt":"2020-03-31T11:00:48Z","silentRounds":3,"language":"GDScript","topics":[],"aiDims":["非AI-好玩","游戏","学习资源"],"zone":"创意","funScore":0.5599999999999999,"domainTags":["Git教程","独立游戏","命令行学习"],"domainKey":"Git教程","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/git-learning-game/oh-my-git","ts":"2026-09-19T22:19:04.205Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"michelpereira/awesome-open-source-games","owner":"michelpereira","name":"awesome-open-source-games","desc":"","summaryCn":"想学习游戏开发却找不到适手的代码参考？michelpereira/awesome-open-sourc","reasonCn":"仓库的分类逻辑非常细致，不局限于「开放世界」或「核心梗」，而是按照游戏机制和技术栈进行了双重维度划分。内容涵盖了从「浏览器基于」的轻量级交互，到「原生」平台上的动作、平台跳跃、策略模拟等重头戏，甚至包括「编程」类沙盒和「点击器」等休闲类型。","stars":3154,"starGrowth":0,"createdAt":"2021-12-26T03:01:58Z","silentRounds":2,"language":"","topics":[],"aiDims":["非AI-好玩","学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["开源游戏","Browser-Based","Game Engine","Source Code"],"domainKey":"开源游戏","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1}],"aiScore":0.38,"source":"search","url":"https://github.com/michelpereira/awesome-open-source-games","ts":"2026-09-19T22:19:04.205Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NUKnightLab/TimelineJS3","owner":"NUKnightLab","name":"TimelineJS3","desc":"","summaryCn":"用 JavaScript 打造的网页动态时间轴故事制作工具。","reasonCn":"为了方便不同技术背景的用户，项目针对各大主流的内容管理系统开发了专门的嵌入插件，支持包括 Wordpress、MediaWiki 以及 Drupal 在内的多个主流平台。这意味着无论是个人博客还是大型维基站点，都可以非常平滑地将这种视觉化的时间轴内容集成到现有的网站架构当中。","stars":3220,"starGrowth":0,"createdAt":"2014-07-29T16:24:03Z","silentRounds":3,"language":"JavaScript","topics":[],"aiDims":["可视化","创意工具"],"zone":"工具","funScore":0.44,"domainTags":["时间轴","网页故事","可视化","JavaScript"],"domainKey":"时间轴","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/NUKnightLab/TimelineJS3","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apify/crawlee-python","owner":"apify","name":"crawlee-python","desc":"Crawlee—A web scraping and browser automation library for Python to build reliable crawlers. Extract data for AI, LLMs, RAG, or GPTs. Download HTML, PDF, JPG, PNG, and other files from websites. Works with Parsel, BeautifulSoup, Playwright, and raw HTTP. Both headful and headless mode. With proxy rotation.","summaryCn":"AI的数据捕手，自动抓取全网信息喂给大模型","reasonCn":"Crawlee是Python界的爬虫框架，专为AI设计。它自带智能重试、代理轮换和浏览器自动化，能稳定抓取网页、PDF和图片。特别适合给RAG、GPTs或LLM提供高质量数据源，支持多种输出格式，上手简单且稳定，是AI应用开发者的数据管道利器。","stars":9544,"starGrowth":1,"createdAt":"2024-01-10T08:44:47Z","silentRounds":2,"language":"Python","topics":["automation"],"aiDims":["RAG与知识","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["网页爬虫","浏览器自动化","数据提取","RAG","Python"],"domainKey":"RAG","pushedAt":"2026-09-23T12:31:38Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/apify/crawlee-python","ts":"2026-09-23T12:31:38Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AIEraDev/Clypra","owner":"AIEraDev","name":"Clypra","desc":"A modern video editor built with Tauri, React, and TypeScript. Focus on building free capabilities of premium capcut functionalities","summaryCn":"用它取代昂贵的剪映会员，打造属于你的本地化智能视频剪辑工作流。","reasonCn":"这款基于Tauri与React开发的视频编辑器，旨在复刻剪映的核心付费功能。通过集成先进的AI算法，实现自动化剪辑、字幕生成及特效处理。其轻量级架构保证了在本地即可完成复杂的视频流处理，无需依赖昂贵的云端订阅，为视频创作者提供了一个免费、隐私且高度可定制的高性能剪辑平台。","stars":3267,"starGrowth":9,"createdAt":"2026-04-26T14:31:10Z","silentRounds":0,"language":"TypeScript","topics":["editor","video-editor"],"aiDims":["创意工具","AI应用"],"zone":"创意","funScore":0,"domainTags":["视频剪辑","本地化创作","自动化转场","字幕生成","Tauri桌面应用"],"domainKey":"视频剪辑","pushedAt":"2026-09-23T21:43:54Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"video-editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/AIEraDev/Clypra","ts":"2026-09-23T21:43:54Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.042317355371900824,"copyOk":false},{"repo":"antimatter15/splat","owner":"antimatter15","name":"splat","desc":"","summaryCn":"目前的 3D 场景重建技术如 NeRF 往往需要高性能 GPU 才能加载和交互，阻碍了","reasonCn":"该库的核心亮点在于其极高的渲染效率。它本质上是点云渲染的扩展，利用高斯分布来表达场景中的细节，使得在普通图形硬件上运行成为可能。相比以往的同类技术，splat 提供了更少的延迟和更平滑的视野过渡，用户只需调整鼠标即可实现 360 度全景观察，视觉冲击力极强，完全不需要编写","stars":3072,"starGrowth":0,"createdAt":"2023-09-11T02:24:56Z","silentRounds":4,"language":"JavaScript","topics":[],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["WebGL渲染","3D高斯泼溅","实时渲染器"],"domainKey":"WebGL渲染","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/antimatter15/splat","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"truelockmc/streambert","owner":"truelockmc","name":"streambert","desc":"A cross-platform Electron Desktop App to stream and download any Movie, TV Series or Anime in the World. Zero Ads and Tracking","summaryCn":"一键跨平台观看和下载全球影视动漫，零广告零追踪的桌面播放器。","reasonCn":"StreamBert 基于 Electron 打包跨平台桌面客户端，内置多源爬虫和解析引擎，可自动检索并聚合全球各大视频平台资源，支持在线观看、批量下载并自动转码，全部流程免广告、无用户追踪，界面简洁可自定义搜索过滤，兼容 Windows、macOS 与 Linux，适配常见视频格式和字幕。","stars":6130,"starGrowth":2,"createdAt":"2026-02-17T17:40:47Z","silentRounds":0,"language":"JavaScript","topics":["anime"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["影视聚合搜索","视频下载器","跨平台桌面应用"],"domainKey":"影视聚合搜索","pushedAt":"2026-09-11T17:06:36Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/truelockmc/streambert","ts":"2026-09-11T17:06:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"flyway/flyway","owner":"flyway","name":"flyway","desc":"Flyway by Redgate • Database Migrations Made Easy.","summaryCn":"数据库迁移神器，让版本控制从代码延伸到你的数据结构","reasonCn":"Flyway是一款由Redgate开发的开源数据库迁移工具，支持Java、.NET和命令行三种方式。它通过管理数据库版本的SQL脚本实现 schema 的自动迁移和回滚，支持超过25种主流数据库，包括MySQL、PostgreSQL、Oracle、SQL Server等。Flyway的核心优势是简单直观的版本管理方式和与CI/CD流水线的无缝集成，是数据库版本控制领域最成熟的工具之一。","stars":10105,"starGrowth":2,"createdAt":"2012-12-14T18:43:22Z","silentRounds":0,"language":"Java","topics":["devops","database"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["数据库迁移","版本控制","CI/CD"],"domainKey":"数据库迁移","pushedAt":"2026-09-15T09:58:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.48,"source":"search","url":"https://github.com/flyway/flyway","ts":"2026-09-15T09:58:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rohitg00/agentmemory","owner":"rohitg00","name":"agentmemory","desc":"#1 Persistent memory for AI coding agents based on real-world benchmarks","summaryCn":"给你的AI编程助手装上一个永不掉线的超级大脑","reasonCn":"在技术实现上，它并不只是简单的对话历史存储，而是构建了一套基于基准测试优化的持久化记忆机制。它能将复杂的项目上下文、用户的特定偏好以及之前解决 Bug 的路径进行结构化存储。当 AI 再次面对类似任务时，它会通过检索增强生成（RAG）的方式，迅速从记忆库中提取相关片段，从","stars":28718,"starGrowth":33,"createdAt":"2026-02-25T07:32:52Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","RAG与知识","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["AI记忆","RAG","编程助手"],"domainKey":"RAG","pushedAt":"2026-09-21T06:13:05Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/rohitg00/agentmemory","ts":"2026-09-21T06:13:05Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.12293677832718156,"copyOk":false},{"repo":"e2b-dev/E2B","owner":"e2b-dev","name":"E2B","desc":"Open-source, secure environment with real-world tools for enterprise-grade agents.","summaryCn":"给AI Agent建安全沙箱，像虚拟机一样跑任意代码的神器","reasonCn":"E2B是一个开源的云端沙箱平台，专门为AI Agent提供隔离的执行环境。它支持Python、Bash、Node.js等语言，Agent可以在里面安全地运行代码、调用API、操作文件系统，就像拥有一个专属的云端电脑。开发者只需几行代码就能部署，支持GPU加速和持久化存储，非常适合需要动态执行代码的AI应用。","stars":13923,"starGrowth":16,"createdAt":"2023-03-04T13:41:18Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["Agent沙箱","容器隔离","代码执行","云端运行时"],"domainKey":"Agent沙箱","pushedAt":"2026-09-22T13:36:25Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/e2b-dev/E2B","ts":"2026-09-22T13:36:25Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.05144753285929757,"copyOk":false},{"repo":"Taiizor/ReaLTaiizor","owner":"Taiizor","name":"ReaLTaiizor","desc":"ReaLTaiizor is a .NET WinForms control library that offers a wide range of components and is user-friendly and design-focused.","summaryCn":"用C#快速搭建漂亮桌面应用的万能控件库，设计感拉满","reasonCn":"ReaLTaiizor是一款专为.NET WinForms设计的现代化UI控件库，包含按钮、面板、导航、通知等丰富的UI组件。采用扁平化设计风格，API简洁易用，支持主题自定义。相比原生WinForms控件，它让桌面应用界面焕然一新，开发者无需手写大量样式代码即可构建专业级UI。","stars":2348,"starGrowth":0,"createdAt":"2020-04-09T22:53:14Z","silentRounds":4,"language":"C#","topics":["design"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WinForms","桌面UI","C#","控件库"],"domainKey":"WinForms","pushedAt":"2026-09-23T03:06:22Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/Taiizor/ReaLTaiizor","ts":"2026-09-23T03:06:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"massgravel/Microsoft-Activation-Scripts","owner":"massgravel","name":"Microsoft-Activation-Scripts","desc":"Open-source Windows and Office activator featuring HWID, Ohook, TSforge, and Online KMS activation methods, along with advanced troubleshooting.","summaryCn":"Windows系统强力激活神器，支持多种黑科技方案","reasonCn":"这是一个开源的Windows和Office激活工具，支持HWID、Ohook、TSforge和在线KMS等多种激活方式，还能解决系统疑难杂症。界面友好，支持多种版本，让普通用户无需复杂操作即可轻松完成系统激活，实用性强，但涉及系统底层修改，使用需谨慎并遵守相关版权规定。","stars":191611,"starGrowth":99,"createdAt":"2020-01-12T23:03:34Z","silentRounds":0,"language":"Batchfile","topics":["hwid","kms38","kms","activator","massgrave","massgravel","windows","windows-10","windows-11","office","ohook","microsoft","microsoft365","office365","powershell","tsforge"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Windows激活","Office激活","批处理脚本","系统工具"],"domainKey":"Windows激活","pushedAt":"2026-08-07T04:07:05.660Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"hwid","source":"github","weight":0.7},{"name":"kms38","source":"github","weight":0.7},{"name":"kms","source":"github","weight":0.7},{"name":"activator","source":"github","weight":0.7},{"name":"massgrave","source":"github","weight":0.7},{"name":"massgravel","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"windows-10","source":"github","weight":0.7},{"name":"windows-11","source":"github","weight":0.7},{"name":"office","source":"github","weight":0.7},{"name":"ohook","source":"github","weight":0.7},{"name":"microsoft","source":"github","weight":0.7},{"name":"microsoft365","source":"github","weight":0.7},{"name":"office365","source":"github","weight":0.7},{"name":"powershell","source":"github","weight":0.7},{"name":"tsforge","source":"github","weight":0.7},{"name":"Batchfile","source":"language","weight":0.5}],"aiScore":0.52,"source":"trending","url":"https://github.com/massgravel/Microsoft-Activation-Scripts","ts":"2026-08-07T04:07:05.660Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.2985500153957758,"copyOk":false},{"repo":"AdventDevInc/kudu","owner":"AdventDevInc","name":"kudu","desc":"Free Windows, Mac and Linux cleaner, scanner, and more.","summaryCn":"跨平台系统清理利器，一键扫描垃圾文件让电脑焕然一新","reasonCn":"基于原生界面构建的免费跨平台系统清理工具，支持 Windows、macOS 和 Linux 三大平台。提供缓存清理、日志清理、回收站清空、浏览器历史清理等多种深度扫描功能，界面简洁现代化，操作直观，无需安装即可运行，适合追求轻量高效系统环境的个人用户和开发者。","stars":3562,"starGrowth":7,"createdAt":"2026-03-14T17:38:51Z","silentRounds":0,"language":"TypeScript","topics":["privacy"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["系统清理","缓存清理","跨平台"],"domainKey":"系统清理","pushedAt":"2026-09-22T12:18:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/AdventDevInc/kudu","ts":"2026-09-22T12:18:32Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03227467714766984,"copyOk":true},{"repo":"zebbern/claude-code-guide","owner":"zebbern","name":"claude-code-guide","desc":"Claude Code Guide - Setup, Commands, workflows, agents, skills & tips-n-tricks from beginner to power user!","summaryCn":"手把手教你从零玩转Claude Code，保姆级实战宝典","reasonCn":"这是一个涵盖Claude Code全方位配置、指令集、工作流与Agent技能的权威指南。项目从新手入门到专家级技巧深度解析，提供从环境搭建到复杂工作流配置的完整路径，帮助开发者快速掌握Claude Code的高级用法与最佳实践。","stars":4631,"starGrowth":2,"createdAt":"2025-06-21T20:43:25Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","教程"],"zone":"资源","funScore":0,"domainTags":["Claude Code","配置指南","工作流","Agent","教程"],"domainKey":"Claude Code","pushedAt":"2026-09-23T00:56:04Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/zebbern/claude-code-guide","ts":"2026-09-23T00:56:04Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"neuml/txtai","owner":"neuml","name":"txtai","desc":"💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows","summaryCn":"txtai把向量数据库和LLM能力打包成一个简单易用的Python库，让AI搜索和知识","reasonCn":"简单来说，以前搭建语义搜索系统需要懂Embedding模型、向量数据库、检索算法，现在txtai把这些都封装好了，几行代码就能跑起来。你可以用它做文档问答、语义检索、知识图谱构建等多种应用，无需深入理解底层原理。 技术架构上，txtai支持多种后端：向量数据库、搜索引擎以及LLM编排","stars":12968,"starGrowth":2,"createdAt":"2020-08-09T19:14:59Z","silentRounds":0,"language":"Python","topics":["vector-db","rag","agents","transformers"],"aiDims":["AI搜索","向量数据库","RAG与知识"],"zone":"AI","funScore":0.5,"domainTags":["语义搜索","向量数据库","RAG","LLM编排"],"domainKey":"向量数据库","pushedAt":"2026-09-23T20:43:52Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/neuml/txtai","ts":"2026-09-23T20:43:52Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"WebKit/WebKit","owner":"WebKit","name":"WebKit","desc":"Home of the WebKit project, the browser engine used by Safari, Mail, App Store and many other applications on macOS, iOS and Linux.","summaryCn":"苹果Safari浏览器的核心引擎，开源世界里的浏览器性能天花板。","reasonCn":"WebKit是一个开源的浏览器渲染引擎，负责解析HTML/CSS并执行JavaScript，被Safari、Mail、App Store等应用广泛使用。它拥有成熟的JavaScript引擎和高效的渲染管线，在处理复杂网页和动画时表现优异。项目采用C++/JavaScript开发，是Web技术栈的重要基石。","stars":10173,"starGrowth":0,"createdAt":"2020-12-11T17:11:43Z","silentRounds":2,"language":"JavaScript","topics":["webkit","browser","web","javascript","macos","ios","gtk"],"aiDims":["非AI-实用","开发工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["浏览器引擎","C++","渲染技术"],"domainKey":"浏览器引擎","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"webkit","source":"github","weight":0.7},{"name":"browser","source":"github","weight":0.7},{"name":"web","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"ios","source":"github","weight":0.7},{"name":"gtk","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/WebKit/WebKit","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JeffersonQin/lightnovel_epub","owner":"JeffersonQin","name":"lightnovel_epub","desc":"🍭 epub generator for (light)novels (轻)小说 epub 生成器，支持站点：轻之国度、轻小说文库","summaryCn":"一键抓取轻小说章节，自动生成排版整齐的 EPUB，省时省力","reasonCn":"lightnovel_epub 基于 Python 实现，能抓取轻之国度、轻小说文库等站点的章节，自动解析目录与正文，转换为带章节标题和图片的 EPUB，用户只需提供小说链接，一键生成可在各类阅读器使用的电子书。支持批量处理，可自定义封面和元数据，生成后直接在手机或电脑阅读器同步阅读。","stars":348,"starGrowth":0,"createdAt":"2021-09-05T02:38:21Z","silentRounds":9,"language":"Python","topics":["ebook"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["轻小说","EPUB生成","网页爬虫","电子书制作","批量转换"],"domainKey":"轻小说","pushedAt":"2026-09-10T00:18:56Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/JeffersonQin/lightnovel_epub","ts":"2026-09-10T00:18:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TheAlgorithms/Algorithms-Explanation","owner":"TheAlgorithms","name":"Algorithms-Explanation","desc":"","summaryCn":"该项目解决的是算法学习门槛高、文档语言受限的痛点。","reasonCn":"为了让维护者方便扩展，项目采用了标准化的目录命名规范。根据文档说明，如果需要添加新语言，建议直接使用该语言的 ISO 639-1 两位字符代码作为文件夹名称，例如使用“hi”代表印地语。这种规范化设计使得多语言社区贡献变得简单且有序，极大地降低了协作成本，让不懂代码的解释工作","stars":3470,"starGrowth":0,"createdAt":"2017-09-29T05:25:25Z","silentRounds":3,"language":"","topics":[],"aiDims":["学习资源"],"zone":"资源","funScore":0.2,"domainTags":["算法解释","多语言文档","基础算法","编程学习"],"domainKey":"算法解释","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.38,"source":"tier","url":"https://github.com/TheAlgorithms/Algorithms-Explanation","ts":"2026-09-19T18:31:48.187Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"elizaOS/eliza","owner":"elizaOS","name":"eliza","desc":"Open source agentic operating system","summaryCn":"开源AI智能体操作系统，让机器人自己社交聊天搞创作","reasonCn":"elizaOS 是一个开源的 AI 智能体操作系统，能够自主运行多个 AI 角色。它支持连接到各种平台（如 Discord、Telegram、Twitter），可以自动回复、生成内容、甚至与其他 AI 互动。对于喜欢探索 AI Agent 能力、想要构建自己的 AI 社区或自动化工作流的玩家来说，这是一个非常好玩且强大的项目。","stars":19409,"starGrowth":3,"createdAt":"2024-07-09T07:55:40Z","silentRounds":0,"language":"TypeScript","topics":["rag","chatbot"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent框架","多智能体系统","RAG","社交媒体机器人","角色扮演AI"],"domainKey":"RAG","pushedAt":"2026-09-23T21:48:54Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/elizaOS/eliza","ts":"2026-09-23T21:48:54Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ValueCell-ai/ClawX","owner":"ValueCell-ai","name":"ClawX","desc":"ClawX is a desktop app that provides a graphical interface for OpenClaw AI agents. It turns CLI-based AI orchestration into a desktop experience without using the terminal. China website is https://clawx.com.cn.","summaryCn":"给AI智能体装上图形界面，再也不用对着黑框敲命令了","reasonCn":"为OpenClaw AI Agent打造的桌面端客户端，采用TypeScript开发。它将复杂的命令行编排流程转化为直观的图形化操作，支持多智能体协作的视觉化管理，让用户在无需接触终端的情况下即可部署和运行复杂的AI Agent工作流。","stars":7610,"starGrowth":0,"createdAt":"2026-02-05T13:58:36Z","silentRounds":4,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI代理","桌面客户端","工作流可视化","OpenClaw"],"domainKey":"AI代理","pushedAt":"2026-09-23T12:51:58Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/ValueCell-ai/ClawX","ts":"2026-09-23T12:51:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"freeCodeCamp/devdocs","owner":"freeCodeCamp","name":"devdocs","desc":"API Documentation Browser","summaryCn":"聚合全网API文档的本地浏览器，断网也能查文档","reasonCn":"freeCodeCamp开发的API文档浏览器，支持数百种编程语言和框架的官方文档离线使用。一键下载全量文档到本地，内建搜索引擎，无需联网即可快速查找API用法。涵盖JavaScript、Python、Ruby等主流技术栈，非常适合开发时快速查阅。","stars":39478,"starGrowth":2,"createdAt":"2013-10-24T18:16:07Z","silentRounds":0,"language":"Ruby","topics":["developer-tools"],"aiDims":["非AI-实用","学习资源","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["API文档","离线查询","开发者工具"],"domainKey":"API文档","pushedAt":"2026-09-23T16:52:13Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.28,"source":"search","url":"https://github.com/freeCodeCamp/devdocs","ts":"2026-09-23T16:52:13Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"drduh/macOS-Security-and-Privacy-Guide","owner":"drduh","name":"macOS-Security-and-Privacy-Guide","desc":"Community guide to securing and improving privacy on macOS.","summaryCn":"macOS安全配置完全指南，守护你的隐私安全","reasonCn":"这是由社区维护的macOS安全与隐私配置指南，涵盖系统设置、网络防护、应用权限、文件系统等多个层面的安全加固建议。内容从基础的系统更新到高级的防火墙配置都有涉及，是一份非常实用的安全参考手册。虽然不涉及AI技术，但对于需要保护数据安全的开发者和普通用户来说是一份必备指南。","stars":22528,"starGrowth":1,"createdAt":"2015-08-31T03:36:35Z","silentRounds":3,"language":"","topics":["security","privacy"],"aiDims":["安全工具","效率工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["macOS安全","隐私加固","系统配置","防火墙","开发者安全"],"domainKey":"macOS安全","pushedAt":"2026-09-20T22:30:00Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7}],"aiScore":0.35,"source":"search","url":"https://github.com/drduh/macOS-Security-and-Privacy-Guide","ts":"2026-09-20T22:30:00Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Marker-Inc-Korea/AutoRAG","owner":"Marker-Inc-Korea","name":"AutoRAG","desc":"AutoRAG: Now your agent can find anything in your computer. It gets smarter if you are using it frequently.","summaryCn":"让AI助手自动翻遍你电脑里所有文件的全能检索大脑","reasonCn":"它的核心技术在于将本地文件理解与大语言模型的检索增强生成完美结合。它用TypeScript编写，能够高效处理各种复杂的本地文件解析。和普通的搜索工具不同，它具备学习进化能力，会根据你平时查阅的习惯和反馈，自动调整检索策略，让你越用它觉得它越聪明。","stars":5104,"starGrowth":2,"createdAt":"2024-01-10T12:25:00Z","silentRounds":0,"language":"TypeScript","topics":["rag"],"aiDims":["RAG与知识","AI Agent"],"zone":"AI","funScore":0,"domainTags":["RAG","本地文件检索","知识管理","TypeScript","自适应检索"],"domainKey":"RAG","pushedAt":"2026-09-23T14:46:43Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Marker-Inc-Korea/AutoRAG","ts":"2026-09-23T14:46:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aurelia/aurelia","owner":"aurelia","name":"aurelia","desc":"Aurelia 2, a standards-based, front-end framework designed for high-performing, ambitious applications.","summaryCn":"前端圈的清流，不追React不追Vue，用标准JS就能写出高性能应用","reasonCn":"Aurelia 2是一个基于Web Components标准的前端框架，采用装饰器语法让代码更简洁。它不依赖虚拟DOM，而是直接操作真实DOM，性能表现优异。支持TypeScript原生开发，编译时就能检查类型错误。采用插件式架构，按需引入功能模块，打包体积可以控制得很小。与传统框架相比，它更接近浏览器原生能力，学习曲线平缓。","stars":1543,"starGrowth":0,"createdAt":"2018-02-15T05:22:58Z","silentRounds":4,"language":"TypeScript","topics":["web-framework","frontend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["前端框架","Web Components","TypeScript"],"domainKey":"前端框架","pushedAt":"2026-09-21T00:36:53Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/aurelia/aurelia","ts":"2026-09-21T00:36:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"googleapis/mcp-toolbox","owner":"googleapis","name":"mcp-toolbox","desc":"MCP Toolbox for Databases is an open source MCP server for databases.","summaryCn":"官方出品的数据库MCP工具箱，让AI直接查库像查询接口一样简单","reasonCn":"Google官方推出的MCP服务器，支持MySQL、PostgreSQL等多种数据库。通过标准MCP协议，大模型能直接执行SQL、读取表结构，无需写代码即可连接数据源。让AI Agent拥有持久化数据存储和查询能力，极大降低了数据库集成的复杂度，是构建数据智能应用的基础设施。","stars":16477,"starGrowth":2,"createdAt":"2024-06-07T20:52:54Z","silentRounds":0,"language":"Go","topics":["agents"],"aiDims":["工具调用","AI Agent","开发者工具"],"zone":"AI","funScore":0,"domainTags":["MCP","数据库","AI工具"],"domainKey":"MCP","pushedAt":"2026-09-23T21:47:50Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/googleapis/mcp-toolbox","ts":"2026-09-23T21:47:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jooby-project/jooby","owner":"jooby-project","name":"jooby","desc":"The modular web framework for Java and Kotlin","summaryCn":"Java和Kotlin开发者的轻量级Web框架，模块化设计很灵活","reasonCn":"Jooby是一个轻量级Java/Kotlin Web框架，采用模块化架构，启动速度快、内存占用低。支持REST API、WebSocket、静态文件等服务，配置简洁，适合构建中小型Web应用和微服务，对于追求简单高效的Java开发者是个不错的选择。","stars":1780,"starGrowth":0,"createdAt":"2014-10-20T02:03:16Z","silentRounds":22,"language":"Java","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web框架","Java开发","微服务"],"domainKey":"Web框架","pushedAt":"2026-09-21T00:10:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/jooby-project/jooby","ts":"2026-09-21T00:10:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dqev/reicon","owner":"dqev","name":"reicon","desc":"Reicon — Open-Source Icon Library for Designers & Developers","summaryCn":"设计师和开发者的开源图标宝库，海量精选一键复制超方便","reasonCn":"Reicon是一个开源图标资源聚合平台，汇集了多个流行图标库，如Lucide、Heroicons等。它提供了强大的搜索和筛选功能，支持SVG格式直接复制使用。对于设计师和前端开发者来说，这相当于一个免费的图标素材仓，省去了到处找图标的麻烦。虽然这个项目与AI无关，但作为实用型工具，在创意设计和前端开发领域非常有价值。GitHub star 1214，热度中等，开源协议友好，适合日常开发使用。","stars":1570,"starGrowth":3,"createdAt":"2026-05-05T06:27:45Z","silentRounds":0,"language":"HTML","topics":["design"],"aiDims":["非AI-实用","开发者工具","创意工具"],"zone":"资源","funScore":0,"domainTags":["图标库","SVG资源","前端设计"],"domainKey":"图标库","pushedAt":"2026-09-19T10:40:50Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/dqev/reicon","ts":"2026-09-19T10:40:50Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ytmdesktop/ytmdesktop","owner":"ytmdesktop","name":"ytmdesktop","desc":"A Desktop App for YouTube Music","summaryCn":"把YouTube Music装成桌面应用，追歌听歌从此告别浏览器标签页","reasonCn":"技术实现上，它基于Electron框架，将Chromium浏览器引擎打包成独立桌面应用。核心优势包括：原生系统集成，支持macOS菜单栏、Windows通知和Linux桌面集成；离线缓存功能，可以在有网络时缓存歌曲供无网时收听；系统托盘常驻，后台播放不影响其他工作；键盘快捷键支持，空格键播放暂","stars":6529,"starGrowth":4,"createdAt":"2018-10-19T16:20:49Z","silentRounds":0,"language":"TypeScript","topics":["desktop","music"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["音乐播放器","Electron","YouTube Music","桌面应用"],"domainKey":"音乐播放器","pushedAt":"2026-09-13T16:43:42Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"music","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/ytmdesktop/ytmdesktop","ts":"2026-09-13T16:43:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tashfeenahmed/freellmapi","owner":"tashfeenahmed","name":"freellmapi","desc":"7.4 billion tokens per month. 34 free LLM providers. 635 free model endpoints. All behind one /v1 endpoint, plus any custom OpenAI-compatible endpoint. Smart routing, automatic failover, encrypted keys. Personal experimentation only.","summaryCn":"把28个免费AI接口的额度打包，一次调用全搞定","reasonCn":"这是一个OpenAI兼容的API代理，聚合了28个免费大模型服务商的免费额度，每月总计约40亿tokens。通过一个统一的/v1端点暴露所有接口，内置智能路由与自动故障转移，支持自定义端点接入任何OpenAI兼容服务，开发者无需逐一注册各平台。","stars":28083,"starGrowth":332,"createdAt":"2026-04-21T13:38:59Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI基础设施","大语言模型","AI应用"],"zone":"AI","funScore":0,"domainTags":["API聚合","大模型","OpenAI兼容"],"domainKey":"API聚合","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/tashfeenahmed/freellmapi","ts":"2026-09-19T04:31:19.347Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.40255955560303386,"copyOk":false},{"repo":"prometheus/prometheus","owner":"prometheus","name":"prometheus","desc":"The Prometheus monitoring system and time series database.","summaryCn":"帮你盯着服务器健康状况的监控神器，像给电脑装个体检医生。","reasonCn":"Prometheus是一个开源的监控系统和时序数据库，专门用来收集、存储和分析服务器指标。它通过Pull模式主动拉取数据，支持强大的PromQL查询语言，能实时监控CPU、内存、网络流量等关键指标，并提供可视化大屏和报警功能，广泛应用于云原生和微服务架构。","stars":66194,"starGrowth":24,"createdAt":"2012-11-24T11:14:12Z","language":"Go","topics":["monitoring"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["系统监控","时序数据库","PromQL","指标采集","云原生"],"domainKey":"系统监控","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/prometheus/prometheus","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07308771187720942,"copyOk":false},{"repo":"DidierStevens/DidierStevensSuite","owner":"DidierStevens","name":"DidierStevensSuite","desc":"","summaryCn":"这是一个在网络安全与恶意软件分析领域享有盛誉的实用工具合集，由资深安全","reasonCn":"在实际的日常安全运营与应急响应工作中，安全分析师常常需要面对形态各异的畸形文件和混淆代码。该项目中的工具凭借其极高的灵活性和精准性，成为了许多安全团队不可或缺的排查利器。无论是在分析钓鱼邮件中携带的恶意 PDF 文档，还是在解构复杂的宏病毒代码时，这套工具都能提供直","stars":2540,"starGrowth":1,"createdAt":"2015-05-08T11:21:00Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["安全工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["安全分析","逆向工程","实用脚本","密码学工具"],"domainKey":"安全分析","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/DidierStevens/DidierStevensSuite","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mljar/mercury","owner":"mljar","name":"mercury","desc":"Impress your boss and turn a Jupyter notebook into a beautiful, shareable web app — no callbacks, no frontend, no rewrite.","summaryCn":"你有没有过这种经历：花了两天时间调模型、跑数据，结果要汇报给老板或业务","reasonCn":"技术上，Mercury是一个Python库，它解析Jupyter Notebook的结构，自动提取其中的输入单元（代码）和输出单元（图表、文本），然后生成一个基于Dash或Streamlit的Web界面。它的架构非常灵活，支持多种主题皮肤，你可以通过简单的注释配置来调整UI布局。","stars":4353,"starGrowth":0,"createdAt":"2021-12-20T13:43:44Z","silentRounds":22,"language":"Python","topics":["data-visualization","data-science"],"aiDims":[],"zone":"工具","funScore":0,"domainTags":["数据可视化","Jupyter","数据科学","Web应用","报告生成"],"domainKey":"数据可视化","pushedAt":"2026-09-14T08:52:34Z","tags":[{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/mljar/mercury","ts":"2026-09-14T08:52:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Attnam/ivan","owner":"Attnam","name":"ivan","desc":"Iter Vehemens ad Necem - a continuation of the graphical roguelike by members of http://attnam.com","summaryCn":"在传统的文字 Roguelike 基础上，IVAN 用像素画面重新演绎了《Iter Vehemen","reasonCn":"在传统的文字 Roguelike 基础上，IVAN 用像素画面重新演绎了《Iter Vehemens ad Necem》的核心乐趣。玩家进入随机生成的地下城，遭遇各种怪物和陷阱，需要在血腥的回合制战斗中合理分配行动点、使用技能、管理背包，才能活下来。","stars":350,"starGrowth":0,"createdAt":"2013-10-02T22:27:48Z","silentRounds":7,"language":"C++","topics":["roguelike"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["roguelike","回合制策略","像素游戏"],"domainKey":"roguelike","pushedAt":"2026-09-09T09:32:06Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"roguelike","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/Attnam/ivan","ts":"2026-09-09T09:32:06Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sxyazi/yazi","owner":"sxyazi","name":"yazi","desc":"💥 Blazing fast terminal file manager written in Rust, based on async I/O.","summaryCn":"用 Rust 写的终端文件管理器，速度飞起，刷新流畅得让人上瘾","reasonCn":"这是一个基于 Rust 语言开发的终端文件管理器，以异步 I/O 技术实现极致的性能。相比传统工具如 ranger 或 nnn，它的刷新速度极快，尤其在大文件夹下优势明显。支持插件系统、图像预览和模糊搜索，让命令行操作文件变得既高效又具视觉享受。","stars":42364,"starGrowth":20,"createdAt":"2023-07-08T11:45:55Z","silentRounds":0,"language":"Rust","topics":["developer-tools","command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Rust","终端工具","文件管理"],"domainKey":"Rust","pushedAt":"2026-09-21T23:25:15Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/sxyazi/yazi","ts":"2026-09-21T23:25:15Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06141629685582098,"copyOk":false},{"repo":"swisskyrepo/PayloadsAllTheThings","owner":"swisskyrepo","name":"PayloadsAllTheThings","desc":"A list of useful payloads and bypass for Web Application Security and Pentest/CTF","summaryCn":"渗透测试和CTF必备的payload绕过手册，安全圈公认的武林秘籍","reasonCn":"这个仓库收录了800多种Web漏洞利用payload和绕过技巧，涵盖SQL注入、XSS、命令注入、反序列化等常见攻击场景。作者按漏洞类型分类整理，每个漏洞都给出多种攻击代码示例和绕过方法，是渗透测试人员和CTF选手的必备参考手册，实用性极强。","stars":81119,"starGrowth":65,"createdAt":"2016-10-18T07:29:07Z","silentRounds":0,"language":"Python","topics":["security","cheatsheet"],"aiDims":["安全工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["渗透测试","漏洞利用","CTF安全","Web安全","绕过技巧"],"domainKey":"渗透测试","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/swisskyrepo/PayloadsAllTheThings","ts":"2026-09-07T02:44:20.540Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1974038757874234,"copyOk":false},{"repo":"Epistates/pmetal","owner":"Epistates","name":"pmetal","desc":"PMetal: high-performance Apple Silicon framework for local LLM inference, LoRA/QLoRA fine-tuning, serving, quantization, and MLX/Metal acceleration.","summaryCn":"简单来说，这个项目解决了Mac用户想跑大模型却苦于没有专业显卡的痛点。","reasonCn":"在技术架构上，pMetal底层使用了MLX框架，这是苹果官方推出的机器学习框架，专为优化Apple Silicon设计。它利用了Metal的高效并行计算能力，实现了对模型权重的量化处理。这意味着原本需要几十GB显存的模型，经过4位量化后，可以压缩到几GB甚至更小，直接在内存中就能跑起来。","stars":319,"starGrowth":0,"createdAt":"2026-01-12T17:27:23Z","silentRounds":2,"language":"Rust","topics":["transformers"],"aiDims":["推理引擎","模型部署","微调"],"zone":"AI","funScore":1,"domainTags":["本地推理","Apple Silicon","量化","推理框架"],"domainKey":"本地推理","pushedAt":"2026-09-23T17:06:42Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Epistates/pmetal","ts":"2026-09-23T17:06:42Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jackwener/xiaohongshu-cli","owner":"jackwener","name":"xiaohongshu-cli","desc":"","summaryCn":"它不仅支持基础的笔记搜索与详情查看，还贴心地提供了短索引导航功能，例如","reasonCn":"在身份认证方面，该项目支持自动提取浏览器 Cookie 或通过二维码登录，确保了操作的便捷性与安全性。用户登录后可以查看个人主页、查看评论与回复，实现与在网页端几乎一致的功能体验，同时又享受到了命令行程序那种轻量级、无广告、即时响应的快感。","stars":2622,"starGrowth":0,"createdAt":"2026-03-08T05:34:05Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["命令行工具","小红书","API逆向"],"domainKey":"命令行工具","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/jackwener/xiaohongshu-cli","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"googlecreativelab/coder","owner":"googlecreativelab","name":"coder","desc":"","summaryCn":"Coder 项目最初的目标是提供一个极简的 Web 编程环境，通过浏览器直接实现代","reasonCn":"随着开源生态的发展，该项目目前已被官方标记为只读的存档状态，这意味着开发者无法通过提议补丁或反馈问题来获取官方支持。尽管处于封存状态，但其内部实现的 Web 组件和模块化设计依然具有参考价值，特别是对于那些想要研究如何构建基于浏览器的轻量级 IDE 的开发者来说，阅读其","stars":2422,"starGrowth":0,"createdAt":"2013-06-19T20:14:10Z","silentRounds":2,"language":"JavaScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Web开发","代码编辑器","JavaScript"],"domainKey":"Web开发","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/googlecreativelab/coder","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"baptisteArno/typebot.io","owner":"baptisteArno","name":"typebot.io","desc":"💬 Typebot is a powerful chatbot builder that you can self-host.","summaryCn":"拖拽积木就能搭建高颜值聊天机器人，支持私有化部署。","reasonCn":"技术上，它采用TypeScript开发，前端界面精致现代，用户体验媲美主流SaaS软件。后端架构支持高并发，能够轻松对接各种Webhook和外部API。更厉害的是它支持大模型接入，你可以把ChatGPT或者本地模型塞进机器人大脑里，让它不再只会死板地回答固定台词，而是具备真正的理解和生成能力","stars":10357,"starGrowth":3,"createdAt":"2021-11-19T09:15:31Z","silentRounds":0,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI应用","工具调用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["聊天机器人构建器","可视化流程编排","自托管","Webhook集成","TypeScript"],"domainKey":"聊天机器人构建器","pushedAt":"2026-09-23T19:18:45Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/baptisteArno/typebot.io","ts":"2026-09-23T19:18:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gptme/gptme","owner":"gptme","name":"gptme","desc":"Your agent in your terminal, equipped with local tools: writes code, uses the terminal, browses the web. Make your own persistent autonomous agent on top!","summaryCn":"把你终端变成AI管家，自动帮你写代码、跑命令、甚至浏览网页，省力神器。","reasonCn":"这是一个运行在终端里的自主智能体，不仅能实时编写代码，还能直接接管你的系统终端权限。它自带丰富的本地工具集，能够根据需求自动执行Shell命令、访问Web资源并完成复杂任务链。作为一个具备持久化记忆的智能代理，它彻底改变了开发者的操作范式，将枯燥的命令行交互升级为由AI驱动的高效自动化协同，特别适合喜欢全键盘操作的极客与重度开发者使用。","stars":4421,"starGrowth":2,"createdAt":"2023-03-24T16:25:01Z","silentRounds":0,"language":"Python","topics":["chatbot","rag"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["AI代理","终端","自动化"],"domainKey":"AI代理","pushedAt":"2026-09-23T21:16:02Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/gptme/gptme","ts":"2026-09-23T21:16:02Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"pdone/lx-music-source","owner":"pdone","name":"lx-music-source","desc":"洛雪音乐源","summaryCn":"一个开源的洛雪音乐平台后端，提供海量免费音乐源，轻松接入播放器。","reasonCn":"lx-music-source 是洛雪音乐播放器的后端数据源插件，聚合多家公开音乐平台的 API，提供统一的搜索、播放列表、歌词和封面接口。项目采用 Node.js 编写，支持自定义源、缓存加速和跨平台部署，开发者可直接在本地或服务器上运行，快速为播放器补齐缺失的曲库。","stars":9065,"starGrowth":29,"createdAt":"2024-11-15T06:42:06Z","silentRounds":0,"language":"JavaScript","topics":["lx-music","lx-music-desktop","lx-music-mobile"],"aiDims":["非AI-实用","非AI-好玩"],"zone":"工具","funScore":0.6,"domainTags":["音乐聚合","API代理","洛雪音乐","Node.js","源插件"],"domainKey":"音乐聚合","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"lx-music","source":"github","weight":0.7},{"name":"lx-music-desktop","source":"github","weight":0.7},{"name":"lx-music-mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/pdone/lx-music-source","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0965973524544953,"copyOk":false},{"repo":"ocropus-archive/DUP-ocropy","owner":"ocropus-archive","name":"DUP-ocropy","desc":"","summaryCn":"OCRopus 最鲜明的特征是它“不装傻”的极客定位。","reasonCn":"在文档中，项目特别列举了一些常被使用者忽视但极具科研价值的功能。例如，它附带了一群用于“ground truth editing and correction”的脚本，以及专门用来“measuring error rates, determining confusion matrices”的深度测量工具。","stars":3464,"starGrowth":0,"createdAt":"2014-10-28T20:26:34Z","silentRounds":4,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","文档理解"],"zone":"资源","funScore":0.2,"domainTags":["OCR","Jupyter Notebook","文档分析","模型训练"],"domainKey":"OCR","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/ocropus-archive/DUP-ocropy","ts":"2026-09-19T18:31:48.187Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liriliri/aya","owner":"liriliri","name":"aya","desc":"","summaryCn":"AYA 是一款专为 Android 开发者设计的桌面应用程序，其核心定位是 ADB 的图","reasonCn":"在功能层面，AYA 提供了全方位的调试工具集。屏幕镜像功能允许用户将手机画面实时投射到电脑上，方便进行 UI 验证和交互测试；文件资源管理器则解决了电脑上直接访问手机存储的痛点，支持文件的拖拽传输与删除。对于深入调试，它集成了进程监控、CPU 和内存性能监测，甚至能够实时","stars":5476,"starGrowth":1,"createdAt":"2024-10-14T10:39:33Z","silentRounds":2,"language":"TypeScript","topics":["adb","electron","electron-app","android"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Android ADB","屏幕镜像","布局检查","日志查看"],"domainKey":"Android ADB","pushedAt":"2026-09-16T18:07:39.903Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"adb","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"electron-app","source":"github","weight":0.7},{"name":"android","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/liriliri/aya","ts":"2026-09-16T18:07:39.903Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Significant-Gravitas/AutoGPT","owner":"Significant-Gravitas","name":"AutoGPT","desc":"AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.","summaryCn":"能自己思考自己干活的AI智能体，设定目标它自动完成","reasonCn":"技术架构上，AutoGPT采用了经典的Agent循环：感知环境、制定计划、执行动作、评估结果。它连接多个大语言模型作为\"大脑\"，配备了一套工具系统，包括网页搜索、文件读写、代码执行、内存存储等。核心创新在于任务分解和记忆机制，能把复杂目标拆解成可执行的小步骤，并通过向量存储","stars":187501,"starGrowth":17,"createdAt":"2023-03-16T09:21:07Z","silentRounds":0,"language":"Python","topics":["llm","ai","agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI智能体","LLM应用","自主Agent","任务自动化"],"domainKey":"AI智能体","pushedAt":"2026-09-23T21:46:12Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/Significant-Gravitas/AutoGPT","ts":"2026-09-23T21:46:12Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05127199854934107,"copyOk":false},{"repo":"alphaXiv/OpenResearch","owner":"alphaXiv","name":"OpenResearch","desc":"Turn your coding agents into research agents","summaryCn":"像军队一样分工协作的AI科研军团，一键启动","reasonCn":"这是一个开源的并行研究Agent框架，旨在解决科研流程碎片化的问题。它允许用户将任何大模型接入，并通过编排多个独立Agent来并行处理文献综述、实验设计、代码生成和结果分析等科研任务。项目基于Rust开发，利用高效并发机制，让AI像专业团队一样协作，极大提升科研效率。","stars":5571,"starGrowth":75,"createdAt":"2026-06-07T01:51:39Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["AI Agent","AI搜索","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["多智能体","科研自动化","LLM框架","并发架构"],"domainKey":"多智能体","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/alphaXiv/OpenResearch","ts":"2026-09-19T04:31:19.347Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.318465266558966,"copyOk":true},{"repo":"IBM/plex","owner":"IBM","name":"plex","desc":"The package of IBM’s typeface, IBM Plex.","summaryCn":"IBM家出品的颜值超高字体家族，设计感拉满。","reasonCn":"这项目就是IBM Plex字体的源代码包，包含字体文件和设计规格。它是开源的，设计感极强，既适合做网页标题也适合做代码注释。虽然本身不是AI工具，但作为设计资源，它能让你的代码界面或演示文稿颜值瞬间提升一个档次，非常实用且好用。","stars":11653,"starGrowth":2,"createdAt":"2017-10-03T20:58:50Z","silentRounds":0,"language":"CSS","topics":["design"],"aiDims":["非AI-好玩","可视化","创意工具"],"zone":"资源","funScore":0,"domainTags":["字体设计","开源字体","UI设计资源"],"domainKey":"字体设计","pushedAt":"2026-09-22T11:32:26Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/IBM/plex","ts":"2026-09-22T11:32:26Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CorentinTh/it-tools","owner":"CorentinTh","name":"it-tools","desc":"Collection of handy online tools for developers, with great UX. ","summaryCn":"程序员必备工具箱，50多个实用工具网页版免费用","reasonCn":"这是一个汇集了50多种开发者常用工具的开源在线平台，包括JSON格式化、正则表达式测试、Base64编解码、UUID生成、时间戳转换、图片压缩、颜色选择器等。全部在浏览器端运行，无需上传数据，隐私安全。界面美观，响应迅速，是开发者日常工作的瑞士军刀。","stars":40667,"starGrowth":7,"createdAt":"2020-04-05T11:50:24Z","silentRounds":0,"language":"Vue","topics":["developer-tools","frontend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["开发者工具箱","JSON格式化","正则测试"],"domainKey":"开发者工具箱","pushedAt":"2026-09-16T17:00:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/CorentinTh/it-tools","ts":"2026-09-16T17:00:22Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02151638920992451,"copyOk":true},{"repo":"GyulyVGC/sniffnet","owner":"GyulyVGC","name":"sniffnet","desc":"Comfortably monitor your Internet traffic 🕵️‍♂️","summaryCn":"你有没有这样的经历——电脑突然变卡了，怀疑是有人在偷网速？","reasonCn":"技术上，SniffNet采用Rust语言开发，天生具备内存安全和高性能优势。它利用libpcap（Unix系）和Npcap（Windows）等底层网络抓包库捕获流量，然后解析出协议、IP地址、端口、数据量等关键信息。与传统命令行工具如tcpdump相比，SniffNet的最大特点是提供了精美的图形界面——实时流量","stars":41194,"starGrowth":12,"createdAt":"2022-07-31T21:55:56Z","silentRounds":0,"language":"Rust","topics":["security"],"aiDims":["安全工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":1,"domainTags":["网络流量分析","数据包捕获","网络可视化","Rust GUI","网络安全监控"],"domainKey":"网络流量分析","pushedAt":"2026-09-19T06:04:10Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/GyulyVGC/sniffnet","ts":"2026-09-19T06:04:10Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03687391367674904,"copyOk":false},{"repo":"hi-godot/godot-ai","owner":"hi-godot","name":"godot-ai","desc":"Production-grade MCP server and AI tools for the Godot engine. A Snap to install. Totally free and fun.","summaryCn":"在游戏开发中，给NPC写逻辑是最累人的工作之一。","reasonCn":"核心亮点在于它完美兼容 MCP（Model Context Protocol）标准，这意味着它可以轻松连接各种模型资源。此外，它还提供了针对 Godot 内部场景树（Scene Tree）的深度优化，能够识别游戏中的节点和属性。这意味着 AI 不仅能聊天，还能直接帮你调整游戏里物体的坐标、颜色甚至是生成新的","stars":2551,"starGrowth":12,"createdAt":"2026-04-12T02:23:43Z","silentRounds":0,"language":"GDScript","topics":["game-development","godot"],"aiDims":["游戏","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["游戏开发","NPC AI","MCP协议","Godot引擎"],"domainKey":"游戏开发","pushedAt":"2026-09-23T11:41:31Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/hi-godot/godot-ai","ts":"2026-09-23T11:41:31Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0601345354762838,"copyOk":false},{"repo":"WillCodeForCats/solaredge-modbus-multi","owner":"WillCodeForCats","name":"solaredge-modbus-multi","desc":"The Complete Local (Modbus/TCP) SolarEdge Integration for Home Assistant","summaryCn":"一个用于智能家居系统的Modbus/TCP模块，用于连接和管理太阳能板","reasonCn":"该项目提供了一个用于管理太阳能板的Modbus/TCP接口，使得家庭自动化系统能够更好地监控和控制这些设备。它还集成了AI代理来优化能源管理和预测性维护。","stars":327,"starGrowth":0,"createdAt":"2021-10-21T15:53:33Z","silentRounds":6,"language":"Python","topics":["home-automation"],"aiDims":["Home-Automation","AI Agent"],"zone":"工具","funScore":0,"domainTags":["家庭自动化","太阳能监控","Modbus TCP","Home Assistant集成","能源管理"],"domainKey":"家庭自动化","pushedAt":"2026-09-22T21:37:30Z","tags":[{"name":"Home-Automation","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":1,"source":"search","url":"https://github.com/WillCodeForCats/solaredge-modbus-multi","ts":"2026-09-22T21:37:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hatoo/oha","owner":"hatoo","name":"oha","desc":"Ohayou(おはよう), HTTP load generator, inspired by rakyll/hey with tui animation.","summaryCn":"压力测试工具还能看动画，让枯燥的压测变得生动有趣","reasonCn":"技术层面，它是用Rust写的，性能很高，能产生大量并发请求。基于tokio异步运行时，单进程就能打出很高的QPS。支持多种HTTP方法、自定义header、认证信息，还能设置延迟时间模拟真实用户行为。输出结果可以用JSON格式方便地集成到监控系统或CI流程中。","stars":10553,"starGrowth":1,"createdAt":"2020-03-02T13:26:35Z","silentRounds":3,"language":"Rust","topics":["command-line"],"aiDims":["非AI-好玩","开发者工具"],"zone":"工具","funScore":0.7,"domainTags":["HTTP压力测试","终端动画","Rust","CLI工具","性能测试"],"domainKey":"HTTP压力测试","pushedAt":"2026-09-10T04:43:05Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/hatoo/oha","ts":"2026-09-10T04:43:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"luckjiawei/frpc-desktop","owner":"luckjiawei","name":"frpc-desktop","desc":"Cross-platform desktop client for FRP, visual configuration, easily achieve intranet penetration!","summaryCn":"可视化内网穿透桌面客户端，不用敲命令也能把家里服务暴露到公网","reasonCn":"这是一个基于 Electron 开发的 FRP 客户端桌面应用，用可视化界面替代了命令行配置 frpc 的过程。支持 Windows、macOS、Linux 跨平台运行，自带配置生成器、状态监控、日志查看等实用功能。虽然和 AI 没啥关系，但作为网络工具非常实用，尤其适合需要内网穿透但不想折腾配置文件的开发者。","stars":6895,"starGrowth":0,"createdAt":"2023-11-27T06:53:22Z","silentRounds":1,"language":"TypeScript","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["内网穿透","网络代理","桌面客户端","FRP配置","跨平台应用"],"domainKey":"内网穿透","pushedAt":"2026-09-16T01:47:10Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/luckjiawei/frpc-desktop","ts":"2026-09-16T01:47:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dhjz/hanzi-study","owner":"dhjz","name":"hanzi-study","desc":"儿童学习汉字, 学前教育, 小朋友学习汉字拼音, 通过小游戏闯关的方式学习汉字, 听音识字, 笔画书写, 学习数学加减法, 十以内或者百以内加减法,学习小学古诗词, 古诗词朗读","summaryCn":"小朋友通过闯关小游戏，轻松学汉字与拼音，算数也能玩起来","reasonCn":"hanzi-study 是一款面向儿童的汉字学习平台，采用多种小游戏闯关方式，让孩子在玩乐中掌握汉字笔画、拼音、读音与基础算数。项目支持音频提示、书写练习、古诗朗诵等模块，数据通过本地缓存，适合无网络环境。代码基于 JavaScript，结构清晰，易于在网页或移动端嵌入。","stars":226,"starGrowth":1,"createdAt":"2025-06-21T05:44:53Z","silentRounds":3,"language":"JavaScript","topics":["learning"],"aiDims":["非AI-好玩","学习资源"],"zone":"创意","funScore":1,"domainTags":["儿童汉字教育","拼音学习","笔画书写","数学启蒙","古诗词游戏化","离线可用Web应用"],"domainKey":"儿童汉字教育","pushedAt":"2026-09-10T03:31:55Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/dhjz/hanzi-study","ts":"2026-09-10T03:31:55Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mex-memory/mex","owner":"mex-memory","name":"mex","desc":"Team memory for engineers and their AI agents. Lives in your repo. Shared through Git.","summaryCn":"直接长在代码仓库里的AI团队共享记忆库，通过Git协同进化。","reasonCn":"技术上，它用TypeScript编写，创新性地把AI的记忆载体设计为明文文件，直接跟随项目的Git仓库进行版本控制。这意味着记忆的变更和代码修改一样享有分支、合并和回滚的能力，彻底解决了传统云端知识库数据同步冲突和隐私安全的问题。","stars":1682,"starGrowth":16,"createdAt":"2026-03-21T10:34:18Z","silentRounds":0,"language":"TypeScript","topics":["documentation"],"aiDims":["AI Agent","开发者工具","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI记忆","团队协作","Git集成"],"domainKey":"AI记忆","pushedAt":"2026-09-22T06:35:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"documentation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/mex-memory/mex","ts":"2026-09-22T06:35:45Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0918449464922711,"copyOk":false},{"repo":"dmtrKovalenko/fff","owner":"dmtrKovalenko","name":"fff","desc":"The fastest and the most accurate file search SDK for AI agents, Neovim, Rust, C, Python, Bun and NodeJS","summaryCn":"让AI Agent一秒定位目标文件的闪电搜索神器","reasonCn":"fff是用Rust编写的高性能文件搜索SDK，专为AI Agent设计，支持Python、Node.js、Rust、C、Bun等多种语言绑定，也可在Neovim中调用。核心优势是速度极快且搜索精准，能帮助AI Agent在海量文件中秒级定位目标文件，解决Agent执行任务时找不到文件的痛点。","stars":10848,"starGrowth":43,"silentRounds":0,"language":"Rust","topics":["filesearch","lua","neovim","neovim-plugin","rust","bun","fff","fzf","grep","nodejs","python"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["文件搜索","AI Agent","Rust SDK","内存映射"],"domainKey":"文件搜索","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"filesearch","source":"github","weight":0.7},{"name":"lua","source":"github","weight":0.7},{"name":"neovim","source":"github","weight":0.7},{"name":"neovim-plugin","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"bun","source":"github","weight":0.7},{"name":"fff","source":"github","weight":0.7},{"name":"fzf","source":"github","weight":0.7},{"name":"grep","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/dmtrKovalenko/fff","ts":"2026-09-23T07:35:20.006Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.140891592920354,"copyOk":false},{"repo":"AniMathIO/AniMathIO","owner":"AniMathIO","name":"AniMathIO","desc":"AniMathIO revolutionizes the creation of mathematical videos, tailored for educators, students, and professionals seeking to bring complex concepts to life. ","summaryCn":"把枯燥数学公式变成动画视频，老师学生都爱看","reasonCn":"AniMathIO是基于 TypeScript 的开源平台，提供 DSL 与可视化编辑器，将 LaTeX 公式、图形、代码块等自动转化为时间轴动画，并支持导出 MP4 或 Web 交互式播放器，帮助教学者快速生成高质量数学教学视频。并可嵌入网页，随时分享。","stars":12,"starGrowth":0,"createdAt":"2024-04-08T08:30:49Z","silentRounds":10,"language":"TypeScript","topics":["video-editor"],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":0.8,"domainTags":["数学动画制作","LaTeX可视化","教学视频生成","DSL动画引擎","WebGL渲染"],"domainKey":"数学动画制作","pushedAt":"2026-09-10T07:29:56Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/AniMathIO/AniMathIO","ts":"2026-09-10T07:29:56Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MHSanaei/3x-ui","owner":"MHSanaei","name":"3x-ui","desc":"Supporting multi-protocol multi-user(Vmess, Vless, Trojan, ShadowSocks, Wireguard, Hysteria, Tunnel, Mixed, HTTP, Tun, MTProto، AmneziaWG)","summaryCn":"多协议代理服务器管理面板，一人多用户流量管理全包了","reasonCn":"基于Xray内核的可视化Web管理面板，支持VMess、VLESS、Trojan、ShadowSocks、Wireguard等十几种协议。提供多用户管理、流量统计、过期时间控制、IP限制等功能。Go语言开发，自带现代UI界面，操作友好。适合需要部署多用户代理服务的个人和小型团队，省去了命令行配置的麻烦。","stars":46792,"starGrowth":48,"createdAt":"2023-02-09T19:16:57Z","silentRounds":0,"language":"Go","topics":["trojan","vless","vmess","shadowsocks2022","reality","http","shadowsocks","wireguard","tls","post-quantum","x25519","mixed","tunnel","ech","hysteria2","tun","fail2ban","xhttp","nodes","mtproto"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["代理服务器","Xray","多协议","流量管理"],"domainKey":"代理服务器","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"trojan","source":"github","weight":0.7},{"name":"vless","source":"github","weight":0.7},{"name":"vmess","source":"github","weight":0.7},{"name":"shadowsocks2022","source":"github","weight":0.7},{"name":"reality","source":"github","weight":0.7},{"name":"http","source":"github","weight":0.7},{"name":"shadowsocks","source":"github","weight":0.7},{"name":"wireguard","source":"github","weight":0.7},{"name":"tls","source":"github","weight":0.7},{"name":"post-quantum","source":"github","weight":0.7},{"name":"x25519","source":"github","weight":0.7},{"name":"mixed","source":"github","weight":0.7},{"name":"tunnel","source":"github","weight":0.7},{"name":"ech","source":"github","weight":0.7},{"name":"hysteria2","source":"github","weight":0.7},{"name":"tun","source":"github","weight":0.7},{"name":"fail2ban","source":"github","weight":0.7},{"name":"xhttp","source":"github","weight":0.7},{"name":"nodes","source":"github","weight":0.7},{"name":"mtproto","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/MHSanaei/3x-ui","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.14707744913660453,"copyOk":true},{"repo":"markel1974/Symphony","owner":"markel1974","name":"Symphony","desc":"An open-source emulation framework designed as a virtual laboratory for exploring, building, and modifying classic computer systems in real-time.","summaryCn":"把老式电脑装进浏览器，随点即玩、随改即测的复古实验室。","reasonCn":"Symphony 是用 Go 编写的开源仿真框架，提供实时的虚拟实验室，让用户可以在网页或本地环境中快速搭建、运行并修改经典计算机系统的硬件模型，如 6502、Z80 等，支持指令级调试、内存监控和外设挂载，适合教学、研究和爱好者探索复古计算机架构的细节。","stars":16,"starGrowth":0,"createdAt":"2024-07-11T12:29:44Z","silentRounds":8,"language":"Go","topics":["retrocomputing"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["复古计算","硬件仿真","计算机教育","WebAssembly"],"domainKey":"复古计算","pushedAt":"2026-09-17T06:42:07Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/markel1974/Symphony","ts":"2026-09-17T06:42:07Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Sudharsanselvaraj/Token-Print","owner":"Sudharsanselvaraj","name":"Token-Print","desc":"Interactive 3D visualization platform for exploring transformer architectures, tensors, and real-time LLM inference.","summaryCn":"实时 3D 可视化大模型真实前向推理全过程。","reasonCn":"TokenPrint 不是静态架构图，而是把 Qwen2.5-0.5B 等模型真实前向传播接入 3D 场景，注意力头、RoPE 旋转、残差流每一帧数值均来自真实 named_parameters 或 GGUF 二进制头，构建脚本更会拦截任何伪造张量。研究者可直观看到每层注意力如何聚焦、位置编码如何旋转，调试推理异常不再盲猜。","stars":197,"starGrowth":1,"createdAt":"2026-07-17T13:17:04Z","silentRounds":2,"language":"TypeScript","topics":["transformers"],"aiDims":["可视化","大语言模型","模型与训练"],"zone":"创意","funScore":1,"domainTags":["Transformer可视化","张量流可视化","大模型解释性","WebGL可视化","交互式学习工具"],"domainKey":"Transformer可视化","pushedAt":"2026-09-23T08:42:54Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Sudharsanselvaraj/Token-Print","ts":"2026-09-23T08:42:54Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"cloudflare/quiche","owner":"cloudflare","name":"quiche","desc":"🥧 Savoury implementation of the QUIC transport protocol and HTTP/3","summaryCn":"让网页加载快如闪电的底层黑科技，正在悄悄改变互联网","reasonCn":"Cloudflare 用 Rust 编写的 QUIC 传输协议和 HTTP/3 实现，直接绕过传统 TCP，显著提升网络传输效率。它解决了拥塞控制和丢包重传的痛点，支持 HTTP/3 多路复用，显著减少网页加载延迟和加载失败率，是现代网络基础设施的核心组件。","stars":12531,"starGrowth":49,"language":"Rust","topics":["quic","http3","network-programming","protocol","rust"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["QUIC协议","HTTP/3","网络传输","Rust"],"domainKey":"QUIC协议","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"quic","source":"github","weight":0.7},{"name":"http3","source":"github","weight":0.7},{"name":"network-programming","source":"github","weight":0.7},{"name":"protocol","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/cloudflare/quiche","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.15873090734977255,"copyOk":false},{"repo":"sadanandpai/frontend-mini-challenges","owner":"sadanandpai","name":"frontend-mini-challenges","desc":"Collection of frontend challenges for learning and interviews","summaryCn":"前端开发者的每日小练习，练技术顺便准备面试","reasonCn":"这个仓库就像前端学习的健身馆，每天给你一个具体题目练手。从简单的CSS布局到复杂的JS算法，循序渐进提升技能。\n每个挑战都是独立的小项目，有清晰的题目描述和验收标准。提供基础代码框架，做完还能看参考答案对比。","stars":2413,"starGrowth":0,"createdAt":"2021-06-28T17:36:25Z","silentRounds":25,"language":"JavaScript","topics":["frontend"],"aiDims":["学习资源","开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["前端学习","编程练习","面试准备"],"domainKey":"前端学习","pushedAt":"2026-09-22T09:49:32Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/sadanandpai/frontend-mini-challenges","ts":"2026-09-22T09:49:32Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"diafygi/webrtc-ips","owner":"diafygi","name":"webrtc-ips","desc":"","summaryCn":"diafygi/webrtc-ips 的核心价值在于揭示了一个常被忽视的前端隐私漏洞：通过","reasonCn":"更令人警惕的是该项目指出的追踪风险。由于这些 STUN 请求不经过标准的 XMLHttpRequest 流程，它们对于 Firebug、Developer Console 等调试工具是“隐形”的。这意味着普通用户甚至部分开发者在使用浏览器时，可能完全 unaware 自己的 IP 正在被外部服务器记录。","stars":3455,"starGrowth":0,"createdAt":"2015-01-26T05:55:12Z","silentRounds":4,"language":"HTML","topics":[],"aiDims":["非AI-好玩","安全工具"],"zone":"工具","funScore":0.5599999999999999,"domainTags":["WebRTC","IP地址检测","STUN服务器","前端安全","浏览器指纹"],"domainKey":"WebRTC","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/diafygi/webrtc-ips","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TapXWorld/ChinaTextbook","owner":"TapXWorld","name":"ChinaTextbook","desc":"所有小初高、大学PDF教材。","summaryCn":"从小一到大学全覆盖的中文教材PDF大合集","reasonCn":"这是一个收录了小学到大学全阶段中文教材的PDF资源库，涵盖语文、数学、英语、物理、化学等各科内容。项目作者整理并开源了这些教材的电子版，方便学生自学和教师备课。接近8万的star数证明需求量很大。纯文档聚合，无复杂技术实现，价值在于资源完整性和可访问性。","stars":82250,"starGrowth":28,"createdAt":"2020-01-05T04:11:06Z","silentRounds":0,"language":"Roff","topics":[],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["中文教材","PDF资源","教育合集","学科分类"],"domainKey":"中文教材","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Roff","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/TapXWorld/ChinaTextbook","ts":"2026-08-09T00:58:42.474Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08502127659574468,"copyOk":false},{"repo":"ayghri/i-have-adhd","owner":"ayghri","name":"i-have-adhd","desc":"A skill to stop your coding agent from burying the answer. ADHD-friendly output.","summaryCn":"让编程助手直击重点，先给方案再列步骤，告别空话套话。","reasonCn":"这是一个给 AI 编程助手用的技能包，核心是改变回答的结构和语气。它强制助手先给出具体行动方案，再把步骤用数字编号，同时禁止使用“希望这有帮助”之类的客套话。安装方式很简单，只需复制粘贴命令或查看仓库说明。非常适合觉得 AI 回答冗长、难以抓住重点的开发者。","stars":50303,"starGrowth":736,"createdAt":"2026-05-13T20:42:06Z","silentRounds":0,"language":"Python","topics":["developer-tools"],"aiDims":["AI Agent","提示工程","开发者工具"],"zone":"AI","funScore":0.7,"domainTags":["AI提示词工程","编码助手","ADHD","输出格式"],"domainKey":"AI提示词工程","pushedAt":"2026-09-19T16:44:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/ayghri/i-have-adhd","ts":"2026-09-19T16:44:46Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.41267280281494145,"copyOk":true},{"repo":"agent-substrate/substrate","owner":"agent-substrate","name":"substrate","desc":"Agent Substrate: the core system","summaryCn":"给AI Agent搭建底层地基的Go语言核心框架","reasonCn":"Agent Substrate 是一个用 Go 写的 Agent 核心系统，专注于为智能体提供稳定的运行底座。它不直接提供复杂的模型能力，而是解决 Agent 在执行任务时的状态管理、记忆持久化和工具调度等基础问题。对于想要从零构建自主 Agent 系统的开发者来说，这是一个非常实用的基础设施选择，让复杂的 Agent 逻辑变得可管理、可扩展，不再是个零散的脚本堆砌。","stars":3449,"starGrowth":560,"language":"Go","topics":[],"aiDims":["AI Agent","AI基础设施","开发者工具"],"zone":"AI","funScore":0,"domainTags":["Agent框架","状态管理","记忆持久化","Go语言"],"domainKey":"Agent框架","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/agent-substrate/substrate","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.787097709481009,"copyOk":false},{"repo":"Codennnn/vue-color-avatar","owner":"Codennnn","name":"vue-color-avatar","desc":"","summaryCn":"它是一个基于 Vue 的矢量风格头像生成器，核心卖点在于将头像拆解为可组合的","reasonCn":"除了精细手动调节，它还提供了一键随机生成和撤销/重做功能。当你不想费时慢调时，点击随机按钮就能瞬间出一张意想不到的怪脸组合；如果不满意，Undo/Redo 让你能轻松回退或重试，没有任何心理负担。这种“探索式”的交互比固定模板更接近“创作”的体验。","stars":3768,"starGrowth":0,"createdAt":"2021-10-16T08:38:11Z","silentRounds":4,"language":"Vue","topics":["vue3","vite","avatar","avatar-generator","avatar-maker"],"aiDims":["创意工具","非AI-好玩"],"zone":"创意","funScore":0.5599999999999999,"domainTags":["向量头像","Vue3","随机生成","批量生成"],"domainKey":"向量头像","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"vue3","source":"github","weight":0.7},{"name":"vite","source":"github","weight":0.7},{"name":"avatar","source":"github","weight":0.7},{"name":"avatar-generator","source":"github","weight":0.7},{"name":"avatar-maker","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/Codennnn/vue-color-avatar","ts":"2026-09-18T22:32:16.581Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tauri-apps/tauri","owner":"tauri-apps","name":"tauri","desc":"Build smaller, faster, and more secure desktop and mobile applications with a web frontend.","summaryCn":"用网页技术造原生App的轻量框架，包体积只有Electron的十分之一","reasonCn":"Tauri是一个基于Rust的跨平台桌面应用开发框架，允许开发者用HTML/CSS/JavaScript构建真正的原生应用。它不使用Chromium内核，而是调用系统自带的WebView，因此安装包体积极小，内存占用远低于Electron。支持Windows、macOS、Linux，并正在拓展移动端能力。","stars":111273,"starGrowth":27,"createdAt":"2019-07-13T09:09:37Z","silentRounds":0,"language":"Rust","topics":["desktop-app","high-performance","mobile-app","native-app","rust","web-frontend","webview"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["桌面应用","Rust","跨平台","前端集成"],"domainKey":"桌面应用","pushedAt":"2026-08-17T20:01:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop-app","source":"github","weight":0.7},{"name":"high-performance","source":"github","weight":0.7},{"name":"mobile-app","source":"github","weight":0.7},{"name":"native-app","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"web-frontend","source":"github","weight":0.7},{"name":"webview","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/tauri-apps/tauri","ts":"2026-08-17T20:01:17Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08172793939230541,"copyOk":false},{"repo":"homotopy-io/homotopy-rs","owner":"homotopy-io","name":"homotopy-rs","desc":"A Rust/WASM implementation of homotopy.io","summaryCn":"用Rust重写的在线协作白板，多人实时画框框聊思路","reasonCn":"技术层面，它用纯Rust编写核心逻辑，编译为WASM后在浏览器运行，省去了原版JavaScript的性能瓶颈。多人同步靠的是CRDT（无冲突复制数据类型）算法，保证分布式环境下数据一致性。形状渲染、光标追踪、实时通信全部在前端完成，不需要重型后端服务。","stars":122,"starGrowth":0,"createdAt":"2020-09-17T10:01:55Z","silentRounds":12,"language":"Rust","topics":["mathematics"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"工具","funScore":0.5,"domainTags":["实时协作白板","CRDT算法","Rust WASM","分布式同步","可视化编辑器"],"domainKey":"实时协作白板","pushedAt":"2026-09-20T02:53:03Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/homotopy-io/homotopy-rs","ts":"2026-09-20T02:53:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"coze-dev/coze-loop","owner":"coze-dev","name":"coze-loop","desc":"Next-generation AI Agent Optimization Platform: Cozeloop addresses challenges in AI agent development by providing full-lifecycle management capabilities from development, debugging, and evaluation to monitoring.  ","summaryCn":"你可能玩过很多AI Agent，写的时候感觉挺顺手，但一上线就发现它们经常“翻车","reasonCn":"技术核心在于其强大的“回环优化”架构，它不仅能记录Agent的每一次调用路径，还能通过一套严谨的指标体系对效果进行打分。你可以清楚地看到哪一步骤消耗了时间，哪一个工具调用拖了后腿。它通过集成多种评测方法，让AI自动帮你检查AI的表现，这种“AI监控AI”的方式极大地缩短了调优周","stars":5744,"starGrowth":1,"createdAt":"2025-06-24T00:26:28Z","silentRounds":2,"language":"Go","topics":["monitoring"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI Agent","模型优化","开发者工具"],"domainKey":"AI Agent","pushedAt":"2026-09-23T08:07:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/coze-dev/coze-loop","ts":"2026-09-23T08:07:22Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"IvanMurzak/Unity-MCP","owner":"IvanMurzak","name":"Unity-MCP","desc":"AI Skills, MCP Tools, and CLI for Unity Engine. Full AI develop and test loop. Use cli for quick setup. Efficient token usage, advanced tools. Any C# method may be turned into a tool by a single line. Works with Claude Code, Gemini, Copilot, Cursor and any other absolutely for free.","summaryCn":"让AI直接接管游戏引擎进行全自动开发与测试的Unity工具集","reasonCn":"技术亮点在于它优雅的架构设计和极低的接入成本。它通过C#实现了与AI大模型的深度打通，最惊艳的是开发者只需要在任意一个C#方法上面加一行简单的属性标记，这个方法就能瞬间变成AI可以安全调用的高级工具。这意味着你可以随心所欲地扩展AI的能力边界。","stars":4320,"starGrowth":1,"createdAt":"2025-04-02T09:57:01Z","silentRounds":1,"language":"C#","topics":["game-development"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":1,"domainTags":["游戏开发","Unity","AI辅助"],"domainKey":"游戏开发","pushedAt":"2026-09-23T20:45:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/IvanMurzak/Unity-MCP","ts":"2026-09-23T20:45:42Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AvaloniaUI/Avalonia","owner":"AvaloniaUI","name":"Avalonia","desc":"Develop Desktop, Embedded, Mobile and WebAssembly apps with C# and XAML. The future of .NET UI","summaryCn":"用一套代码搞定跨平台桌面软件的神器，C#选手必备","reasonCn":"基于.NET框架的跨平台UI工具包，支持Windows、Linux、macOS及WebAssembly，使用XAML声明式语法构建界面。相比Flutter依赖Dart，它让熟悉的.NET开发者无需学习新语言即可开发高性能UI，适合构建企业级桌面应用或跨端工具。","stars":31546,"starGrowth":9,"createdAt":"2013-12-05T20:13:21Z","silentRounds":0,"language":"C#","topics":["mobile","desktop"],"aiDims":["开发者工具","非AI-好玩","可视化"],"zone":"工具","funScore":0,"domainTags":["跨平台UI","C#","XAML"],"domainKey":"跨平台UI","pushedAt":"2026-09-23T18:31:17Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/AvaloniaUI/Avalonia","ts":"2026-09-23T18:31:17Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02785589298167755,"copyOk":true},{"repo":"speed47/spectre-meltdown-checker","owner":"speed47","name":"spectre-meltdown-checker","desc":"","summaryCn":"CPU 硬件漏洞如 Spectre 和 Meltdown 自 2018 年披露以来持续演变，衍生出 ","reasonCn":"该工具的亮点在于其全面性和可执行性。它不仅检测最初的 Spectre v1/v2 和 Meltdown，还覆盖了大量后续发现的 CVE，如 CVE-2018-3615 (Foreshadow) 和 CVE-2018-12126 (Fallout)。脚本会自动读取系统的内核参数和 CPU 标志位，判断特定的缓解措施（如 IBRS、STIBP）是否已启用。","stars":3952,"starGrowth":0,"createdAt":"2018-01-07T14:00:14Z","silentRounds":2,"language":"Shell","topics":["mitigation","meltdown","spectre","cve-2017-5753","cve-2017-5715","cve-2017-5754","cve-2018-3640","cve-2018-3639","foreshadow","cve-2018-3615","cve-2018-3620","cve-2018-3646","zombieload","cve-2019-11135","downfall","reptar"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["Spectre","Meltdown","漏洞检测","系统安全"],"domainKey":"Spectre","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mitigation","source":"github","weight":0.7},{"name":"meltdown","source":"github","weight":0.7},{"name":"spectre","source":"github","weight":0.7},{"name":"cve-2017-5753","source":"github","weight":0.7},{"name":"cve-2017-5715","source":"github","weight":0.7},{"name":"cve-2017-5754","source":"github","weight":0.7},{"name":"cve-2018-3640","source":"github","weight":0.7},{"name":"cve-2018-3639","source":"github","weight":0.7},{"name":"foreshadow","source":"github","weight":0.7},{"name":"cve-2018-3615","source":"github","weight":0.7},{"name":"cve-2018-3620","source":"github","weight":0.7},{"name":"cve-2018-3646","source":"github","weight":0.7},{"name":"zombieload","source":"github","weight":0.7},{"name":"cve-2019-11135","source":"github","weight":0.7},{"name":"downfall","source":"github","weight":0.7},{"name":"reptar","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/speed47/spectre-meltdown-checker","ts":"2026-09-18T20:05:54.376Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sngyai/Sequoia-X","owner":"sngyai","name":"Sequoia-X","desc":"A股自动选股系统 — 多种技术形态自动扫描，收盘后自动运行并推送飞书","summaryCn":"收盘后自动扫遍全A股，符合条件的股票直接推送到你的飞书","reasonCn":"基于Python开发的A股自动选股工具，支持MACD金叉死叉、均线排列、突破形态等多种技术扫描策略。收盘后自动执行全市场筛选，结果通过飞书机器人推送，内置多因子策略模板，可自定义条件组合，支持对接主流行情数据源，代码结构清晰易于扩展。","stars":7513,"starGrowth":42,"createdAt":"2018-07-20T16:33:48Z","silentRounds":0,"language":"Python","topics":["ta-lib","python","turtle-trade","a-shares","pandas","akshare","baostock","stock-screening","trading"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0.7,"domainTags":["A股","选股","技术分析","飞书"],"domainKey":"A股","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ta-lib","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"turtle-trade","source":"github","weight":0.7},{"name":"a-shares","source":"github","weight":0.7},{"name":"pandas","source":"github","weight":0.7},{"name":"akshare","source":"github","weight":0.7},{"name":"baostock","source":"github","weight":0.7},{"name":"stock-screening","source":"github","weight":0.7},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/sngyai/Sequoia-X","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.14277093038732863,"copyOk":false},{"repo":"lmnr-ai/lmnr","owner":"lmnr-ai","name":"lmnr","desc":"Laminar - open-source observability platform purpose-built for AI agents. YC S24.","summaryCn":"开发Agent时，最让人头疼的是一旦系统出BUG，你根本不知道它是哪一步逻辑出","reasonCn":"技术上，它利用了先进的追踪机制，能够记录Agent在多次递归、循环调用过程中的完整上下文。它不仅支持记录输入输出，还能追踪每一次调用背后的元数据，与现有的AI开发框架无缝集成。其独特的架构保证了在高频调用下的低延迟，不会成为你应用的性能瓶颈。","stars":3276,"starGrowth":1,"createdAt":"2024-08-29T03:45:28Z","silentRounds":2,"language":"TypeScript","topics":["monitoring"],"aiDims":["AI Agent","监控","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["可观测性","Agent监控","日志追踪","调试工具"],"domainKey":"可观测性","pushedAt":"2026-09-23T21:08:09Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"监控","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/lmnr-ai/lmnr","ts":"2026-09-23T21:08:09Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"commaai/openpilot","owner":"commaai","name":"openpilot","desc":"openpilot is an operating system for robotics. Currently, it upgrades the driver assistance system on 300+ supported cars.","summaryCn":"简单来说，这个项目就像是给汽车换上了一个开源的灵魂。","reasonCn":"技术架构上，它采用了类似端到端神经网络的设计思路，抛弃了传统复杂的规则堆砌，直接通过摄像头捕捉的画面和传感器数据输入模型，由深度学习算法直接输出方向盘转角和油门刹车指令。它支持实时监控驾驶员状态，对硬件算力和功耗控制做到了极致优化，在车载低功耗设备上也能流畅运","stars":63697,"starGrowth":1,"createdAt":"2016-11-24T01:33:30Z","silentRounds":3,"language":"Python","topics":["robotics"],"aiDims":["AI Agent","非AI-实用"],"zone":"AI","funScore":1,"domainTags":["自动驾驶","神经网络","汽车改装"],"domainKey":"自动驾驶","pushedAt":"2026-09-23T21:19:14Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/commaai/openpilot","ts":"2026-09-23T21:19:14Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"p1neappleXpress/OpenFlux","owner":"p1neappleXpress","name":"OpenFlux","desc":"Network stack research tool. TCP tunnel with pluggable transports.","summaryCn":"可插拔传输层的TCP隧道研究工具，支持zstd压缩与双出口模式。","reasonCn":"OpenFlux 是一款面向网络栈研究的 TCP 隧道工具，核心亮点在于可插拔传输层设计与分批 zstd 编解码，能在吞吐与延迟间取得平衡。项目提供两种出口后端：L3 层原始包转发适合极致性能场景，L4 层 gVisor 代理则兼容性更强、部署更便捷。完全非商业、无隐藏订阅，适合网络研究员、安全从业者或需要自定义隧道协议的开发者深度定制与实验。","stars":1803,"starGrowth":20,"createdAt":"2026-06-21T22:13:11Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["TCP隧道","网络协议","安全研究","VPN","插件框架"],"domainKey":"TCP隧道","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/p1neappleXpress/OpenFlux","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.11193344425956737,"copyOk":true},{"repo":"google/brax","owner":"google","name":"brax","desc":"Massively parallel rigidbody physics simulation on accelerator hardware.","summaryCn":"在加速器上超高速模拟刚体物理，让机器人仿真更真实","reasonCn":"BraX基于JAX实现了大规模并行刚体物理引擎，能够在GPU/TPU等加速硬件上一次性模拟上万物体碰撞、约束与动力学，提供Python接口并支持自动微分，便于在强化学习和机器人控制中直接嵌入物理梯度。同时支持多种边界条件和自定义材料属性，方便科研与工业仿真。","stars":3238,"starGrowth":1,"createdAt":"2021-06-02T21:45:19Z","silentRounds":5,"language":"Jupyter Notebook","topics":["robotics"],"aiDims":["非AI-实用"],"zone":"AI","funScore":0.7,"domainTags":["物理仿真","强化学习","JAX"],"domainKey":"物理仿真","pushedAt":"2026-09-15T09:02:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/google/brax","ts":"2026-09-15T09:02:32Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Ziforge/ableton-liveapi-tools","owner":"Ziforge","name":"ableton-liveapi-tools","desc":"Thread-safe Python Remote Script for Ableton Live with 220 LiveAPI tools","summaryCn":"第一段：很多音乐人和现场表演者想用代码控制 Ableton Live，却总被 Live 自","reasonCn":"第一段：很多音乐人和现场表演者想用代码控制 Ableton Live，却总被 Live 自带的 Python 脚本的线程限制卡住。Ableton Live API Tools 把这些限制全部搬走，提供一个线程安全的远程控制层，让你可以在外部 IDE、Jupyter Notebook 或自定义软件里直接操控 Live，像玩乐高一样堆叠音","stars":16,"starGrowth":0,"createdAt":"2025-10-20T01:09:59Z","silentRounds":9,"language":"Python","topics":["music-production"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["音乐制作软件","Python脚本","现场表演自动化","MIDI控制","音频API"],"domainKey":"音乐制作软件","pushedAt":"2026-09-09T05:44:37Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music-production","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/Ziforge/ableton-liveapi-tools","ts":"2026-09-09T05:44:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pbakaus/impeccable","owner":"pbakaus","name":"impeccable","desc":"The design language that makes your AI harness better at design.","summaryCn":"让AI应用长出高颜值，设计感直接拉满的神器","reasonCn":"这是一个专为AI应用打造的设计语言系统，提供完整的UI组件库和设计规范。包含按钮、表单、卡片、导航栏等常见元素的标准化设计，帮助开发者快速构建专业级AI界面。采用JavaScript实现，支持自定义主题，让AI应用不再千篇一律，而是拥有统一且精致的设计语言。","stars":70283,"starGrowth":287,"language":"JavaScript","topics":[],"aiDims":["AI应用","创意工具"],"zone":"工具","funScore":0,"domainTags":["UI组件库","设计系统","前端开发"],"domainKey":"UI组件库","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/pbakaus/impeccable","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3122504730873753,"copyOk":false},{"repo":"netease-youdao/LobsterAI","owner":"netease-youdao","name":"LobsterAI","desc":"Open-source, desktop-grade AI agent that gets real work done — data analysis, slides, docs, video & web research. Built on OpenClaw; runs tools on your real desktop and takes commands from your phone via WeChat, Feishu, DingTalk & Telegram.","summaryCn":"你的电脑就是AI代理的工位，远程下命令它帮你搞定一切","reasonCn":"LobsterAI是网易有道开源的桌面级AI智能体，基于OpenClaw构建，能在你的真实电脑上执行数据分析、生成PPT和文档、视频制作以及网页研究等任务。它不只是聊天机器人，而是能真正操作电脑、调用工具的自主Agent，还能通过手机远程发指令控制，实现了从PC到移动端的无缝协作。","stars":6064,"starGrowth":3,"createdAt":"2026-02-12T09:56:20Z","silentRounds":0,"language":"TypeScript","topics":["agent","ai-agent","ai-assistant","automation","autonomous-agents","chatbot","copilot","cross-platform","desktop-app","electron","llm","mcp","office-automation","openclaw","productivity","react","tyepscript"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["AI代理","桌面自动化","多设备协同","远程指令","任务执行"],"domainKey":"AI代理","pushedAt":"2026-09-23T14:20:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai-assistant","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"autonomous-agents","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"desktop-app","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"office-automation","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"tyepscript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"bigbro","url":"https://github.com/netease-youdao/LobsterAI","ts":"2026-09-23T14:20:46Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"OS-1337/OS1337","owner":"OS-1337","name":"OS1337","desc":"OS/1337 Project ","summaryCn":"复古黑客风格的操作系统内核，玩代码的乐趣","reasonCn":"在技术实现上，它采用了极简的架构设计，完全用C语言编写，没有复杂的依赖。它模拟了早期的DOS环境，包括命令行解释器（Shell）、简单的文件系统以及内存管理。这种复古风格不仅体现在界面和交互上，更体现在代码风格上，充满了那个时代程序员们的硬核美学。","stars":32,"starGrowth":0,"createdAt":"2023-08-24T04:22:52Z","silentRounds":9,"language":"Shell","topics":["retrocomputing"],"aiDims":["非AI-好玩","创意工具","retrocomputing"],"zone":"创意","funScore":1,"domainTags":["操作系统内核","复古计算","C语言","命令行模拟器"],"domainKey":"操作系统内核","pushedAt":"2026-09-09T20:12:43Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"retrocomputing","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/OS-1337/OS1337","ts":"2026-09-09T20:12:43Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mouredev/python-web","owner":"mouredev","name":"python-web","desc":"Curso para aprender desarrollo frontend Web con Python puro desde cero. Elaborado durante las emisiones en directo desde Twitch de MoureDev.","summaryCn":"纯Python也能做前端！从零到精通Web开发的完整实战教程","reasonCn":"这是MoureDev直播期间整理的Python前端开发教程，从HTML、CSS到JavaScript全覆盖。用PyScript等新技术实现只用Python写前端。GitHub 4700多星标，教程结构完整，适合Python开发者转型前端或想尝试无JS开发的初学者。","stars":4765,"starGrowth":1,"createdAt":"2023-10-03T11:35:14Z","silentRounds":3,"language":"HTML","topics":["frontend","tutorial"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["Python前端","Web教程","PyScript"],"domainKey":"Python前端","pushedAt":"2026-09-15T06:55:16Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"tutorial","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/mouredev/python-web","ts":"2026-09-15T06:55:16Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xms0g/rtow-dos","owner":"xms0g","name":"rtow-dos","desc":"Ray Tracing in One Weekend for MS-DOS","summaryCn":"用C语言在MS-DOS上完整实现射线追踪，复古视觉大作","reasonCn":"· 基本光照模型：漫反射、镜面反射与雾化；\n· 支持纹理映射与法线映射，让物体表面细节更逼真；\n· 交互式视角切换，支持键盘控制相机；\n· 输出BMP或直接在屏幕上绘制像素，兼容DOSBox或真实硬件。 核心技术：\n- 采用CPU单线程实现光线与三角形的相交测试，使用BVH加速；\n- 利用整数","stars":17,"starGrowth":0,"createdAt":"2025-02-19T13:29:12Z","silentRounds":9,"language":"C","topics":["retrocomputing"],"aiDims":["非AI-好玩","可视化"],"zone":"创意","funScore":1,"domainTags":["光线追踪","MS-DOS","复古计算","图形学演示","BVH加速"],"domainKey":"光线追踪","pushedAt":"2026-09-09T02:55:32Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/xms0g/rtow-dos","ts":"2026-09-09T02:55:32Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TencentCloud/CubeSandbox","owner":"TencentCloud","name":"CubeSandbox","desc":"Instant, Concurrent, Secure & Lightweight Sandbox for AI Agents.","summaryCn":"让AI Agent安全跑代码的沙箱，支持并发隔离，腾讯开源","reasonCn":"CubeSandbox是腾讯云开源的轻量级沙箱，专为AI Agent设计，支持高并发、隔离执行、资源限制。基于gVisor实现容器级安全隔离，自带代码执行环境，让Agent可以安全地写代码、跑脚本、调用工具而不担心泄露或破坏宿主环境。","stars":12658,"starGrowth":9,"createdAt":"2026-04-10T03:00:31Z","silentRounds":0,"language":"Go","topics":["agents"],"aiDims":["AI Agent","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI沙箱","Agent安全","容器隔离","gVisor"],"domainKey":"AI沙箱","pushedAt":"2026-09-23T13:59:02Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/TencentCloud/CubeSandbox","ts":"2026-09-23T13:59:02Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03495964607362932,"copyOk":true},{"repo":"janhq/jan","owner":"janhq","name":"jan","desc":"Jan is an open source alternative to ChatGPT that runs 100% offline on your computer.","summaryCn":"笔记本离线版ChatGPT，隐私安全又好玩的本机AI助手","reasonCn":"Jan是一款开源的本地大模型运行客户端，主打完全离线、隐私安全。它让用户无需联网、无需付费订阅，就能在自己的电脑上流畅运行Llama、ChatGLM等主流大模型。界面简洁直观，像安装普通软件一样简单，真正实现了“把AI装进电脑里”，是隐私敏感用户和AI爱好者的福音。","stars":44605,"starGrowth":13,"createdAt":"2023-08-17T02:17:10Z","silentRounds":0,"language":"TypeScript","topics":["llm"],"aiDims":["AI应用","大语言模型","效率工具"],"zone":"AI","funScore":0,"domainTags":["本地大模型","离线运行","LLM","桌面应用"],"domainKey":"本地大模型","pushedAt":"2026-09-23T07:14:54Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/janhq/jan","ts":"2026-09-23T07:14:54Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03987434144154243,"copyOk":false},{"repo":"carpentries-incubator/SDC-BIDS-fMRI","owner":"carpentries-incubator","name":"SDC-BIDS-fMRI","desc":"fMRI Imaging Analysis","summaryCn":"基于Jupyter的fMRI分析工作流，快速上手神经影像处理","reasonCn":"功能性磁共振成像（fMRI）数据处理往往涉及繁杂的预处理步骤、统计模型构建以及结果可视化，传统上需要熟悉多个命令行工具。SDC-BIDS-fMRI 项目把这些步骤封装在 Jupyter Notebook 中，遵循 BIDS（Brain Imaging Data Structure）标准，让研究者只需准备符合规范的数据文件，即可在","stars":49,"starGrowth":0,"createdAt":"2019-03-14T14:37:45Z","silentRounds":11,"language":"Jupyter Notebook","topics":["english"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["fMRI","神经影像","生物信息学","Jupyter","数据分析"],"domainKey":"fMRI","pushedAt":"2026-09-22T02:59:09Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"english","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/carpentries-incubator/SDC-BIDS-fMRI","ts":"2026-09-22T02:59:09Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"P5-wrapper/next","owner":"P5-wrapper","name":"next","desc":"A NextJS specific library for the @P5-Wrapper/react project.","summaryCn":"Next.js里做创意编程的神器，canvas艺术轻松实现","reasonCn":"这个项目把著名的创意编程库p5.js封装成了Next.js专用的React组件，让你不用手写复杂Canvas代码就能做交互式动画和生成艺术。核心设计是提供声明式API，在Next.js的SSR环境中也能安全地渲染创意效果，解决了原生p5.js在框架中兼容性差的问题。","stars":53,"starGrowth":0,"createdAt":"2023-04-19T11:47:41Z","silentRounds":9,"language":"TypeScript","topics":["creative-coding"],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":0.8,"domainTags":["创意编程","Next.js","Canvas动画","p5.js封装"],"domainKey":"创意编程","pushedAt":"2026-09-08T23:20:36Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/P5-wrapper/next","ts":"2026-09-08T23:20:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"denoland/celld","owner":"denoland","name":"celld","desc":"self-hosted, distributed Durable Objects","summaryCn":"在本地搭建自己的分布式状态管理器，告别云端绑定的自由度","reasonCn":"deno团队的开源实现，将Cloudflare Durable Objects的核心能力本地化。基于Rust开发，支持多种存储后端如SQLite、PostgreSQL，可灵活部署在任何环境。提供了完整的状态隔离、跨实例通信和持久化机制，适合构建需要状态感知的分布式应用。","stars":4776,"starGrowth":13,"createdAt":"2025-04-25T12:27:48Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Durable Objects","分布式状态管理","Rust","SQLite","PostgreSQL"],"domainKey":"Durable Objects","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/denoland/celld","ts":"2026-09-18T10:23:42.801Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.047165829145728644,"copyOk":false},{"repo":"tsinghua-fib-lab/AgentSociety","owner":"tsinghua-fib-lab","name":"AgentSociety","desc":"AgentSociety 2 is a modern, LLM-native agent simulation platform designed for social science research and experimental design. It provides a flexible framework for creating and managing intelligent agents in simulated environments.","summaryCn":"模拟AI社会运行的沙盒平台，仿佛在看科幻电影一样","reasonCn":"这个项目的核心魅力在于其“社会模拟”的属性。它不仅仅是让AI聊天，而是构建了一个包含经济、政治、法律等多种维度的复杂环境。每个智能体都有独立的动机和决策逻辑，它们之间的交互会产生涌现行为，有时甚至会让研究者感到惊讶。","stars":1301,"starGrowth":1,"createdAt":"2025-02-06T05:09:18Z","silentRounds":1,"language":"Python","topics":["simulation"],"aiDims":["AI Agent","AI基础设施","非AI-好玩"],"zone":"AI","funScore":0.8,"domainTags":["AI 社会模拟","实验设计","群体智能","社会科学","Agent 交互"],"domainKey":"AI 社会模拟","pushedAt":"2026-09-23T13:45:12Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/tsinghua-fib-lab/AgentSociety","ts":"2026-09-23T13:45:12Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rowboatlabs/rowboat","owner":"rowboatlabs","name":"rowboat","desc":"AI coworker with memory and collaboration","summaryCn":"带记忆的AI同事，帮你干活还记着你干过啥。","reasonCn":"它的核心在于“记忆”。Rowboat 内置了向量数据库和持久化存储机制，能把对话中的关键信息提取出来并保存。当你下次问它问题时，它会检索之前的记忆，给出连贯的回答。此外，它还支持工具调用，这意味着它不仅能聊天，还能帮你执行具体的任务，比如查邮件、写代码片段或者生成文档。","stars":17951,"starGrowth":11,"createdAt":"2025-01-13T09:30:40Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI助手","向量数据库","记忆系统","生产力工具"],"domainKey":"向量数据库","pushedAt":"2026-09-23T10:45:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/rowboatlabs/rowboat","ts":"2026-09-23T10:45:40Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03483833769706423,"copyOk":false},{"repo":"tidwall/jj","owner":"tidwall","name":"jj","desc":"","summaryCn":"jj 工具的核心设计哲学是“极速”，它底层由 GJSON 和 SJSON 两个高性能库驱动","reasonCn":"上手使用 jj 就像在终端里做手术一样精准。用户通过简单的路径语法即可提取字符串、获取 JSON 块或按索引访问数组元素，甚至能直接修改文档中的特定值。例如，查询不存在的键会返回空值而不是报错，获取原始字符串值也有专门的参数，这种细粒度的控制让数据处理变得异常灵活，无需","stars":2059,"starGrowth":0,"silentRounds":0,"language":"Go","topics":[],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.32,"domainTags":["JSON编辑","命令行工具","Go语言","数据提取"],"domainKey":"JSON编辑","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/tidwall/jj","ts":"2026-09-23T18:23:13.861Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jcubic/ascii-globe","owner":"jcubic","name":"ascii-globe","desc":"Zero dependencies isomorphic ASCII-Art globe renderer in JavaScript","summaryCn":"在终端里转动的ASCII球，零依赖随手玩神器","reasonCn":"ascii-globe 用纯 TypeScript 实现跨平台的 ASCII 地球仪渲染器，完全不依赖任何第三方库，既能在 Node 环境运行，也可直接在浏览器通过 ES 模块使用。它利用字符矩阵和经纬度投影算法，把三维球面转换为二维字符画，实现旋转、缩放并自动适配终端宽高。开发者只需调用 render 函数并传入容器或 stdout，即可看到实时旋转的地球，代码行数不足百行，适合作为学习同构渲染和数学投影的案例。","stars":10,"starGrowth":0,"createdAt":"2026-05-01T12:55:50Z","silentRounds":8,"language":"TypeScript","topics":["ascii-art"],"aiDims":["非AI-好玩","可视化"],"zone":"工具","funScore":1,"domainTags":["ASCII艺术","地球仪","零依赖库","跨平台渲染","终端可视化"],"domainKey":"ASCII艺术","pushedAt":"2026-09-10T23:17:42Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ascii-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.52,"source":"search","url":"https://github.com/jcubic/ascii-globe","ts":"2026-09-10T23:17:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"oxc-project/oxc","owner":"oxc-project","name":"oxc","desc":"⚓ A collection of high-performance JavaScript tools.","summaryCn":"用Rust重写的JS工具全家桶，速度吊打老牌选手","reasonCn":"Oxc是用Rust从零编写的高性能JavaScript和TypeScript工具链，涵盖解析器、Linter、压缩器、转换器等多项核心能力，号称要全面超越ESLint、Babel等传统工具，以极快的速度大幅降低大型项目的编译等待时间，让前端开发流程焕然一新。","stars":22851,"starGrowth":10,"createdAt":"2023-02-09T05:46:51Z","silentRounds":0,"language":"Rust","topics":["compiler","javascript","linter","minifier","parser","typescript","transpiler"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["JavaScript工具链","Linter","编译器","Rust"],"domainKey":"JavaScript工具链","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"compiler","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"linter","source":"github","weight":0.7},{"name":"minifier","source":"github","weight":0.7},{"name":"parser","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"transpiler","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/oxc-project/oxc","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.031312852829197844,"copyOk":true},{"repo":"zenstory-ai/oh-story-claudecode","owner":"zenstory-ai","name":"oh-story-claudecode","desc":"网文/小说写作 skill 包，覆盖长篇与短篇网络小说的扫榜、拆文、写作、去AI味、封面图全流程 | An all-in-one skill pack for long- and short-form web fiction.","summaryCn":"一款网文小说写作工具，覆盖长篇与短篇网络小说的扫榜、拆文、写作全流程","reasonCn":"zenstory-ai是一个专注于网文小说写作的工具，它通过提供一站式的服务，帮助作家快速完成创作任务，提高写作效率。","stars":7057,"starGrowth":9,"createdAt":"2026-04-22T16:35:39Z","silentRounds":0,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","AI写作"],"zone":"AI","funScore":0,"domainTags":["网文写作","AI辅助","写作模板"],"domainKey":"网文写作","pushedAt":"2026-09-23T08:31:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/zenstory-ai/oh-story-claudecode","ts":"2026-09-23T08:31:20Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0369911860563979,"copyOk":false},{"repo":"iOfficeAI/OfficeCLI","owner":"iOfficeAI","name":"OfficeCLI","desc":"OfficeCLI is the first and best Office suite  purpose-built for AI agents to read, edit, and automate Word, Excel, and PowerPoint files. Free, open-source, single binary, no Office installation required.","summaryCn":"大家日常办公最头疼的就是处理成堆的Excel报表或改不完的PPT格式，如果想用","reasonCn":"核心优势在于其轻量化的单二进制部署模式，不需要依赖庞大的图形界面环境，在服务器端运行极为稳健。它将文档的复杂格式进行了深度解耦，支持高精度的读写操作，能够保持文档结构的完整性。相比于调用传统Office API，它的响应速度更快，且因为支持CLI模式，能非常方便地集成到现有","stars":31056,"starGrowth":47,"createdAt":"2026-03-15T07:39:12Z","silentRounds":0,"language":"C#","topics":["ai"],"aiDims":["AI应用","文档理解","效率工具"],"zone":"AI","funScore":0.5,"domainTags":["Office","文档处理","CLI","自动化","AI Agent"],"domainKey":"AI Agent","pushedAt":"2026-09-21T20:19:02Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/iOfficeAI/OfficeCLI","ts":"2026-09-21T20:19:02Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.17464822256568777,"copyOk":false},{"repo":"mapeditor/tiled","owner":"mapeditor","name":"tiled","desc":"Flexible level editor","summaryCn":"做游戏关卡的瑞士军刀，拖拽就能搭出奇妙世界","reasonCn":"Tiled 是专业的 2D 地图编辑器，支持多层图块、对象层和碰撞区域设置。采用 XML 格式存储，插件系统可扩展，能无缝集成到 Unity、Godot 等主流游戏引擎。它支持等距网格、瓦片动画和脚本触发器，可通过 Qt 插件接口添加自定义功能。输出格式包括 JSON、CSV 和 XML，方便程序读取。","stars":12904,"starGrowth":1,"createdAt":"2011-02-27T15:18:22Z","silentRounds":2,"language":"C++","topics":["game"],"aiDims":["游戏","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["2D地图编辑","游戏关卡设计","图块编辑"],"domainKey":"2D地图编辑","pushedAt":"2026-09-21T20:30:04Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/mapeditor/tiled","ts":"2026-09-21T20:30:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sonosaurus/sonobus","owner":"sonosaurus","name":"sonobus","desc":"Source code for SonoBus, a real-time network audio streaming collaboration tool.","summaryCn":"把麦克风接到网上，让朋友隔空和你同步演奏的小工具","reasonCn":"它用底层音频引擎搭建实时网络房间，把本地麦克风、系统声音和浏览器播放端连成一条低延迟总线。支持多人同场、分路混音、广播频道与独立收听，能把网络排练变成像面对面一样的协作。对做即兴音乐、直播演示、远程合奏的人很顺手，也方便给小团队快速搭起小音频桥。","stars":2191,"starGrowth":0,"createdAt":"2020-07-30T04:17:14Z","silentRounds":2,"language":"C++","topics":["music"],"aiDims":["创意工具","非AI-好玩","非AI-实用"],"zone":"工具","funScore":0.7,"domainTags":["网络音频","实时协作","音乐工具"],"domainKey":"网络音频","pushedAt":"2026-09-04T21:37:12Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.58,"source":"search","url":"https://github.com/sonosaurus/sonobus","ts":"2026-09-04T21:37:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ppy/osu-web","owner":"ppy","name":"osu-web","desc":"the browser-facing portion of osu!","summaryCn":"浏览器里直接玩osu!节奏音游，PHP开发的Web版本，免安装。","reasonCn":"这是osu!音乐节奏游戏的Web端开源实现，采用PHP后端框架，支持用户注册、歌曲下载、排行榜系统、多人对战等完整功能。项目维护了游戏的在线平台部分，包括社交系统、好友列表、即时消息等功能，让玩家不安装客户端直接通过浏览器参与游戏。","stars":1156,"starGrowth":1,"createdAt":"2015-08-26T07:14:32Z","silentRounds":3,"language":"PHP","topics":["backend","frontend"],"aiDims":["游戏"],"zone":"工具","funScore":0,"domainTags":["音乐游戏后端","Web游戏平台","Laravel框架","实时排行榜","多人匹配系统"],"domainKey":"音乐游戏后端","pushedAt":"2026-09-19T15:43:20Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ppy/osu-web","ts":"2026-09-19T15:43:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lhh010/dsh-ui-whale","owner":"lhh010","name":"dsh-ui-whale","desc":"【求⭐】🐋DSH Web UI 全手绘像素鲸鱼伙伴插件：会话标题栏常驻，平时眨眼/偶尔摆尾/动胸鳍，思考运行时持续动起来，回合完成头顶喷水，点击还会冒爱心，不工作时还会偷懒睡觉，零核心改动。 【喜欢的话就点点star⭐吧~】","summaryCn":"在网页会话标题栏常驻一只会眨眼、摆尾还会偷懒睡觉的全手绘像素鲸鱼。","reasonCn":"技术实现上，它采用TypeScript编写，主打一个零核心代码修改的非侵入式插件架构。开发者可以通过简单的样式注入和组件挂载，把这只小鲸鱼安放在各种Web UI的标题栏中。它通过精妙的帧动画和状态监听，能够和系统的运行状态实时联动——AI在思考时它会兴奋地动起来，回合一完成它就会","stars":33,"starGrowth":0,"createdAt":"2026-08-13T13:10:16Z","silentRounds":6,"language":"TypeScript","topics":["pixel-art"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["像素艺术","UI插件","动画交互","Web增强"],"domainKey":"像素艺术","pushedAt":"2026-09-23T16:48:44Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/lhh010/dsh-ui-whale","ts":"2026-09-23T16:48:44Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"awesome-selfhosted/awesome-selfhosted","owner":"awesome-selfhosted","name":"awesome-selfhosted","desc":"A list of Free Software network services and web applications which can be hosted on your own servers","summaryCn":"自托管神器合集，把自己的服务器变成私人云服务仓库","reasonCn":"这个项目收集了大量可以自托管的免费软件，包括网盘、日记、笔记、监控、自动化等实用工具。你只需一台服务器就能搭建私人云盘、个人博客、家庭NAS等，完全掌控数据，star超过31万，是DIY爱好者和隐私保护者的宝藏清单。","stars":321064,"starGrowth":282,"createdAt":"2015-06-01T02:33:17Z","silentRounds":0,"language":"","topics":["self-hosted","awesome","privacy"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"资源","funScore":0,"domainTags":["自托管","开源软件","隐私保护","列表导航","DIY"],"domainKey":"自托管","pushedAt":"2026-09-23T17:59:29Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7}],"aiScore":0.65,"source":"trending","url":"https://github.com/awesome-selfhosted/awesome-selfhosted","ts":"2026-09-23T17:59:29Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3026349886626965,"copyOk":true},{"repo":"p1ngul1n0/blackbird","owner":"p1ngul1n0","name":"blackbird","desc":"An OSINT tool to search for accounts by username and email in social networks.","summaryCn":"查用户名和邮箱在哪些社交平台存在过的侦查利器","reasonCn":"一款强大的OSINT开源情报工具，通过用户名或邮箱地址批量查询几十个主流社交平台的存在情况，包括GitHub、Twitter、Instagram、Reddit等。结果可直接导出为HTML报告，适合安全研究人员、渗透测试人员和调查记者快速获取公开信息。","stars":8525,"starGrowth":14,"createdAt":"2022-05-06T14:25:44Z","silentRounds":0,"language":"Python","topics":["osint","pentesting","cybersecurity","python"],"aiDims":["安全工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["社交媒体账号检测","OSINT","渗透测试","批量查询"],"domainKey":"社交媒体账号检测","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"osint","source":"github","weight":0.7},{"name":"pentesting","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.58,"source":"trending","url":"https://github.com/p1ngul1n0/blackbird","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0469266862170088,"copyOk":true},{"repo":"vim-test/vim-test","owner":"vim-test","name":"vim-test","desc":"Run your tests at the speed of thought","summaryCn":"在 Vim 里一键运行单元测试，省去切窗切命令的麻烦。","reasonCn":"vim-test 为 Vim/Neovim 提供统一的测试键，自动识别 Jest、Mocha、RSpec、pytest 等框架，执行时在分割窗或浮动窗即时显示通过或失败信息，支持并行和自定义命令，让开发者写代码时无需离开编辑器即可完成完整的测试循环，提升调试效率。","stars":3165,"starGrowth":0,"createdAt":"2014-12-15T17:34:43Z","silentRounds":8,"language":"Vim Script","topics":["testing"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Vim插件","单元测试","测试框架集成","开发者工具","Neovim"],"domainKey":"Vim插件","pushedAt":"2026-09-16T09:50:02Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"Vim Script","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/vim-test/vim-test","ts":"2026-09-16T09:50:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"RobLoach/raylib-aseprite","owner":"RobLoach","name":"raylib-aseprite","desc":"Load Aseprite files for animated sprites in raylib.","summaryCn":"在 2D 游戏开发中，动画素材的处理往往是最繁琐的部分之一。","reasonCn":"这个插件的核心功能就是“桥接”。它让 raylib 能够理解 Aseprite 生成的 .ase 和 .aseprite 文件格式。以前，开发者需要手动将 Aseprite 的动画切片导出为单独的图片文件，然后在代码里加载。现在，有了这个插件，你只需要在代码中加载一个文件，它就会自动解析出所有的帧、图层和调","stars":128,"starGrowth":0,"createdAt":"2021-04-11T05:04:21Z","silentRounds":7,"language":"C","topics":["raylib"],"aiDims":["创意工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["raylib","像素艺术","游戏开发","动画 sprites","C语言"],"domainKey":"raylib","pushedAt":"2026-09-11T20:52:04Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"raylib","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/RobLoach/raylib-aseprite","ts":"2026-09-11T20:52:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"titanwings/distilly","owner":"titanwings","name":"distilly","desc":"Distilly — Distill how they think into reusable Skills for any Agent or Bot. Formerly Colleague Skill（原同事 Skill）.","summaryCn":"把大模型变成懂业务的超级助手，一键部署到你的机器人里","reasonCn":"Distilly 能够把复杂的大模型推理过程拆解成可复用的技能包，支持多种 Agent 框架和 LLM 提供商。它通过管道化处理，让 Agent 可以像调用工具一样调用这些技能，显著提升复杂任务的执行效率。目前支持 TypeScript 开发，生态兼容性好。","stars":24970,"starGrowth":13,"createdAt":"2026-03-30T05:22:23Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI技能","Agent框架","TypeScript","提示工程","自动化工作流"],"domainKey":"提示工程","pushedAt":"2026-09-22T14:01:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/titanwings/distilly","ts":"2026-09-22T14:01:08Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0486742490989187,"copyOk":false},{"repo":"jenkinsci/jenkins","owner":"jenkinsci","name":"jenkins","desc":"Jenkins automation server","summaryCn":"老牌CI/CD自动化服务器，软件开发者天天用它自动构建部署代码","reasonCn":"技术上基于Java编写，采用插件式架构，拥有超过1800个插件可覆盖各种场景。支持从Git拉取代码后自动触发构建流程，能集成测试工具、代码质量检查、容器构建等各个环节。Pipeline功能用Groovy脚本定义，灵活度极高，支持复杂的多阶段流程编排。","stars":26578,"starGrowth":3,"createdAt":"2010-11-22T21:21:23Z","silentRounds":0,"language":"Java","topics":["devops"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["CI/CD","持续集成","自动化构建","DevOps"],"domainKey":"CI/CD","pushedAt":"2026-09-23T20:21:37Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/jenkinsci/jenkins","ts":"2026-09-23T20:21:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Gitlawb/openclaude","owner":"Gitlawb","name":"openclaude","desc":"runs anywhere. uses anything","summaryCn":"能让AI在所有平台上跑，还能调用各种工具的万能助手","reasonCn":"基于TypeScript构建的跨平台AI Agent框架，支持接入多种大语言模型后端。内置强大的工具调用系统，可自动调用API、操作文件、执行命令等。采用插件化架构设计，支持本地和云端灵活部署，实现复杂任务的多步骤自动化执行。","stars":33487,"starGrowth":15,"createdAt":"2026-04-01T03:15:49Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","ai"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.8,"domainTags":["AI Agent","TypeScript","自动化","工具调用"],"domainKey":"AI Agent","pushedAt":"2026-09-22T03:54:15Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Gitlawb/openclaude","ts":"2026-09-22T03:54:15Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.055612566070415384,"copyOk":false},{"repo":"palantir/blueprint","owner":"palantir","name":"blueprint","desc":"A React-based UI toolkit for the web","summaryCn":"Palantir开源的React UI组件库，大厂级设计美学","reasonCn":"Palantir开源的企业级React UI组件库，包含100多个精心设计的组件，涵盖表单、表格、导航等常用场景。采用TypeScript编写，风格与Palantir内部产品一致，提供完整的design token系统和主题定制能力，适合构建专业级后台管理系统。","stars":22085,"starGrowth":2,"createdAt":"2016-10-25T21:17:50Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["UI组件库","React","TypeScript"],"domainKey":"UI组件库","pushedAt":"2026-09-23T21:23:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/palantir/blueprint","ts":"2026-09-23T21:23:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/Web-Dev-For-Beginners","owner":"microsoft","name":"Web-Dev-For-Beginners","desc":"24 Lessons, 12 Weeks, Get Started as a Web Developer","summaryCn":"微软官方的Web开发入门课，12周带你从零成为前端开发者","reasonCn":"这是微软官方出品的Web开发入门课程，包含24个精心设计的课时，覆盖HTML、CSS、JavaScript等核心技术。课程设计遵循学习曲线，每节课配有练习和项目实践，让零基础学习者也能系统掌握前端技能。虽然不直接涉及AI，但掌握Web开发是构建AI应用的前置能力，对于想开发AI Web项目的读者来说是很好的基础储备。","stars":96728,"starGrowth":2,"createdAt":"2020-11-10T02:44:00Z","silentRounds":0,"language":"JavaScript","topics":["learning"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["Web开发","前端","教程"],"domainKey":"Web开发","pushedAt":"2026-09-19T06:14:01Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/microsoft/Web-Dev-For-Beginners","ts":"2026-09-19T06:14:01Z","category":"learning","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"modular/modular","owner":"modular","name":"modular","desc":"The Modular Platform (includes MAX & Mojo)","summaryCn":"一站式AI平台，从训练到部署全覆盖，让大模型开发如虎添翼。","reasonCn":"Modular 平台集成了 MAX 推理引擎和 Mojo 编程语言，提供从模型训练、优化到部署的完整工具链。支持主流框架（PyTorch、TensorFlow）和硬件（GPU、NPU、边缘设备），性能优化接近手写 CUDA。适合追求极致性能的 AI 工程师、研究者和生产环境部署团队。","stars":29855,"starGrowth":8,"createdAt":"2023-04-28T22:17:24Z","silentRounds":0,"language":"Mojo","topics":["ml","ai"],"aiDims":["AI基础设施","模型训练","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["AI推理","模型部署","Mojo语言","高性能计算"],"domainKey":"模型部署","pushedAt":"2026-09-23T06:42:57Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Mojo","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/modular/modular","ts":"2026-09-23T06:42:57Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0248038854463239,"copyOk":false},{"repo":"flyteorg/flyte","owner":"flyteorg","name":"flyte","desc":"Dynamic, resilient AI orchestration. Coordinate data, models, and compute as you build AI workflows.","summaryCn":"想象一下你正在研发一个复杂的AI模型，这需要先从数据库抓取数据，清洗成千","reasonCn":"技术上，它建立在Kubernetes之上，采用声明式架构，将计算逻辑与底层基础设施完全解耦。其核心优势在于强类型系统，确保了数据在不同步骤间流转时的准确性，同时支持容器化执行，能够轻松调度GPU集群。与Airflow相比，它更懂机器学习的特殊需求，如自动化的缓存机制和超强的资源管","stars":7552,"starGrowth":5,"createdAt":"2019-10-21T17:40:04Z","silentRounds":0,"language":"Go","topics":["data-science"],"aiDims":["AI基础设施","模型与训练"],"zone":"工具","funScore":0,"domainTags":["机器学习流水线","工作流编排","GPU调度","Kubernetes","Python SDK"],"domainKey":"机器学习流水线","pushedAt":"2026-09-23T20:15:04Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/flyteorg/flyte","ts":"2026-09-23T20:15:04Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.016986228813559323,"copyOk":false},{"repo":"PostHog/posthog.com","owner":"PostHog","name":"posthog.com","desc":"Official docs, website, and handbook for PostHog.","summaryCn":"PostHog产品分析平台的官方文档和资源网站","reasonCn":"PostHog是开源的产品分析平台，这个仓库包含其官方文档、网站和产品手册。PostHog提供类似Mixpanel和Amplitude的产品分析功能，但核心优势是开源可自部署，数据完全自主可控。支持事件追踪、用户分群、A/B测试和产品反馈收集，无需担心数据外泄，适合对数据隐私敏感的企业团队。","stars":1143,"starGrowth":1,"createdAt":"2020-05-01T20:17:52Z","silentRounds":24,"language":"TypeScript","topics":["design"],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["产品分析","自部署","事件追踪","开发者文档"],"domainKey":"产品分析","pushedAt":"2026-09-23T21:41:29Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/PostHog/posthog.com","ts":"2026-09-23T21:41:29Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LmeSzinc/StarRailCopilot","owner":"LmeSzinc","name":"StarRailCopilot","desc":"","summaryCn":"《崩坏：星穹铁道》游戏自动化脚本，自动完成日常任务和副本。","reasonCn":"这个脚本的核心亮点在于它覆盖了游戏的主要日常玩法。打本功能可以自动规划角色养成路线并完成副本；收获功能能自动完成每日任务、收派委托和收取奖励；模拟宇宙功能支持刷取遗器和开拓力；后台托管功能可以自动启动模拟器并在后台完成清体力操作，通过仪表盘了解资源情况。","stars":4196,"starGrowth":1,"createdAt":"2023-05-16T15:53:36Z","silentRounds":2,"language":"Python","topics":["honkai-star-rail"],"aiDims":["工具调用"],"zone":"工具","funScore":0.38,"domainTags":["游戏自动化","模拟器","脚本工具"],"domainKey":"游戏自动化","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"honkai-star-rail","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/LmeSzinc/StarRailCopilot","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ttezel/twit","owner":"ttezel","name":"twit","desc":"","summaryCn":"该库的一个关键细节在于其对认证上下文的严格区分。","reasonCn":"在功能覆盖上，twit 同时支持 REST 和 Streaming 两种 API 模式。REST 适合用于查询特定用户、关键词或时间范围内的静态数据，而 Streaming 则允许程序保持长连接，实时接收符合条件的海量数据流。这种双模支持使得开发者可以根据业务场景，选择在历史数据分析和实时监控之间灵活切","stars":4280,"starGrowth":0,"createdAt":"2011-11-30T13:15:31Z","silentRounds":4,"language":"JavaScript","topics":[],"aiDims":["学习资源","开发者工具"],"zone":"资源","funScore":0.2,"domainTags":["Twitter API","Node.js","REST接口","流式数据","应用认证"],"domainKey":"Twitter API","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/ttezel/twit","ts":"2026-09-18T00:01:25.926Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MadsLorentzen/ai-job-search","owner":"MadsLorentzen","name":"ai-job-search","desc":"The job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.","summaryCn":"让AI帮你自动投简历写求职信，找工作变成全自动流水线","reasonCn":"这是一个基于Claude Code构建的AI求职框架，能自动评估职位匹配度、定制简历、写求职信，还能帮你准备面试。全程在你的本地机器上运行，不用把个人信息上传到第三方平台，数据安全有保障，让找工作这件事变得自动化又高效。","stars":43673,"starGrowth":261,"createdAt":"2026-03-18T14:33:04Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["AI Agent","效率工具","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI求职","简历生成","求职自动化","Claude Code"],"domainKey":"AI求职","pushedAt":"2026-09-21T06:41:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/MadsLorentzen/ai-job-search","ts":"2026-09-21T06:41:22Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3815144368374052,"copyOk":false},{"repo":"microsoft/qlib","owner":"microsoft","name":"qlib","desc":"Qlib is an AI-oriented Quant investment platform that aims to use AI tech to empower Quant Research, from exploring ideas to implementing productions. Qlib supports diverse ML modeling paradigms, including supervised learning, market dynamics modeling, and RL, and is now equipped with https://github.com/microsoft/RD-Agent to automate R&D process.","summaryCn":"微软AI量化投资平台，从策略研究到实盘交易一条龙","reasonCn":"qlib的架构很完整：数据层支持多种金融数据源（Yahoo Finance、Tushare、Binance等）；因子层提供大量预置因子和自定义因子引擎；模型层支持LightGBM、XGBoost、MLP、LSTM等多种机器学习方法；回测引擎可以模拟历史交易表现；交易执行层可以对接实盘系统。","stars":48760,"starGrowth":32,"createdAt":"2020-08-14T06:46:00Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","效率工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["量化投资","机器学习","回测引擎","因子挖掘","金融AI"],"domainKey":"量化投资","pushedAt":"2026-09-22T05:57:23Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/microsoft/qlib","ts":"2026-09-22T05:57:23Z","category":"tool","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.09796882690730106,"copyOk":true},{"repo":"virtualgeoff/sunclock","owner":"virtualgeoff","name":"sunclock","desc":"24 hr sun clock","summaryCn":"显示当地太阳位置和升降时间的极简时钟工具","reasonCn":"这是一个实用的前端天气可视化工具，将抽象的天气API数据转化为直观的24小时太阳轨迹图。它实时计算并展示当前时刻太阳在地平线上的位置、高度角以及日出日落时间，帮助用户理解光照变化。界面简洁清新，无需复杂交互，适合嵌入个人主页或作为桌面挂件，兼具功能性与审美价值，让日历更有“温度”。","stars":65,"starGrowth":0,"createdAt":"2022-05-27T04:13:16Z","silentRounds":11,"language":"JavaScript","topics":["weather"],"aiDims":["非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["太阳位置计算","日出日落可视化","Web组件","天文算法","前端Widget"],"domainKey":"太阳位置计算","pushedAt":"2026-09-07T06:11:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"weather","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/virtualgeoff/sunclock","ts":"2026-09-07T06:11:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Suwayomi/Suwayomi-Server","owner":"Suwayomi","name":"Suwayomi-Server","desc":"A rewrite of Tachiyomi for the Desktop","summaryCn":"把手机看漫画的能力搬到桌面，本地服务器架构管理你的漫画库","reasonCn":"技术上，它采用服务器-客户端分离架构。服务端负责下载、缓存、管理漫画源，客户端负责展示。支持多源插件系统，可以接入各种漫画网站。核心亮点是跨设备同步，你在手机看到哪了，打开电脑自动接续。它还支持离线下载、分类管理、阅读历史记录。","stars":7739,"starGrowth":4,"createdAt":"2020-12-23T20:31:38Z","silentRounds":0,"language":"Java","topics":["desktop"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0,"domainTags":["漫画阅读","自托管","跨平台同步","插件系统","NAS"],"domainKey":"漫画阅读","pushedAt":"2026-09-23T14:56:35Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/Suwayomi/Suwayomi-Server","ts":"2026-09-23T14:56:35Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"feross/simple-peer","owner":"feross","name":"simple-peer","desc":"","summaryCn":"这款库的核心价值在于将原本极其复杂的WebRTC握手过程抽象为简洁的API接口。","reasonCn":"它的设计目标是让浏览器能够轻松实现点对点通信，无需中间服务器的深度参与即可传输视频流或原始数据。这对于需要低延迟交互的应用至关重要，例如在浏览器内部构建多人协同编辑工具或实时游戏客户端。由于其对各种现代浏览器兼容性做了深度的磨合测试，开发者不必担心在不同设备上","stars":7803,"starGrowth":1,"createdAt":"2014-06-26T18:46:23Z","silentRounds":3,"language":"JavaScript","topics":["webrtc","p2p","nodejs","browser","data-channels","peer-connection","javascript"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["WebRTC","实时通信","浏览器通信","P2P"],"domainKey":"WebRTC","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"webrtc","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"browser","source":"github","weight":0.7},{"name":"data-channels","source":"github","weight":0.7},{"name":"peer-connection","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/feross/simple-peer","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"goruck/home-generative-agent","owner":"goruck","name":"home-generative-agent","desc":"AI agent for Home Assistant — talk to your home, create automations in plain language, analyze cameras with face recognition, and get proactive anomaly alerts. Cloud LLMs or fully local via Ollama.","summaryCn":"很多智能家居系统所谓的“智能”，其实就是一堆僵硬的定时任务，而这个项目旨","reasonCn":"技术上，它通过将Home Assistant的API接口包装给LLM，使其能够理解复杂的家庭环境状态。它支持高度定制，既可以利用云端大模型的强大推理能力实现深度思考，也能通过本地Ollama部署实现纯本地化的隐私保护，确保家里数据不外流。","stars":306,"starGrowth":1,"createdAt":"2024-08-09T12:02:17Z","silentRounds":2,"language":"Python","topics":["home-automation"],"aiDims":["AI Agent","非AI-实用"],"zone":"AI","funScore":0.7,"domainTags":["智能家居","Home Assistant","AI管家","自动化"],"domainKey":"智能家居","pushedAt":"2026-09-23T13:08:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/goruck/home-generative-agent","ts":"2026-09-23T13:08:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"melgarafael/DeskcommCRM","owner":"melgarafael","name":"DeskcommCRM","desc":"Open-source AI sales OS — self-hosted CRM with native AI agents + WhatsApp (WAHA). Open alternative to Kommo, Octadesk & Intercom for any business that sells by chat. MCP-ready, multi-tenant, LGPD.","summaryCn":"自托管AI销售系统，自动陪客户聊 WhatsApp 谈生意","reasonCn":"这是一个开源的CRM系统，核心卖点是原生集成了AI智能体，能自动通过WhatsApp处理销售对话。它支持多租户架构，适合中小企业主自用或部署给团队。内置MCP就绪和LGPD合规，算是比较完善的企业级替代方案，让销售自动化变得触手可及。","stars":3502,"starGrowth":91,"createdAt":"2026-04-28T20:00:33Z","silentRounds":0,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["CRM","WhatsApp","AI-Agent","Sales-Automation"],"domainKey":"CRM","pushedAt":"2026-09-23T21:20:25Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/melgarafael/DeskcommCRM","ts":"2026-09-23T21:20:25Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.42114654483152486,"copyOk":true},{"repo":"id-Software/DOOM","owner":"id-Software","name":"DOOM","desc":"","summaryCn":"《毁灭战士》源代码的发布并非单纯的代码共享，它有着明确的使用门槛。","reasonCn":"在平台支持方面，这个开源项目有着显著的局限。文档中明确提到“the code only compiles and runs on linux”，官方未能发布DOS或Windows版本的代码，原因是当时使用了受版权保护的声音库。这一历史遗留问题使得该项目成为Linux游戏开发和计算机历史爱好者眼中的珍品，它促使玩家和开","stars":19595,"starGrowth":6,"createdAt":"2012-01-31T21:28:06Z","silentRounds":0,"language":"C++","topics":[],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":0.6,"domainTags":["游戏引擎","C++","FPS","Linux游戏开发"],"domainKey":"游戏引擎","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.44,"source":"tier","url":"https://github.com/id-Software/DOOM","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018918601684103087,"copyOk":false},{"repo":"reactchartjs/react-chartjs-2","owner":"reactchartjs","name":"react-chartjs-2","desc":"React components for Chart.js, the most popular charting library","summaryCn":"React开发者的图表库，拖几行代码就能画出炫酷数据图","reasonCn":"这是 Chart.js 的 React 封装版，支持折线图、柱状图、饼图、雷达图等十余种图表类型。基于 TypeScript 开发，类型提示友好，可高度定制颜色、动画和交互事件。配合 React 组件化开发，直接通过 props 传入数据即可渲染图表，适合快速搭建数据可视化面板和仪表盘。","stars":6940,"starGrowth":0,"createdAt":"2016-05-06T22:01:08Z","silentRounds":22,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["React图表组件","数据可视化","Chart.js封装"],"domainKey":"React图表组件","pushedAt":"2026-09-19T09:25:05Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/reactchartjs/react-chartjs-2","ts":"2026-09-19T09:25:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MovingBlocks/Terasology","owner":"MovingBlocks","name":"Terasology","desc":"Terasology - open source voxel world","summaryCn":"体素世界沙盒游戏，无限地图探索建造，开放源码的我的世界类神作。","reasonCn":"Terasology是一款开源的体素风格沙盒生存游戏，拥有无限程序生成的地图和复杂的生态模拟系统。玩家可以探索、采集资源、建造基地并与生物互动，因其出色的画质表现和模块化设计，成为游戏开发者和玩家喜爱的开源项目之一。","stars":3896,"starGrowth":0,"createdAt":"2011-03-04T03:49:19Z","silentRounds":4,"language":"Java","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["体素游戏","沙盒生存","程序化生成","开源游戏","生态模拟"],"domainKey":"体素游戏","pushedAt":"2026-09-21T13:36:39Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/MovingBlocks/Terasology","ts":"2026-09-21T13:36:39Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"a2aproject/A2A","owner":"a2aproject","name":"A2A","desc":"Agent2Agent (A2A) is an open protocol enabling communication and interoperability between opaque agentic applications.","summaryCn":"让不同AI助手能互相对话协作的开放协议正在重新定义AI生态","reasonCn":"Agent2Agent是一个开放通信协议，让不同厂商的AI agent能够互相协作。想象你的Claude Code能直接调用代码审查工具，或者你的AI助手能帮你在另一个AI工具里操作，这就是A2A协议的目标。2.5万star说明它已成为AI agent生态的基础设施级项目。","stars":25894,"starGrowth":9,"createdAt":"2025-03-25T18:44:21Z","silentRounds":0,"language":"Shell","topics":["agents"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["Agent通信协议","多Agent互操作","AI基础设施"],"domainKey":"Agent通信协议","pushedAt":"2026-09-22T13:28:54Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/a2aproject/A2A","ts":"2026-09-22T13:28:54Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.028042712597512936,"copyOk":false},{"repo":"KunAgent/Kun","owner":"KunAgent","name":"Kun","desc":"Local-first AI agent workspace for coding, writing, design, research, and automation — one runtime for desktop GUI and TUI.","summaryCn":"本地运行的全能AI助手，写代码剪视频搞自动化一键搞定","reasonCn":"Kun是一个本地优先的AI Agent工作空间，支持编码、写作、设计、研究和自动化五大核心场景。采用TypeScript开发，同时提供桌面GUI和终端TUI两种交互模式。最大亮点是完全本地运行，隐私数据不出本机，集成主流大模型API，还能通过工具调用实现复杂工作流自动化，是一个真正的AI生产力平台。","stars":6322,"starGrowth":0,"createdAt":"2026-05-21T13:22:06Z","silentRounds":1,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","AI应用","效率工具"],"zone":"AI","funScore":0.7,"domainTags":["AI代理","自动化","本地优先"],"domainKey":"AI代理","pushedAt":"2026-09-23T13:56:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/KunAgent/Kun","ts":"2026-09-23T13:56:17Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chart-kit/react-native-chart-kit","owner":"chart-kit","name":"react-native-chart-kit","desc":"Beautiful charts for React Native. Line, area, bar, pie, donut, progress, and contribution heatmaps for dashboards, reports, and data-rich mobile apps.","summaryCn":"React Native专属图表库，多种炫酷图表轻松搞定","reasonCn":"专为React Native打造的图表组件库，提供折线、柱状、饼图等多种类型。基于SVG渲染，性能优秀且样式灵活，无需原生依赖即可集成。适合移动端数据可视化需求，安装简单，API清晰，是RN开发者构建仪表盘和数据报表的实用工具。","stars":3113,"starGrowth":0,"createdAt":"2017-12-20T02:30:08Z","silentRounds":21,"language":"TypeScript","topics":["mobile","data-visualization"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["React Native图表","SVG渲染","移动端数据可视化","TypeScript组件库"],"domainKey":"React Native图表","pushedAt":"2026-09-06T20:00:28Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/chart-kit/react-native-chart-kit","ts":"2026-09-06T20:00:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opensolon/solon","owner":"opensolon","name":"solon","desc":"🔥 Java enterprise application development framework for full scenario: Restrained, Efficient, Open, Ecologicalll!!! 700% higher concurrency 50% memory savings Startup is 10 times faster. Packing 90% smaller; Compatible with java8 ~ java26; Supports LTS. (Replaceable spring)","summaryCn":"Java框架里的性能怪兽，启动快9倍内存省一半还打包小","reasonCn":"Solon是新一代Java企业级开发框架，主打极致轻量化和高性能。相比Spring Boot，启动速度快10倍、内存占用少50%、包体积压缩90%，并发能力提升700%。采用零依赖核心架构，按需加载模块，支持注解驱动和API编程双模式，兼容主流Java生态。","stars":2790,"starGrowth":2,"createdAt":"2018-07-07T13:23:50Z","silentRounds":0,"language":"Java","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["Java Web框架","高性能微服务","轻量级框架"],"domainKey":"Java Web框架","pushedAt":"2026-09-13T22:16:37Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/opensolon/solon","ts":"2026-09-13T22:16:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"t8y2/dbx","owner":"t8y2","name":"dbx","desc":"25 MB lightweight cross-platform database client for 90+ databases, including MySQL, PostgreSQL, SQLite, Redis, MongoDB, DuckDB, SQL Server, and Dameng. Built-in AI, MCP Server, CLI, desktop and Docker. | 轻量级跨平台数据库管理工具，支持 MySQL、PostgreSQL、SQLite、Redis、MongoDB、达梦等 90+ 数据库，提供桌面端、Docker、CLI、内置 AI 助手和 MCP Server。","summaryCn":"内置AI的轻量数据库客户端，支持70种数据库，电脑手机都能用","reasonCn":"用Rust开发的跨平台数据库客户端，支持MySQL、PostgreSQL、SQLite、Redis、MongoDB等70多种数据库。内置AI助手辅助SQL编写，集成MCP Server协议，提供命令行、桌面客户端和Docker部署三种使用方式，体积仅20MB极其轻量。","stars":20548,"starGrowth":90,"createdAt":"2026-04-29T03:00:36Z","language":"Rust","topics":["database","docker"],"aiDims":["AI应用","开发者工具","效率工具"],"zone":"AI","funScore":0.7,"domainTags":["数据库客户端","AI助手","MCP Server","Rust"],"domainKey":"数据库客户端","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/t8y2/dbx","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3397679579521122,"copyOk":true},{"repo":"nextcloud/desktop","owner":"nextcloud","name":"desktop","desc":"💻 Desktop sync client for Nextcloud","summaryCn":"你的私有云盘客户端，所有设备文件实时同步的神。","reasonCn":"这是Nextcloud的官方桌面同步客户端，用C++和Qt开发，性能很稳。它能帮你把电脑上的文件夹和Nextcloud服务器实时同步，支持版本回溯。对于注重隐私、不想把数据放在国外大厂服务器的开发者或极客来说，这是最靠谱的本地数据管家。","stars":3872,"starGrowth":0,"createdAt":"2017-09-27T11:38:04Z","silentRounds":5,"language":"C++","topics":["desktop"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文件同步","私有云","数据管理","跨平台"],"domainKey":"文件同步","pushedAt":"2026-09-23T19:58:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/nextcloud/desktop","ts":"2026-09-23T19:58:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hexabot-ai/Hexabot","owner":"hexabot-ai","name":"Hexabot","desc":"Hexabot v3 is an AI workflow automation platform, combining workflows, actions, agents, and conversational channels in one runtime.","summaryCn":"如果你想打造一个能自动处理客户咨询、后台订单，甚至联动公司内部数据库的","reasonCn":"技术架构上，Hexabot v3 采用了模块化设计。它的核心是一个强大且可扩展的事件引擎，能够同时处理多节点的并发任务。它支持深度定制的 Actions，允许 AI 直接调用外部 API 来操作业务，这使得它从简单的对话机器人升级为“拥有手脚的”智能代理。","stars":1242,"starGrowth":2,"createdAt":"2024-09-09T18:57:42Z","silentRounds":0,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI agent","工作流自动化","聊天机器人","Docker","MCP"],"domainKey":"AI agent","pushedAt":"2026-09-23T09:54:10Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/hexabot-ai/Hexabot","ts":"2026-09-23T09:54:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"3DCoded/3MS","owner":"3DCoded","name":"3MS","desc":"Modular Multimaterial System for Klipper 3D Printers","summaryCn":"给3D打印机加装模块化多材料系统的黑科技","reasonCn":"从技术角度看，3MS利用Klipper的插件架构，实现了对多挤出器热床的精细控制。它不只是简单地控制喷头移动，而是通过复杂的算法协调不同材料的挤出路径，确保它们在接触时能完美融合，而不是分层或错位。这就好比给打印机装上了一套“精密外科手术刀”，让它能处理以前只能靠手工完成","stars":157,"starGrowth":0,"createdAt":"2024-06-18T15:07:50Z","silentRounds":7,"language":"Shell","topics":["3d-printing"],"aiDims":["非AI-实用","3d-printing"],"zone":"工具","funScore":0.6,"domainTags":["3D 打印","Klipper","多材料","固件插件"],"domainKey":"3D 打印","pushedAt":"2026-09-11T18:25:45Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"3d-printing","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/3DCoded/3MS","ts":"2026-09-11T18:25:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"knadh/listmonk","owner":"knadh","name":"listmonk","desc":"High performance, self-hosted, newsletter and mailing list manager with a modern dashboard. Single binary app.","summaryCn":"不用折腾，一条命令建起自己的邮件订阅平台","reasonCn":"listmonk是一个高性能的自托管邮件列表和新闻通讯管理系统，基于Go语言编译成单二进制文件运行。它提供现代化的管理仪表盘，支持MySQL/PostgreSQL数据库，内置A/B测试、邮件模板引擎、详细发送统计和订阅者管理功能。无需依赖微服务架构，单个二进制即可完整部署。","stars":23531,"starGrowth":12,"createdAt":"2019-06-26T05:08:39Z","silentRounds":0,"language":"Go","topics":["self-hosted"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["邮件服务","自托管","订阅管理","营销工具","数据库"],"domainKey":"邮件服务","pushedAt":"2026-09-13T21:49:56Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/knadh/listmonk","ts":"2026-09-13T21:49:56Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03752989673197059,"copyOk":false},{"repo":"flame-engine/flame","owner":"flame-engine","name":"flame","desc":"A Flutter based game engine.","summaryCn":"用Flutter引擎打造跨平台游戏，手机电脑网页都能跑","reasonCn":"它架构上分成了渲染、物理、输入、资源管理等模块，采用组件系统让你像搭积木一样拼装游戏逻辑。核心亮点是它的确定性游戏循环，保证不同设备上的同步行为。物理系统内置了AABB碰撞检测，配合自定义碰撞处理很灵活。","stars":10765,"starGrowth":2,"createdAt":"2017-10-22T03:01:50Z","silentRounds":0,"language":"Dart","topics":["game"],"aiDims":["游戏"],"zone":"工具","funScore":0.6,"domainTags":["游戏引擎","Flutter","跨平台游戏"],"domainKey":"游戏引擎","pushedAt":"2026-09-23T15:30:58Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/flame-engine/flame","ts":"2026-09-23T15:30:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kevinraymond/fosfora","owner":"kevinraymond","name":"fosfora","desc":"Free, open-source real-time audio & music visualizer / VJ engine, with audio-reactive GPU shaders, particles, and layer compositing, driven by WGSL and controlled with MIDI, OSC, or your phone. Cross-platform (Vulkan/Metal), with NDI, Spout, Syphon, and virtual-camera output.","summaryCn":"能用手机或MIDI设备实时操控的音乐视觉化与酷炫VJ舞台引擎","reasonCn":"在底层技术上，它完全基于WebGPU和WGSL着色器语言进行构建，这意味着它能榨干你显卡的每一滴性能，渲染出极其丝滑的百万级粒子和复杂的视觉图层。更厉害的是，它不只是自嗨，还支持标准的MIDI控制器、OSC协议，甚至能直接用你的手机作为遥控器，在实时演出中随心所欲地切换视觉效果","stars":45,"starGrowth":2,"createdAt":"2026-02-26T04:22:24Z","silentRounds":0,"language":"Rust","topics":["creative-coding"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["音乐可视化","实时渲染","WebGPU","VJ引擎"],"domainKey":"音乐可视化","pushedAt":"2026-09-23T04:29:59Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/kevinraymond/fosfora","ts":"2026-09-23T04:29:59Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"doobidoo/mcp-memory-service","owner":"doobidoo","name":"mcp-memory-service","desc":"Open-source persistent memory for AI agent pipelines (LangGraph, CrewAI, AutoGen) and Claude. REST API + knowledge graph + autonomous consolidation.","summaryCn":"给AI大脑装个长期记忆，再也不怕对话断片儿了","reasonCn":"这是一个专为AI智能体设计的持久化记忆服务，支持LangGraph、CrewAI等多种框架。它通过REST API提供存储接口，结合知识图谱技术自动整理信息，让AI在多轮对话中记住上下文并自动合并冗余知识，大幅提升智能体的连贯性和实用性。","stars":1953,"starGrowth":0,"createdAt":"2024-12-26T10:15:44Z","silentRounds":3,"language":"Python","topics":["vector-db"],"aiDims":["AI Agent","知识图谱","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["AI Agent记忆","知识图谱","向量数据库","RAG","持久化存储"],"domainKey":"知识图谱","pushedAt":"2026-09-23T20:06:18Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/doobidoo/mcp-memory-service","ts":"2026-09-23T20:06:18Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"slowkow/harmonypy","owner":"slowkow","name":"harmonypy","desc":"🎼 Integrate multiple high-dimensional datasets with fuzzy k-means and locally linear adjustments.","summaryCn":"用模糊聚类搞生物信息学，把不同来源的高维数据像编舞一样对齐整合","reasonCn":"做生物信息学的同僚都知道，最头疼的不是分析一个数据集，而是当手里有单细胞RNA-seq、蛋白质组学、临床数据等多个“方言”不同的 dataset 时，怎么把它们拼成一张完整的拼图？harmonypy 就是那个“翻译官”和“协调员”。","stars":284,"starGrowth":0,"createdAt":"2019-12-19T17:25:59Z","silentRounds":10,"language":"Python","topics":["bioinformatics"],"aiDims":["非AI-实用","开发者工具","学习资源"],"zone":"工具","funScore":0,"domainTags":["生物信息学","数据整合","模糊聚类"],"domainKey":"生物信息学","pushedAt":"2026-09-16T18:09:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"bioinformatics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/slowkow/harmonypy","ts":"2026-09-16T18:09:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"svenstaro/miniserve","owner":"svenstaro","name":"miniserve","desc":"🌟 For when you really just want to serve some files over HTTP right now!","summaryCn":"一行命令开启文件服务器，分享文件就这么简单","reasonCn":"它的核心就是一个极简的HTTP文件服务器。Rust编写，编译出来不到10MB的单二进制文件，没有运行时依赖，复制到任何机器上都能跑。功能上支持目录列表、文件上传、ZIP下载，还有美观的自定义主题和浏览器内预览功能。最重要的是它支持Basic认证和HTTPS，安全性有保障。","stars":7870,"starGrowth":2,"createdAt":"2018-04-26T09:52:56Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["HTTP文件服务器","反向代理","内网穿透","自托管","Rust"],"domainKey":"HTTP文件服务器","pushedAt":"2026-09-18T22:44:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/svenstaro/miniserve","ts":"2026-09-18T22:44:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"presenton/presenton","owner":"presenton","name":"presenton","desc":"Open-Source AI Presentation Generator and API (Gamma, Canva, Beautiful AI, Decktopus, Presentations AI Alternative)","summaryCn":"输入主题自动生成PPT的AI神器，让演示文稿制作不再是苦差事","reasonCn":"核心技术上，它本质上是一个AI Agent工作流。你输入主题或上传文档作为输入，系统会调用大语言模型来生成大纲和详细内容，然后结合模板引擎自动排版，再调用图片生成服务或从图库中检索配图。整个过程自动化完成，用户几乎不用手动操作。","stars":10702,"starGrowth":28,"createdAt":"2025-05-10T14:12:46Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI演示文稿","PPT生成","Agent工作流"],"domainKey":"AI演示文稿","pushedAt":"2026-09-22T10:37:12Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/presenton/presenton","ts":"2026-09-22T10:37:12Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.09184900018688097,"copyOk":true},{"repo":"nextcloud/android","owner":"nextcloud","name":"android","desc":"📱 Nextcloud Android app","summaryCn":"手机里的私人云盘管家，文件自动同步备份，数据自己掌握","reasonCn":"Nextcloud官方Android客户端，支持文件上传下载、文档预览、日历通讯录同步，兼容WebDAV和CalDAV/CardDAV协议。提供端到端加密和两步验证，可将手机照片自动备份到私有云服务器，实现跨设备无缝协作，是替代商业云存储的开源方案。","stars":5590,"starGrowth":1,"createdAt":"2016-06-06T21:23:36Z","silentRounds":2,"language":"Kotlin","topics":["mobile"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["云存储","Android应用","文件同步","私有云"],"domainKey":"云存储","pushedAt":"2026-09-23T16:51:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/nextcloud/android","ts":"2026-09-23T16:51:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ahujasid/ableton-mcp","owner":"ahujasid","name":"ableton-mcp","desc":"Control Ableton Live with any LLM: create tracks, arrange clips & compose music via MCP","summaryCn":"用任何AI大模型直接指挥音乐软件，动动嘴就能自动编曲创作。","reasonCn":"基于Anthropic最新的Model Context Protocol技术，无缝打通了AI大模型与专业音乐制作软件Ableton Live。让你能够直接通过自然语言下达指令，自动化创建音轨、编排音乐片段并进行复杂的编曲创作，彻底打破了技术与音乐之间的壁垒。","stars":3086,"starGrowth":5,"createdAt":"2025-03-19T17:00:09Z","silentRounds":0,"language":"Python","topics":["music"],"aiDims":["AI Agent","工具调用","创意工具"],"zone":"创意","funScore":1,"domainTags":["音乐制作","Ableton Live","MCP","AI应用","自动编曲"],"domainKey":"音乐制作","pushedAt":"2026-09-22T10:37:13Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/ahujasid/ableton-mcp","ts":"2026-09-22T10:37:13Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.019860661049902785,"copyOk":true},{"repo":"microg/GmsCore","owner":"microg","name":"GmsCore","desc":"Free implementation of Play Services","summaryCn":"没装Google的安卓手机也能用需要Play服务的App","reasonCn":"这是一个免费的Google Play Services开源替代实现，专为没有预装GMS的安卓设备设计。它通过虚拟服务框架让依赖Google移动服务的第三方应用（如游戏、地图、推送通知等）能够正常运行，无需root权限。核心亮点是轻量级架构，相比完整GMS占用更少内存和电量，支持主流安卓版本，且可自由安装卸载。项目采用开源协议，社区持续维护，是国内用户刷机后的必备工具之一，也为全球开源Android生态贡献了重要力量。","stars":14662,"starGrowth":4,"createdAt":"2014-04-23T19:28:47Z","silentRounds":0,"language":"Java","topics":["mobile","maps"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Android 系统","Google Play 服务","移动开发"],"domainKey":"Android 系统","pushedAt":"2026-09-23T10:06:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"maps","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/microg/GmsCore","ts":"2026-09-23T10:06:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AUK9527/Are-u-ok","owner":"AUK9527","name":"Are-u-ok","desc":"","summaryCn":"该项目详细列出了针对IPQ807x和MT798x等芯片平台的适配方案，涵盖了小米AX9","reasonCn":"项目维护者对插件包进行了长期的版本迭代，提供从22.03到25.12等多个时间节点的版本选择。这种持续的维护确保了即便在固件版本更新后，用户依然能找到匹配的插件包，从而保持网络环境的稳定与功能的持续可用。对于那些喜欢将小主机或高性能路由器进行深度定制的玩家来说，这无疑提","stars":14176,"starGrowth":3,"createdAt":"2021-12-10T07:57:34Z","silentRounds":0,"language":"","topics":[],"aiDims":["开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["路由器","固件适配","iStore","网络设备"],"domainKey":"路由器","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1}],"aiScore":0.44,"source":"search","url":"https://github.com/AUK9527/Are-u-ok","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"uutils/coreutils","owner":"uutils","name":"coreutils","desc":"Cross-platform Rust rewrite of the GNU coreutils","summaryCn":"用Rust重写的Linux工具箱，跨平台又快速，Linux玩家必备","reasonCn":"用现代Rust语言完整重写GNU coreutils，兼容所有常用命令如ls、cat、grep等。支持Windows、macOS等Linux发行版，速度更快且内存占用更低。对于经常跨平台开发或追求系统性能优化的极客来说，是提升日常命令行体验的优质替代方案。","stars":24152,"starGrowth":6,"createdAt":"2013-08-02T16:22:56Z","silentRounds":0,"language":"Rust","topics":["rust","coreutils","gnu-coreutils","busybox","cross-platform","command-line-tool"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Rust","命令行工具","系统工具","GNU兼容","跨平台"],"domainKey":"Rust","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"rust","source":"github","weight":0.7},{"name":"coreutils","source":"github","weight":0.7},{"name":"gnu-coreutils","source":"github","weight":0.7},{"name":"busybox","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"command-line-tool","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/uutils/coreutils","ts":"2026-09-18T10:23:42.801Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018745279894004638,"copyOk":true},{"repo":"GMGNAI/gmgn-skills","owner":"GMGNAI","name":"gmgn-skills","desc":"GMGN OpenAPI skills for AI Agent — query tokens, wallets, and market data, and execute on-chain trades across Solana, BSC, and Base.","summaryCn":"专为AI智能体打造的链上交易外挂，打通多链数据与自动化。 (29字)","reasonCn":"技术架构上，它专门封装了GMGN平台的各类开放接口，采用轻量级Python编写。通过标准化的接口规范，能够让任何主流的AI Agent框架（如LangChain、AutoGPT等）无缝调用。它不仅能毫秒级查询各个公链上的代币流动性、K线数据和巨鲸钱包动态，还能通过安全签名机制在Solana、BSC等链上","stars":558,"starGrowth":0,"createdAt":"2026-03-12T01:57:18Z","silentRounds":1,"language":"Python","topics":["trading"],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","区块链","自动化交易","技能包","Web3"],"domainKey":"AI Agent","pushedAt":"2026-09-23T09:30:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/GMGNAI/gmgn-skills","ts":"2026-09-23T09:30:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kwaroran/Risuai","owner":"kwaroran","name":"Risuai","desc":"Make your own story. User-friendly software for LLM roleplaying","summaryCn":"专为大模型角色扮演打造的神级前端，定制你的专属二次元对话世界。","reasonCn":"在技术实现上，它是一个纯前端的TypeScript应用，支持部署到网页端或打包成桌面客户端。它对各种大模型后端有着顶级的兼容性，不仅能接OpenAI、Claude等商业API，还能完美连上各种本地运行的开源大模型。它内置了强大的宏定义、世界书（Lorebook）系统和正则表达式过滤，能让AI记住","stars":1691,"starGrowth":0,"createdAt":"2023-04-12T13:08:52Z","silentRounds":3,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI应用","大语言模型","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["角色扮演","对话模拟","LLM","RPG"],"domainKey":"角色扮演","pushedAt":"2026-09-23T12:31:07Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/kwaroran/Risuai","ts":"2026-09-23T12:31:07Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lichess-org/mobile","owner":"lichess-org","name":"mobile","desc":"Lichess mobile app","summaryCn":"免费开源的国际象棋APP，随时随地和全球棋手过招","reasonCn":"Lichess移动端是基于Flutter开发的开源国际象棋应用，支持iOS和Android双平台。提供在线实时对弈、残局训练、棋局回放和Stockfish引擎分析等核心功能。界面简洁流畅，操作体验优秀，是移动棋类应用中的标杆之作，完全免费无广告。","stars":2554,"starGrowth":0,"createdAt":"2022-05-01T10:26:50Z","silentRounds":1,"language":"Dart","topics":["mobile","chess"],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["国际象棋","移动应用","实时对弈","棋类"],"domainKey":"国际象棋","pushedAt":"2026-09-23T16:21:43Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"chess","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/lichess-org/mobile","ts":"2026-09-23T16:21:43Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"owncloud/android","owner":"owncloud","name":"android","desc":":phone: The ownCloud Android App","summaryCn":"自己搭建私有网盘，手机随时同步文件，告别云端隐私泄露","reasonCn":"ownCloud Android客户端，支持文件同步、分享和协作，兼容WebDAV协议。采用Kotlin编写，界面简洁流畅，支持大文件分片上传、断点续传。可连接自建或公共ownCloud服务器，提供端到端加密保障数据安全，是企业和个人私有云存储的理想移动端方案。","stars":4171,"starGrowth":0,"createdAt":"2012-08-25T10:35:56Z","silentRounds":3,"language":"Kotlin","topics":["mobile"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["私有云","文件同步","WebDAV","Android"],"domainKey":"私有云","pushedAt":"2026-09-23T16:22:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/owncloud/android","ts":"2026-09-23T16:22:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zcreativelabs/react-simple-maps","owner":"zcreativelabs","name":"react-simple-maps","desc":"Composable SVG map charts for data visualization in React","summaryCn":"用React轻松绘制交互式SVG世界地图，数据一键可视化","reasonCn":"react-simple-maps 基于 D3-geo 与 topojson 实现可组合的 SVG 地图组件，支持投影转换、缩放平移、交互事件和自定义样式，配合 React Hooks 能快速在页面嵌入世界、国家或地区地图，并通过 GeoJSON 数据绑定实现数据驱动的色彩或气泡可视化。","stars":3355,"starGrowth":3,"createdAt":"2017-03-03T18:17:19Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization","maps"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["数据可视化","React","地图","SVG"],"domainKey":"数据可视化","pushedAt":"2026-09-12T15:26:55Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"maps","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.46,"source":"search","url":"https://github.com/zcreativelabs/react-simple-maps","ts":"2026-09-12T15:26:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JCodesMore/ai-website-cloner-template","owner":"JCodesMore","name":"ai-website-cloner-template","desc":"Clone any website with one command using AI coding agents","summaryCn":"一个命令用AI自动克隆任意网站，轻松获取前端代码模板","reasonCn":"这个项目利用AI coding agent自动分析目标网站的整体结构和前端代码，一键克隆出完整的HTML、CSS和JS文件。无需手动扒代码，输入网址就能获得可运行的前端模板，非常适合快速搭建原型、学习优秀设计，或者把喜欢的网站风格复用到自己的项目里。","stars":34790,"starGrowth":62,"createdAt":"2026-03-13T11:14:39Z","silentRounds":0,"language":"TypeScript","topics":["developer-tools","automation","ai"],"aiDims":["AI Agent","开发者工具","效率工具"],"zone":"AI","funScore":0.7,"domainTags":["AI编码代理","网站克隆","前端开发","自动化","代码生成"],"domainKey":"AI编码代理","pushedAt":"2026-09-20T09:02:52Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/JCodesMore/ai-website-cloner-template","ts":"2026-09-20T09:02:52Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.22961563667720608,"copyOk":false},{"repo":"OpenBB-finance/OpenBB","owner":"OpenBB-finance","name":"OpenBB","desc":"Open Data Platform for analysts, quants and AI agents.","summaryCn":"面向分析师和AI的智能金融数据平台，海量市场数据随手可得","reasonCn":"OpenBB是一个开源金融数据终端，覆盖股票、加密货币、宏观经济、期权等全品类市场数据。它提供Python SDK和REST API，支持本地部署和云端调用。项目架构模块化，数据源可插拔，内置大量分析工具。最关键的是它专门为AI agents设计，大模型可以通过标准接口直接调用金融数据，无需自己写爬虫或买数据。","stars":73386,"starGrowth":24,"createdAt":"2020-12-20T10:46:38Z","silentRounds":0,"language":"Python","topics":["ml","ai"],"aiDims":["AI应用","工具调用","非AI-实用"],"zone":"AI","funScore":0.5,"domainTags":["金融","量化","数据终端","AI Agent","开源"],"domainKey":"AI Agent","pushedAt":"2026-09-23T20:49:42Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/OpenBB-finance/OpenBB","ts":"2026-09-23T20:49:42Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07298111356389501,"copyOk":false},{"repo":"apache/cordova-android","owner":"apache","name":"cordova-android","desc":"Apache Cordova Android","summaryCn":"用Web技术开发原生App的跨平台老将，至今仍在维护","reasonCn":"Apache基金会维护的Android平台cordova实现，允许开发者使用HTML、CSS和JavaScript构建原生移动应用。支持丰富的插件生态系统，可访问相机、定位、传感器等原生能力，适合前端开发者快速转向移动端开发，维护稳定但技术路线相对传统。","stars":3788,"starGrowth":1,"createdAt":"2012-11-28T08:00:22Z","silentRounds":24,"language":"JavaScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Apache Cordova","跨平台开发","Android容器","WebView","移动端开发"],"domainKey":"Apache Cordova","pushedAt":"2026-09-18T20:44:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/apache/cordova-android","ts":"2026-09-18T20:44:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"openshift/origin","owner":"openshift","name":"origin","desc":"Conformance test suite for OpenShift","summaryCn":"OpenShift 的官方验收测试工具，确保部署符合标准","reasonCn":"Origin 是 OpenShift 的官方合规测试套件，用于验证各版本的 OpenShift 发行版是否遵循上游规范。对于使用 OpenShift 的企业运维和 DevOps 团队来说，这是保证集群稳定性和兼容性的核心工具，确保你的部署不出问题。","stars":8687,"starGrowth":0,"createdAt":"2014-07-30T20:04:05Z","silentRounds":3,"language":"Go","topics":["devops","kubernetes"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["OpenShift","Kubernetes","合规测试","DevOps"],"domainKey":"OpenShift","pushedAt":"2026-09-23T14:25:36Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/openshift/origin","ts":"2026-09-23T14:25:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mcinglis/c-style","owner":"mcinglis","name":"c-style","desc":"","summaryCn":"在系统级编程中，如何写出优雅、正确且易于维护的 C 语言代码一直是程序员们","reasonCn":"该项目的核心亮点在于其独特的编程价值观：它坚定地将正确性、可读性、简单性和可维护性排在运行速度之前，警惕过早优化带来的隐患。虽然随着时间的推移，作者在仓库中加了温馨提示，表示这份文档已年代久远并转而推荐更权威的工业标准（如 SEI CERT C 编码标准和编译选项加固指南","stars":2148,"starGrowth":1,"createdAt":"2013-09-21T22:33:50Z","silentRounds":1,"language":"","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["C语言编程规范","代码风格","最佳实践"],"domainKey":"C语言编程规范","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1}],"aiScore":0.26,"source":"search","url":"https://github.com/mcinglis/c-style","ts":"2026-09-22T22:51:50.275Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ruc-datalab/DeepAnalyze","owner":"ruc-datalab","name":"DeepAnalyze","desc":"DeepAnalyze is the first agentic LLM for autonomous data science. 🎈你的AI数据分析师，自动分析大量数据，一键生成专业分析报告！","summaryCn":"你的AI数据分析师，自动生成专业分析报告。","reasonCn":"DeepAnalyze 是首个面向数据科学的自主智能体 LLM。它能够自主连接数据库、清洗数据、进行可视化分析，并最终生成专业的分析报告。用户只需上传数据或给出分析目标，它就能像资深分析师一样完成繁琐的 ETL 流程，直接输出可落地的结论，极大地降低了数据分析的门槛。","stars":4648,"starGrowth":0,"createdAt":"2025-10-11T11:19:21Z","silentRounds":1,"language":"Python","topics":["data-visualization","chatbot"],"aiDims":["AI Agent","AI应用","非AI-实用"],"zone":"AI","funScore":0.8,"domainTags":["Data Science","LLM Agent","Automation","Data Visualization","Report Generation"],"domainKey":"Data Science","pushedAt":"2026-09-23T13:41:15Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/ruc-datalab/DeepAnalyze","ts":"2026-09-23T13:41:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fennaraOfficial/fennara-godot-ai","owner":"fennaraOfficial","name":"fennara-godot-ai","desc":"AI chat and agent tooling for Godot, with both MCP support and in-built native chat window.","summaryCn":"把大模型和智能体直接塞进Godot游戏引擎的开发神器","reasonCn":"这是一个用Rust编写的高性能Godot引擎插件，集成了原生AI聊天窗口与完整的模型上下文协议支持。它不仅能让独立开发者在编辑器里直接和AI协同写代码，还能让游戏里的NPC具备实时运行的Agent决策能力。通过深度的引擎底层绑定，它成功打通了游戏开发工具链与现代大语言模型的双向赋能。","stars":294,"starGrowth":0,"createdAt":"2026-06-01T06:16:55Z","silentRounds":3,"language":"Rust","topics":["godot"],"aiDims":["AI Agent","游戏","开发者工具"],"zone":"AI","funScore":0.7,"domainTags":["游戏开发","AI集成","Godot插件"],"domainKey":"游戏开发","pushedAt":"2026-09-23T12:04:06Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/fennaraOfficial/fennara-godot-ai","ts":"2026-09-23T12:04:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ohsusannamarie/technical-interview-prep-toolkit","owner":"ohsusannamarie","name":"technical-interview-prep-toolkit","desc":"Candidate-friendly technical interview prep resources, checklists, coding practice links, and recruiter-shareable guides.","summaryCn":"技术面试全套备考资源，清单题库复盘指南全齐","reasonCn":"想冲刺面试却苦于资料不齐全？这套工具包把面试准备拆解成清晰的清单、题库、复盘笔记，所有内容用 Markdown 写成，既方便本地阅读，也可直接部署到 GitHub Pages。\\n核心功能包括：面试清单模板（覆盖系统设计、算法、行为面试）；题目列表按公司/技术栈分类；答案解析与最佳实践","stars":18,"starGrowth":0,"createdAt":"2019-11-06T21:41:43Z","silentRounds":8,"language":"","topics":["interview"],"aiDims":["学习资源"],"zone":"资源","funScore":0,"domainTags":["技术面试","面试准备","题库"],"domainKey":"技术面试","pushedAt":"2026-09-10T06:44:16Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7}],"aiScore":0.5,"source":"search","url":"https://github.com/ohsusannamarie/technical-interview-prep-toolkit","ts":"2026-09-10T06:44:16Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vapor/vapor","owner":"vapor","name":"vapor","desc":"💧 A server-side Swift HTTP web framework.","summaryCn":"用Swift写后端API的瑞士军刀，让服务端开发像搭积木一样简单","reasonCn":"Vapor是基于Swift的开源服务端Web框架，提供类型安全的路由系统、中间件管道、ORM数据库操作和WebSocket支持。可与PostgreSQL、Redis、MongoDB等无缝集成，适合构建高性能RESTful API和实时应用，是苹果生态后端开发的首选方案之一。","stars":26214,"starGrowth":0,"createdAt":"2016-01-18T22:37:52Z","silentRounds":2,"language":"Swift","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Swift","Web框架","后端开发","API"],"domainKey":"Swift","pushedAt":"2026-09-23T13:05:58Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/vapor/vapor","ts":"2026-09-23T13:05:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Jakubantalik/transitions.dev","owner":"Jakubantalik","name":"transitions.dev","desc":"Collection of the most essential transitions for web apps, skill for agents and Refine tool for agents","summaryCn":"网页动效精选库，顺带给了AI智能体一些实用技能包","reasonCn":"这个仓库整理了网页开发中最常用的CSS过渡动画效果，设计精美且开箱即用。同时它还包含AI智能体的技能模块和Refine工具，帮助开发者快速构建带流畅动效的智能体交互界面。前端设计爱好者和AI应用开发者都能从中受益，毕竟好看的动画能让产品体验提升一大截。","stars":4310,"starGrowth":24,"createdAt":"2026-04-21T21:01:08Z","silentRounds":0,"language":"HTML","topics":["design"],"aiDims":["非AI-好玩","非AI-实用","工具调用"],"zone":"创意","funScore":1,"domainTags":["Web 动效","CSS 过渡","AI 智能体","交互设计"],"domainKey":"Web 动效","pushedAt":"2026-09-21T13:24:58Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/Jakubantalik/transitions.dev","ts":"2026-09-21T13:24:58Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.10644640371229697,"copyOk":true},{"repo":"xbtlin/ai-berkshire","owner":"xbtlin","name":"ai-berkshire","desc":"AI 时代的伯克希尔：基于 Claude Code / Codex 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built for Claude Code / Codex. 4 masters' methodologies + multi-agent adversarial analysis.","summaryCn":"巴菲特方法论+AI多Agent，投资研究也能智能化","reasonCn":"技术实现上，项目基于Claude Code和OpenAI Codex两大编程Agent平台，构建了多Agent协作系统。每个Agent被赋予不同的投资分析角色，通过结构化提示词调用各自的能力，最终整合形成完整的研究结论。框架中包含DCF估值模型、财务指标分析、护城河评估等经典投资分析工具，全部用代码实","stars":16488,"starGrowth":9,"createdAt":"2026-04-07T11:19:51Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["价值投资","多Agent","投资研究","Claude Code","量化分析"],"domainKey":"价值投资","pushedAt":"2026-09-22T17:00:12Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/xbtlin/ai-berkshire","ts":"2026-09-22T17:00:12Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.034365065502183406,"copyOk":true},{"repo":"vladmandic/sdnext","owner":"vladmandic","name":"sdnext","desc":"SD.Next: All-in-one WebUI for AI generative image and video creation, captioning and processing","summaryCn":"本地部署AI绘画神器，图片视频生成一站式搞定","reasonCn":"技术上，它基于Stable Diffusion Core搭建，支持ControlNet、IP-Adapter这些高级功能。ControlNet能让你控制生成的构图和姿态，IP-Adapter可以保持人物一致性。你可以通过Web界面上传图片、调整参数、生成图像，还能做视频生成。","stars":7342,"starGrowth":1,"createdAt":"2022-12-24T20:59:11Z","silentRounds":2,"language":"Python","topics":["transformers"],"aiDims":["多模态Agent","AI应用","创意工具"],"zone":"AI","funScore":0,"domainTags":["图像生成","视频生成","Stable Diffusion","ControlNet","IP-Adapter"],"domainKey":"图像生成","pushedAt":"2026-09-23T09:42:40Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/vladmandic/sdnext","ts":"2026-09-23T09:42:40Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gofr-dev/gofr","owner":"gofr-dev","name":"gofr","desc":"An opinionated GoLang framework for accelerated microservice development. Built in support for databases and observability.","summaryCn":"用Go写微服务就像搭积木，这个框架让后端开发快人一步","reasonCn":"gofr 是一个 opinionated 的 Go 微服务框架，内置了数据库（PostgreSQL、MySQL、MongoDB、Redis）、可观测性（Prometheus、Grafana、OpenTelemetry）、日志、HTTP 服务器等核心组件，开箱即用，省去重复配置。提供代码生成、统一错误处理、JWT 中间件等开发能力，大幅降低微服务搭建成本。框架设计简洁，开发者只需专注业务逻辑，不用在基建上浪费时间。","stars":20899,"starGrowth":0,"createdAt":"2023-10-24T10:14:48Z","silentRounds":24,"language":"Go","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["微服务","Go语言","后端框架","可观测性","数据库连接池"],"domainKey":"微服务","pushedAt":"2026-09-23T11:54:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/gofr-dev/gofr","ts":"2026-09-23T11:54:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nuxt-community/axios-module","owner":"nuxt-community","name":"axios-module","desc":"Secure and easy axios integration for Nuxt 2","summaryCn":"Nuxt 2 的 axios 神器，一行配置搞定前后端请求","reasonCn":"专为 Nuxt 2 打造的 axios 集成模块，自动处理服务端渲染环境下的请求差异，内置请求拦截器和响应处理。通过 nuxt.config.js 简单配置即可全局使用 axios，无需手动封装，省去大量样板代码，特别适合需要频繁请求 API 的 Nuxt 项目。","stars":1186,"starGrowth":0,"createdAt":"2017-08-13T11:20:04Z","silentRounds":22,"language":"JavaScript","topics":["backend"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["Nuxt2","HTTP客户端","SSR请求","Axios封装"],"domainKey":"Nuxt2","pushedAt":"2026-09-16T07:46:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/nuxt-community/axios-module","ts":"2026-09-16T07:46:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Adrian104/Dungeon-Generator","owner":"Adrian104","name":"Dungeon-Generator","desc":"Free and open-source BSP-based procedural 2D dungeon map generator.","summaryCn":"很多独立开发者和桌面Roguelike制作者都苦于没有现成的地下城地图生成工具，","reasonCn":"项目核心采用二叉空间划分（BSP）技术，递归把整个地图空间切分为若干房间，再通过走廊把房间连接起来，保证每个房间可达且结构自然。代码全部用现代 C++17 实现，使用标准库和轻量级的随机数生成器，没有外部依赖，支持自定义房间最小/最大尺寸、走廊宽度、种子以及生成次数等参数","stars":16,"starGrowth":0,"createdAt":"2021-02-20T20:20:43Z","silentRounds":9,"language":"C++","topics":["roguelike"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":0,"domainTags":["程序化生成","地下城地图","BSP算法","Roguelike开发","C++游戏工具"],"domainKey":"程序化生成","pushedAt":"2026-09-09T20:19:04Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"roguelike","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/Adrian104/Dungeon-Generator","ts":"2026-09-09T20:19:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kedacore/keda","owner":"kedacore","name":"keda","desc":"KEDA is a Kubernetes-based Event Driven Autoscaling component. It provides event driven scale for any container running in Kubernetes","summaryCn":"让服务器自动根据消息量伸缩，省电又省钱的神器","reasonCn":"KEDA 是一个基于 Kubernetes 的云原生事件驱动自动扩缩容组件，它能够监听各种消息队列和事件源，根据实时负载动态调整应用副本数，从而在保证服务稳定性的同时大幅降低资源成本。该项目设计初衷是为了解决云原生环境下的动态资源调度难题，让运维人员无需编写复杂的脚本即可实现精细化的弹性伸缩策略。","stars":10546,"starGrowth":6,"createdAt":"2019-02-13T00:12:33Z","silentRounds":0,"language":"Go","topics":["kubernetes"],"aiDims":["AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","自动扩缩容","事件驱动","云原生","消息队列"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T16:35:51Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/kedacore/keda","ts":"2026-09-23T16:35:51Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.019706808268537834,"copyOk":false},{"repo":"simstudioai/sim","owner":"simstudioai","name":"sim","desc":"Build, deploy, and orchestrate AI agents. Sim is the central intelligence layer for your AI workforce.","summaryCn":"零代码编排AI Agent，多人协作处理复杂任务，企业级智能体工作台","reasonCn":"Sim是一个AI Agent编排平台，让你用可视化方式把多个AI智能体串联起来自动完成任务，内置RAG检索增强和丰富的工作流引擎，支持REST API和SDK接入，适合需要多Agent协作处理复杂流程的企业场景，目前已有近三万Star，是Agent编排领域较受关注的开源项目。","stars":29706,"starGrowth":5,"createdAt":"2025-01-05T22:47:49Z","silentRounds":0,"language":"TypeScript","topics":["rag","ai","chatbot","agents"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["AI Agent","多智能体协作","工作流编排","RAG"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:50:57Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/simstudioai/sim","ts":"2026-09-23T21:50:57Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.015504948495253483,"copyOk":false},{"repo":"symfony/console","owner":"symfony","name":"console","desc":"Eases the creation of beautiful and testable command line interfaces","summaryCn":"PHP开发者的命令行瑞士军刀，轻松打造美观的终端命令工具","reasonCn":"Symfony Console是PHP生态中最流行的命令行框架，提供了完整的指令定义、参数解析、选项处理、进度条和交互式输入等能力。支持自动补全、颜色输出、表格格式化输出，能与Composer无缝集成，被Laravel等框架底层广泛使用，成熟稳定。","stars":9813,"starGrowth":0,"createdAt":"2011-02-22T18:00:31Z","silentRounds":9,"language":"PHP","topics":["command-line"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["命令行框架","PHP开发","CLI工具构建","终端交互"],"domainKey":"命令行框架","pushedAt":"2026-09-18T16:09:14Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/symfony/console","ts":"2026-09-18T16:09:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"radareorg/radare2","owner":"radareorg","name":"radare2","desc":"UNIX-like reverse engineering framework and command-line toolset","summaryCn":"黑客眼中的瑞士军刀，逆向分析二进制文件的开源神器","reasonCn":"这是一个功能强大的逆向工程框架，支持反汇编、调试、分析和脚本编写。它不依赖图形界面，通过命令行即可深入剖析可执行文件内部结构。虽然与AI无直接关联，但作为安全领域的经典工具，其插件生态可辅助恶意代码分析。","stars":24855,"starGrowth":11,"createdAt":"2012-07-03T07:42:26Z","silentRounds":0,"language":"C","topics":["security"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["逆向工程","二进制分析","安全研究","CTF","反汇编"],"domainKey":"逆向工程","pushedAt":"2026-09-22T11:04:25Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/radareorg/radare2","ts":"2026-09-22T11:04:25Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03432770066385033,"copyOk":true},{"repo":"SheerSt/pokewilds","owner":"SheerSt","name":"pokewilds","desc":"","summaryCn":"一款以经典宝可梦为灵感的沙盒生存游戏，带你进入一个全新的野生冒险世界。","reasonCn":"PokeWilds是一款极具创意的宝可梦衍生沙盒游戏，它打破了传统捕捉战斗的单一模式，引入了生存、采集与建造等元素。在这个广阔的世界中，玩家可以探索地图、收集资源并与野生精灵互动，体验独特的冒险旅程。尽管目前项目仍处于Alpha测试阶段，但其独特的玩法机制和自由度高的游戏体验，为喜爱宝可梦题材又追求新鲜感的玩家提供了极佳的选择，是一款让人忍不住想投入时间探索的趣味项目。","stars":2892,"starGrowth":1,"createdAt":"2021-02-26T06:09:00Z","silentRounds":3,"language":"Assembly","topics":[],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0.5599999999999999,"domainTags":["开源游戏","沙盒生存","角色扮演","宝可梦衍生"],"domainKey":"开源游戏","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Assembly","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/SheerSt/pokewilds","ts":"2026-09-19T22:19:04.205Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"labring/sealos","owner":"labring","name":"sealos","desc":"Deploy real projects from GitHub or your AI coding agent, then keep them running with AI-powered operations.","summaryCn":"对于很多开发者来说，好不容易写好了一个AI应用或者开源项目，结果在准备发","reasonCn":"技术架构上，它基于Kubernetes内核进行了深度魔改与创新，将底层服务器资源抽象为标准化的计算和存储单元。其独特的亮点在于强大的自动化运维能力，不仅能一键将GitHub上的代码转化为在线服务，还能结合AI智能体实时监控应用运行状态，自动进行故障诊断和弹性扩缩容，彻底解放了运","stars":18353,"starGrowth":1,"createdAt":"2018-08-15T12:22:50Z","silentRounds":3,"language":"TypeScript","topics":["docker","kubernetes"],"aiDims":["AI基础设施","模型部署","开发者工具"],"zone":"AI","funScore":0,"domainTags":["Kubernetes","云原生","AI运维","K8s","自动化"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T12:29:38Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/labring/sealos","ts":"2026-09-23T12:29:38Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vxcontrol/pentagi","owner":"vxcontrol","name":"pentagi","desc":"Fully autonomous AI Agents system capable of performing complex penetration testing tasks","summaryCn":"让AI全自动帮你搞渗透测试的黑客助手，安全研究员必备玩具","reasonCn":"这是一个完全自主的AI Agent系统，专为执行复杂的渗透测试任务而设计。基于Go语言构建，能够自动规划并执行漏洞扫描、利用和报告生成。它代表了AI在网络安全领域的深度应用，将人工渗透测试流程自动化，极大提升了红队作战的效率。","stars":24871,"starGrowth":8,"createdAt":"2025-01-06T22:22:48Z","silentRounds":0,"language":"Go","topics":["self-hosted"],"aiDims":["AI Agent","安全工具"],"zone":"AI","funScore":0,"domainTags":["自动化渗透测试","AI Agent","网络安全","漏洞利用"],"domainKey":"AI Agent","pushedAt":"2026-09-23T12:43:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/vxcontrol/pentagi","ts":"2026-09-23T12:43:46Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.024964979293152667,"copyOk":false},{"repo":"ReactiveX/RxJava","owner":"ReactiveX","name":"RxJava","desc":"","summaryCn":"在最新的4.0版本中，ReactiveX/RxJava最引人注目的改变是彻底移除了所有第三","reasonCn":"其核心设计理念是利用`java.util.concurrent.Flow`接口进行底层构建，这种基于Java标准库的实现方式让RxJava与现代JVM生态无缝融合。它通过声明式的运算符将异步事件流组合在一起，把低级的线程调度、同步机制和安全问题全部隐藏在抽象层之下。","stars":48191,"starGrowth":0,"createdAt":"2013-01-08T20:11:48Z","silentRounds":6,"language":"Java","topics":["java","rxjava","flow","reactive-streams"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["响应式编程","异步流处理","Java生态","操作符链","背压处理"],"domainKey":"响应式编程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"java","source":"github","weight":0.7},{"name":"rxjava","source":"github","weight":0.7},{"name":"flow","source":"github","weight":0.7},{"name":"reactive-streams","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.44,"source":"tier","url":"https://github.com/ReactiveX/RxJava","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"htlin222/gym-course","owner":"htlin222","name":"gym-course","desc":"Evidence-graded strength-training course curated from 316 public YouTube demos — 11 chapters, 44 units, 31 hours. Every link verified, every movement category checked against PubMed: only 2 of 65 evidence claims reach \"strong\". Course content in Traditional Chinese.","summaryCn":"全网最硬核的健身课程，每招每式都有医学论文背书，科学健身不瞎练。","reasonCn":"整个课程分为11个章节，涵盖了从基础力量到高级进阶的各个方面。项目维护者花费了大量时间，不仅核对视频内容，还查阅了PubMed上的医学文献，确保推荐的动作是安全有效的。对于健身新手来说，它是一个极佳的入门向导；对于有一定基础的爱好者来说，它也是一个检查自己动作是否标准","stars":12,"starGrowth":0,"createdAt":"2026-07-26T17:00:04Z","silentRounds":9,"language":"JavaScript","topics":["fitness"],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":1,"domainTags":["力量训练","循证医学","运动科学","健身教程","健身课程"],"domainKey":"力量训练","pushedAt":"2026-09-09T09:06:12Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/htlin222/gym-course","ts":"2026-09-09T09:06:12Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/parquet-java","owner":"apache","name":"parquet-java","desc":"","summaryCn":"这是一个由Apache基金会维护的底层开源项目，主要用于提供Apache Parquet数","reasonCn":"在技术实现上，该项目严格遵循Dremel论文中描述的记录粉碎与组装算法，从而能够完美表示和处理复杂的嵌套数据结构。开发团队在历年的技术演进中不断优化其性能，确保它在面对大规模批量数据时依然能够保持出色的吞吐表现，满足各类企业级大数据分析场景的严苛要求。","stars":3081,"starGrowth":0,"createdAt":"2014-06-10T07:00:07Z","silentRounds":3,"language":"Java","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Java","数据存储","列式存储","Hadoop"],"domainKey":"Java","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/apache/parquet-java","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"prisma/orm","owner":"prisma","name":"orm","desc":"Next-generation ORM for Node.js & TypeScript | PostgreSQL, MySQL, MariaDB, SQL Server, SQLite, MongoDB and CockroachDB","summaryCn":"下一代 Node.js/TS ORM，数据库操作像写代码一样丝滑。","reasonCn":"Prisma 是一款现代化的 ORM（对象关系映射）库，专为 Node.js 和 TypeScript 设计。它通过自动生成类型安全的数据库客户端，让开发者能够用类和对象的方式操作数据库，而不再需要编写繁琐的 SQL 语句。它支持 PostgreSQL、MySQL、SQLite 等多种主流数据库，极大地提升了开发效率和代码安全性。","stars":47660,"starGrowth":7,"createdAt":"2019-06-20T13:33:47Z","silentRounds":0,"language":"TypeScript","topics":["database"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["数据库","ORM","TypeScript"],"domainKey":"数据库","pushedAt":"2026-09-23T08:26:50Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/prisma/orm","ts":"2026-09-23T08:26:50Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02144062106588334,"copyOk":false},{"repo":"argoproj/argo-workflows","owner":"argoproj","name":"argo-workflows","desc":"Workflow Engine for Kubernetes","summaryCn":"帮你把复杂的AI任务像搭积木一样串起来，自动跑完所有步骤。","reasonCn":"Argo Workflows 是基于 Kubernetes 的下一代工作流引擎，专为自动化任务编排设计。它采用声明式定义，让用户能以 YAML 形式直观描述任务流转逻辑。核心优势在于强大的状态管理、自动重试、并行执行以及丰富的插件生态，完美解决大规模数据处理、模型训练流水线等复杂场景下的调度难题。","stars":16999,"starGrowth":1,"createdAt":"2017-08-21T18:50:44Z","silentRounds":2,"language":"Go","topics":["ml"],"aiDims":["开发者工具","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes工作流","CI/CD","流水线编排","声明式配置","Go"],"domainKey":"Kubernetes工作流","pushedAt":"2026-09-23T08:43:24Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/argoproj/argo-workflows","ts":"2026-09-23T08:43:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"t1m0thyj/WinDynamicDesktop","owner":"t1m0thyj","name":"WinDynamicDesktop","desc":"Port of macOS Mojave Dynamic Desktop feature to Windows","summaryCn":"macOS动态桌面搬上Windows，壁纸自动跟着时间变","reasonCn":"这是一个将macOS Mojave动态桌面功能移植到Windows的开源项目，用C#编写。它通过读取系统时间和地理位置，自动切换对应时段的高清壁纸，支持手动和自动两种模式，可以从Unsplash等来源获取图片，支持多显示器，系统托盘管理很方便。","stars":5030,"starGrowth":4,"createdAt":"2018-06-08T23:32:27Z","silentRounds":0,"language":"C#","topics":["desktop"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["动态壁纸","桌面美化","macOS移植"],"domainKey":"动态壁纸","pushedAt":"2026-09-18T01:08:31Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/t1m0thyj/WinDynamicDesktop","ts":"2026-09-18T01:08:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gregmalcolm/python_koans","owner":"gregmalcolm","name":"python_koans","desc":"","summaryCn":"用禅宗公案式的测试题，边改错边学Python语法，寓教于乐。","reasonCn":"这是一套通过编写测试用例来学习Python的互动式教程。你需要运行测试，然后逐行修复代码中的错误填空，直到所有测试通过。它模仿了Ruby Koans的风格，把语法知识拆解成一个个小挑战，适合有编程基础想巩固Python细节的初学者。","stars":5116,"starGrowth":1,"createdAt":"2010-03-01T17:04:31Z","silentRounds":5,"language":"Python","topics":["python","koans"],"aiDims":["学习资源"],"zone":"资源","funScore":0.5,"domainTags":["Python","编程学习","测试驱动"],"domainKey":"Python","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"python","source":"github","weight":0.7},{"name":"koans","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/gregmalcolm/python_koans","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"0x90n/InfoSec-Black-Friday","owner":"0x90n","name":"InfoSec-Black-Friday","desc":"","summaryCn":"InfoSec-Black-Friday 是一个专注于信息安全领域的年度黑色星期五及网络星期","reasonCn":"这个项目的核心价值在于其垂直领域的专业性。它不混杂普通的电子产品折扣，而是聚焦于杀毒软件、防火墙、渗透测试框架、安全认证课程以及风险处理平台等专业工具。对于预算有限的学生、初级安全工程师或独立研究者来说，这里往往是发现超值安全装备的最佳场所。","stars":4371,"starGrowth":0,"createdAt":"2019-11-27T20:57:04Z","silentRounds":5,"language":"","topics":[],"aiDims":["非AI-实用","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["安全工具折扣","黑五促销","信息安全","优惠目录"],"domainKey":"安全工具折扣","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.38,"source":"tier","url":"https://github.com/0x90n/InfoSec-Black-Friday","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"usememos/memos","owner":"usememos","name":"memos","desc":"Open-source, self-hosted note-taking tool built for quick capture. Markdown-native, lightweight, and fully yours.","summaryCn":"你是不是有时候灵感来了，想快速记点什么，但打开一个重量级笔记软件又觉得","reasonCn":"Memos的设计理念是'快'——快速记录、快速访问、快速检索。它采用Markdown原生格式，界面简洁，没有多余的复杂功能。关键是它是自托管的，数据完全存在你自己服务器上，不用担心隐私泄露或被平台监控。 安装部署非常简单。","stars":63258,"starGrowth":25,"createdAt":"2021-12-08T15:30:18Z","silentRounds":0,"language":"Go","topics":["note-taking","self-hosted"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["自托管笔记","Markdown编辑器","知识管理"],"domainKey":"自托管笔记","pushedAt":"2026-09-23T00:19:12Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/usememos/memos","ts":"2026-09-23T00:19:12Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07618562079104618,"copyOk":false},{"repo":"CharlesPikachu/musicdl","owner":"CharlesPikachu","name":"musicdl","desc":"Musicdl: A lightweight music downloader written in pure python. (轻量级无损音乐下载器，支持数十个音乐/有声读物平台，例如网易云音乐，QQ音乐，酷狗音乐，酷我音乐，咪咕音乐，千千静听，汽水音乐，Bilibili，街声，喜马拉雅，懒人听书，荔枝FM，蜻蜓FM，JOOX，TIDAL，YouTube，Apple Music，Spotify，Qobuz，SoundCloud等主流音乐平台)","summaryCn":"很多时候我们想收藏一首歌，却发现有的歌在网易云听，有的在QQ音乐听，还要","reasonCn":"技术上，它是通过爬虫和逆向音频流获取接口实现的。作者非常细致地处理了不同平台特有的加密签名和鉴权逻辑，确保了下载过程的顺畅。它支持多种音频质量选择，能够智能解析解析复杂的媒体流，并自动补全歌曲的元数据，比如歌手信息、封面图等，让下载后的文件分类整齐、易于管理，","stars":6288,"starGrowth":1,"createdAt":"2018-06-26T21:38:42Z","silentRounds":1,"language":"Python","topics":["music"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["音乐下载","无损音频","爬虫","音频流解析"],"domainKey":"音乐下载","pushedAt":"2026-09-23T03:11:15Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/CharlesPikachu/musicdl","ts":"2026-09-23T03:11:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bottlepy/bottle","owner":"bottlepy","name":"bottle","desc":"bottle.py is a fast and simple micro-framework for python web-applications.","summaryCn":"Python最轻量的Web框架，几行代码就能搭个网站","reasonCn":"Bottle是一个零依赖的单文件Python Web框架，非常适合快速开发小型应用和API接口。虽然它本身不涉及AI能力，但作为后端基础，常被用于构建AI应用的Web服务层。代码简洁，上手极快，是学习Python Web开发的经典入门项目。","stars":8790,"starGrowth":1,"createdAt":"2009-06-30T17:51:30Z","silentRounds":6,"language":"Python","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Python","Web框架","微框架","WSGI"],"domainKey":"Python","pushedAt":"2026-09-18T09:07:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/bottlepy/bottle","ts":"2026-09-18T09:07:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"anderspitman/awesome-tunneling","owner":"anderspitman","name":"awesome-tunneling","desc":"List of ngrok, Cloudflare Tunnel, Tailscale, and ZeroTier alternatives and other tunneling software and services. Focus on self-hosting.","summaryCn":"ngrok等内网穿透工具的完整清单，自托管场景下的网络穿透指南","reasonCn":"这是一个汇总ngrok、Cloudflare Tunnel、Tailscale、ZeroTier等多种内网穿透工具的Awesome列表，专注于自托管方案。star两万一千多，是开发者和中小企业搭建远程访问通道的实用参考指南，帮助快速找到合适的隧道解决方案。","stars":21856,"starGrowth":10,"createdAt":"2020-09-16T20:38:27Z","silentRounds":0,"language":"","topics":["self-hosted"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"资源","funScore":0,"domainTags":["内网穿透","网络穿透","自托管","Awesome列表","开发工具"],"domainKey":"内网穿透","pushedAt":"2026-09-15T13:38:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7}],"aiScore":0.38,"source":"search","url":"https://github.com/anderspitman/awesome-tunneling","ts":"2026-09-15T13:38:27Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.03137262079062957,"copyOk":false},{"repo":"TheWaWaR/simple-http-server","owner":"TheWaWaR","name":"simple-http-server","desc":"","summaryCn":"simple-http-server 本质上是一个基于 Rust 生态高效构建的单文件静态文件服","reasonCn":"该项目最吸引人的地方在于其高度集成的中间件能力。它打破了传统简易服务器只“放文件”的局限。从功能列表可见，它内置了“CORS 跨域请求”、“压缩传输”以及“基于范围的请求”。对于前端工程师而言，这意味着你不用再写一堆复杂的 Nginx 反向代理规则，或者不用配置繁琐的 Webpack 开发","stars":3457,"starGrowth":1,"createdAt":"2017-05-27T10:52:10Z","silentRounds":4,"language":"Rust","topics":[],"aiDims":["开发者工具","安全工具"],"zone":"工具","funScore":0.2,"domainTags":["静态文件服务","Rust开发","HTTPS加密","跨域处理"],"domainKey":"静态文件服务","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/TheWaWaR/simple-http-server","ts":"2026-09-19T18:31:48.187Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"inkandswitch/backstitch","owner":"inkandswitch","name":"backstitch","desc":"Backstitch: Real-Time Version Control for Godot","summaryCn":"简单来说，这个工具解决了游戏开发者多年来的一个痛点：多人一起做游戏时，","reasonCn":"在技术架构上，它由大名鼎鼎的Ink & Switch实验室操刀，底层核心采用高性能的Rust语言编写，保证了极低延迟和并发处理能力。它创新性地运用了CRDT（无冲突复制数据类型）技术，能够在没有中央服务器强制锁定文件的情况下，自动在本地和远程之间同步并合并游戏场景的修改。","stars":254,"starGrowth":0,"createdAt":"2024-12-04T09:32:09Z","silentRounds":4,"language":"Rust","topics":["godot"],"aiDims":["开发者工具","创意工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["实时版本控制","游戏引擎","协作工具"],"domainKey":"实时版本控制","pushedAt":"2026-09-23T21:13:11Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/inkandswitch/backstitch","ts":"2026-09-23T21:13:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kgoedecke/doop","owner":"kgoedecke","name":"doop","desc":"The open-source alternative to Paper.design. A multiplayer design canvas where humans and AI agents design together, live. MCP built in.","summaryCn":"能和AI一起在画布上实时协作设计的黑科技画板","reasonCn":"其核心技术基于现代化的多人实时同步架构，并大胆引入了 MCP（Model Context Protocol）。这意味着 AI Agent 不再是简单的文本生成器，而是拥有了对画布状态的感知能力和操作权限。它可以实时读取你当前选中的元素，根据上下文理解你的设计意图，并调用工具直接在画布上创建或修改","stars":742,"starGrowth":5,"createdAt":"2026-08-22T15:40:51Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["AI Agent","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["设计协作","AI绘画","MCP协议"],"domainKey":"设计协作","pushedAt":"2026-09-22T16:25:10Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/kgoedecke/doop","ts":"2026-09-22T16:25:10Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.0528234501347709,"copyOk":false},{"repo":"gravitational/teleport","owner":"gravitational","name":"teleport","desc":"The easiest, and most secure way to access and protect all of your infrastructure.","summaryCn":"一个让你用浏览器就能安全访问所有服务器、数据库和K8s集群的瑞士军刀","reasonCn":"Teleport 是一个现代化的基础设施访问平台，它用动态颁发的短期证书替代传统 SSH 密钥，解决了密码共享、密钥管理和访问审计等老问题。支持 Kubernetes、数据库、远程桌面和 SSH，内置 Web 终端，还能做细粒度权限控制，star 超过两万，是企业级安全访问的热门选择。","stars":20936,"starGrowth":7,"createdAt":"2015-03-02T19:36:00Z","silentRounds":0,"language":"Go","topics":["security","kubernetes"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["基础设施访问","SSH","Kubernetes","零信任"],"domainKey":"基础设施访问","pushedAt":"2026-09-17T15:49:47Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/gravitational/teleport","ts":"2026-09-17T15:49:47Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02200305693542224,"copyOk":false},{"repo":"apernet/hysteria","owner":"apernet","name":"hysteria","desc":"Hysteria is a powerful, lightning fast and censorship resistant proxy.","summaryCn":"速度快又抗封锁的网络代理工具，技术发烧友的网络加速器","reasonCn":"基于QUIC协议的抗封锁代理工具，速度极快且能绕过网络限制。支持混淆、认证和多种传输协议，专为高延迟和不稳定网络环境优化。对于需要稳定快速网络连接的技术用户，特别是经常访问受限内容的场景，是一款不可多得的实用工具。","stars":22547,"starGrowth":10,"createdAt":"2020-04-21T00:01:34Z","silentRounds":0,"language":"Go","topics":["golang","proxy","reliable-udp","quic","censorship-circumvention","http-proxy","socks5","relay","hysteria","vpn","tun"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网络代理","QUIC协议","抗封锁"],"domainKey":"网络代理","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"golang","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"reliable-udp","source":"github","weight":0.7},{"name":"quic","source":"github","weight":0.7},{"name":"censorship-circumvention","source":"github","weight":0.7},{"name":"http-proxy","source":"github","weight":0.7},{"name":"socks5","source":"github","weight":0.7},{"name":"relay","source":"github","weight":0.7},{"name":"hysteria","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"tun","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/apernet/hysteria","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03133055395396283,"copyOk":true},{"repo":"yewstack/yew","owner":"yewstack","name":"yew","desc":"Rust / Wasm framework for creating reliable and efficient web applications","summaryCn":"用Rust写网页，编译成Wasm速度超快，一个项目两栖开发","reasonCn":"Yew是用Rust构建WebAssembly前端框架，让开发者能用Rust语言写高性能Web应用。它支持组件化开发，通过虚拟DOM实现高效渲染，同时享受Rust编译期类型安全和零成本抽象。与React等JS框架相比，Yew能复用后端Rust业务逻辑，避免前后端语言割裂，适合构建高性能单页应用、实时协作工具和复杂数据可视化场景。","stars":32817,"starGrowth":2,"createdAt":"2017-12-16T14:19:02Z","silentRounds":0,"language":"Rust","topics":["web-framework"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Rust","WebAssembly","前端框架","组件化","虚拟DOM"],"domainKey":"Rust","pushedAt":"2026-09-18T03:19:36Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/yewstack/yew","ts":"2026-09-18T03:19:36Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pingcap/tidb","owner":"pingcap","name":"tidb","desc":"TiDB is built for agentic workloads that grow unpredictably, with ACID guarantees and native support for transactions, analytics, and vector search. No data silos. No noisy neighbors. No infrastructure ceiling.","summaryCn":"MySQL兼容的分布式数据库，能跑交易也能跑AI向量搜索，一站式搞定","reasonCn":"TiDB是PingCAP开源的分布式数据库，兼容MySQL协议，支持在线事务处理（OLTP）和在线分析处理（OLAP）混合负载，原生支持向量搜索功能，适合AI应用的数据层。水平扩展能力强，无单点故障，不用改应用代码就能从单机MySQL平滑迁移。","stars":40572,"starGrowth":5,"createdAt":"2015-09-06T04:01:52Z","silentRounds":0,"language":"Go","topics":["database","ai"],"aiDims":["向量数据库","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","向量搜索","HTAP","MySQL兼容"],"domainKey":"分布式数据库","pushedAt":"2026-09-23T21:45:50Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/pingcap/tidb","ts":"2026-09-23T21:45:50Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015369713102632358,"copyOk":true},{"repo":"kiyotakali/Miru","owner":"kiyotakali","name":"Miru","desc":"An open-source AI companion that actually remembers you — runs entirely on your own Mac or server. Desktop pet · long-term memory · proactive companionship · multi-device sync.","summaryCn":"能记住你所有习惯的AI伴侣，在本地Mac上运行","reasonCn":"Miru的技术架构非常巧妙。它把AI Agent的几个关键能力拆解得很细致。首先是“长期记忆”，它会把重要的对话和你的个人信息存储起来，让AI能持续“理解”你。其次是“主动陪伴”，它不是等你说什么才回答，而是会主动发消息问候你，提醒你日程，甚至在你心情不好的时候逗你开心。","stars":160,"starGrowth":1,"createdAt":"2026-07-03T09:10:18Z","silentRounds":4,"language":"Python","topics":["anime"],"aiDims":["AI Agent","多模态Agent","AI应用"],"zone":"AI","funScore":1,"domainTags":["AI伴侣","本地部署","长期记忆","Mac应用"],"domainKey":"AI伴侣","pushedAt":"2026-09-23T13:33:14Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/kiyotakali/Miru","ts":"2026-09-23T13:33:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dialogic-godot/dialogic","owner":"dialogic-godot","name":"dialogic","desc":"💬 Create Dialogs, Visual Novels, RPGs, and manage Characters with Godot to create your Game!","summaryCn":"可视化对话系统插件，零代码制作RPG和视觉小说","reasonCn":"Dialogic是Godot 4.x的对话系统插件，通过可视化节点编辑器创建复杂对话树、条件分支和角色动画。支持时间轴编辑器、变量绑定和自定义事件，让游戏开发者无需编写代码就能实现专业的剧情对话系统。对于想制作视觉小说、RPG或任何需要丰富对话的游戏项目来说，这是目前Godot生态中最成熟的解决方案，5000多star证明了它的受欢迎程度。","stars":6016,"starGrowth":3,"createdAt":"2020-06-07T12:49:43Z","silentRounds":0,"language":"GDScript","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"工具","funScore":0.7,"domainTags":["对话系统","游戏开发","Godot插件"],"domainKey":"对话系统","pushedAt":"2026-08-30T13:29:27Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/dialogic-godot/dialogic","ts":"2026-08-30T13:29:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"gogs/gogs","owner":"gogs","name":"gogs","desc":"The painless way to host your own Git service","summaryCn":"轻量自托管Git服务器，几分钟搭建私有代码仓库","reasonCn":"Gogs是一个用Go语言编写的轻量级Git服务，自托管方案比GitLab更省资源。支持私有仓库、代码审查、Issue管理和Webhook等完整功能，单机即可部署，特别适合小团队和个人开发者。已有近5万star，安装简单资源占用极低。","stars":47820,"starGrowth":0,"createdAt":"2014-02-12T01:57:08Z","silentRounds":1,"language":"Go","topics":["self-hosted","raspberry-pi"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Git","自托管","代码仓库","版本控制","轻量级"],"domainKey":"Git","pushedAt":"2026-09-12T20:28:48Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/gogs/gogs","ts":"2026-09-12T20:28:48Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"visit1985/mdp","owner":"visit1985","name":"mdp","desc":"A command-line based markdown presentation tool.","summaryCn":"一个基于C语言的命令行工具，用于演示和教学。","reasonCn":"mdp是一个开源项目，旨在提供一个易于学习和理解的AI概念演示平台。","stars":5280,"starGrowth":1,"createdAt":"2014-09-14T13:18:50Z","silentRounds":7,"language":"C","topics":["command-line"],"aiDims":["命令行工具","学习资源"],"zone":"工具","funScore":0,"domainTags":["Markdown","命令行","演示","教学"],"domainKey":"Markdown","pushedAt":"2026-09-09T07:32:34Z","tags":[{"name":"命令行工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/visit1985/mdp","ts":"2026-09-09T07:32:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alibaba/MNN","owner":"alibaba","name":"MNN","desc":"MNN: A blazing-fast, lightweight inference engine battle-tested by Alibaba, powering high-performance on-device LLMs and Edge AI.","summaryCn":"手机端也能秒开大模型，阿里开源的轻量级极速推理引擎。","reasonCn":"技术亮点在于它的底层是用纯粹的C++编写，裁剪得极其轻量。它针对移动端各种芯片做了深度优化，不仅支持CPU，还能完美调用GPU和NPU进行硬件加速。在处理大语言模型时，它通过独特的量化和内存复用技术，把庞大的模型瘦身到手机也能轻松带动的程度，运行速度和显存占用表现极为惊艳","stars":16130,"starGrowth":3,"createdAt":"2019-04-15T07:40:18Z","silentRounds":0,"language":"C++","topics":["ml"],"aiDims":["推理引擎","模型部署","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["移动端推理","大模型部署","边缘计算","模型量化"],"domainKey":"移动端推理","pushedAt":"2026-09-23T09:26:40Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/alibaba/MNN","ts":"2026-09-23T09:26:40Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"arc53/DocsGPT","owner":"arc53","name":"DocsGPT","desc":"Private AI platform for agents, assistants and enterprise search. Built-in Agent Builder, Deep research, Document analysis, Multi-model support, and API connectivity for agents.","summaryCn":"用AI自建私有知识库和Agent，告别数据上传云端","reasonCn":"DocsGPT是由arc53团队开源的一款企业级AI平台，核心价值是让你用AI自建私有知识库和智能助手，而且所有数据都在自己服务器上跑，再也不用担心把敏感信息传到云端了。\n技术架构上，DocsGPT提供了完整的AI应用开发工具链。","stars":18284,"starGrowth":2,"createdAt":"2023-02-02T11:03:23Z","silentRounds":0,"language":"Python","topics":["rag","agents","ml","transformers"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["RAG","AI Agent","私有化部署","企业知识库","文档分析"],"domainKey":"RAG","pushedAt":"2026-09-22T15:45:43Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/arc53/DocsGPT","ts":"2026-09-22T15:45:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"voidcosmos/npkill","owner":"voidcosmos","name":"npkill","desc":"List any node_modules 📦 dir in your system and how heavy they are. You can then select which ones you want to erase to free up space 🧹","summaryCn":"扫描清理 node_modules 目录，释放磁盘空间","reasonCn":"这个项目用 TypeScript 编写，核心思路很简单：递归扫描你指定的目录，找到所有 node_modules 文件夹，计算每个的磁盘占用大小，然后以表格形式展示出来。你可以选择性地删除任何一个文件夹，或者一键清理全部。它支持多种界面模式，包括命令行交互界面和 JSON 输出，方便集成到自","stars":9452,"starGrowth":1,"createdAt":"2019-06-30T12:26:26Z","silentRounds":2,"language":"TypeScript","topics":["command-line","developer-tools"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["磁盘清理","node_modules","CLI工具","Node.js开发","磁盘空间管理"],"domainKey":"磁盘清理","pushedAt":"2026-09-08T10:43:40Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/voidcosmos/npkill","ts":"2026-09-08T10:43:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"scrapy/scrapy","owner":"scrapy","name":"scrapy","desc":"Scrapy, a fast high-level web crawling & scraping framework for Python.","summaryCn":"专治各种网站数据抓取需求，自动爬取效率惊人","reasonCn":"Scrapy是Python生态中最成熟、最强大的开源爬虫框架，专为大规模网页数据采集而生。它内置异步并发引擎，可同时抓取数千个页面；支持自动重试、去重、深度控制等核心功能。项目提供了完整的数据管道（Pipeline）机制，爬虫得到的数据可以直接清洗、验证并保存到数据库或文件。相比手写爬虫，Scrapy让你从「写几行代码就开爬」变成「系统化采集」，适合从个人研究到企业级数据工程的各类场景。","stars":64446,"starGrowth":8,"createdAt":"2010-02-22T02:01:14Z","silentRounds":0,"language":"Python","topics":["python","scraping","crawling","framework","crawler","hacktoberfest","web-scraping","web-scraping-python"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Python","爬虫","数据采集","异步框架"],"domainKey":"Python","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"python","source":"github","weight":0.7},{"name":"scraping","source":"github","weight":0.7},{"name":"crawling","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"crawler","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"web-scraping","source":"github","weight":0.7},{"name":"web-scraping-python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/scrapy/scrapy","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02437240480402197,"copyOk":false},{"repo":"NG-ZORRO/ng-zorro-antd","owner":"NG-ZORRO","name":"ng-zorro-antd","desc":"Angular UI Component Library based on Ant Design","summaryCn":"阿里巴巴蚂蚁出品的Angular组件库，开箱即用颜值高","reasonCn":"技术上，Ng-Zorro完整实现了Ant Design的设计规范，提供70多个组件，覆盖表格、表单、按钮、弹窗、导航、布局等所有常用场景。每个组件都有严格的TypeScript类型定义，IDE智能提示体验很好。支持主题定制，可以通过修改变量一键切换颜色风格。","stars":9173,"starGrowth":0,"createdAt":"2017-08-08T14:59:33Z","silentRounds":24,"language":"TypeScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Angular","UI组件库","企业后台","Ant Design"],"domainKey":"Angular","pushedAt":"2026-09-23T01:40:15Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/NG-ZORRO/ng-zorro-antd","ts":"2026-09-23T01:40:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"varun29ankuS/shodh-memory","owner":"varun29ankuS","name":"shodh-memory","desc":"Local, LLM-free memory for AI agents. A single offline Rust binary — deterministic and auditable — that learns from use, forgets the irrelevant, and strengthens what matters. No cloud, no API keys.","summaryCn":"离线运行的Agent记忆库，保护隐私的神器","reasonCn":"它的最大优势在于“离线”和“安全”。传统的记忆方案往往依赖云服务或OpenAI的API，而Shodh Memory完全运行在你的本地机器上。它通过向量嵌入技术，将对话内容转化为数值向量，并存储在本地数据库中。当Agent需要回忆信息时，它会在这个本地数据库中进行快速检索，无需任何网络请求。","stars":289,"starGrowth":1,"createdAt":"2025-12-03T13:39:53Z","silentRounds":2,"language":"Rust","topics":["vector-db"],"aiDims":["AI Agent","工具调用","向量数据库"],"zone":"AI","funScore":0.7,"domainTags":["Agent记忆","向量检索","本地AI"],"domainKey":"Agent记忆","pushedAt":"2026-09-23T18:30:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/varun29ankuS/shodh-memory","ts":"2026-09-23T18:30:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"autogluon/autogluon","owner":"autogluon","name":"autogluon","desc":"Fast and Accurate ML in 3 Lines of Code","summaryCn":"三行代码，自动给你调出一个能用的机器学习模型","reasonCn":"它把表格、图像、文本、时间序列和多模态任务做成统一入口，内部会自动选择随机森林、梯度提升树、神经网络和预训练模型，进行数据切分、超参搜索、早停与模型融合。这样普通工程师也能在几小时内拿到接近专业调参的效果，同时保留导出训练配置、复现和继续优化的能力，适合业务落地与研究基线。","stars":10734,"starGrowth":7,"createdAt":"2019-07-29T18:51:24Z","silentRounds":0,"language":"Python","topics":["ml","deep-learning"],"aiDims":["模型与训练","开发者工具","AI应用"],"zone":"AI","funScore":0,"domainTags":["自动机器学习","模型训练","多模态预测","数据科学"],"domainKey":"模型训练","pushedAt":"2026-09-23T21:31:47Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/autogluon/autogluon","ts":"2026-09-23T21:31:47Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.022956400223588597,"copyOk":true},{"repo":"fail2ban/fail2ban","owner":"fail2ban","name":"fail2ban","desc":"Daemon to ban hosts that cause multiple authentication errors","summaryCn":"给服务器装个智能保安，谁暴力破解就自动拉黑谁","reasonCn":"fail2ban通过监控SSH、HTTP等服务的日志，自动识别反复登录失败的恶意IP，并调用iptables或firewalld等防火墙规则将其封禁。支持多种action和filter配置，可扩展性强，是Linux服务器安全防护的经典方案，运维圈知名度极高。","stars":18662,"starGrowth":3,"createdAt":"2011-09-28T16:24:20Z","silentRounds":0,"language":"Python","topics":["security","monitoring"],"aiDims":["安全工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["服务器安全","入侵防护","日志监控"],"domainKey":"服务器安全","pushedAt":"2026-09-11T20:05:13Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/fail2ban/fail2ban","ts":"2026-09-11T20:05:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"avhz/RustQuant","owner":"avhz","name":"RustQuant","desc":"Rust library for quantitative finance.","summaryCn":"用Rust写的金融量化库，帮你快速做策略回测和风险分析","reasonCn":"RustQuant 是基于 Rust 语言的开源金融量化框架，提供时间序列、回测、风险模型、资产配置等模块，支持高性能计算和并发执行。它实现了常见的因子模型、均值方差优化、蒙特卡罗模拟，兼容 CSV、Parquet 等数据源，方便研究员快速搭建交易策略并进行严谨统计检验。","stars":1819,"starGrowth":0,"createdAt":"2022-08-31T10:03:24Z","silentRounds":6,"language":"Rust","topics":["mathematics","trading","finance"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["量化金融","策略回测","风险建模","Rust生态"],"domainKey":"量化金融","pushedAt":"2026-09-12T12:20:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"trading","source":"github","weight":0.7},{"name":"finance","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/avhz/RustQuant","ts":"2026-09-12T12:20:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aster94/logic-analyzer","owner":"aster94","name":"logic-analyzer","desc":"Logic Analyzer, for Arduino, AVR, ESP8266 and STM32 with a very nice working processing interface, you could run it also on any Android device.","summaryCn":"手机变逻辑分析仪，单线测出电路信号时序秘密","reasonCn":"基于Arduino等单片机的轻量级逻辑分析仪，核心是用软件算法在资源有限的MCU上捕获数字信号。它避开了昂贵硬件，通过优化采样和数据处理，让树莓派或安卓手机成为便携示波器，解决了嵌入式开发中调试时序难题的痛点。","stars":396,"starGrowth":0,"createdAt":"2016-11-25T10:28:23Z","silentRounds":6,"language":"C++","topics":["arduino"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.7,"domainTags":["逻辑分析仪","嵌入式调试","Arduino","Android可视化"],"domainKey":"逻辑分析仪","pushedAt":"2026-09-12T10:10:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/aster94/logic-analyzer","ts":"2026-09-12T10:10:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Capsize-Games/airunner","owner":"Capsize-Games","name":"airunner","desc":"Offline inference engine for art, real-time voice conversations, LLM powered chatbots and automated workflows","summaryCn":"本地一键部署AI绘画和语音对话的超级工作台","reasonCn":"Airunner 的核心架构非常巧妙。底层它依赖 Python 生态，利用 PyTorch 进行模型加载和推理，但通过 PyQt5 构建了高性能的桌面 GUI。它支持多种主流的生成式模型，包括 Stable Diffusion 1.5/2.1、Kandinsky 2 等，能处理图像生成、图像编辑、局部重绘等任务。","stars":1317,"starGrowth":1,"createdAt":"2023-03-21T19:14:03Z","silentRounds":16,"language":"Python","topics":["chatbot"],"aiDims":["AI基础设施","AI应用","多模态Agent"],"zone":"AI","funScore":1,"domainTags":["本地部署","AI 绘画","语音交互"],"domainKey":"本地部署","pushedAt":"2026-09-23T05:21:51Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Capsize-Games/airunner","ts":"2026-09-23T05:21:51Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"carla-simulator/carla","owner":"carla-simulator","name":"carla","desc":"Open-source simulator for autonomous driving research.","summaryCn":"简单来说，这个项目就像是一个专门给自动驾驶汽车开的、极其逼真的电脑游戏","reasonCn":"技术架构上，它采用了经典的服务器-客户端模式。底层基于虚幻引擎4进行物理和图形渲染，能高度逼真地模拟激光雷达、摄像头、毫米波雷达等各种传感器的物理特性。它支持OpenDRIVE标准的地图格式，能导入真实的城市道路数据。","stars":14420,"starGrowth":3,"createdAt":"2017-10-24T09:06:23Z","silentRounds":0,"language":"C++","topics":["deep-learning"],"aiDims":["AI基础设施","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["自动驾驶","仿真器","传感器模拟"],"domainKey":"自动驾驶","pushedAt":"2026-09-23T17:30:08Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/carla-simulator/carla","ts":"2026-09-23T17:30:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dharmx/walls","owner":"dharmx","name":"walls","desc":"","summaryCn":"一个涵盖二十多种细分风格的自动随机展示高清壁纸资源仓库，方便快速浏览。","reasonCn":"这个仓库并不展示所有文件，而是通过随机脚本算法，每次只从每个风格分类中提取前两张图片进行展示。这种设计极大地降低了海量素材的浏览压力，让静态的资源页面具备了动态探索的趣味性。用户无需注册或安装客户端，直接在网页上刷新即可发现不同配色与构图的高清屏幕背景图。它适合想要寻找个性化桌面搭配、或者纯粹享受视觉浏览过程的普通用户，是一个轻量级且体验友好的数字美学集。","stars":9054,"starGrowth":11,"createdAt":"2021-09-23T18:58:34Z","silentRounds":0,"language":"","topics":["wallpapers","nord","anime","dracula","portrait","everblush","everforest","gruvbox","material","ocean","radium","consious","javacafe"],"aiDims":["非AI-好玩"],"zone":"资源","funScore":1,"domainTags":["壁纸","随机展示","桌面美学","图片浏览"],"domainKey":"壁纸","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"wallpapers","source":"github","weight":0.7},{"name":"nord","source":"github","weight":0.7},{"name":"anime","source":"github","weight":0.7},{"name":"dracula","source":"github","weight":0.7},{"name":"portrait","source":"github","weight":0.7},{"name":"everblush","source":"github","weight":0.7},{"name":"everforest","source":"github","weight":0.7},{"name":"gruvbox","source":"github","weight":0.7},{"name":"material","source":"github","weight":0.7},{"name":"ocean","source":"github","weight":0.7},{"name":"radium","source":"github","weight":0.7},{"name":"consious","source":"github","weight":0.7},{"name":"javacafe","source":"github","weight":0.7}],"aiScore":0.38,"source":"search","url":"https://github.com/dharmx/walls","ts":"2026-09-14T23:04:28.868Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.036644797879390324,"copyOk":true},{"repo":"ExistentialAudio/BlackHole","owner":"ExistentialAudio","name":"BlackHole","desc":"","summaryCn":"让 macOS 应用间零延迟互通音频的驱动程序。","reasonCn":"BlackHole 是一款运行在 macOS 上的虚拟音频驱动，它的核心作用是让不同的应用程序能够像接驳了物理设备一样直接交换声音。它能消除音频传输过程中的额外延迟，解决用户在多应用间传音时的卡顿痛点，特别适合需要将电脑发出的声音实时导入录音软件、虚拟机或其他应用的用户。","stars":19798,"starGrowth":6,"createdAt":"2019-09-27T18:40:46Z","silentRounds":0,"language":"C","topics":["audio","driver","macos","loopback","virtual-audio-cable"],"aiDims":["非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["虚拟音频驱动","macOS","音频路由","直播","零配置"],"domainKey":"虚拟音频驱动","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"audio","source":"github","weight":0.7},{"name":"driver","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"loopback","source":"github","weight":0.7},{"name":"virtual-audio-cable","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.41,"source":"tier","url":"https://github.com/ExistentialAudio/BlackHole","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01890918274573189,"copyOk":true},{"repo":"SeleniumHQ/docker-selenium","owner":"SeleniumHQ","name":"docker-selenium","desc":"Provides a simple way to run Selenium Grid with Chrome, Firefox, and Edge using Container Platform, making it easier to perform browser automation at scale","summaryCn":"一键部署浏览器自动化集群，告别环境配置烦恼，大规模测试更简单","reasonCn":"Selenium官方推出的Docker化解决方案，将Chrome、Firefox、Edge等浏览器打包成标准化容器。你可以一键启动一个完整的Selenium Grid集群，支持多浏览器、多版本并行测试。无需本地安装复杂的环境依赖，在CI/CD流水线中直接使用，大幅提升自动化测试的部署效率和稳定性。","stars":8653,"starGrowth":0,"createdAt":"2014-11-17T17:39:22Z","silentRounds":5,"language":"Go","topics":["devops","kubernetes"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["浏览器自动化","Docker","CI/CD"],"domainKey":"浏览器自动化","pushedAt":"2026-09-18T02:39:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/SeleniumHQ/docker-selenium","ts":"2026-09-18T02:39:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Avarok-Cybersecurity/atlas","owner":"Avarok-Cybersecurity","name":"atlas","desc":"Pure Rust Inference Engine","summaryCn":"想象一下，你开发了一个AI模型，但它在电脑上跑起来像老牛拉破车，这时你就","reasonCn":"技术上，Atlas利用了Rust的零成本抽象和无垃圾回收特性，解决了Python生态在推理时的性能瓶颈。它对Transformer架构进行了针对性的算子优化，通过精简内存路径和高效的SIMD指令集利用，使模型推导延迟大幅缩减。这种设计比传统的PyTorch运行时更小巧，且在跨平台编译上有着天然优势","stars":698,"starGrowth":0,"createdAt":"2026-05-05T16:59:33Z","silentRounds":2,"language":"Rust","topics":["transformers"],"aiDims":["推理引擎","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["推理引擎","Transformer","Rust"],"domainKey":"推理引擎","pushedAt":"2026-09-23T21:41:52Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/Avarok-Cybersecurity/atlas","ts":"2026-09-23T21:41:52Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fuxicodex/Fuxi","owner":"fuxicodex","name":"Fuxi","desc":"FuXi is a fast, self-contained AI coding agent that lives in your terminal — edit code, run commands, and drive tools, with cost-aware routing across LLM providers.","summaryCn":"藏在黑底白字终端里既聪明又懂得精打细算的独立编程助手","reasonCn":"在底层架构上，它采用了高度内聚的自包含设计，不需要依赖复杂的外部图形界面。其最亮眼的技术特性是具备成本感知的智能路由机制。它能够根据你当前的任务难度、紧急程度以及不同大模型API的实时价格，自动把请求分发给最划算、最合适的模型，既不耽误干活，又能帮你把花销精打细算","stars":3357,"starGrowth":0,"createdAt":"2026-08-04T07:32:11Z","silentRounds":1,"language":"Python","topics":["ai-agent","productivity"],"aiDims":["代码助手","AI Agent","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI 编程助手","终端工具","代码编辑","成本感知路由","Agent"],"domainKey":"Agent","pushedAt":"2026-09-23T06:24:31Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/fuxicodex/Fuxi","ts":"2026-09-23T06:24:31Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"originalankur/maptoposter","owner":"originalankur","name":"maptoposter","desc":"Transform your favorite cities into beautiful, minimalist designs. MapToPoster lets you create and export visually striking map posters with code.","summaryCn":"用代码把喜欢的城市变成极简艺术海报的超酷工具","reasonCn":"基于OpenStreetMap数据，用Python脚本将城市地图简化为矢量SVG海报。支持自定义配色、线条风格，内置巴黎东京等热门城市模板。可导出SVG或PNG高清格式，GitHub获1.3万星标，社区活跃持续迭代，是设计师和开发者的创意利器。","stars":14147,"starGrowth":7,"createdAt":"2026-01-08T12:29:35Z","silentRounds":0,"language":"Python","topics":["design"],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["城市地图","矢量海报","生成艺术","Python"],"domainKey":"城市地图","pushedAt":"2026-08-06T07:57:04Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/originalankur/maptoposter","ts":"2026-08-06T07:57:04Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.026981296387926772,"copyOk":false},{"repo":"octobercms/october","owner":"octobercms","name":"october","desc":"Self-hosted CMS platform based on the Laravel PHP Framework.","summaryCn":"自己搭建的WordPress替代方案，灵活可控免被坑","reasonCn":"October CMS是一个基于Laravel PHP框架的自托管内容管理系统，适合不想被WordPress插件生态绑架、希望完全掌控代码和数据的技术用户。它采用插件和主题分离架构，支持自定义字段类型、后端权限管理和REST API，界面简洁现代化。","stars":11147,"starGrowth":1,"createdAt":"2013-11-26T23:30:43Z","silentRounds":4,"language":"PHP","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["CMS","内容管理","PHP框架","自托管"],"domainKey":"CMS","pushedAt":"2026-09-22T21:04:53Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/octobercms/october","ts":"2026-09-22T21:04:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sindresorhus/awesome","owner":"sindresorhus","name":"awesome","desc":"😎 Awesome lists about all kinds of interesting topics","summaryCn":"全球最大的Awesome列表合集，收集了无数精选资源主题","reasonCn":"这是sindresorhus维护的awesome列表总仓库，汇集了数千个awesome列表，涵盖编程、AI、设计、生活等几乎所有领域。每个子列表都是社区精选的高质量资源集合，是程序员必备的资源导航站，star接近50万，是最热门的开源仓库之一。","stars":509011,"starGrowth":313,"createdAt":"2014-07-11T13:42:37Z","silentRounds":0,"language":"","topics":["awesome"],"aiDims":["学习资源"],"zone":"资源","funScore":0,"domainTags":["Awesome列表","资源索引","导航站","精选合集"],"domainKey":"Awesome列表","pushedAt":"2026-09-02T02:10:47Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.4,"source":"trending","url":"https://github.com/sindresorhus/awesome","ts":"2026-09-02T02:10:47Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30184475384618403,"copyOk":false},{"repo":"VoltAgent/awesome-claude-code-subagents","owner":"VoltAgent","name":"awesome-claude-code-subagents","desc":"A collection of 100+ specialized Claude Code subagents covering a wide range of development use cases","summaryCn":"为 Claude Code 量身定做的专业领域小助手","reasonCn":"该项目提供了 100 多个针对 Claude Code 定制的子 Agent 配置，涵盖了从前端调优到后端架构的各种开发场景。通过定义专门的角色职责和操作规程，将复杂的开发流程拆解为多个专业化子任务，显著增强了 AI 处理特定垂直领域问题的专业度。","stars":25269,"starGrowth":27,"createdAt":"2025-07-30T12:31:51Z","silentRounds":0,"language":"Shell","topics":["awesome"],"aiDims":["AI Agent","代码助手","提示工程"],"zone":"AI","funScore":0.5,"domainTags":["AI编程","子代理","代码辅助"],"domainKey":"AI编程","pushedAt":"2026-09-21T08:24:27Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/VoltAgent/awesome-claude-code-subagents","ts":"2026-09-21T08:24:27Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.08420550872610709,"copyOk":false},{"repo":"TanStack/ai","owner":"TanStack","name":"ai","desc":"🤖 Type-safe, provider-agnostic TypeScript AI SDK for streaming chat, tool calling, agents, and multimodal apps across OpenAI, Anthropic, Gemini, React, Vue, Svelte, and Solid.","summaryCn":"做人工智能应用时，前端工程师常被同一件事折磨：模型接口各家不同，回答一","reasonCn":"技术上，它最吸引人的地方是类型友好和接口统一。你可以把不同模型服务商接到同一套抽象里，聊天、流式、工具调用、代理逻辑都能按项目结构组织。对 React、Vue、Svelte、Solid 这类前端场景很友好，因为它关心的不只是“能调模型”，而是“能在真实界面里稳定跑起来”。","stars":3130,"starGrowth":1,"createdAt":"2025-10-08T17:52:48Z","silentRounds":1,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","大语言模型","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI SDK","模型集成","前端"],"domainKey":"AI SDK","pushedAt":"2026-09-23T16:24:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.86,"source":"search","url":"https://github.com/TanStack/ai","ts":"2026-09-23T16:24:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alyssaxuu/screenity","owner":"alyssaxuu","name":"screenity","desc":"The free and privacy-friendly screen recorder with no limits 🎥","summaryCn":"免费无广告无限制的屏幕录制工具，保护隐私不上传数据","reasonCn":"screenity是一款专为追求隐私和极简主义用户设计的屏幕录制工具，完全免费开源无广告无水印。采用原生JavaScript开发直接在浏览器运行无需安装任何软件，支持区域、全屏、窗口三种录制模式。最大亮点是零数据追踪所有录制内容本地处理不上传服务器，导出MP4格式兼容主流播放器。","stars":18719,"starGrowth":3,"createdAt":"2020-11-01T15:16:31Z","silentRounds":0,"language":"JavaScript","topics":["design","editor"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["屏幕录制","浏览器扩展","隐私保护","开源工具"],"domainKey":"屏幕录制","pushedAt":"2026-09-22T15:42:24Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"editor","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/alyssaxuu/screenity","ts":"2026-09-22T15:42:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"0xInfection/Awesome-WAF","owner":"0xInfection","name":"Awesome-WAF","desc":"Everything about Web Application Firewalls (WAFs) from Security Standpoint! 🔥","summaryCn":"网络安全防火墙的开源资源大全，7千多星的安全爱好者必看清单","reasonCn":"这个项目系统收集了Web应用防火墙的完整开源生态资源，涵盖各类WAF工具的源码实现、配置文件模板、OWASP规则集和最佳实践文档。从ModSecurity到云厂商WAF方案，从入门教程到企业级部署指南，一站式满足安全团队的学习和搭建需求。适合渗透测试人员、运维工程师和安全研究人员参考。","stars":7629,"starGrowth":2,"createdAt":"2019-01-08T04:57:06Z","silentRounds":0,"language":"Python","topics":["awesome"],"aiDims":["安全工具","学习资源"],"zone":"资源","funScore":0,"domainTags":["Web应用防火墙","安全工具","渗透测试","安全资源","OWASP"],"domainKey":"Web应用防火墙","pushedAt":"2026-08-26T12:48:34Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/0xInfection/Awesome-WAF","ts":"2026-08-26T12:48:34Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"YouROK/TorrServer","owner":"YouROK","name":"TorrServer","desc":"Torrent stream server","summaryCn":"把种子文件变成在线流媒体，边下边看体验拉满","reasonCn":"TorrServer是一个BT种子流媒体服务器，支持边下载边播放，让你在线看电影电视剧而不用等整个文件下载完。兼容kodi、potplayer等主流播放器，支持多种索引源自动查找资源。Go语言编写，性能出色，可以部署在树莓派等低功耗设备上，是小巧但强大的家庭媒体中心解决方案。","stars":3057,"starGrowth":5,"createdAt":"2018-08-29T09:21:00Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["BT流媒体","种子下载","家庭媒体","流媒体服务器"],"domainKey":"BT流媒体","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/YouROK/TorrServer","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01990677134445535,"copyOk":false},{"repo":"roboflow/supervision","owner":"roboflow","name":"supervision","desc":"We write your reusable computer vision tools. 💜","summaryCn":"一行代码搞定AI视觉检测结果的漂亮可视化","reasonCn":"这个Python库专门解决计算机视觉项目中的可视化难题，能自动绘制检测框、跟踪轨迹、分类标签和置信度分数，支持YOLO等多种主流模型。它把复杂的绘图逻辑封装成简洁接口，让开发者不用手写matplotlib就能输出专业级结果展示。","stars":50995,"starGrowth":10,"createdAt":"2022-11-28T14:08:44Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["计算机视觉","目标检测可视化","视频追踪","YOLO","Python库"],"domainKey":"计算机视觉","pushedAt":"2026-09-22T15:42:26Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/roboflow/supervision","ts":"2026-09-22T15:42:26Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030588292969899008,"copyOk":true},{"repo":"pi-hole/pi-hole","owner":"pi-hole","name":"pi-hole","desc":"A black hole for Internet advertisements","summaryCn":"给路由器装个“黑洞”，全网广告自动拦截。","reasonCn":"它的核心优势在于“全网拦截”和“极简部署”。它基于 Linux，可以运行在树莓派、VPS 或路由器上。它内置了几个庞大的广告域名列表（如 StevenBlack、Firebog），并支持从 GitHub 获取社区维护的列表。它还提供了一个 Web 控制面板，让你实时查看被拦截的请求数量、Top 域名等统计数据","stars":61015,"starGrowth":0,"createdAt":"2014-06-08T15:02:55Z","silentRounds":1,"language":"Shell","topics":["raspberry-pi"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["DNS拦截","广告过滤","网络管理","树莓派","隐私保护"],"domainKey":"DNS拦截","pushedAt":"2026-09-23T16:43:53Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/pi-hole/pi-hole","ts":"2026-09-23T16:43:53Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kane50613/takumi","owner":"kane50613","name":"takumi","desc":"Render OG images and paged PDFs from JSX, HTML, and CSS. No headless browser. Runs on Node.js, Cloudflare Workers, browsers, and Rust.","summaryCn":"用JSX写界面，秒生成OG图片和PDF，无需浏览器，多平台都能跑","reasonCn":"Takumi是一个基于Rust的渲染引擎，能将JSX、HTML和CSS直接转换为OG图片和分页PDF，完全不需要无头浏览器。支持Node.js、Cloudflare Workers、浏览器和Rust环境运行，跨平台能力极强，部署成本低、资源占用少，特别适合需要批量生成图片的团队。","stars":3020,"starGrowth":5,"createdAt":"2025-06-01T08:19:38Z","silentRounds":0,"language":"Rust","topics":["satori","jsx","og-image","cloudflare-workers","gif","html-to-image","image-generation","opengraph","rust","wasm","pdf","pdf-generation","html-to-pdf","invoice","svg","opengraph-images","pdf-generator","puppeteer-alternative","social-cards"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["PDF生成","服务端渲染","Rust引擎","JSX转换"],"domainKey":"PDF生成","pushedAt":"2026-08-09T21:14:47.416Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"satori","source":"github","weight":0.7},{"name":"jsx","source":"github","weight":0.7},{"name":"og-image","source":"github","weight":0.7},{"name":"cloudflare-workers","source":"github","weight":0.7},{"name":"gif","source":"github","weight":0.7},{"name":"html-to-image","source":"github","weight":0.7},{"name":"image-generation","source":"github","weight":0.7},{"name":"opengraph","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"pdf","source":"github","weight":0.7},{"name":"pdf-generation","source":"github","weight":0.7},{"name":"html-to-pdf","source":"github","weight":0.7},{"name":"invoice","source":"github","weight":0.7},{"name":"svg","source":"github","weight":0.7},{"name":"opengraph-images","source":"github","weight":0.7},{"name":"pdf-generator","source":"github","weight":0.7},{"name":"puppeteer-alternative","source":"github","weight":0.7},{"name":"social-cards","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/kane50613/takumi","ts":"2026-08-09T21:14:47.416Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.019966887417218544,"copyOk":false},{"repo":"meowtec/imgo","owner":"meowtec","name":"imgo","desc":"🖼️ Image optimization app for macOS, Windows and Linux.","summaryCn":"图片压缩神器，一键无损缩小体积，界面友好上手快","reasonCn":"imgo 是一款跨平台图片优化工具，支持 macOS、Windows 和 Linux。核心亮点是批量处理能力强，拖拽即可导入多张图片自动压缩。内置无损/有损双模式，用户可按需调整质量参数。界面简洁现代，适合设计师和开发者日常使用，也能批量导出 WebP/AVIF 等现代格式。","stars":4420,"starGrowth":0,"createdAt":"2017-05-20T15:42:44Z","silentRounds":8,"language":"TypeScript","topics":["desktop"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["图片压缩","桌面应用","图片优化","跨平台"],"domainKey":"图片压缩","pushedAt":"2026-09-19T07:43:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/meowtec/imgo","ts":"2026-09-19T07:43:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"oracle/dotnet-db-samples","owner":"oracle","name":"dotnet-db-samples","desc":".NET code samples for Oracle database developers #OracleDotNet","summaryCn":"官方出品的C#连接甲骨文数据库开发示例合集，新手写代码不再抓瞎。","reasonCn":"技术层面来看，它不仅仅是简单的增删改查，还囊括了最新版本的ODP.NET驱动特性。支持异步编程模型、连接池管理、JSON数据处理、高级安全特性以及Entity Framework Core的高级映射。它通过清晰的项目结构展示了依赖注入和最佳实践，避免了新手常踩的内存泄漏和连接未关闭等性能坑。","stars":434,"starGrowth":0,"createdAt":"2014-07-31T17:26:32Z","silentRounds":8,"language":"C#","topics":["tutorial"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["数据库","C#","示例代码"],"domainKey":"数据库","pushedAt":"2026-09-11T19:17:07Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/oracle/dotnet-db-samples","ts":"2026-09-11T19:17:07Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vectordotdev/vector","owner":"vectordotdev","name":"vector","desc":"A high-performance observability data pipeline.","summaryCn":"把零散日志变成超快查询大杀器，运维神器！","reasonCn":"Vector 是一个高性能的观测性数据处理管道，用 Rust 编写，专为收集、转换和加载数据而设计。它能在几秒钟内处理数百万条日志事件，并将它们无缝推送到 Elasticsearch、Prometheus 等目标系统。Vector 支持丰富的数据源和输出端，能极大简化日志聚合、监控和告警的搭建过程。","stars":22602,"starGrowth":3,"createdAt":"2018-08-27T16:57:34Z","silentRounds":0,"language":"Rust","topics":["monitoring"],"aiDims":["AI基础设施","开发工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["日志监控","数据处理","Rust","可观测性"],"domainKey":"日志监控","pushedAt":"2026-09-23T21:52:51Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/vectordotdev/vector","ts":"2026-09-23T21:52:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zaproxy/zaproxy","owner":"zaproxy","name":"zaproxy","desc":"The ZAP by Checkmarx Core project","summaryCn":"一个强大的安全工具，用于保护网络流量","reasonCn":"zaproxy是网络安全领域的一个重要开源项目，它提供了全面的网络安全解决方案，包括防火墙、入侵检测和预防等。","stars":15807,"starGrowth":3,"createdAt":"2015-06-03T16:55:01Z","silentRounds":0,"language":"Java","topics":["security"],"aiDims":["AI基础设施"],"zone":"工具","funScore":0,"domainTags":["网络安全","漏洞扫描","渗透测试","代理工具","安全审计"],"domainKey":"网络安全","pushedAt":"2026-09-23T16:42:40Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/zaproxy/zaproxy","ts":"2026-09-23T16:42:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DanielMartensson/Open-SAE-J1939","owner":"DanielMartensson","name":"Open-SAE-J1939","desc":"SAE J1939 protocol free to use for embedded systems or PC with CAN-bus","summaryCn":"开源J1939 CAN库，能在Arduino或PC上实现汽车总线通信。","reasonCn":"Open‑SAE‑J1939是用纯C实现的SAE J1939协议栈，支持PGN、SPN解析及多帧传输，可直接移植到Arduino、STM32等MCU，亦可在带CAN接口的PC上运行，提供报文构建、发送与接收API，帮助开发者快速搭建车载网络调试或诊断工具。","stars":651,"starGrowth":0,"createdAt":"2021-07-14T17:54:40Z","silentRounds":9,"language":"C","topics":["arduino"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["SAE J1939","CAN总线","车载网络","嵌入式C","汽车诊断"],"domainKey":"SAE J1939","pushedAt":"2026-09-11T17:38:20Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/DanielMartensson/Open-SAE-J1939","ts":"2026-09-11T17:38:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"STranslate/STranslate","owner":"STranslate","name":"STranslate","desc":"","summaryCn":"STranslate 2.0是一款基于WPF框架开发的Windows桌面应用程序，主要聚焦于自","reasonCn":"从技术实现来看，项目采用C#语言开发，利用WPF技术栈实现了良好的跨版本Windows系统兼容性。虽然底层依赖于第三人方的翻译API和OCR服务，但STranslate通过统一的界面封装，屏蔽了后端接口的复杂性，用户只需在设置中填入相应的密钥（如需要）或选择预设的引擎，即可开始使用。","stars":8092,"starGrowth":8,"createdAt":"2022-12-17T13:40:24Z","silentRounds":0,"language":"C#","topics":["wpf","ocr","plugin","translate","stranslate"],"aiDims":["非AI-实用","效率工具","工具调用"],"zone":"工具","funScore":0,"domainTags":["翻译","OCR","截图翻译","WPF","桌面应用"],"domainKey":"翻译","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"wpf","source":"github","weight":0.7},{"name":"ocr","source":"github","weight":0.7},{"name":"plugin","source":"github","weight":0.7},{"name":"translate","source":"github","weight":0.7},{"name":"stranslate","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/STranslate/STranslate","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02696589223924864,"copyOk":false},{"repo":"VueTorrent/VueTorrent","owner":"VueTorrent","name":"VueTorrent","desc":"The sleekest looking WEBUI for qBittorrent made with Vuejs!","summaryCn":"颜值超高的BT下载Web界面用Vuejs打造让下载管理变得赏心悦目","reasonCn":"技术上VueTorrent采用Vue.js框架构建单页面应用，通过WebSocket与qBittorrent-nox后端通信实现实时数据同步。核心功能包括下载列表管理、分类标签、搜索集成、队列优先级控制、速度限制、IP过滤等。界面支持深色和浅色主题，布局采用响应式设计适配各种屏幕尺寸。","stars":6969,"starGrowth":1,"createdAt":"2019-04-15T14:51:45Z","silentRounds":2,"language":"Vue","topics":["frontend"],"aiDims":["非AI-实用","效率工具","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["BT下载","文件传输","Web界面"],"domainKey":"BT下载","pushedAt":"2026-09-20T03:31:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/VueTorrent/VueTorrent","ts":"2026-09-20T03:31:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ray-project/ray","owner":"ray-project","name":"ray","desc":"Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.","summaryCn":"给AI模型训练加速的分布式引擎，多GPU多线程一把梭","reasonCn":"Ray是专为AI设计的分布式计算框架，底层Ray Core负责跨机器任务调度，上层提供Ray Train分布式训练、Ray Serve模型部署、Ray RLlib强化学习、Ray Data数据处理等组件。独特优势是Python代码自动并行化无需大改，原生支持LLM推理加速，是AI基础设施领域的核心项目之一。","stars":43898,"starGrowth":5,"createdAt":"2016-10-25T19:38:30Z","silentRounds":0,"language":"Python","topics":["ml","llm"],"aiDims":["AI基础设施","模型与训练","大语言模型"],"zone":"工具","funScore":0,"domainTags":["分布式计算","AI框架","机器学习"],"domainKey":"分布式计算","pushedAt":"2026-09-23T21:02:58Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/ray-project/ray","ts":"2026-09-23T21:02:58Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01534170121645633,"copyOk":false},{"repo":"lbjlaq/Antigravity-Manager","owner":"lbjlaq","name":"Antigravity-Manager","desc":"Professional Antigravity Account Manager &amp; Switcher. One-click seamless account switching for Antigravity Tools. Built with Tauri v2 + React (Rust).专业的 Antigravity 账号管理与切换工具。为 Antigravity 提供一键无缝账号切换功能。","summaryCn":"一键搞定 Antigravity 账号无缝切换，连登党必备的神器","reasonCn":"Antigravity 是一款功能强大的浏览器扩展，但多账号管理常令人头疼。本项目基于 Rust 和 Tauri v2 开发，专为解决账号切换繁琐问题打造。通过一键无缝切换，极大提升了多账号操作效率，特别适合需要频繁切换账号的用户。其高性能架构确保了切换的流畅性，是提升工作效率的实用工具。","stars":31673,"starGrowth":60,"silentRounds":0,"language":"Rust","topics":["account-manager","antigravity"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["浏览器扩展","账号管理","自动化工具"],"domainKey":"浏览器扩展","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"account-manager","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/lbjlaq/Antigravity-Manager","ts":"2026-09-23T07:35:20.006Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.18568307391153346,"copyOk":false},{"repo":"nmichlo/ruck","owner":"nmichlo","name":"ruck","desc":"🧬 Modularised Evolutionary Algorithms For Python with Optional JIT and Multiprocessing (Ray) support. Inspired by PyTorch Lightning","summaryCn":"第一段：在很多参数优化、特征搜索甚至神经网络结构搜索的场景里，传统的梯","reasonCn":"第一段：在很多参数优化、特征搜索甚至神经网络结构搜索的场景里，传统的梯度下降方法往往受限于局部最优，进化算法以其全局搜索的特性成为重要补充。但实现一个既高效又易用的进化框架并不容易，常常需要自己写循环、管理并行、记录日志。","stars":52,"starGrowth":0,"createdAt":"2021-09-21T12:04:27Z","silentRounds":9,"language":"Python","topics":["raylib"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["进化算法","遗传算法","超参数优化","分布式计算","Numba加速"],"domainKey":"进化算法","pushedAt":"2026-09-13T17:29:11Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"raylib","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/nmichlo/ruck","ts":"2026-09-13T17:29:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0jonjo/calcpace","owner":"0jonjo","name":"calcpace","desc":"A Ruby gem for runners: pace, time, and distance calculations, unit conversions, race predictions, GPS track analysis, age grading, VO2max estimation, and training zones.","summaryCn":"跑步党必备的 Ruby 库，算配速、测 VO2max 就靠它。","reasonCn":"它的核心能力包括：基于时间和距离计算配速，支持区间划分；分析 GPS 轨迹文件，计算海拔变化和爬升距离；根据年龄和成绩进行年龄分级评分，看看自己在同年龄段里处于什么水平；估算最大摄氧量，评估心肺功能。甚至还能预测比赛配速，让你在比赛前心里有底。","stars":14,"starGrowth":0,"createdAt":"2022-01-20T16:02:46Z","silentRounds":11,"language":"Ruby","topics":["fitness"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["跑步","配速计算","体能评估"],"domainKey":"跑步","pushedAt":"2026-09-06T20:31:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/0jonjo/calcpace","ts":"2026-09-06T20:31:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"clshortfuse/renodx","owner":"clshortfuse","name":"renodx","desc":"Renovation Engine for DirectX Games","summaryCn":"老游戏在新系统上也能流畅跑，DirectX增强引擎让帧率起飞","reasonCn":"它的核心技术是在DirectX API层做拦截和转发，把老游戏的渲染请求转换成现代图形API可以高效执行的指令。着色器部分进行了重写和优化，减少了兼容层带来的性能损耗。相比同类工具，它在光追支持和特定游戏优化上做了专项改进，针对一些 popular 的老游戏有实测帧率提升数据，效果经","stars":4314,"starGrowth":4,"createdAt":"2024-02-05T15:39:11Z","silentRounds":0,"language":"HLSL","topics":["directx","games","hdr","mod"],"aiDims":["非AI-好玩","游戏","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["DirectX","游戏兼容性","图形增强","老游戏修复"],"domainKey":"DirectX","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"directx","source":"github","weight":0.7},{"name":"games","source":"github","weight":0.7},{"name":"hdr","source":"github","weight":0.7},{"name":"mod","source":"github","weight":0.7},{"name":"HLSL","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/clshortfuse/renodx","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"geekjourneyx/md2wechat-skill","owner":"geekjourneyx","name":"md2wechat-skill","desc":"Markdown to WeChat CLI | 一键排版发布到微信公众号：支持 40+ 排版样式和专业主题 、AI 配图 、批量发布 、多账号管理","summaryCn":"Markdown一键转微信排版，AI配图还能批量发稿的神器","reasonCn":"这是一个基于Go开发的命令行工具，能将Markdown文本一键排版并发布到微信公众号。它内置了40多种专业排版样式，支持AI自动生成配图，还具备批量发布和多账号管理功能。对于自媒体运营者和开发者来说，它解决了排版繁琐和发布效率低的痛点。","stars":3667,"starGrowth":0,"createdAt":"2026-01-11T06:36:13Z","silentRounds":1,"language":"Go","topics":["ai-agent"],"aiDims":["AI应用","AI写作"],"zone":"AI","funScore":0.7,"domainTags":["微信公众号","Markdown排版","AI Agent","内容创作"],"domainKey":"AI Agent","pushedAt":"2026-09-23T14:24:44Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/geekjourneyx/md2wechat-skill","ts":"2026-09-23T14:24:44Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"openai/codex-security","owner":"openai","name":"codex-security","desc":"OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https://www.npmjs.com/package/@openai/codex-security","summaryCn":"OpenAI出品的代码安全警察，专治各种漏洞，让你写代码不再背锅","reasonCn":"这是OpenAI官方推出的代码安全扫描工具，集成Codex AI能力，自动识别代码中的安全漏洞并提供修复建议。支持CLI和SDK两种使用方式，开发者可直接集成到CI/CD流程中。相比传统安全工具，它更智能，不仅能发现问题，还能生成修复代码，显著降低安全漏洞风险。","stars":10826,"starGrowth":6,"createdAt":"2026-07-13T22:00:13Z","silentRounds":0,"language":"TypeScript","topics":["ai-security","application-security","cli","code-scanning","codex","codex-security","cybersecurity","devsecops","nodejs","npm","openai","sdk","security","typescript","vulnerability-scanning"],"aiDims":["安全工具","AI Agent","代码助手"],"zone":"AI","funScore":0,"domainTags":["代码安全","漏洞扫描","AI辅助","OpenAI","DevSecOps"],"domainKey":"代码安全","pushedAt":"2026-09-23T21:43:23Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-security","source":"github","weight":0.7},{"name":"application-security","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"code-scanning","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"codex-security","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"devsecops","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"npm","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"sdk","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"vulnerability-scanning","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"bigbro","url":"https://github.com/openai/codex-security","ts":"2026-09-23T21:43:23Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.029493995935710327,"copyOk":false},{"repo":"gophish/gophish","owner":"gophish","name":"gophish","desc":"Open-Source Phishing Toolkit","summaryCn":"钓鱼攻击是企业安全最大的漏洞之一。这个项目是一个功能完整的开源钓鱼测试","reasonCn":"技术上它支持高度自定义的钓鱼邮件模板，可以模仿任意品牌。核心功能包括链接追踪、表单收集、多语言支持和详细攻击报告。相比商业方案，它完全免费开源，可私有化部署。 安装需要Go环境和基础Web服务器。通过Web界面创建钓鱼活动、配置模板、发送测试邮件，系统自动追踪点击和提交","stars":14244,"starGrowth":4,"createdAt":"2013-11-18T23:26:43Z","silentRounds":0,"language":"Go","topics":["gophish","phishing","golang","security"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["安全测试","钓鱼模拟","渗透测试","Web平台"],"domainKey":"安全测试","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"gophish","source":"github","weight":0.7},{"name":"phishing","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/gophish/gophish","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"elkowar/eww","owner":"elkowar","name":"eww","desc":"ElKowars wacky widgets","summaryCn":"用Rust写配置就能做出超酷桌面小部件，极客必备","reasonCn":"基于Rust开发的桌面小部件引擎，通过简洁的配置文件即可创建实时显示系统状态、时间、音乐播放等信息的桌面面板。支持自定义布局和样式，可绑定任意命令输出作为动态内容，性能极高且资源占用极低，是Linux桌面美化爱好者的利器。","stars":12680,"starGrowth":3,"createdAt":"2020-09-19T20:07:39Z","silentRounds":0,"language":"Rust","topics":["desktop"],"aiDims":["非AI-好玩","可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["桌面小部件","Wayland","Linux","系统监控"],"domainKey":"桌面小部件","pushedAt":"2026-07-17T07:29:44Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/elkowar/eww","ts":"2026-07-17T07:29:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mahmoud/awesome-python-applications","owner":"mahmoud","name":"awesome-python-applications","desc":"💿 Free software that works great, and also happens to be open-source Python. ","summaryCn":"Python写的开源应用软件大全， games到设计工具应有尽有","reasonCn":"这是一个精心整理的Python开源应用合集，收录了各种用Python开发的实用软件和有趣项目，涵盖游戏、设计工具、生产力工具、数据处理等多个类别。每个项目都有简介和链接，是发现Python生态优秀应用的宝藏目录，适合寻找灵感或学习参考。","stars":18053,"starGrowth":3,"createdAt":"2018-11-14T07:36:44Z","silentRounds":0,"language":"Jupyter Notebook","topics":["game"],"aiDims":["非AI-好玩","学习资源","非AI-实用"],"zone":"资源","funScore":0.7,"domainTags":["Python应用","开源软件","应用合集","编程生态"],"domainKey":"Python应用","pushedAt":"2026-07-11T16:46:22Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/mahmoud/awesome-python-applications","ts":"2026-07-11T16:46:22Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google-research/timesfm","owner":"google-research","name":"timesfm","desc":"TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.","summaryCn":"Google开源的时间序列预测基础模型，提供大模型级预测能力。","reasonCn":"TimesFM是Google Research推出的时间序列预测基础模型。它解决了传统模型难以处理长序列数据的问题，能通过预训练掌握海量时间数据规律，支持从短到中等长度的精准预测。该模型不仅开源，还已被集成到BigQuery、Google Sheets等企业级产品中，为不同规模的数据分析提供支持。","stars":33483,"starGrowth":72,"createdAt":"2024-04-29T21:26:26Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","推理引擎","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["时间序列预测","大模型","数据科学"],"domainKey":"时间序列预测","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/google-research/timesfm","ts":"2026-09-18T10:23:42.801Z","category":"learning","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.22245103485350776,"copyOk":true},{"repo":"manticoresoftware/manticoresearch","owner":"manticoresoftware","name":"manticoresearch","desc":"Open-source search database for full-text, vector, and hybrid search with real-time indexing and SQL.","summaryCn":"能在一秒内从海量数据里精准捞出想要内容的数据库","reasonCn":"它的底层是用纯C++写的，把性能压榨到了极致。它最大的优势是把传统的文字检索和现代的向量检索（也就是AI常用来找语义相似度的技术）完美结合在了一起。你既可以用传统的关键词去搜，也可以用大模型生成的向量去做语义模糊搜索。","stars":12029,"starGrowth":2,"createdAt":"2017-06-28T01:15:30Z","silentRounds":0,"language":"C++","topics":["database","api"],"aiDims":["向量数据库","AI搜索"],"zone":"工具","funScore":0,"domainTags":["搜索数据库","向量检索","全文检索"],"domainKey":"搜索数据库","pushedAt":"2026-09-23T14:20:53Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/manticoresoftware/manticoresearch","ts":"2026-09-23T14:20:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OshekharO/awesome-entertainment","owner":"OshekharO","name":"awesome-entertainment","desc":"List of awesome apps and website for entertainment.","summaryCn":"这个开源库为你整理了各种宝藏娱乐网站与神仙追剧摸鱼软件资源。","reasonCn":"该项目是一个汇集了全球优质数字娱乐资源的精选清单，涵盖在线动漫观赏、免费影视流媒体、数字音乐播放器、独立小游戏及各类电子书阅读工具等。所有推荐项目均经过分类归纳，并附带访问链接与开源替代方案，帮助用户迅速挖掘有趣高效的休闲娱乐应用。","stars":250,"starGrowth":1,"createdAt":"2022-02-24T12:29:07Z","silentRounds":3,"language":"CSS","topics":["anime"],"aiDims":["学习资源"],"zone":"资源","funScore":0,"domainTags":["娱乐网站","动漫","游戏","电子书"],"domainKey":"娱乐网站","pushedAt":"2026-09-17T11:03:42Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/OshekharO/awesome-entertainment","ts":"2026-09-17T11:03:42Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"joelbqz/writer-computer","owner":"joelbqz","name":"writer-computer","desc":"Local-first desktop markdown editor","summaryCn":"本地优先Markdown编辑器，专注沉浸式写作无干扰","reasonCn":"基于Electron构建的本地优先桌面Markdown编辑器，所有数据完全存储在本地设备，不依赖云端服务。支持实时预览、暗色主题、自定义CSS样式，提供流畅的写作体验。适合技术文档撰写、笔记整理、博客创作，注重隐私安全的用户福音。","stars":1581,"starGrowth":2,"createdAt":"2026-04-30T15:14:09Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Markdown编辑器","本地优先","写作工具"],"domainKey":"Markdown编辑器","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/joelbqz/writer-computer","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"integrations/slack","owner":"integrations","name":"slack","desc":"","summaryCn":"现代软件开发往往依赖于多款协作工具的高效配合，其中GitHub作为代码托管的","reasonCn":"除了基础的通知推送功能外，该指南还详细讲解了如何进行组织或仓库级别的订阅管理、自定义通知过滤规则、处理部署提醒以及在聊天界面中直接对GitHub内容采取行动。它支持复杂的团队权限控制企业网格环境，并提供了定时提醒等实用功能，帮助研发人员更好地安排项目进度，促进跨部门","stars":3458,"starGrowth":1,"createdAt":"2017-08-15T23:32:28Z","silentRounds":4,"language":"","topics":[],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["GitHub集成","Slack通知","团队协作","项目管理"],"domainKey":"GitHub集成","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1}],"aiScore":0.32,"source":"search","url":"https://github.com/integrations/slack","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"WorldFlowAI/everything-claude-code","owner":"WorldFlowAI","name":"everything-claude-code","desc":"Claude Code toolkit - agents, commands, skills, rules, and hooks for productive AI-assisted development","summaryCn":"Claude Code的瑞士军刀，装了这个插件全家桶效率翻倍","reasonCn":"这是一个Claude Code的完整工具包，包含智能体、自定义命令、技能模块、规则文件和钩子系统。把Claude Code从单一对话工具升级成能自主执行任务的AI开发助手，支持自动化工作流和复杂项目开发，是AI辅助编程的进阶配置方案。","stars":3473,"starGrowth":37,"createdAt":"2026-01-23T12:41:36Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI编程","Claude Code","技能包"],"domainKey":"AI编程","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/WorldFlowAI/everything-claude-code","ts":"2026-09-21T04:10:58.868Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.1715530089260006,"copyOk":true},{"repo":"waooAI/waoowaoo","owner":"waooAI","name":"waoowaoo","desc":"首家工业级全流程 AI 影视生产平台。Industry-first professional AI Agent platform for controllable film & video production. From shorts to live-action with Hollywood-standard workflows.","summaryCn":"AI拍电影从剧本到成片，全流程自动化搞定","reasonCn":"waoowaoo是业界首个工业级AI影视生产平台，支持从剧本创作到成片输出的全流程自动化。采用多模态AI Agent协作，覆盖分镜生成、视频制作、特效处理、剪辑合成等环节。提供好莱坞标准工作流，支持短片到长片的可控生成。结合专业影视制作规范和AI能力，让普通用户也能制作专业级影视内容。","stars":14252,"starGrowth":20,"createdAt":"2026-01-22T12:09:36Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["多模态Agent","创意工具","AI应用"],"zone":"AI","funScore":0.8,"domainTags":["AI 视频","影视制作","Agent","视频生成"],"domainKey":"Agent","pushedAt":"2026-09-21T10:39:25Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/waooAI/waoowaoo","ts":"2026-09-21T10:39:25Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.07705192253718775,"copyOk":false},{"repo":"lanyeeee/bilibili-video-downloader","owner":"lanyeeee","name":"bilibili-video-downloader","desc":"哔哩哔哩 bilibili B站 视频 下载器，图形界面 + nfo刮削 + 广告标记 + 字幕下载 + 弹幕下载，轻松将视频加入emby等媒体库","summaryCn":"B站视频下载神器，图形界面一键操作，自动整理媒体库超省心","reasonCn":"Rust开发的高效B站视频下载工具，提供GUI和命令行两种操作方式。支持nfo元数据自动刮削生成，能识别并标记视频中的广告片段，同时可下载字幕和弹幕数据。下载完成后自动整理文件结构，方便直接导入Emby、Jellyfin等媒体服务器，省去手动归类的时间成本。","stars":2089,"starGrowth":3,"createdAt":"2025-07-08T12:34:21Z","silentRounds":0,"language":"Rust","topics":["bilibili","download","downloader","gui","naive-ui","rust","tauri","tauri-app","vue"],"aiDims":["非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["B站下载","视频下载","媒体库"],"domainKey":"B站下载","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"bilibili","source":"github","weight":0.7},{"name":"download","source":"github","weight":0.7},{"name":"downloader","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"naive-ui","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"tauri","source":"github","weight":0.7},{"name":"tauri-app","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/lanyeeee/bilibili-video-downloader","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deepnight/ldtk","owner":"deepnight","name":"ldtk","desc":"Modern, lightweight and efficient 2D level editor","summaryCn":"用拖拽就能造游戏的关卡编辑器，零代码也能做独立游戏","reasonCn":"基于 Haxe 引擎打造的多平台 2D 关卡编辑器，采用独特的层级网格系统，支持自动图块拼接、瓦片映射和程序化内容生成。导出格式广泛兼容 Unity、Godot、Phaser、LÖVE 等主流游戏引擎，内置碰撞体配置、触发区域、自定义元数据层，以及完整的 PNG/Tiled 格式支持，独立开发者的关卡制作效率提升数倍。","stars":4198,"starGrowth":3,"createdAt":"2020-05-29T09:39:23Z","silentRounds":0,"language":"Haxe","topics":["2d","electron","game-development","haxe","haxe-libs","level-editor"],"aiDims":["游戏","创意工具","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["2D关卡设计","游戏开发工具","瓦片映射","独立游戏制作","可视化编辑"],"domainKey":"2D关卡设计","pushedAt":"2026-07-12T13:55:51Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"2d","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"haxe","source":"github","weight":0.7},{"name":"haxe-libs","source":"github","weight":0.7},{"name":"level-editor","source":"github","weight":0.7},{"name":"Haxe","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/deepnight/ldtk","ts":"2026-07-12T13:55:51Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SagerNet/sing-box","owner":"SagerNet","name":"sing-box","desc":"The universal proxy platform","summaryCn":"新一代全能代理神器，支持多种协议，速度快配置灵活","reasonCn":"sing-box是一个全能型代理平台，支持V2Ray、Trojan、Shadowsocks、Hysteria等多种协议。它采用Go语言编写，性能强劲，配置简洁，支持DNS分流、路由规则等功能。相比旧版V2Ray，sing-box速度更快、资源占用更低，是翻墙爱好者的新选择。","stars":38216,"starGrowth":17,"createdAt":"2022-06-30T13:27:22Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["代理软件","网络加速","协议支持","网络工具","网络安全"],"domainKey":"代理软件","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/SagerNet/sing-box","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05233451957295374,"copyOk":true},{"repo":"google/adk-go","owner":"google","name":"adk-go","desc":"An open-source, code-first Go toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.","summaryCn":"谷歌开源的开发套件，让写Go语言的程序员能丝滑构建并管控AI智能体。","reasonCn":"这是谷歌推出的一个以代码为优先的Go工具包，专为构建、评估与部署复杂AI智能体设计。它将AI能力深度集成至Go生态中，提供高度灵活的控制流与任务调度功能，特别适合需要企业级稳定性与高性能表现的工程团队。通过规范化的API设计，开发者可以更轻松地管理智能体的状态、上下文与推理流程，确保在大规模生产环境下的可控性与扩展能力，是后端开发者进军AI Agent领域的利器。","stars":8816,"starGrowth":0,"createdAt":"2025-05-05T17:16:26Z","silentRounds":2,"language":"Go","topics":["agents"],"aiDims":["AI Agent","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["AI智能体","Go语言","Agent框架","工具包"],"domainKey":"AI智能体","pushedAt":"2026-09-23T20:09:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/google/adk-go","ts":"2026-09-23T20:09:42Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"cli/cli","owner":"cli","name":"cli","desc":"GitHub’s official command line tool","summaryCn":"GitHub 官方命令行神器，少敲几行命令就能逛代码圈。","reasonCn":"GitHub 官方维护的命令行工具，提供了核心功能的 CLI 版本。它支持交互式浏览、提交、代码审查和仓库管理等操作，让开发者无需打开浏览器就能完成大部分日常开发任务。通过 Pager（分页器），它还能让繁杂的输出更易读，完美替代了原本生硬的网页操作。","stars":46376,"starGrowth":15,"createdAt":"2019-10-03T15:24:53Z","silentRounds":0,"language":"Go","topics":["git"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["GitHub CLI","命令行工具","代码审查","仓库管理","API 封装"],"domainKey":"GitHub CLI","pushedAt":"2026-09-23T21:21:36Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/cli/cli","ts":"2026-09-23T21:21:36Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04597032948076591,"copyOk":true},{"repo":"nginx/nginx","owner":"nginx","name":"nginx","desc":"The official NGINX Open Source repository.","summaryCn":"互联网的守门大将，扛住亿级流量的网页守护神","reasonCn":"全球最流行的Web服务器和反向代理，能扛住亿级并发请求。支持HTTPS加密、负载均衡、动静分离、缓存加速，还能做API网关和微服务网关。性能远超Apache，配置灵活，资源占用极低，单台机器就能支撑百万级并发连接，是互联网基础设施的基石，云原生时代也离不开它。","stars":31710,"starGrowth":7,"createdAt":"2015-06-23T10:26:27Z","silentRounds":0,"language":"C","topics":["security"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web服务器","反向代理","负载均衡","高并发"],"domainKey":"Web服务器","pushedAt":"2026-09-16T12:13:03Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/nginx/nginx","ts":"2026-09-16T12:13:03Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02166225165562914,"copyOk":true},{"repo":"wealthfolio/wealthfolio","owner":"wealthfolio","name":"wealthfolio","desc":"A beautiful, private, local-first personal finance tracker. Investments, net worth, spending, and simulations.","summaryCn":"私人财务管家，本地存储你的投资、支出和净资产","reasonCn":"基于Rust开发的个人理财追踪工具，支持投资组合管理、净资产计算、支出追踪和财务模拟。采用本地优先架构，数据完全存储在用户设备上，无需上传云端。提供可视化图表和多种金融数据源接入。与Mint、YNAB等商业产品不同，它是免费开源的，适合关注隐私和数据安全的用户。","stars":9027,"starGrowth":9,"createdAt":"2024-05-27T14:44:18Z","silentRounds":0,"language":"Rust","topics":["macos-app","portfolio-tracker","tauri-app","self-hosted"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["个人理财","投资组合管理","本地优先","Rust"],"domainKey":"个人理财","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"macos-app","source":"github","weight":0.7},{"name":"portfolio-tracker","source":"github","weight":0.7},{"name":"tauri-app","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/wealthfolio/wealthfolio","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.029991026919242273,"copyOk":false},{"repo":"ipfs/kubo","owner":"ipfs","name":"kubo","desc":"IPFS implementation in Go: a daemon that stores and serves content-addressed data, with a CLI, HTTP Gateway, and RPC API","summaryCn":"Kubo 是全球第一个星际文件系统（IPFS）实现方案，也是目前使用最广泛的底层","reasonCn":"在底层架构设计上，它采用了严格的内容寻址规范。无论文件放置在网络的哪个角落，系统都会为其计算出唯一的哈希标识与有向无环图，确保只要内容没有改变，寻址结果就绝对一致。为了实现最大化的设备间互操作，它集成了用于文件系统整理的 UnixFS 规范、用于浏览器直连访问的 HTTP ","stars":17139,"starGrowth":0,"createdAt":"2014-06-26T08:14:34Z","silentRounds":2,"language":"Go","topics":["ipfs"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["IPFS","分布式存储","内容寻址","去中心化网络","节点服务"],"domainKey":"IPFS","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ipfs","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.41,"source":"trending","url":"https://github.com/ipfs/kubo","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepspeedai/DeepSpeed","owner":"deepspeedai","name":"DeepSpeed","desc":"DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.","summaryCn":"让训练大模型不再烧钱，多张显卡串起来秒变超级计算机","reasonCn":"DeepSpeed是微软开发的深度学习优化库，核心是ZeRO优化技术，能把模型参数、梯度和优化器状态分片存储到多张显卡上，大幅降低单卡显存需求。支持数据并行、张量并行和流水线并行的3D并行策略，让千亿参数模型的训练变得可行，被LLaMA等知名项目广泛采用。","stars":43154,"starGrowth":6,"createdAt":"2020-01-23T18:35:18Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["模型与训练","AI基础设施","大语言模型"],"zone":"AI","funScore":0,"domainTags":["深度学习","分布式训练","大模型","ZeRO优化","PyTorch扩展"],"domainKey":"深度学习","pushedAt":"2026-09-23T15:38:19Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/deepspeedai/DeepSpeed","ts":"2026-09-23T15:38:19Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01841711081243917,"copyOk":true},{"repo":"cathrynlavery/diagram-design","owner":"cathrynlavery","name":"diagram-design","desc":"38 editorial diagram types for Claude Code, Codex, and Pi. Self-contained HTML + SVG. No shadows. No Mermaid slop.","summaryCn":"Claude Code专用的超精美图表设计器，告别丑陋流程图","reasonCn":"它的设计理念是“编辑级质量”。普通的图表生成器只是简单的代码转图片，而Diagram Design生成的SVG是可编辑的，你可以用Illustrator、Figma等工具进一步修改。它支持29种图表类型，包括流程图、时序图、甘特图等，每种类型都有精心设计的配色和布局，避免了传统工具的“工程感”。","stars":42011,"starGrowth":183,"createdAt":"2026-04-16T18:01:16Z","silentRounds":0,"language":"HTML","topics":["data-visualization"],"aiDims":["创意工具","可视化","AI写作"],"zone":"工具","funScore":0,"domainTags":["图表生成","SVG编辑","AI辅助设计"],"domainKey":"图表生成","pushedAt":"2026-09-19T17:47:52Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/cathrynlavery/diagram-design","ts":"2026-09-19T17:47:52Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3756816071981148,"copyOk":true},{"repo":"libgit2/libgit2sharp","owner":"libgit2","name":"libgit2sharp","desc":"","summaryCn":"LibGit2Sharp 允许 .NET 开发者直接访问 Git 的底层架构，无需调用外部命令","reasonCn":"该项目的核心亮点在于其极高的执行速度，它直接复用了 libgit2 这一高性能的底层实现，保障了在进行大型仓库操作时的稳定性和高效性。对于想要在现有业务中集成版本管理功能的程序员来说，直接通过 NuGet 包管理器引用即可快速部署，项目通过 StackOverflow 社区维护，能为开发者提","stars":3458,"starGrowth":0,"createdAt":"2011-02-26T16:44:49Z","silentRounds":3,"language":"C#","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Git","版本控制","C#",".NET"],"domainKey":"Git","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/libgit2/libgit2sharp","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tinyhttp/tinyhttp","owner":"tinyhttp","name":"tinyhttp","desc":"🦄 0-legacy, tiny & fast web framework as a replacement of Express","summaryCn":"Express的超轻量替代品，几KB就搞定Web服务，快得飞起","reasonCn":"这是一个超轻量的TypeScript Web框架，体积只有Express的零头，性能却更强。它摒弃了Express的历史包袱和中间件依赖，原生支持Promise和async-await语法，路由写法极其简洁。内置了现代化的API设计，支持类型安全。适合追求高性能和精简依赖的现代Node.js项目，尤其是微服务和边缘计算场景，是Express用户的优质替代品。","stars":2904,"starGrowth":0,"createdAt":"2020-06-13T01:19:08Z","silentRounds":22,"language":"TypeScript","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web框架","TypeScript","Node.js","HTTP服务器","后端开发"],"domainKey":"Web框架","pushedAt":"2026-09-11T00:56:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/tinyhttp/tinyhttp","ts":"2026-09-11T00:56:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"giraffe-fsharp/Giraffe","owner":"giraffe-fsharp","name":"Giraffe","desc":"A native functional ASP.NET Core web framework for F# developers.","summaryCn":"用F#写Web应用的函数式框架，代码优雅性能好","reasonCn":"Giraffe是专为F#开发者打造的ASP.NET Core函数式Web框架，采用纯函数风格构建HTTP处理管道。它将路由、中间件和控制器组合成可复用的高阶函数链，代码更简洁且类型安全。相比传统MVC模式，Giraffe避免了繁琐的控制器类，通过Route模块和管道组合实现更优雅的API和Web服务开发，同时保留ASP.NET Core全部功能。","stars":2255,"starGrowth":0,"createdAt":"2017-01-07T17:31:24Z","silentRounds":24,"language":"F#","topics":["web-framework"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["F#","Web框架","函数式编程","ASP.NET Core","API开发"],"domainKey":"F#","pushedAt":"2026-09-22T12:59:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"F#","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/giraffe-fsharp/Giraffe","ts":"2026-09-22T12:59:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tailcallhq/forgecode","owner":"tailcallhq","name":"forgecode","desc":"AI enabled pair programmer for Claude, GPT, O Series, Grok, Deepseek, Gemini and 300+ models","summaryCn":"一个终端里的AI编程搭档，支持Claude GPT Deepseek等300多个模型任你选","reasonCn":"技术上用Rust编写，性能好启动快，支持300多个AI模型，从OpenAI到国产的Deepseek都能用。它不仅能回答代码问题，还能帮你生成代码、解释bug、优化性能，功能相当全面。关键是它理解上下文，不是每次提问都从零开始，而是会记住之前的对话内容。","stars":7636,"starGrowth":0,"createdAt":"2024-12-08T20:33:37Z","silentRounds":3,"language":"Rust","topics":["command-line"],"aiDims":["代码助手","AI Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI编程助手","终端工具","多模型支持"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T01:56:58Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/tailcallhq/forgecode","ts":"2026-09-23T01:56:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"athasdev/athas","owner":"athasdev","name":"athas","desc":"A lightweight, cross-platform code editor, built with Tauri (Rust and React) with Git support, AI agents, vim keybindings.","summaryCn":"用Rust和前端技术打造的轻量级现代代码编辑器，自带AI和Vim。","reasonCn":"技术上，它站在了巨人的肩膀上：底层采用Tauri框架，结合了Rust的极致性能安全与前端React的高效生态。相比那些动辄吃掉几个GB内存的传统Electron编辑器，它不仅内存占用极低、启动速度秒开，而且由于Rust的加持，在处理大规模文本和高并发插件时依然稳如老狗。","stars":3073,"starGrowth":1,"createdAt":"2025-06-21T13:14:47Z","silentRounds":4,"language":"TypeScript","topics":["editor"],"aiDims":["开发者工具","AI Agent","代码助手"],"zone":"工具","funScore":0.7,"domainTags":["代码编辑器","Tauri","AI编程"],"domainKey":"代码编辑器","pushedAt":"2026-09-22T23:41:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/athasdev/athas","ts":"2026-09-22T23:41:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"opensteno/plover","owner":"opensteno","name":"plover","desc":"","summaryCn":"在许多医疗记录、法庭速录或极客追求的场景中，传统键盘的 QWERTY 布局已触","reasonCn":"Plover 并非孤立存在，它是 Open Steno Project 的核心软件组件。该项目的愿景是彻底开放机器速记领域，从提供免费的软件（如 Plover）、指导使用廉价的 DIY 硬件，到整理全套自学资源，让任何人都能独立掌握这项看似神秘的技术。","stars":2591,"starGrowth":0,"createdAt":"2012-06-13T14:50:49Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["速记输入","桌面应用","开源硬件","打字加速"],"domainKey":"速记输入","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/opensteno/plover","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"facebookarchive/pfff","owner":"facebookarchive","name":"pfff","desc":"","summaryCn":"基于OCaml的多语言抽象语法树解析与风格保留重构工具包。","reasonCn":"Pfff是Meta开源的静态分析基础设施，它通过解析代码构建抽象语法树，执行保持原有代码风格的自动重构。支持与C、Java、JavaScript等语言集成，能精准处理C++、Rust等复杂度高的语法结构。它为大型代码库的批量修改和规范检查提供了底层引擎能力，避免手动修改带来的格式混乱与逻辑错误。","stars":2433,"starGrowth":0,"createdAt":"2010-09-23T05:13:55Z","silentRounds":2,"language":"OCaml","topics":[],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["静态分析","代码重构","多语言解析","OCaml"],"domainKey":"静态分析","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"OCaml","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/facebookarchive/pfff","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"katiefenn/parker","owner":"katiefenn","name":"parker","desc":"","summaryCn":"Parker 旨在解决前端开发中样式表缺乏量化评估手段的问题。","reasonCn":"项目提供了灵活的测量方式，除了针对本地文件外，还特别支持使用 Curl 测量远程样式表。这意味着开发团队可以直接对生产环境或预发布环境的 CSS 资源进行即时体检，无需下载文件即可获取复杂度报告。这种远程分析能力使得性能评估更加贴近真实部署场景，能够快速发现线上样式表存在","stars":2433,"starGrowth":0,"createdAt":"2014-02-09T22:21:09Z","silentRounds":2,"language":"JavaScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["样式表分析","CSS 复杂度","静态分析工具"],"domainKey":"样式表分析","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/katiefenn/parker","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"academic/awesome-datascience","owner":"academic","name":"awesome-datascience","desc":":memo: An awesome Data Science repository to learn and apply for real world problems.","summaryCn":"数据科学学习资源大合集，从入门到精通一站式搞定","reasonCn":"这是一个数据科学领域的学习资源汇总项目，汇集了从入门到进阶的全方位学习资料，包括编程语言基础、统计学、机器学习、深度学习、数据可视化、大数据处理等各个方向。 项目内容非常丰富，分为多个板块。Python方向有Pandas、NumPy、Scikit-learn等核心库的学习资源；R语言方向有g","stars":30055,"starGrowth":11,"createdAt":"2014-07-05T07:07:43Z","silentRounds":0,"language":"","topics":["ml"],"aiDims":["学习资源","AI基础设施","教程"],"zone":"资源","funScore":0,"domainTags":["数据科学","学习资源","机器学习","教程"],"domainKey":"数据科学","pushedAt":"2026-09-23T21:20:45Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7}],"aiScore":0.78,"source":"search","url":"https://github.com/academic/awesome-datascience","ts":"2026-09-23T21:20:45Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03409798702378972,"copyOk":false},{"repo":"eli64s/readme-ai","owner":"eli64s","name":"readme-ai","desc":"README file generator, powered by AI.","summaryCn":"丢进项目地址就能自动生成高大上说明文档的AI小能手。","reasonCn":"技术实现上，它巧妙地结合了代码静态分析与大语言模型的文本生成能力。它会扫描项目中的文件树、依赖配置文件和核心源码，提取出关键信息并发送给大模型进行语义润色。用户不仅可以选择不同的AI模型提供商，还能通过自定义模版来控制输出文档的风格，满足开源项目或商业闭源项目的","stars":2989,"starGrowth":1,"createdAt":"2023-01-16T19:47:17Z","silentRounds":3,"language":"Python","topics":["documentation"],"aiDims":["AI应用","代码助手","开发者工具"],"zone":"AI","funScore":0,"domainTags":["README生成","文档自动化","AI工具"],"domainKey":"README生成","pushedAt":"2026-09-23T02:48:35Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"documentation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/eli64s/readme-ai","ts":"2026-09-23T02:48:35Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"isno/theByteBook","owner":"isno","name":"theByteBook","desc":"⭐ 【出版书籍】京东购买链接 https://item.jd.com/14531549.html  深入讲解内核网络、Kubernetes、ServiceMesh、容器等云原生相关技术。经历实践检验的“大规模分布式系统”开发指南。","summaryCn":"京东可买的云原生技术神书，搞定K8s和微服务。","reasonCn":"全书内容架构清晰，从底层的 Linux 内核网络优化，到上层的 Kubernetes 集群管理、Service Mesh 服务网格落地，再到容器化部署的最佳实践，形成了一个完整的知识闭环。对于正在从事微服务架构转型的技术团队，或者是希望深入理解 K8s 底层原理的工程师来说，这本书提供了非常宝贵的","stars":8550,"starGrowth":1,"createdAt":"2022-10-08T04:44:22Z","silentRounds":24,"language":"JavaScript","topics":["devops"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["云原生","Kubernetes","微服务","网络","书籍"],"domainKey":"云原生","pushedAt":"2026-08-14T05:58:03Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/isno/theByteBook","ts":"2026-08-14T05:58:03Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"drashland/drash","owner":"drashland","name":"drash","desc":"A strongly typed, runtime-agnostic web framework for building structured HTTP services in JavaScript, built on Web Standards.","summaryCn":"用TypeScript写网页服务，像搭积木一样简单有趣。","reasonCn":"Drash 的核心魅力在于它的类型安全和结构化设计。它把 Web 开发的各个环节——从路由定义到请求响应——都封装得非常清晰。你不需要在代码里到处找请求参数，框架会自动帮你处理，让你的代码读起来就像写文档一样顺滑。 如果你想试试，安装非常简单。","stars":1064,"starGrowth":0,"createdAt":"2019-02-22T05:27:13Z","silentRounds":24,"language":"TypeScript","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["TypeScript","Web 框架","HTTP 服务","结构化开发"],"domainKey":"TypeScript","pushedAt":"2026-09-19T23:35:04Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/drashland/drash","ts":"2026-09-19T23:35:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"koute/stdweb","owner":"koute","name":"stdweb","desc":"","summaryCn":"在 Web 开发领域，JavaScript 长期占据主导，而 stdweb 的出现让 Rust 语言","reasonCn":"该库最显著的特性是支持在 Rust 代码中直接嵌入 JavaScript 代码，这种混合编程模式极大地增强了灵活性。当浏览器存在原生特性而 Rust 生态尚未完全覆盖时，开发者可以直接调用底层的 JS 代码，无需复杂的 FFI 调度。","stars":3454,"starGrowth":0,"createdAt":"2017-03-12T20:07:37Z","silentRounds":4,"language":"Rust","topics":["rust","javascript","web","gui","asmjs","emscripten","webassembly","webasm"],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Rust绑定","Web API","JS互操作","客户端开发"],"domainKey":"Rust绑定","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"rust","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"web","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"asmjs","source":"github","weight":0.7},{"name":"emscripten","source":"github","weight":0.7},{"name":"webassembly","source":"github","weight":0.7},{"name":"webasm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/koute/stdweb","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"K9i-0/ccpocket","owner":"K9i-0","name":"ccpocket","desc":"Mobile client for Codex and Claude — control coding agents from your phone via WebSocket bridge","summaryCn":"在手机上随时随地远程遥控电脑写代码的AI智能体移动客户端。","reasonCn":"核心技术上，它采用Flutter/Dart跨平台框架开发，保证了iOS和Android端丝滑流畅的体验。底层通过WebSocket安全桥接协议，与运行在桌面端的AI编码代理（如Claude Engineer、Codex等）进行实时双向通信，把电脑端的终端输出、代码修改和AI思考过程无延迟地同步到手机屏幕上。","stars":1074,"starGrowth":4,"createdAt":"2026-02-06T00:18:05Z","silentRounds":0,"language":"Dart","topics":["mobile"],"aiDims":["代码助手","AI Agent","效率工具"],"zone":"AI","funScore":0.7,"domainTags":["AI编程","移动端","Claude","Codex","远程开发"],"domainKey":"AI编程","pushedAt":"2026-09-23T06:34:20Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/K9i-0/ccpocket","ts":"2026-09-23T06:34:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"milvus-io/bootcamp","owner":"milvus-io","name":"bootcamp","desc":"Dealing with all unstructured data, such as reverse image search, audio search, molecular search, video analysis, question and answer systems, NLP, etc.","summaryCn":"把各种非结构化数据喂给AI，像搭积木一样轻松搞定搜图、听音和问答。","reasonCn":"这个项目是个向量数据库实战教程，教你怎么用Milvus处理图片、音频、视频、文档等各种非结构化数据。它通过演示NLP问答、反向图像搜索等场景，让你快速掌握向量数据库的部署和操作技巧，是学习AI数据管理的优质资源。","stars":2445,"starGrowth":1,"createdAt":"2019-08-09T10:00:06Z","silentRounds":26,"language":"Jupyter Notebook","topics":["vector-db"],"aiDims":["向量数据库","RAG与知识","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["向量数据库","非结构化数据","NLP","图像搜索"],"domainKey":"向量数据库","pushedAt":"2026-09-23T05:51:14Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/milvus-io/bootcamp","ts":"2026-09-23T05:51:14Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cheat-engine/cheat-engine","owner":"cheat-engine","name":"cheat-engine","desc":"","summaryCn":"Cheat Engine 是一款面向游戏改装的工具，采用 Pascal 语言开发，项目结构基","reasonCn":"Cheat Engine 通过内存扫描、Hook 注入、脚本插件等技术，允许玩家在单机游戏中修改变量、修复 BUG 或添加自定义功能。它自带可视化脚本编辑器，支持 Lua 脚本和 DLL 插件，使高级用户能够编写复杂的作弊逻辑或与其他工具联动。","stars":19211,"starGrowth":5,"createdAt":"2015-04-18T11:21:57Z","silentRounds":0,"language":"Pascal","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["游戏修改器","内存扫描","逆向工程","Lua脚本"],"domainKey":"游戏修改器","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Pascal","source":"language","weight":0.5}],"aiScore":0.392,"source":"tier","url":"https://github.com/cheat-engine/cheat-engine","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015780802665139763,"copyOk":false},{"repo":"urho3d/urho3d","owner":"urho3d","name":"urho3d","desc":"Game engine","summaryCn":"用C++写的轻量3D游戏引擎，跨平台能跑2D也能跑3D","reasonCn":"Urho3D是一个轻量级跨平台3D游戏引擎，使用C++开发，内置渲染、物理、音频、输入系统等核心模块，支持2D和3D游戏开发。它的特点是零外部依赖、易于嵌入项目，采用Angelscript作为脚本语言，代码结构清晰，适合中小型游戏或原型开发。","stars":4767,"starGrowth":1,"createdAt":"2013-10-02T08:55:35Z","silentRounds":23,"language":"C++","topics":["2d","3d","3d-graphics","angelscript","cpp","cpp17","cross-platform","engine","game","game-development","game-engine","gamedev","open-source","urho3d"],"aiDims":["游戏"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","跨平台开发","C++编程","3D渲染","独立游戏"],"domainKey":"游戏引擎","pushedAt":"2023-01-25T17:29:02Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"2d","source":"github","weight":0.7},{"name":"3d","source":"github","weight":0.7},{"name":"3d-graphics","source":"github","weight":0.7},{"name":"angelscript","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"cpp17","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"engine","source":"github","weight":0.7},{"name":"game","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"urho3d","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/urho3d/urho3d","ts":"2023-01-25T17:29:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"schultek/jaspr","owner":"schultek","name":"jaspr","desc":"Modern web framework for building websites in Dart. Supports SPAs, SSR and SSG.","summaryCn":"用Dart写Web应用的全栈框架，一套代码搞定前后端渲染","reasonCn":"Jaspr是Dart生态中的现代Web框架，支持SPA、SSR和SSG三种渲染模式。它借鉴了Flutter的声明式UI思想，让前端开发像搭积木一样直观。项目采用编译时检查，错误在编写阶段就能被发现，减少运行时bug。相比React或Vue，它更轻量且与Dart语言特性深度整合。","stars":2484,"starGrowth":1,"createdAt":"2022-02-08T12:14:58Z","silentRounds":3,"language":"Dart","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Dart","Web 框架","SSR","全栈"],"domainKey":"Dart","pushedAt":"2026-09-22T08:06:55Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/schultek/jaspr","ts":"2026-09-22T08:06:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"GoogleCloudPlatform/generative-ai","owner":"GoogleCloudPlatform","name":"generative-ai","desc":"Sample code and notebooks for Generative AI on Google Cloud, with Gemini Enterprise Agent Platform","summaryCn":"谷歌云官方的生成式AI大宝库，海量现成代码和笔记本教你玩转大模型。","reasonCn":"技术层面上，它深度整合了谷歌云Vertex AI生态，涵盖了从基础的提示词工程、大模型微调、RAG知识库搭建，到复杂的企业级AI Agent平台搭建等全方位内容。所有的示例都采用Jupyter Notebook编写，代码逻辑清晰，注释详细，不仅告诉你怎么调API，还深入展示了如何在生产环境中保证大模","stars":17747,"starGrowth":6,"createdAt":"2023-05-05T12:31:07Z","silentRounds":0,"language":"Jupyter Notebook","topics":["agents"],"aiDims":["AI Agent","学习资源","模型与训练"],"zone":"资源","funScore":0,"domainTags":["生成式 AI","Gemini","提示词工程","RAG","企业级 Agent"],"domainKey":"生成式 AI","pushedAt":"2026-09-23T16:03:49Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/GoogleCloudPlatform/generative-ai","ts":"2026-09-23T16:03:49Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.019014255930579815,"copyOk":true},{"repo":"supabase/supabase","owner":"supabase","name":"supabase","desc":"The Postgres development platform. Supabase gives you a dedicated Postgres database to build your web, mobile, and AI applications.","summaryCn":"后端开发者的神器，一键搞定数据库和认证，让AI应用落地更简单","reasonCn":"Supabase是基于PostgreSQL的开源Firebase替代品，提供实时数据库、用户认证、存储等全套后端能力。对AI开发者而言，它是构建AI应用最稳定的数据底座，支持向量存储和函数扩展，能轻松对接LangChain等AI框架，省去重复造轮子的麻烦，让开发者专注模型与交互逻辑。","stars":110632,"starGrowth":128,"createdAt":"2019-10-12T05:56:49Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI基础设施","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["PostgreSQL","后端框架","实时数据库","开发者工具"],"domainKey":"PostgreSQL","pushedAt":"2026-09-23T21:51:40Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/supabase/supabase","ts":"2026-09-23T21:51:40Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.30347096680887986,"copyOk":false},{"repo":"sensepost/objection","owner":"sensepost","name":"objection","desc":"📱 objection - runtime mobile exploration","summaryCn":"手机App运行时调试神器，绕过SSL、Hook函数、分析二进制漏洞","reasonCn":"Objection是一款开源的移动应用运行时探索工具，支持Android和iOS平台。内置SSL卸载、Hook函数、进程注入、文件系统访问等多种功能，可绕过常见的安全防护机制。提供Python扩展接口，让渗透测试效率大幅提升，在移动安全圈口碑极佳。","stars":9401,"starGrowth":0,"createdAt":"2017-06-29T11:21:38Z","silentRounds":3,"language":"Python","topics":["mobile","security"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["移动安全","渗透测试","Frida","运行时调试","SSL卸载"],"domainKey":"移动安全","pushedAt":"2026-09-17T07:20:15Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/sensepost/objection","ts":"2026-09-17T07:20:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"all-contributors/allcontributors.org","owner":"all-contributors","name":"allcontributors.org","desc":"✨ The all-contributors bot website and documentation. Recognize all contributors, not just the ones who push code ✨","summaryCn":"开源项目.contributors文件自动生成的贴心小工具","reasonCn":"它最核心的功能是根据项目里的PR、Issue、commit等数据，自动识别不同类型的贡献者，然后生成格式化的贡献者列表。不只是代码贡献者，文档撰写者、设计者、翻译、甚至提供资金支持的，都会被正确标记和归类。它还提供了一个命令行工具和GitHub Action，可以集成到你的发布流程中，","stars":8102,"starGrowth":0,"createdAt":"2016-02-28T07:29:02Z","silentRounds":7,"language":"MDX","topics":["acknowledgements","all-contributors","contributors","open-source-tooling","opensource","opensource-management","recognition"],"aiDims":["效率工具","开发者工具","工具调用"],"zone":"工具","funScore":0,"domainTags":["开源工具","贡献者管理","自动化","社区建设","文档生成"],"domainKey":"开源工具","pushedAt":"2026-08-17T09:20:42Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"acknowledgements","source":"github","weight":0.7},{"name":"all-contributors","source":"github","weight":0.7},{"name":"contributors","source":"github","weight":0.7},{"name":"open-source-tooling","source":"github","weight":0.7},{"name":"opensource","source":"github","weight":0.7},{"name":"opensource-management","source":"github","weight":0.7},{"name":"recognition","source":"github","weight":0.7},{"name":"MDX","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/all-contributors/allcontributors.org","ts":"2026-08-17T09:20:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dimforge/rapier.js","owner":"dimforge","name":"rapier.js","desc":"Official JavaScript bindings for the Rapier physics engine ⚠️ MIGRATED TO https://github.com/dimforge/rapier/tree/master/typescript ⚠️","summaryCn":"让网页也能跑物理模拟的引擎，游戏开发者的福音","reasonCn":"Rapier物理引擎的JavaScript绑定版本，专为Web端2D和3D物理模拟设计。采用Rust编写编译为WebAssembly，性能远超纯JS方案，支持刚体动力学、碰撞检测和约束求解。无论你是做网页游戏还是物理可视化，它都能让物理模拟又快又稳。","stars":697,"starGrowth":0,"createdAt":"2020-07-07T21:08:18Z","silentRounds":23,"language":"TypeScript","topics":[],"aiDims":["非AI-好玩","游戏","可视化"],"zone":"工具","funScore":0,"domainTags":["物理引擎","WebAssembly","2D/3D模拟","游戏开发"],"domainKey":"物理引擎","pushedAt":"2026-07-12T16:55:48Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/dimforge/rapier.js","ts":"2026-07-12T16:55:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"meilisearch/meilisearch","owner":"meilisearch","name":"meilisearch","desc":"A lightning-fast search engine API bringing AI-powered hybrid search to your sites and applications.","summaryCn":"比搜索引擎更快更智能的本地搜索方案，一行代码就能接入","reasonCn":"基于Rust开发的轻量级搜索引擎，支持全文检索和向量语义搜索双管齐下。内置混合检索算法，能把关键词匹配和语义理解结合起来，搜索精度远超传统方案。安装部署极其简单，几行命令即可运行，API友好，开发者两小时就能上线一个生产级搜索服务。","stars":59371,"starGrowth":7,"createdAt":"2018-04-23T11:45:28Z","silentRounds":0,"language":"Rust","topics":["database","vector-db","ai"],"aiDims":["AI搜索","向量数据库","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["搜索引擎","向量检索","混合搜索","数据库"],"domainKey":"搜索引擎","pushedAt":"2026-09-23T14:00:43Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"vector-db","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/meilisearch/meilisearch","ts":"2026-09-23T14:00:43Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.021353708039278437,"copyOk":false},{"repo":"k8sgpt-ai/k8sgpt","owner":"k8sgpt-ai","name":"k8sgpt","desc":"Giving Kubernetes Superpowers to everyone","summaryCn":"给Kubernetes装上AI超能力，自动诊断排障让运维变轻松","reasonCn":"k8sgpt是将AI能力引入Kubernetes运维的开源工具，能自动分析集群状态、诊断问题并给出修复建议。它支持多种LLM后端，能理解K8s事件和错误信息。开发者只需运行一条命令就能获得AI辅助的诊断报告。适合K8s运维人员、SRE团队，大幅降低排障门槛。","stars":8196,"starGrowth":0,"createdAt":"2023-03-21T19:58:16Z","silentRounds":3,"language":"Go","topics":["devops"],"aiDims":["AI基础设施","AI应用","开发者工具"],"zone":"工具","funScore":0.7,"domainTags":["Kubernetes","DevOps","AI运维"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T17:25:02Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/k8sgpt-ai/k8sgpt","ts":"2026-09-23T17:25:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ipfs/ipfs-desktop","owner":"ipfs","name":"ipfs-desktop","desc":"An unobtrusive and user-friendly desktop application for IPFS on Windows, Mac and Linux. ","summaryCn":"IPFS的图形化入口，帮你轻松搭建去中心化存储节点，告别命令行恐惧","reasonCn":"这是IPFS官方的桌面客户端，专为不想敲命令的用户设计。它支持Windows、Mac和Linux，提供直观的界面来管理本地存储节点、上传文件并生成内容哈希。相比纯命令行工具，它大幅降低了去中心化存储的使用门槛，适合需要本地节点管理但追求体验的开发者或普通用户。","stars":6587,"starGrowth":0,"createdAt":"2015-05-30T23:42:36Z","silentRounds":3,"language":"JavaScript","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["IPFS","去中心化存储","桌面客户端"],"domainKey":"IPFS","pushedAt":"2026-09-17T04:56:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ipfs/ipfs-desktop","ts":"2026-09-17T04:56:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CISOfy/lynis","owner":"CISOfy","name":"lynis","desc":"Lynis - Security auditing tool for Linux, macOS, and UNIX-based systems. Assists with compliance testing (HIPAA/ISO27001/PCI DSS) and system hardening. Agentless, and installation optional.","summaryCn":"给服务器做全面体检的硬核扫描器，自动找漏洞还能生成合规报告","reasonCn":"Lynis是一款开源的系统审计与加固工具，专为Linux、macOS及Unix系统打造。它无需安装代理，即可对系统配置、软件版本、权限设置进行深度扫描，自动识别潜在的安全漏洞与弱口令风险，并支持HIPAA、ISO27001、PCI DSS等合规性检查。","stars":16365,"starGrowth":2,"createdAt":"2013-12-10T14:00:49Z","silentRounds":0,"language":"Shell","topics":["devops"],"aiDims":["非AI-实用","开发者工具","安全工具"],"zone":"工具","funScore":0,"domainTags":["安全审计","系统加固","合规检查","DevOps"],"domainKey":"安全审计","pushedAt":"2026-09-16T21:41:02Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/CISOfy/lynis","ts":"2026-09-16T21:41:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"skilion/onedrive","owner":"skilion","name":"onedrive","desc":"","summaryCn":"在 Linux 系统上，官方 OneDrive 客户端往往带有较重的桌面环境依赖，甚至有","reasonCn":"项目的核心亮点在于其实时同步能力。它利用了 Linux 内核的 Inotify 子系统，能够即时捕捉本地文件的变更并触发上传，同时也支持 OneDrive 共享文件夹功能。此外，对于大文件传输，项目内置了断点续传机制，确保在网络波动时不会前功尽弃。","stars":3947,"starGrowth":0,"createdAt":"2015-09-22T18:58:17Z","silentRounds":4,"language":"D","topics":["linux","onedrive","dlang"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["OneDrive","Linux","文件同步","命令行工具"],"domainKey":"OneDrive","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"linux","source":"github","weight":0.7},{"name":"onedrive","source":"github","weight":0.7},{"name":"dlang","source":"github","weight":0.7},{"name":"D","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/skilion/onedrive","ts":"2026-09-18T20:05:54.376Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"BasedHardware/omi","owner":"BasedHardware","name":"omi","desc":"AI that sees your screen, listens to your conversations and tells you what to do","summaryCn":"像私人助理一样的AI硬件，能看屏幕、听对话，自动帮你处理日常任务","reasonCn":"Omi是一个开源的AI硬件设备，运行在Android手机上，能实时监听你的对话、查看你的屏幕内容，并根据上下文主动提醒或执行任务。它是一个真正的本地化多模态AI Agent，旨在成为你的第二大脑，保护隐私的同时提供无缝的智能辅助。","stars":13534,"starGrowth":2,"createdAt":"2024-03-22T18:12:25Z","silentRounds":0,"language":"Python","topics":["mobile"],"aiDims":["AI Agent","AI应用","多模态Agent"],"zone":"AI","funScore":1,"domainTags":["AI Agent","智能家居","隐私计算"],"domainKey":"AI Agent","pushedAt":"2026-09-23T20:23:39Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/BasedHardware/omi","ts":"2026-09-23T20:23:39Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"introfini/ZotSeek","owner":"introfini","name":"ZotSeek","desc":"AI semantic search for Zotero, with a built-in MCP server for AI agents (Claude Code, Codex). Find papers by meaning. 100% local and private.","summaryCn":"当你手头积累了成百上千篇PDF论文时，想找某一个具体观点往往像大海捞针。","reasonCn":"它的技术核心是将你的Zotero本地数据库转换为向量索引，利用轻量级嵌入模型（Embedding）进行语义映射。最厉害的是它内置了MCP协议，这意味着你可以把它无缝接入任何支持MCP的AI开发工具，让你的AI代理直接查阅你的文献，实现基于本地知识库的自动摘要和观点溯源。","stars":212,"starGrowth":1,"createdAt":"2025-12-26T19:37:29Z","silentRounds":3,"language":"TypeScript","topics":["transformers"],"aiDims":["AI搜索","RAG与知识","AI Agent"],"zone":"AI","funScore":0.6,"domainTags":["Zotero","语义搜索","MCP","文献管理","本地化"],"domainKey":"Zotero","pushedAt":"2026-09-22T13:49:08Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/introfini/ZotSeek","ts":"2026-09-22T13:49:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Avaiga/taipy","owner":"Avaiga","name":"taipy","desc":"Turns Data and AI algorithms into production-ready web applications in no time.","summaryCn":"把数据算法秒变Web应用，无需写前端代码，Python就能搞定全栈","reasonCn":"Taipy是一个让Python开发者能快速将数据分析和AI算法转化为生产级Web应用的框架。你只需写几行Python代码，就能自动生成带交互界面的Dashboard、图表和控制面板。支持多线程、高并发，还能直接部署到云端。省去了写HTML/CSS/JS的麻烦，让数据科学家能独立完成应用开发。","stars":19434,"starGrowth":0,"createdAt":"2022-02-18T15:55:45Z","silentRounds":10,"language":"Python","topics":["developer-tools","data-visualization","automation"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Python框架","数据可视化","Web应用","全栈开发"],"domainKey":"Python框架","pushedAt":"2026-08-10T02:47:18Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/Avaiga/taipy","ts":"2026-08-10T02:47:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kemalcr/kemal","owner":"kemalcr","name":"kemal","desc":"Fast, Effective, Simple Web Framework","summaryCn":"Crystal语言里最简洁高效的Web框架，几行代码搭出高性能服务器","reasonCn":"基于Krystal库构建的轻量级Web框架，语法类似Sinatra但底层是编译型Crystal语言，运行速度远超Ruby和Python同类框架。支持路由、中间件、文件上传、WebSocket和JSON处理，GitHub Star接近4000，是Crystal生态中最成熟的Web框架之一。","stars":3917,"starGrowth":0,"createdAt":"2015-10-23T16:56:29Z","silentRounds":5,"language":"Crystal","topics":["web-framework","api"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Web框架","Crystal语言","高性能API","REST API","中间件"],"domainKey":"Web框架","pushedAt":"2026-09-15T11:22:13Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"Crystal","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/kemalcr/kemal","ts":"2026-09-15T11:22:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"twbs/rfs","owner":"twbs","name":"rfs","desc":"✩ Automates responsive resizing ✩","summaryCn":"让网页字体自动跟着屏幕大小变形，不用再手撸媒体查询","reasonCn":"这个项目是Bootstrap团队出的CSS工具，核心是利用clamp()函数实现响应式字体缩放。开发者只需引入一个CSS文件，所有设置好的元素就会根据视口宽度自动计算字体大小，从手机端到桌面端平滑过渡，省去繁琐的@media规则编写。","stars":3369,"starGrowth":0,"createdAt":"2017-06-04T13:49:56Z","silentRounds":22,"language":"CSS","topics":["frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["响应式字体","CSS clamp","前端工具库"],"domainKey":"响应式字体","pushedAt":"2026-09-10T12:30:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/twbs/rfs","ts":"2026-09-10T12:30:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alibaba/zvec","owner":"alibaba","name":"zvec","desc":"A lightweight, lightning-fast, in-process vector database","summaryCn":"阿里开源的极速嵌入式向量数据库，RAG 检索快得离谱","reasonCn":"这是阿里巴巴开源的一个轻量级、高性能向量数据库，专为 AI 应用中的向量检索需求设计。它无需外部依赖，可直接嵌入到现有进程中，提供亚毫秒级的查询延迟，是构建 RAG 系统和语义搜索功能的理想底层组件，特别适合资源受限的边缘场景。","stars":15992,"starGrowth":4,"createdAt":"2025-12-05T07:45:38Z","silentRounds":0,"language":"C++","topics":["rag","vector-db"],"aiDims":["向量数据库","AI基础设施","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["向量数据库","RAG","语义搜索"],"domainKey":"向量数据库","pushedAt":"2026-09-23T06:47:50Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"vector-db","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/alibaba/zvec","ts":"2026-09-23T06:47:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mobile-next/mobile-mcp","owner":"mobile-next","name":"mobile-mcp","desc":"Model Context Protocol Server for Mobile Automation and Scraping (iOS, Android, Emulators, Simulators and Real Devices)","summaryCn":"让AI大模型远程控制手机和模拟器的开源神器","reasonCn":"基于MCP协议构建的移动设备自动化服务，让AI Agent通过标准接口控制iOS和Android设备。支持物理手机、模拟器、仿真器等多种终端，可实现界面操作、数据抓取、自动化测试等功能。用TypeScript开发，5800多星，热度很高。","stars":6812,"starGrowth":1,"createdAt":"2025-03-28T16:21:06Z","silentRounds":1,"language":"TypeScript","topics":["mobile","emulator"],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["MCP协议","移动自动化","AI Agent","设备控制"],"domainKey":"AI Agent","pushedAt":"2026-09-23T15:38:57Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"emulator","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/mobile-next/mobile-mcp","ts":"2026-09-23T15:38:57Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"RustPython/RustPython","owner":"RustPython","name":"RustPython","desc":"A Python Interpreter written in Rust","summaryCn":"用Rust重写Python解释器，带你从底层读懂Python怎么运行","reasonCn":"这不是一个AI项目，而是一个用Rust语言从头实现的Python解释器，目标是兼容CPython的语法和行为。它让开发者可以深入理解Python字节码执行、对象模型、GIL机制等底层原理。代码组织清晰，注释丰富，是学习语言实现和Rust系统编程的优质教材，2万2千star说明社区认可度很高。","stars":22351,"starGrowth":0,"createdAt":"2018-05-28T19:27:01Z","silentRounds":2,"language":"Rust","topics":["rust","python3","python-language","jit","interpreter","hacktoberfest","language","compiler","wasm"],"aiDims":["非AI-好玩","开发者工具","学习资源"],"zone":"工具","funScore":0,"domainTags":["Python解释器","Rust","语言实现"],"domainKey":"Python解释器","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"rust","source":"github","weight":0.7},{"name":"python3","source":"github","weight":0.7},{"name":"python-language","source":"github","weight":0.7},{"name":"jit","source":"github","weight":0.7},{"name":"interpreter","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"language","source":"github","weight":0.7},{"name":"compiler","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/RustPython/RustPython","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alexballas/go2tv","owner":"alexballas","name":"go2tv","desc":"Cast media files to Smart TVs and Chromecast devices.","summaryCn":"一行命令把视频文件投屏到电视，命令行党的媒体播放神器","reasonCn":"它的核心技术是支持两种投屏协议：DLNA和Chromecast。DLNA是智能电视普遍支持的媒体共享协议，go2tv作为DLNA客户端，能发现和控制网络上的DLNA接收设备。Chromecast则是Google的投屏协议，go2tv通过Chromecast SDK与设备通信。","stars":1110,"starGrowth":0,"createdAt":"2021-02-08T23:30:26Z","silentRounds":2,"language":"Go","topics":["dlna-mediarenderer","upnp","cast","samsumg-tvs","videos","dlna","media","audio","video","golang","streaming-video","chromecast"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["DLNA投屏","Chromecast","媒体流传输","命令行工具"],"domainKey":"DLNA投屏","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"dlna-mediarenderer","source":"github","weight":0.7},{"name":"upnp","source":"github","weight":0.7},{"name":"cast","source":"github","weight":0.7},{"name":"samsumg-tvs","source":"github","weight":0.7},{"name":"videos","source":"github","weight":0.7},{"name":"dlna","source":"github","weight":0.7},{"name":"media","source":"github","weight":0.7},{"name":"audio","source":"github","weight":0.7},{"name":"video","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"streaming-video","source":"github","weight":0.7},{"name":"chromecast","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/alexballas/go2tv","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Golle/Titan","owner":"Golle","name":"Titan","desc":"Titan, Game Engine in .NET 8 C# 11 and DirectX 12","summaryCn":"用C#打造的轻量游戏引擎，DirectX 12加持超带感","reasonCn":"这是一个完全用C#和.NET 8构建的轻量游戏引擎，采用现代C# 11特性和DirectX 12图形API。不同于Unity的笨重，它追求简洁高效，让游戏开发者能用熟悉的C#语言快速搭建游戏原型。对于喜欢折腾引擎、想了解底层原理的开发者来说，这是一个很好的学习和实践项目。","stars":88,"starGrowth":0,"createdAt":"2020-10-04T09:22:07Z","silentRounds":23,"language":"C#","topics":["csharp","dotnet","game-engine","gamedev","indiedev"],"aiDims":["游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","C#开发","DirectX 12","图形编程","跨平台渲染"],"domainKey":"游戏引擎","pushedAt":"2023-12-09T05:04:02Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"dotnet","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"indiedev","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/Golle/Titan","ts":"2023-12-09T05:04:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aandrew-me/tgpt","owner":"aandrew-me","name":"tgpt","desc":"AI Chatbots in terminal for free","summaryCn":"试想一下，你正在写代码写到一半，突然卡壳想查个文档，还得特意打开浏览器","reasonCn":"技术上，它通过模拟网络请求，直接连接各种免费的AI聊天后端，屏蔽了复杂的API协议处理。它最大的亮点是极度的轻量和简洁，没有花哨的界面，就是一个单纯的命令行程序，通过Go语言编写，编译出的二进制文件随处可跑。","stars":3263,"starGrowth":1,"createdAt":"2023-04-17T14:11:25Z","silentRounds":4,"language":"Go","topics":["chatbot"],"aiDims":["AI应用","非AI-好玩","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["终端","AI聊天","Go","命令行工具","效率工具"],"domainKey":"终端","pushedAt":"2026-09-23T14:39:04Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/aandrew-me/tgpt","ts":"2026-09-23T14:39:04Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"2FastLabs/agent-squad","owner":"2FastLabs","name":"agent-squad","desc":"Flexible and powerful framework for managing multiple AI agents and handling complex conversations","summaryCn":"专为苹果生态打造的强大智能体编排框架，让多个AI分工合作处理复杂任务。","reasonCn":"基于Swift语言构建的灵活且强大的多智能体管理框架，专注于复杂对话场景下的任务分工与协同。它提供了一套优雅的API设计，允许开发者在苹果生态中轻松调度多个独立的人工智能代理，实现复杂的业务逻辑处理，填补了Swift在现代人工智能代理编排领域的生态空白。","stars":7769,"starGrowth":1,"createdAt":"2024-07-23T12:48:30Z","silentRounds":1,"language":"Swift","topics":["chatbot","agents"],"aiDims":["AI Agent","多模态Agent","开发者工具"],"zone":"AI","funScore":0,"domainTags":["Swift","多智能体","AI框架","对话编排","Agent协作"],"domainKey":"Swift","pushedAt":"2026-09-23T11:09:25Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/2FastLabs/agent-squad","ts":"2026-09-23T11:09:25Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"flowable/flowable-engine","owner":"flowable","name":"flowable-engine","desc":"","summaryCn":"该引擎的核心优势在于其全面遵循国际主流的业务流程建模标准，这意味着开发","reasonCn":"对于开发人员而言，上手该平台非常友好，因为它提供了丰富的Java和REST API。通过这些接口，开发者能够轻松构建可视化界面，让业务人员直接参与到流程设计中。无论是自动化处理日常的OA办公申请，还是构建精密的金融风控规则引擎，Flowable都能通过配置化方式快速响应业务的变化，","stars":9551,"starGrowth":3,"createdAt":"2016-10-13T07:21:43Z","silentRounds":0,"language":"Java","topics":["bpmn","java","workflow","workflow-engine"],"aiDims":["效率工具"],"zone":"工具","funScore":0,"domainTags":["BPMN","工作流引擎","企业级","业务流程"],"domainKey":"BPMN","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"bpmn","source":"github","weight":0.7},{"name":"java","source":"github","weight":0.7},{"name":"workflow","source":"github","weight":0.7},{"name":"workflow-engine","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/flowable/flowable-engine","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Nukesor/pueue","owner":"Nukesor","name":"pueue","desc":":stars: Manage your shell commands.","summaryCn":"跑命令不用盯着终端了，后台排队随时暂停恢复","reasonCn":"Pueue是Rust编写的命令行队列管理工具，支持在后台运行命令并随时暂停恢复。你可以把编译、训练等耗时任务丢进队列，自己继续干活。每个任务有独立日志和状态，比nohup更灵活。支持任务依赖设置和优先级调整，适合多命令并行工作流。开发者能同时管理多个命令而不用开一堆终端。","stars":6344,"starGrowth":3,"createdAt":"2015-09-04T16:24:23Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行任务队列","后台进程管理","任务调度","Shell自动化","Rust CLI工具"],"domainKey":"命令行任务队列","pushedAt":"2026-09-09T10:57:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Nukesor/pueue","ts":"2026-09-09T10:57:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ain/smartbanner.js","owner":"ain","name":"smartbanner.js","desc":"Customisable and lightweight smart app banner for iOS and Android","summaryCn":"让网页长出会认手机的小横幅，打开就提示装应用","reasonCn":"这是一个面向移动网页的轻量横幅组件，在手机打开站点时显示可配置的应用安装引导。它会根据设备系统和浏览器环境，展示符合智能横幅规范的提示条，支持自定义标题、图标、按钮文字与点击行为。开发者初始化时传入商店地址、截图和样式参数，就能把网页流量引到应用内。","stars":656,"starGrowth":0,"createdAt":"2016-05-29T09:40:13Z","silentRounds":7,"language":"JavaScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["移动网页","应用引导","转化工具","前端组件"],"domainKey":"移动网页","pushedAt":"2026-09-15T07:22:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/ain/smartbanner.js","ts":"2026-09-15T07:22:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lfnovo/open-notebook","owner":"lfnovo","name":"open-notebook","desc":"An Open Source implementation of Notebook LM with more flexibility and features","summaryCn":"把Google NotebookLM搬进本地，不仅能用还能深度定制","reasonCn":"这是一款开源的NotebookLM替代品，用TypeScript构建，主打灵活性和功能丰富。它允许用户上传本地文档，通过RAG技术让AI理解并生成内容，不仅支持基础问答，还能进行长文本总结、思维导图生成等，完全脱离Google云服务。","stars":39389,"starGrowth":32,"createdAt":"2024-10-21T17:58:46Z","silentRounds":0,"language":"TypeScript","topics":["note-taking","learning"],"aiDims":["AI应用","AI写作","文档理解"],"zone":"AI","funScore":0,"domainTags":["RAG","文档问答","知识库","NotebookLM","AI应用"],"domainKey":"RAG","pushedAt":"2026-09-20T21:49:41Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"learning","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/lfnovo/open-notebook","ts":"2026-09-20T21:49:41Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09843722866790221,"copyOk":false},{"repo":"firstcontributions/first-contributions","owner":"firstcontributions","name":"first-contributions","desc":"🚀✨ Help beginners to contribute to open source projects","summaryCn":"手把手教你完成开源社区第一次代码贡献的新手村专属训练场","reasonCn":"这个项目的精妙之处在于它把贡献开源变成了一个极其标准化的游戏化通关任务。它没有深奥的代码逻辑，你只需要按照网页上的图文指引，在它的特定文件里把自己的名字加上去，然后提交一个合并请求。整个过程把原本高大上的开源贡献拆解成了最基础的几步，让小白在没有任何心理负担的","stars":56043,"starGrowth":13,"createdAt":"2016-09-20T14:35:09Z","silentRounds":0,"language":"","topics":["tutorial"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["开源贡献","新手指南","教程"],"domainKey":"开源贡献","pushedAt":"2026-09-23T21:38:38Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/firstcontributions/first-contributions","ts":"2026-09-23T21:38:38Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03969589422407794,"copyOk":true},{"repo":"pink-kinematics/pink","owner":"pink-kinematics","name":"pink","desc":"Python inverse kinematics using Pinocchio and QP solvers","summaryCn":"Pinocchio 逆运动学库，让机器人控制更轻松","reasonCn":"pink 使用 Python 封装 Pinocchio 动力学库，提供快速、精准的逆运动学求解。项目内置 QP 求解器，支持多约束、优先级规划，适用于工业机器人、仿真仿人机学等场景。代码结构模块化，API 简洁，易于在 ROS、PyBullet 等框架集成。","stars":845,"starGrowth":0,"createdAt":"2022-02-06T10:05:13Z","silentRounds":2,"language":"Python","topics":["robotics"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["逆运动学","机器人控制","Pinocchio","QP求解器","运动规划"],"domainKey":"逆运动学","pushedAt":"2026-09-10T08:01:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/pink-kinematics/pink","ts":"2026-09-10T08:01:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hwdsl2/setup-ipsec-vpn","owner":"hwdsl2","name":"setup-ipsec-vpn","desc":"Set up your own IPsec VPN server in just a few minutes, with IPsec/L2TP, Cisco IPsec and IKEv2. Supports Ubuntu, Debian, CentOS/RHEL, Alpine Linux and Raspberry Pi OS. Includes client config and management scripts.","summaryCn":"三分钟自建VPN服务器，手机电脑随时随地安全上网，支持多种操作系统","reasonCn":"这是一个极简的VPN服务器一键搭建脚本，支持IPsec/L2TP、Cisco IPsec和IKEv2三种协议，兼容Ubuntu、Debian、CentOS、Alpine和树莓派系统。安装过程全自动，自动生成客户端配置文件，适合需要自建VPN的个人用户和小团队。","stars":28550,"starGrowth":12,"createdAt":"2016-01-07T16:44:12Z","silentRounds":0,"language":"Shell","topics":["security","raspberry-pi","privacy","self-hosted"],"aiDims":["安全工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["VPN","Linux","网络工具","安全"],"domainKey":"VPN","pushedAt":"2026-09-21T05:23:51Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/hwdsl2/setup-ipsec-vpn","ts":"2026-09-21T05:23:51Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.037260945709281956,"copyOk":false},{"repo":"open-duelyst/duelyst","owner":"open-duelyst","name":"duelyst","desc":"","summaryCn":"2020 年，由 Counterplay Games 开发的热门数字卡牌游戏《Duelyst》停止运营","reasonCn":"与需要联网的在线游戏不同，该项目支持完全本地化的游戏部署。对于怀旧玩家来说，最吸引人的特性是本地模式下的“全解锁”机制：所有曾经需要对战获取的阵营和稀有卡牌都向玩家开放，同时系统会赠送额外的金币（Gold）和精神值（Spirit），让玩家能够无门槛地自由组建和测试各种极端","stars":3905,"starGrowth":0,"createdAt":"2022-03-23T18:21:27Z","silentRounds":3,"language":"JavaScript","topics":[],"aiDims":["非AI-好玩","游戏","卡牌策略"],"zone":"创意","funScore":0.5,"domainTags":["卡牌游戏","回合制策略","数字收藏","本地服务器","Docker"],"domainKey":"卡牌游戏","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"卡牌策略","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/open-duelyst/duelyst","ts":"2026-09-18T20:41:13.636Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ModelEngine-Group/nexent","owner":"ModelEngine-Group","name":"nexent","desc":"Nexent is a zero-code platform for auto-generating production-grade AI agents using Harness Engineering principles — unified tools, skills, memory, and orchestration with built-in constraints, feedback loops, and control planes.","summaryCn":"很多人想做AI Agent，但一看到复杂的Prompt工程和底层代码就头疼。","reasonCn":"它的设计理念非常硬核，借鉴了Harness工程化原则。这意味着它不像市面上的玩具，而是奔着“生产可用”去的。它内置了严格的约束检查机制，比如你可以设定AI只能查阅哪些特定文档，绝对不能超出权限，这种可控性对于企业应用来说是至关重要的。","stars":5884,"starGrowth":2,"createdAt":"2025-04-28T10:44:33Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["Agent编排","RAG","Prompt工程"],"domainKey":"RAG","pushedAt":"2026-09-23T09:47:07Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ModelEngine-Group/nexent","ts":"2026-09-23T09:47:07Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kornia/kornia","owner":"kornia","name":"kornia","desc":"🐍 Geometric Computer Vision Library for Spatial AI","summaryCn":"在传统的计算机视觉中，处理图像旋转、缩放或者三维空间投影通常依赖OpenCV","reasonCn":"核心亮点在于它的所有算子都原生支持PyTorch的张量运算、GPU加速以及自动求导。这意味着你可以把图像几何变换、相机内外参计算、立体匹配等复杂操作直接写进神经网络的损失函数里。和传统的OpenCV相比，它不仅运算速度极快，而且支持反向传播，极大地推动了自动驾驶、AR/VR、机器人","stars":11378,"starGrowth":3,"createdAt":"2018-08-22T10:31:37Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI基础设施","视觉与图像"],"zone":"工具","funScore":0,"domainTags":["计算机视觉","PyTorch","深度学习","图像几何"],"domainKey":"计算机视觉","pushedAt":"2026-09-23T19:49:19Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"视觉与图像","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/kornia/kornia","ts":"2026-09-23T19:49:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Arnold-n/P1P2MQTT","owner":"Arnold-n","name":"P1P2MQTT","desc":"Monitor and control Daikin/Rotex Altherma heat pumps from Home Assistant directly via the 2-wire P1/P2 thermostat interface. Also enables reading other Japanese Home Bus System based standards, a.o. Hitachi H-link, Hisense H-NET, MHI, Daikin DIII-NET (F1/F2) bus, Mitsubishi M-Net bus, Toshiba TCC-Link, Panasonic/Sanyo SIII-Net, Haier, and York.","summaryCn":"很多家庭在使用 Daikin 或 Rotex Altherma 系列的热泵时，想把它们接入 Hom","reasonCn":"很多家庭在使用 Daikin 或 Rotex Altherma 系列的热泵时，想把它们接入 Home Assistant 却苦于缺少官方的网络协议支持，只能依赖昂贵的专用网关或手动编写复杂的 Modbus 解析。P1P2MQTT 正是为了解决这个痛点而生：它通过热泵自带的两线 P1/P2 接口直接读取内部总线信息，并把这些","stars":493,"starGrowth":0,"createdAt":"2019-03-12T08:28:33Z","silentRounds":9,"language":"C++","topics":["home-automation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["热泵控制","Home Assistant集成","MQTT桥接","ESP32","P1/P2总线"],"domainKey":"热泵控制","pushedAt":"2026-09-13T19:55:29Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Arnold-n/P1P2MQTT","ts":"2026-09-13T19:55:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"felixrieseberg/macintosh.js","owner":"felixrieseberg","name":"macintosh.js","desc":"","summaryCn":"该项目是一个完全运行在 Electron 框架内的 Mac OS 8 仿真环境，它成功地在","reasonCn":"它的使用场景非常广泛，既是计算机爱好者的掌上玩具，也是技术人员研究早期 UI 设计的活化石。虽然它是纯粹用 JavaScript 构建的，但在现代主流系统上运行起来意外流畅，界面中的菜单、窗口控制以及简单的程序切换都精准还原了当年的体验。","stars":8497,"starGrowth":1,"createdAt":"2020-07-14T19:45:49Z","silentRounds":3,"language":"JavaScript","topics":["javascript","macintosh","emulator","electron"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["macOS 8","复古仿真","Electron","怀旧","操作系统"],"domainKey":"macOS 8","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"javascript","source":"github","weight":0.7},{"name":"macintosh","source":"github","weight":0.7},{"name":"emulator","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/felixrieseberg/macintosh.js","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MacDownApp/macdown","owner":"MacDownApp","name":"macdown","desc":"","summaryCn":"MacDown是一款专为macOS打造的开源Markdown编辑器，其诞生背景颇为特别，作","reasonCn":"安装过程极其简便，用户可以直接下载应用包将文件拖入应用程序文件夹，或者在终端执行Homebrew Cask命令“brew install --cask macdown”一键完成部署。这种零配置的特性降低了使用门槛，使得即使是非技术背景的普通用户也能快速上手。","stars":9836,"starGrowth":1,"createdAt":"2014-06-22T08:27:06Z","silentRounds":5,"language":"Rich Text Format","topics":["markdown","markdown-editor","markdown-to-html","markdown-convertor","markdown-writer","markdown-viewer","markdown-converter","objective-c","cocoa","macos","xcode"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Markdown","编辑器","macOS","写作","富文本"],"domainKey":"Markdown","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"markdown","source":"github","weight":0.7},{"name":"markdown-editor","source":"github","weight":0.7},{"name":"markdown-to-html","source":"github","weight":0.7},{"name":"markdown-convertor","source":"github","weight":0.7},{"name":"markdown-writer","source":"github","weight":0.7},{"name":"markdown-viewer","source":"github","weight":0.7},{"name":"markdown-converter","source":"github","weight":0.7},{"name":"objective-c","source":"github","weight":0.7},{"name":"cocoa","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"xcode","source":"github","weight":0.7},{"name":"Rich Text Format","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/MacDownApp/macdown","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Skyvern-AI/skyvern","owner":"Skyvern-AI","name":"skyvern","desc":"Automate browser based workflows with AI","summaryCn":"让AI帮你自动操作浏览器，完成各种重复性网页任务","reasonCn":"基于大语言模型理解网页内容，自动生成并执行浏览器操作序列。配合截图分析页面结构，能完成点击、输入、滚动等动作，支持Playwright和Selenium后端，可自动化处理表单填写、数据抓取、登录验证等任务，大幅减少人工重复操作。","stars":23054,"starGrowth":1,"createdAt":"2024-02-28T15:45:19Z","silentRounds":1,"language":"Python","topics":["automation","llm"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["浏览器自动化","网页任务编排","AI Agent","RPA","Playwright"],"domainKey":"AI Agent","pushedAt":"2026-09-23T04:35:25Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Skyvern-AI/skyvern","ts":"2026-09-23T04:35:25Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KeygraphHQ/shannon","owner":"KeygraphHQ","name":"shannon","desc":"Shannon is an AI pentester for web applications and APIs. It analyzes your source code, identifies attack vectors, and executes real exploits to prove vulnerabilities before they reach production.","summaryCn":"用AI自动找Web漏洞，边分析代码边执行真实攻击，比渗透工程师还狠","reasonCn":"核心技术在于AI驱动的漏洞分析引擎。它不是简单扫描端口或特征匹配，而是理解代码逻辑，识别业务层面的安全缺陷。支持Web应用和API两类目标，能生成可复现的PoC证明漏洞存在。相比传统工具如Burp Suite，它的优势在于主动推理而非被动扫描。","stars":48292,"starGrowth":16,"createdAt":"2025-09-27T01:51:47Z","silentRounds":0,"language":"TypeScript","topics":["security"],"aiDims":["安全工具","AI Agent","AI应用"],"zone":"AI","funScore":0.8,"domainTags":["AI Security","Pentesting","Cybersecurity","Automated","Vulnerability"],"domainKey":"AI Security","pushedAt":"2026-09-21T12:58:11Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/KeygraphHQ/shannon","ts":"2026-09-21T12:58:11Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05879274413981611,"copyOk":false},{"repo":"nvm-sh/nvm","owner":"nvm-sh","name":"nvm","desc":"Node Version Manager - POSIX-compliant bash script to manage multiple active node.js versions. $nvm: 3ArcxqLtXMmBnWbbtfwQgVL3MNnDsggzgGDtXMnjpump","summaryCn":"JS开发者的版本管理神器，换版本不用删重装","reasonCn":"这是一个POSIX兼容的bash脚本，用于管理多个活跃的Node.js版本。开发者可以通过它轻松切换不同版本的Node环境，解决依赖冲突，无需卸载重装。支持一键安装、卸载、查看已安装列表，配置文件简单直观，极大提升了前端和全栈开发的工作效率。","stars":95167,"starGrowth":10,"createdAt":"2010-04-15T17:47:47Z","silentRounds":0,"language":"Shell","topics":["nvm","nodejs","node","node-js","version-manager","posix","posix-compliant","bash","zsh","shell","nvmrc","lts","install"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Node.js","版本管理","命令行"],"domainKey":"Node.js","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"nvm","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"node","source":"github","weight":0.7},{"name":"node-js","source":"github","weight":0.7},{"name":"version-manager","source":"github","weight":0.7},{"name":"posix","source":"github","weight":0.7},{"name":"posix-compliant","source":"github","weight":0.7},{"name":"bash","source":"github","weight":0.7},{"name":"zsh","source":"github","weight":0.7},{"name":"shell","source":"github","weight":0.7},{"name":"nvmrc","source":"github","weight":0.7},{"name":"lts","source":"github","weight":0.7},{"name":"install","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/nvm-sh/nvm","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030315235323168745,"copyOk":true},{"repo":"gyrdym/ml_linalg","owner":"gyrdym","name":"ml_linalg","desc":"SIMD-based linear algebra and statistics for data science with dart","summaryCn":"Dart写统计和矩阵时不用憋着等的实用数学小工具","reasonCn":"面向Dart语言的数值计算库，将矩阵运算、线性方程组、分解和统计函数封装成本地代码，并借助SIMD指令提升向量与数组吞吐。适用于移动、桌面和服务端的数据处理、特征工程、概率抽样、优化算法与机器学习原型，能减少手写循环和重复实现基础数学函数的成本，也能让复杂算法在普通设备上更顺手一些。","stars":88,"starGrowth":0,"createdAt":"2017-07-01T08:34:17Z","silentRounds":6,"language":"Dart","topics":["mathematics"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["线性代数","数值计算","Dart语言","SIMD优化","数据科学"],"domainKey":"线性代数","pushedAt":"2026-09-13T17:24:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/gyrdym/ml_linalg","ts":"2026-09-13T17:24:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SwiftOldDriver/iOS-Weekly","owner":"SwiftOldDriver","name":"iOS-Weekly","desc":"🇨🇳 老司机技术 iOS 周报","summaryCn":"iOS开发者的周刊小报，每周速递最新技术动态、库和工具，帮你省时省力。","reasonCn":"这是一份由国内开发者维护的iOS技术周刊，每周整理发布移动端开发领域的优质文章、工具库、开源项目和行业资讯。虽然不涉及AI核心业务，但内容覆盖Swift、UIKit、架构设计等实用领域，是移动端开发者获取行业资讯、跟踪技术趋势的实用信息源。","stars":4984,"starGrowth":0,"createdAt":"2017-12-27T03:11:16Z","silentRounds":22,"language":"JavaScript","topics":["mobile"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["iOS开发","技术周刊","Swift","移动端架构","开源库推荐"],"domainKey":"iOS开发","pushedAt":"2026-09-13T16:15:36Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/SwiftOldDriver/iOS-Weekly","ts":"2026-09-13T16:15:36Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GD4AI/obsidian-llm-wiki","owner":"GD4AI","name":"obsidian-llm-wiki","desc":"Karpathy's LLM Wiki implementation plugin for Obsidian - turns notes and PDFs into a linked, LLM-powered knowledge base with entity pages, concept pages, graph-powered Q&A, and local-first privacy.","summaryCn":"Obsidian本身是一个很强大的笔记工具，但面对大量的笔记，查找和关联确实是","reasonCn":"从技术角度看，它结合了RAG（检索增强生成）和知识图谱技术。它会把你的笔记文本切分成小块，向量化后存入向量数据库。当你提问时，它会先在向量库里找到最相关的片段，然后把这些片段作为上下文喂给大语言模型，让模型基于这些上下文生成答案。","stars":648,"starGrowth":2,"createdAt":"2026-04-26T14:57:39Z","silentRounds":0,"language":"TypeScript","topics":["note-taking","knowledge-management"],"aiDims":["RAG与知识","知识图谱","文档理解"],"zone":"AI","funScore":0,"domainTags":["Obsidian","RAG","知识图谱","本地AI"],"domainKey":"RAG","pushedAt":"2026-09-23T06:15:52Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/GD4AI/obsidian-llm-wiki","ts":"2026-09-23T06:15:52Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"eugeniughelbur/obsidian-second-brain","owner":"eugeniughelbur","name":"obsidian-second-brain","desc":"Persistent memory for Claude Code and 6 other CLI agents, stored as plain markdown in your Obsidian vault. Stop re-explaining your projects, decisions and people every session. 45 commands: hybrid semantic search, self-rewriting notes, key-less web research, and scheduled agents that maintain the vault while you sleep.","summaryCn":"想象一下，你有一个超级聪明的AI助手，但每次开始工作时，你都要花半小时给","reasonCn":"技术上，它通过构建一个混合语义索引引擎，实时同步你的本地Markdown文件，将其转化为AI可直接检索的知识向量。它巧妙地利用了Obsidian作为轻量级本地化知识库的优势，通过Python脚本驱动，实现了45个自定义命令，让AI Agent可以自由调用你的笔记，完成跨文件的复杂逻辑梳理，这在","stars":4581,"starGrowth":13,"createdAt":"2026-03-24T11:58:12Z","silentRounds":0,"language":"Python","topics":["note-taking","knowledge-management"],"aiDims":["AI Agent","效率工具","RAG与知识"],"zone":"AI","funScore":0.7,"domainTags":["AI记忆","知识管理","Obsidian插件"],"domainKey":"AI记忆","pushedAt":"2026-09-21T20:18:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/eugeniughelbur/obsidian-second-brain","ts":"2026-09-21T20:18:32Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.05701611001964636,"copyOk":false},{"repo":"surrealdb/surrealdb","owner":"surrealdb","name":"surrealdb","desc":"A scalable, distributed, collaborative, document-graph database, for the realtime web","summaryCn":"文档加图谱二合一的数据库，存数据和查关系一次搞定，超酷","reasonCn":"SurrealDB是一个创新的混合数据库，同时支持文档存储和图关系查询。用Rust编写，性能出色，支持分布式部署和实时数据同步。开发者可以用一种查询语言同时处理JSON文档和复杂图关系，特别适合需要实时协作的场景，如在线文档、社交网络、即时通讯等应用。","stars":33055,"starGrowth":0,"createdAt":"2021-12-09T15:03:29Z","silentRounds":1,"language":"Rust","topics":["database","developer-tools"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","图数据库","实时协作","Rust"],"domainKey":"分布式数据库","pushedAt":"2026-09-14T23:11:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.38,"source":"trending","url":"https://github.com/surrealdb/surrealdb","ts":"2026-09-14T23:11:22Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"WordPress/gutenberg","owner":"WordPress","name":"gutenberg","desc":"","summaryCn":"WordPress 引入全新区块编辑与协作范式的底层核心开发中心。","reasonCn":"作为 WordPress 颠覆传统发布体验的核心项目，它通过全新的区块编辑器彻底改变了网站构建方式。项目目前正处于横跨编辑、定制、协作及多语言的四个演进阶段中，旨在为全球创作者打造更直观、灵活且现代化的数字内容发布流程。","stars":11756,"starGrowth":0,"createdAt":"2017-02-03T19:34:19Z","silentRounds":4,"language":"JavaScript","topics":["wordpress","gutenberg","wordpress-plugin"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["WordPress","区块编辑器","前端开发","CMS","内容管理"],"domainKey":"WordPress","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"wordpress","source":"github","weight":0.7},{"name":"gutenberg","source":"github","weight":0.7},{"name":"wordpress-plugin","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/WordPress/gutenberg","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aldinokemal/go-whatsapp-web-multidevice","owner":"aldinokemal","name":"go-whatsapp-web-multidevice","desc":"GOWA - WhatsApp REST API with support for UI, Multi Account, Webhooks, and MCP, and Chatwoot. Built with Golang for efficient memory use.","summaryCn":"用代码操控WhatsApp，支持多账号登录和自动化消息收发","reasonCn":"基于WhatsApp Web协议用Go语言实现的REST API服务，支持多设备同时登录、消息收发、文件传输和webhook回调。可接入Chatwoot客服系统和MCP协议，无需官方API审批即可快速搭建WhatsApp自动化机器人，适合个人和小型团队集成开发。","stars":4852,"starGrowth":3,"createdAt":"2022-02-09T16:19:30Z","silentRounds":0,"language":"Go","topics":["whatsapp","whatsapp-api","whatsapp-api-go","golang","golang-whatsapp","golang-whatsapp-api","whatsapp-web-multi-device","bot","go","whatsapp-multi-device","rest","rest-api"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["WhatsApp","REST API","多设备登录","消息自动化","webhook"],"domainKey":"WhatsApp","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"whatsapp","source":"github","weight":0.7},{"name":"whatsapp-api","source":"github","weight":0.7},{"name":"whatsapp-api-go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"golang-whatsapp","source":"github","weight":0.7},{"name":"golang-whatsapp-api","source":"github","weight":0.7},{"name":"whatsapp-web-multi-device","source":"github","weight":0.7},{"name":"bot","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"whatsapp-multi-device","source":"github","weight":0.7},{"name":"rest","source":"github","weight":0.7},{"name":"rest-api","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/aldinokemal/go-whatsapp-web-multidevice","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"databasus/databasus","owner":"databasus","name":"databasus","desc":"PostgreSQL backup tool with Point-In-Time-Recovery and restore verification","summaryCn":"PostgreSQL 数据库的自动备份神器，恢复验证一步到位","reasonCn":"这是一个基于 Go 语言开发的高效备份工具，专门针对 PostgreSQL 数据库设计。它不仅能自动进行 Point-In-Time-Recovery（PITR）备份，还能在恢复完成后自动进行校验，确保数据100%准确。项目以 7900+ 星标的热度证明了其强大实用，让数据库运维变得简单可靠。","stars":8626,"starGrowth":7,"createdAt":"2025-06-05T13:10:03Z","silentRounds":0,"language":"Go","topics":["devops","database"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["PostgreSQL","备份工具","PITR","数据恢复"],"domainKey":"PostgreSQL","pushedAt":"2026-09-22T19:20:44Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/databasus/databasus","ts":"2026-09-22T19:20:44Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.023434500347785766,"copyOk":false},{"repo":"ageerle/ruoyi-ai","owner":"ageerle","name":"ruoyi-ai","desc":"An enterprise AI development framework for building AI agents. It provides unified management of multi-provider LLMs, secure enterprise knowledge bases with high-precision retrieval, visual workflow orchestration and multi-agent coordination. Compatible with mainstream Agent Skill standards, it enables developers to efficiently build production-gra","summaryCn":"如果你是个企业开发者，想给公司业务系统加上AI大脑，却又不想从零搞定模型","reasonCn":"它的技术核心在于对RAG（检索增强生成）流程的重构。不仅支持常见的向量数据库存储，还引入了精细化的知识处理逻辑，确保模型输出不仅聪明而且准确。框架层面采用了Java生态中成熟的配置管理，支持灵活切换不同的LLM厂商，无需大幅修改代码即可实现模型平替。","stars":5719,"starGrowth":1,"createdAt":"2024-01-16T04:16:41Z","silentRounds":1,"language":"Java","topics":["rag"],"aiDims":["AI Agent","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["RAG","Java框架","企业级"],"domainKey":"RAG","pushedAt":"2026-09-23T09:39:00Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ageerle/ruoyi-ai","ts":"2026-09-23T09:39:00Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NanmiCoder/MediaCrawler","owner":"NanmiCoder","name":"MediaCrawler","desc":"小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频 ｜ 评论爬虫、微博帖子 ｜ 评论爬虫、百度贴吧帖子 ｜ 百度贴吧评论回复爬虫 | 知乎问答文章｜评论爬虫","summaryCn":"一键爬取小红书抖音B站数据，社交内容采集利器","reasonCn":"技术上，它采用Playwright自动化浏览器驱动，模拟真实用户登录和浏览行为，避免了传统API请求容易被反爬机制封禁的问题。支持多种平台：小红书可以通过账号密码或二维码登录采集笔记和评论；抖音、快手支持视频元数据和评论抓取；B站可采集视频信息、弹幕和评论；微博支持帖子和评","stars":65552,"starGrowth":26,"createdAt":"2023-06-09T12:14:34Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["社交媒体爬虫","Playwright","数据采集","小红书爬虫","B站爬虫"],"domainKey":"社交媒体爬虫","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/NanmiCoder/MediaCrawler","ts":"2026-08-13T01:37:33.040Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07918989504515499,"copyOk":false},{"repo":"dropwizard/dropwizard","owner":"dropwizard","name":"dropwizard","desc":"A damn simple library for building production-ready RESTful web services.","summaryCn":"Java领域构建生产级RESTful服务的经典极简框架，稳定可靠","reasonCn":"Dropwizard是一个成熟的Java Web框架，将Jersey、Jetty等组件打包，提供开箱即用的生产环境配置。虽然不直接涉及AI模型或智能体，但它是构建后端API服务、集成AI推理接口的绝佳基础设施，适合Java生态下的工程化开发。","stars":8580,"starGrowth":0,"createdAt":"2011-01-19T19:52:29Z","silentRounds":24,"language":"Java","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","工具调用"],"zone":"工具","funScore":0,"domainTags":["Java","RESTful","微服务","监控","Web框架"],"domainKey":"Java","pushedAt":"2026-09-22T00:19:56Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/dropwizard/dropwizard","ts":"2026-09-22T00:19:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hipcityreg/situation-monitor","owner":"hipcityreg","name":"situation-monitor","desc":"","summaryCn":"该项目通过集成底层的状态捕获逻辑，解决了在信息碎片化环境下难以高效追踪","reasonCn":"核心优势在于其对于特定监控目标的轻量化设计，项目采用 TypeScript 构建，具备较好的类型安全性与可维护性。对于有自定义监控需求的技术开发者而言，该代码库提供了一个稳固的逻辑骨架，方便在不同的业务场景中快速复用与适配，无论是用于简单的状态抓取还是复杂的环境预警，都能","stars":4204,"starGrowth":2,"createdAt":"2026-01-07T18:07:09Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["监控系统","TypeScript","环境监测"],"domainKey":"监控系统","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/hipcityreg/situation-monitor","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"photoprism/photoprism","owner":"photoprism","name":"photoprism","desc":"AI-Powered Photos App 🌈💎✨","summaryCn":"AI自动帮你整理照片，告别手动打标签的痛苦","reasonCn":"这是一款基于AI的自托管照片管理应用，使用计算机视觉技术自动识别照片内容、人脸和场景，支持关键词搜索和智能分类。你可以把它想象成开源版的Google Photos，隐私更安全，功能同样强大，是个人和团队管理大量照片的理想选择。","stars":40228,"starGrowth":5,"createdAt":"2018-01-27T12:00:15Z","silentRounds":0,"language":"Go","topics":["ml","ai"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI照片管理","计算机视觉","自托管","TensorFlow","图像识别"],"domainKey":"AI照片管理","pushedAt":"2026-09-23T13:40:32Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/photoprism/photoprism","ts":"2026-09-23T13:40:32Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015372874614696231,"copyOk":true},{"repo":"NVIDIA/SkillSpector","owner":"NVIDIA","name":"SkillSpector","desc":"Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, security risks, prompt injection, data exfiltration, and supply-chain risks in Claude Code, Codex, and MCP skills before you install them.","summaryCn":"AI智能体的安全体检仪，专治各种 Prompt 注入","reasonCn":"这是 NVIDIA 出品的安全扫描工具，专门用来检测 AI Agent 和技能包中的漏洞。它能发现 Prompt 注入、数据泄露、供应链风险等恶意模式，确保 Claude Code、Codex 和 MCP 技能包在上线前是安全的，防止被黑客攻击或滥用。","stars":18109,"starGrowth":109,"createdAt":"2026-03-21T00:28:43Z","silentRounds":0,"language":"Python","topics":["agent-security","agent-skills","agentic-ai","ai-security","claude-code","mcp","prompt-injection","security-scanner","security-tools","security-workflow","supply-chain-security"],"aiDims":["安全工具","AI Agent","AI基础设施"],"zone":"AI","funScore":0.7,"domainTags":["AI安全","Prompt注入检测","MCP","代码审计"],"domainKey":"AI安全","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agent-security","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"ai-security","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"prompt-injection","source":"github","weight":0.7},{"name":"security-scanner","source":"github","weight":0.7},{"name":"security-tools","source":"github","weight":0.7},{"name":"security-workflow","source":"github","weight":0.7},{"name":"supply-chain-security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/NVIDIA/SkillSpector","ts":"2026-09-19T04:31:19.347Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.3816687834778287,"copyOk":false},{"repo":"chronoxor/CppTrader","owner":"chronoxor","name":"CppTrader","desc":"High performance components for building Trading Platform such as ultra fast matching engine, order book processor","summaryCn":"高速C++匹配引擎，帮你轻松搭建交易平台","reasonCn":"CppTrader 提供超低延迟的撮合引擎和高效的订单簿处理模块，基于 lock‑free 数据结构和 SIMD 优化，实现微秒级成交；支持多市场、多品种，具备插件式风控与持久化，模块化设计，可单独部署撮合或行情服务，提供 C++ 与 Python 接口。","stars":1072,"starGrowth":1,"createdAt":"2017-07-19T13:42:08Z","silentRounds":10,"language":"C++","topics":["trading"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["撮合引擎","高频交易","订单簿"],"domainKey":"撮合引擎","pushedAt":"2026-09-09T00:15:24Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/chronoxor/CppTrader","ts":"2026-09-09T00:15:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"antvis/L7","owner":"antvis","name":"L7","desc":"🌎 Large-scale WebGL-powered Geospatial Data Visualization analysis engine.","summaryCn":"用WebGL把海量地理数据在浏览器里渲染成炫酷地图可视化","reasonCn":"AntV旗下的大型地理空间数据可视化引擎，基于WebGL实现，支持在浏览器中高效渲染数百万级地理坐标点、热力图、航线图等多种可视化图层。兼容Mapbox、OpenLayers、Leaflet等主流地图框架，提供React/Vue组件封装，适合做数据大屏、地理分析平台和可视化仪表盘。","stars":4064,"starGrowth":1,"createdAt":"2018-11-21T16:19:07Z","silentRounds":23,"language":"TypeScript","topics":["data-visualization","maps"],"aiDims":["可视化","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["地理数据可视化","WebGL","地图"],"domainKey":"地理数据可视化","pushedAt":"2026-09-08T21:36:22Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"maps","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/antvis/L7","ts":"2026-09-08T21:36:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"go-flutter-desktop/go-flutter","owner":"go-flutter-desktop","name":"go-flutter","desc":"Flutter on Windows, MacOS and Linux - based on Flutter Embedding, Go and GLFW.","summaryCn":"Go程序员用它把Flutter界面变成电脑桌面软件","reasonCn":"这个项目让懂Go语言的人能直接调用Flutter框架，把网页风格的UI打包成Windows、macOS和Linux的原生桌面程序。它绕过了官方限制，让你用熟悉的Go语言开发跨平台GUI应用，非常适合有Go背景的开发者快速构建桌面工具，不用去学Dart也能写出漂亮的界面。","stars":5931,"starGrowth":0,"createdAt":"2018-08-19T22:04:27Z","silentRounds":22,"language":"Go","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["桌面开发","GUI框架","跨平台","Flutter"],"domainKey":"桌面开发","pushedAt":"2026-09-08T18:57:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/go-flutter-desktop/go-flutter","ts":"2026-09-08T18:57:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"codexu/note-gen","owner":"codexu","name":"note-gen","desc":"Capture first. Organize later. A local-first Markdown app that turns scattered records into clear notes with AI.","summaryCn":"本地优先的 AI 笔记助手，碎片化记录秒变清晰笔记。","reasonCn":"技术上，它利用了 RAG（检索增强生成）技术，将你的本地文档向量化和索引，让 AI 能够理解上下文并进行结构化输出。它不需要联网调用 API，所有处理都在本地完成，数据安全完全由你掌控。它支持从剪贴板直接导入文本，也支持从本地文件批量导入。","stars":12830,"starGrowth":1,"createdAt":"2024-08-06T05:20:16Z","silentRounds":1,"language":"TypeScript","topics":["rag"],"aiDims":["AI写作","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["本地笔记","RAG","Markdown编辑器","隐私保护"],"domainKey":"RAG","pushedAt":"2026-09-23T00:49:18Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/codexu/note-gen","ts":"2026-09-23T00:49:18Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Pasta-Devs/Marinara-Engine","owner":"Pasta-Devs","name":"Marinara-Engine","desc":"AI frontend","summaryCn":"想做 AI 应用，但不想被繁琐的前端代码拖累？","reasonCn":"它的核心优势在于“开发者体验”。传统的前端开发需要自己处理 WebSocket 连接、流式输出、Markdown 渲染、错误重试等一堆细节。Marinara 把这些都封装好了，你只需要关注业务逻辑。比如，它内置了流式响应处理，能让 AI 的回答像打字机一样逐字显示；内置了 Markdown 渲染器，让代码","stars":671,"starGrowth":2,"createdAt":"2026-03-04T15:37:29Z","silentRounds":0,"language":"TypeScript","topics":["ai","ai-agents","ai-agents-framework","frontend","local"],"aiDims":["AI应用","工具调用","创意工具"],"zone":"工具","funScore":0.5,"domainTags":["前端框架","TypeScript","流式渲染","Markdown渲染","WebSocket"],"domainKey":"前端框架","pushedAt":"2026-09-22T17:31:17Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-agents-framework","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"local","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/Pasta-Devs/Marinara-Engine","ts":"2026-09-22T17:31:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Facepunch/sbox-public","owner":"Facepunch","name":"sbox-public","desc":"s&box is a modern game engine, built on Valve's Source 2 and the latest .NET technology, it provides a modern intuitive editor for creating games","summaryCn":"Valve官方开源游戏引擎，用C#重写Source 2，做游戏超顺手","reasonCn":"这是Valve亲手打造的现代游戏引擎，基于Source 2底层，用C#和.NET重写，拥有可视化编辑器。支持物理模拟、光照渲染、网络多人游戏等核心游戏开发能力，兼容Steam平台。对想做3D游戏的团队和个人开发者来说，是难得能直接摸到源码、自由修改的高自由度开源项目。","stars":6470,"starGrowth":1,"createdAt":"2025-11-24T08:18:33Z","silentRounds":3,"language":"C#","topics":["gamedev","sbox","source2"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["游戏引擎","Source 2","C#","游戏开发"],"domainKey":"游戏引擎","pushedAt":"2026-08-18T20:59:09Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"gamedev","source":"github","weight":0.7},{"name":"sbox","source":"github","weight":0.7},{"name":"source2","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.52,"source":"bigbro","url":"https://github.com/Facepunch/sbox-public","ts":"2026-08-18T20:59:09Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LanRhyme/MicYou","owner":"LanRhyme","name":"MicYou","desc":"MicYou is a powerful tool that turns your Android device into a high-quality microphone for your PC.","summaryCn":"闲置安卓手机别吃灰，秒变PC外接高音质麦克风利器","reasonCn":"基于Rust开发的跨平台音频采集工具，通过局域网或USB将安卓设备麦克风数据实时传输到电脑。支持低延迟音频流、降噪处理和多种音频格式，无需Root即可使用。对比同类方案如DroidShip、WO Mic等，MicYou在音质和稳定性上有明显优势，且开源免费、持续活跃维护。","stars":3953,"starGrowth":8,"createdAt":"2026-02-10T08:44:24Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["音频采集","Android","Rust","局域网"],"domainKey":"音频采集","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/LanRhyme/MicYou","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03608560586896028,"copyOk":false},{"repo":"javascript-tutorial/uk.javascript.info","owner":"javascript-tutorial","name":"uk.javascript.info","desc":"Modern JavaScript Tutorial in Ukrainian (JavaScript українською 🇺🇦)","summaryCn":"JavaScript的学习往往容易陷入碎片化，很多新手学了一堆框架却不理解语言本","reasonCn":"教程的亮点在于它的严谨性与深度。不同于那些只教快餐式编码的教程，它会深入讲解闭包、原型链、事件循环、微任务队列等核心机制。每一章节都配有丰富的互动练习，通过直接在浏览器里修改代码，让你在实践中掌握JavaScript的每个细微概念。","stars":393,"starGrowth":0,"createdAt":"2019-04-14T14:23:28Z","silentRounds":17,"language":"HTML","topics":["tutorial"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["JavaScript","教程","编程学习","前端"],"domainKey":"JavaScript","pushedAt":"2026-08-31T05:09:18Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/javascript-tutorial/uk.javascript.info","ts":"2026-08-31T05:09:18Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OtterMind/Chat2DB","owner":"OtterMind","name":"Chat2DB","desc":"🔥🔥🔥 AI-driven database tool and SQL client, The hottest GUI client, supporting MySQL, Oracle, PostgreSQL, DB2, SQL Server, DB2, SQLite, H2, ClickHouse, and more.","summaryCn":"AI帮你写SQL查数据库，啥数据库都能连的神器","reasonCn":"Chat2DB是一款AI驱动的数据库管理工具，支持MySQL、PostgreSQL、Oracle等十余种主流数据库。核心亮点是内置AI能力，用户只需输入自然语言描述查询需求，即可自动生成SQL语句，还能智能解释执行结果、优化查询性能。提供跨平台桌面客户端和Web版本，界面现代简洁，极大降低了数据库操作门槛。","stars":28180,"starGrowth":7,"createdAt":"2023-06-20T14:05:26Z","silentRounds":0,"language":"Java","topics":["database","llm"],"aiDims":["AI应用","效率工具","工具调用"],"zone":"AI","funScore":0,"domainTags":["数据库管理","AI辅助","SQL"],"domainKey":"数据库管理","pushedAt":"2026-09-23T09:40:49Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/OtterMind/Chat2DB","ts":"2026-09-23T09:40:49Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.021745209368346348,"copyOk":true},{"repo":"paradedb/paradedb","owner":"paradedb","name":"paradedb","desc":"One Postgres for your application data, full-text search, vector retrieval, and aggregations. Home of the pg_search extension.","summaryCn":"给Postgres装上AI的向量检索和全文搜索功能","reasonCn":"ParadeDB是一个PostgreSQL扩展，让数据库原生支持向量检索和全文搜索。它集成了pgvector和pg_search，让Postgres不仅能做传统SQL查询，还能做语义搜索和混合检索，是构建AI应用的后端基础设施。相比单独部署向量数据库，它免去了数据同步的麻烦。","stars":9298,"starGrowth":2,"createdAt":"2023-06-30T20:21:47Z","silentRounds":0,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI搜索"],"zone":"AI","funScore":0,"domainTags":["向量数据库","PostgreSQL扩展","语义检索","混合搜索","RAG基础设施"],"domainKey":"向量数据库","pushedAt":"2026-09-23T21:50:43Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/paradedb/paradedb","ts":"2026-09-23T21:50:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"antfu-collective/taze","owner":"antfu-collective","name":"taze","desc":"","summaryCn":"Taze 的设计理念在于提供极致的简洁与效率，特别是通过 npx 命令即插即用，","reasonCn":"在功能细节上，它非常注重安全性，始终尊重用户预设的版本号范围，避免了因强制升级导致的破坏性更改。除了常规的 npm 包更新，该工具还支持同步 Node.js 环境的 pinned 版本以及 GitHub Actions 的工作流配置，甚至可以进行 SHA 签名绑定，从而确保 CI/CD 流程的稳定性。","stars":4282,"starGrowth":0,"createdAt":"2020-07-22T05:48:15Z","silentRounds":3,"language":"TypeScript","topics":["dependencies","cli-tool","updates"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["包管理","依赖更新","命令行工具"],"domainKey":"包管理","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"dependencies","source":"github","weight":0.7},{"name":"cli-tool","source":"github","weight":0.7},{"name":"updates","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/antfu-collective/taze","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HFQR/xitca-web","owner":"HFQR","name":"xitca-web","desc":"An HTTP library and web framework for safe Rust ","summaryCn":"用Rust写的极速Web框架，让服务器跑得更快更稳","reasonCn":"xitca-web是基于Rust的异步HTTP框架，专为高性能Web服务设计。它采用零成本抽象理念，利用Tokio运行时实现非阻塞I/O，让开发者能用安全的Rust代码构建高速API服务。相比Actix和Axum，它的API设计更简洁直观，同时保持了Rust内存安全的优势。","stars":1009,"starGrowth":0,"createdAt":"2021-05-09T06:49:54Z","silentRounds":22,"language":"Rust","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Rust","Web框架","HTTP库","异步编程"],"domainKey":"Rust","pushedAt":"2026-09-16T05:04:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/HFQR/xitca-web","ts":"2026-09-16T05:04:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Yoctol/bottender","owner":"Yoctol","name":"bottender","desc":"","summaryCn":"Bottender 的一个核心设计理念是将对话 UI 的复杂性抽象化。","reasonCn":"在多平台支持方面，该框架解决了一个常见的痛点：不同聊天平台的交互能力差异巨大。例如，有些平台支持富文本和交互按钮，有些则仅支持纯文本。Bottender 允许开发者使用“渐进式增强”或“优雅降级”策略，即编写一套核心逻辑，然后根据渠道能力自动适配展示形式，或者在低能力渠道上","stars":4280,"starGrowth":0,"createdAt":"2017-10-16T09:40:49Z","silentRounds":4,"language":"TypeScript","topics":["messaging","messenger","line","slack","telegram","bot","chatbot","framework","viber"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["TypeScript","聊天机器人","声明式框架","多渠道集成","对话UI"],"domainKey":"TypeScript","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"messaging","source":"github","weight":0.7},{"name":"messenger","source":"github","weight":0.7},{"name":"line","source":"github","weight":0.7},{"name":"slack","source":"github","weight":0.7},{"name":"telegram","source":"github","weight":0.7},{"name":"bot","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"viber","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/Yoctol/bottender","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JustVugg/colibri","owner":"JustVugg","name":"colibri","desc":"Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦","summaryCn":"纯C零依赖引擎，让普通电脑也能跑超大MoE模型","reasonCn":"colibri是用纯C实现的轻量化推理引擎，专门用于运行前沿的Mixture‑of‑Experts大模型，零第三方依赖，模型专家参数可从磁盘流式加载，显著降低内存占用，使普通CPU甚至嵌入式设备也能进行推理。兼容主流ONNX和PyTorch导出模型。","stars":37130,"starGrowth":232,"createdAt":"2026-07-01T12:27:49Z","silentRounds":0,"language":"C","topics":[],"aiDims":["推理引擎","AI基础设施","模型部署"],"zone":"AI","funScore":0.6,"domainTags":["MoE模型推理","纯C推理引擎","边缘设备部署","模型流式加载","ONNX兼容"],"domainKey":"MoE模型推理","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/JustVugg/colibri","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.47811742526259093,"copyOk":false},{"repo":"open-compass/opencompass","owner":"open-compass","name":"opencompass","desc":"OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.","summaryCn":"开源大模型评测平台，支持100多个数据集全面测模型","reasonCn":"OpenCompass就是一个这样的大模型评测平台。它支持Llama、Mistral、GPT-4、LLaMa2、Qwen、GLM、Claude等数十种主流模型，在100多个数据集上进行评估。评测维度包括语言理解（如CMU Movie Review）、数学推理（如MATH、GSM8K）、代码能力（如HumanEval）、安全对齐等。","stars":7468,"starGrowth":1,"createdAt":"2023-06-15T12:42:58Z","silentRounds":1,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","模型评估","效率工具"],"zone":"AI","funScore":0.6,"domainTags":["大模型评测","基准测试","评估平台","模型选型"],"domainKey":"大模型评测","pushedAt":"2026-09-23T09:48:34Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型评估","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.87,"source":"search","url":"https://github.com/open-compass/opencompass","ts":"2026-09-23T09:48:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"claration/Impactor","owner":"claration","name":"Impactor","desc":"Cross-platform &amp; feature rich iOS/iPadOS/tvOS sideloading application. Formerly known as PlumeImpactor.","summaryCn":"跨平台给iOS设备装第三方应用的开源神器","reasonCn":"Impactor是一个开源的跨平台应用，让你能在iOS、iPadOS、tvOS设备上侧载和安装第三方应用，无需越狱。支持macOS、Windows、Linux，功能丰富。对于想要安装 TestFlight 应用之外应用的iOS用户来说，这是一个非常有用的工具，让你重新掌控自己的设备。","stars":3318,"starGrowth":4,"createdAt":"2025-10-17T10:48:04Z","silentRounds":0,"language":"Rust","topics":["altstore","ios","ipa","ipados","livecontainer","sideload","sideloading","sideloadly","sidestore","signature","signer","signing","sideloader"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["iOS侧载","证书签名","跨平台桌面应用","应用管理","移动端部署"],"domainKey":"iOS侧载","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"altstore","source":"github","weight":0.7},{"name":"ios","source":"github","weight":0.7},{"name":"ipa","source":"github","weight":0.7},{"name":"ipados","source":"github","weight":0.7},{"name":"livecontainer","source":"github","weight":0.7},{"name":"sideload","source":"github","weight":0.7},{"name":"sideloading","source":"github","weight":0.7},{"name":"sideloadly","source":"github","weight":0.7},{"name":"sidestore","source":"github","weight":0.7},{"name":"signature","source":"github","weight":0.7},{"name":"signer","source":"github","weight":0.7},{"name":"signing","source":"github","weight":0.7},{"name":"sideloader","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/claration/Impactor","ts":"2026-09-07T02:44:20.540Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"K2SOsint/Legendary_OSINT","owner":"K2SOsint","name":"Legendary_OSINT","desc":"A list of OSINT tools &amp; resources for (fraud-)investigators, CTI-analysts, KYC, AML and more.","summaryCn":"侦探级OSINT情报资源库，1600+星标，揭秘互联网公开信息挖掘技巧","reasonCn":"这是一个专门面向欺诈调查员、网络安全情报分析师、KYC反洗钱从业者的开源情报（OSINT）资源合集。项目收录了从被动信息收集、主动渗透测试到恶意软件分析的全套工具链，涵盖域名/IP定位、社交媒体情报挖掘、图像反向搜索、暗网数据追踪等核心能力，是安全调查和合规审查领域的实用工具宝库。","stars":1744,"starGrowth":1,"createdAt":"2025-09-12T14:33:21Z","silentRounds":2,"language":"","topics":["cti","cyberthreatintelligence","osint","osint-tools-list","socmint"],"aiDims":["安全工具","工具调用","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["开源情报","安全研究","威胁情报","合规审查"],"domainKey":"开源情报","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"cti","source":"github","weight":0.7},{"name":"cyberthreatintelligence","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"osint-tools-list","source":"github","weight":0.7},{"name":"socmint","source":"github","weight":0.7}],"aiScore":0.52,"source":"trending","url":"https://github.com/K2SOsint/Legendary_OSINT","ts":"2026-08-08T00:55:52.206Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"gregorik/InstantOrganicCaves","owner":"gregorik","name":"InstantOrganicCaves","desc":"An extremely lightweight, purist C++ engine for UE5 to generate procedural cave & rock structures at runtime.","summaryCn":"UE5里一键生成程序化洞穴，游戏美术师的梦想工具","reasonCn":"为UE5打造的极致轻量级C++引擎，专门用于在运行时实时生成程序化洞穴和岩石结构。不需要预烘焙数据，纯算法生成，性能开销极小。对于想做开放世界、 Roguelike、地牢类游戏的工作室来说，这意味着可以动态生成无限变化的地下场景，大幅减少美术工作量。","stars":41,"starGrowth":0,"createdAt":"2026-01-09T18:49:56Z","silentRounds":9,"language":"C++","topics":["procedural-generation"],"aiDims":["游戏","创意工具"],"zone":"创意","funScore":1,"domainTags":["程序化生成","UE5","洞穴生成"],"domainKey":"程序化生成","pushedAt":"2026-09-08T07:14:26Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/gregorik/InstantOrganicCaves","ts":"2026-09-08T07:14:26Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"1Panel-dev/1Panel","owner":"1Panel-dev","name":"1Panel","desc":"🔥 1Panel is a modern, open-source Linux server management panel and a lightweight AI management platform.","summaryCn":"把服务器变成一块会说话的中控屏，运维不再靠吼","reasonCn":"它用后端语言写成，专注服务器日常运维，把容器、网站、数据库、监控、备份、防火墙和日志管理收进网页界面。除了传统面板能力，它还加入轻量智能应用入口，方便把模型服务、接口应用和开发实例放在同一处维护，减少反复敲命令的麻烦。适合个人站长、小团队和开发者做服务器运维、容器编排与智能应用实验。","stars":37000,"starGrowth":6,"createdAt":"2022-07-19T15:44:52Z","silentRounds":0,"language":"Go","topics":["docker"],"aiDims":["开发者工具","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["服务器管理","Docker","运维","容器编排","Web面板"],"domainKey":"服务器管理","pushedAt":"2026-09-23T10:17:35Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/1Panel-dev/1Panel","ts":"2026-09-23T10:17:35Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.018486486486486486,"copyOk":false},{"repo":"NirDiamant/RAG_Techniques","owner":"NirDiamant","name":"RAG_Techniques","desc":"This repository showcases various advanced techniques for Retrieval-Augmented Generation (RAG) systems. Each technique has a detailed notebook tutorial.","summaryCn":"手把手教你从零搭RAG系统，20多种进阶技巧全部配有实操笔记","reasonCn":"这是一个RAG（检索增强生成）技术的系统性教程仓库，包含Naive RAG、HyDE、Query Rewriting、Re-ranking、Multimodal RAG等二十余种主流和前沿方案的详细实现。每个技巧都配有完整的Jupyter Notebook，从零开始搭建到高级变种全覆盖，代码可直接运行复现。","stars":29586,"starGrowth":14,"createdAt":"2024-07-13T16:08:36Z","silentRounds":0,"language":"Jupyter Notebook","topics":["rag","ml","vector-db","ai","llm"],"aiDims":["RAG与知识","教程","AI应用"],"zone":"资源","funScore":0,"domainTags":["RAG系统","检索增强生成","Jupyter笔记","向量检索"],"domainKey":"RAG系统","pushedAt":"2026-09-21T10:54:13Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"vector-db","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/NirDiamant/RAG_Techniques","ts":"2026-09-21T10:54:13Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.04341959034678564,"copyOk":false},{"repo":"goq/telegram-list","owner":"goq","name":"telegram-list","desc":"List of telegram groups, channels & bots // Список интересных групп, каналов и ботов телеграма // Список чатов для программистов","summaryCn":"程序员Telegram资源宝库，一屏找到所有好群组好频道","reasonCn":"这个项目本质上是一个精心编排的开源资源索引，把Telegram上对程序员有价值的群组、频道和机器人分类整理。内容包括中文和英文社区，按技术栈细分：前端、后端、AI/ML、DevOps、移动端、安全等。每个条目都经过筛选，确保社群质量。","stars":4976,"starGrowth":0,"createdAt":"2016-08-12T16:16:12Z","silentRounds":5,"language":"","topics":["frontend","devops","data-science"],"aiDims":["非AI-实用","学习资源","开发者工具"],"zone":"资源","funScore":0,"domainTags":["资源导航","社群列表","Telegram","开发者社区"],"domainKey":"资源导航","pushedAt":"2026-08-30T10:11:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7}],"aiScore":0.5,"source":"search","url":"https://github.com/goq/telegram-list","ts":"2026-08-30T10:11:27Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KKKKhazix/human-writing","owner":"KKKKhazix","name":"human-writing","desc":"让 AI 写的中文读起来像一个具体的人在说话。通用创作与改稿 Skill，开箱即用。","summaryCn":"你是不是经常遇到这种情况：让 AI 写一篇小红书文案或公众号文章，结果读起","reasonCn":"从技术角度来看，这个项目的核心在于“风格化”和“人性化”的重构。它深入研究了中文的表达习惯，识别出 AI 生成文本中的典型特征，比如过度使用连接词、句式结构单一、情感表达夸张或不自然等。通过特定的算法和提示词工程（Prompt Engineering），它能对原始生成内容进行二次加工，","stars":3780,"starGrowth":5,"createdAt":"2026-08-05T01:19:41Z","silentRounds":0,"language":"Python","topics":["agent-skills","chinese-writing","creative-writing","writing-skill"],"aiDims":[],"zone":"AI","funScore":0.7,"domainTags":["中文写作","提示词工程","AI改写"],"domainKey":"中文写作","pushedAt":"2026-08-11T15:07:45Z","tags":[{"name":"agent-skills","source":"github","weight":0.7},{"name":"chinese-writing","source":"github","weight":0.7},{"name":"creative-writing","source":"github","weight":0.7},{"name":"writing-skill","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"bigbro","url":"https://github.com/KKKKhazix/human-writing","ts":"2026-08-11T15:07:45Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02845238095238095,"copyOk":false},{"repo":"public-api-lists/public-api-lists","owner":"public-api-lists","name":"public-api-lists","desc":"A curated list of free public APIs — searchable, community-maintained, with a free JSON API.","summaryCn":"收录全球免费公开API的宝藏清单，开发者调接口必备手册","reasonCn":"这是一个由全球开发者社区共同维护的免费公开API索引库，收录了数千个可免费调用的接口服务，覆盖天气查询、支付对接、地理定位、图像识别、AI大模型等热门领域。项目提供友好的搜索功能和JSON格式API，方便开发者快速查找和集成。虽然本身不是AI项目，但现代AI应用几乎都需要调用各种外部API来增强功能，比如接入ChatGPT API、调用图像生成服务等，这个清单相当于AI开发者的接口淘宝，帮你快速找到需要的第三方服务。","stars":15873,"starGrowth":11,"createdAt":"2020-01-05T20:34:00Z","silentRounds":0,"language":"","topics":["developer-tools","awesome"],"aiDims":["开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["公共API","开发者工具","API目录","免费接口"],"domainKey":"公共API","pushedAt":"2026-09-14T07:19:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.35,"source":"search","url":"https://github.com/public-api-lists/public-api-lists","ts":"2026-09-14T07:19:57Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.03507900207900208,"copyOk":true},{"repo":"StarTrail-org/LEANN","owner":"StarTrail-org","name":"LEANN","desc":"[MLsys2026]: RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.","summaryCn":"本地跑RAG还能省97%存储，隐私全在自己手里","reasonCn":"LEANN是一个主打本地私有化的RAG框架，能在个人设备上实现快速准确的检索增强生成。核心亮点是97%的存储节省，支持多种向量数据库后端和嵌入模型。无需联网调用云端API，数据安全完全掌握在自己手中，适合对隐私敏感的个人开发者和中小企业。","stars":12955,"starGrowth":5,"createdAt":"2025-06-09T06:52:59Z","silentRounds":0,"language":"Python","topics":["vector-db","rag"],"aiDims":["RAG与知识","向量数据库","AI应用"],"zone":"工具","funScore":0,"domainTags":["RAG","向量数据库","压缩","隐私"],"domainKey":"RAG","pushedAt":"2026-09-22T23:00:56Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/StarTrail-org/LEANN","ts":"2026-09-22T23:00:56Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01615785411038209,"copyOk":false},{"repo":"EKKOLearnAI/hermes-studio","owner":"EKKOLearnAI","name":"hermes-studio","desc":"Web dashboard for Hermes Agent — multi-platform AI chat, session management, scheduled jobs, usage analytics ","summaryCn":"AI Agent的可视化操控台，多平台聊天统一管理就像开公司后台","reasonCn":"Hermes AI Agent的Web管理面板，支持多平台消息接入、会话管理、定时任务配置和使用数据分析。图形化界面让开发者无需写代码即可配置Agent行为，监控运行状态，查看对话历史和使用统计，是部署和维护AI Agent应用的一站式工具。","stars":11177,"starGrowth":12,"createdAt":"2026-04-11T07:38:18Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","AI应用"],"zone":"工具","funScore":0,"domainTags":["AI Agent","会话管理","定时任务"],"domainKey":"AI Agent","pushedAt":"2026-09-22T10:34:39Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/EKKOLearnAI/hermes-studio","ts":"2026-09-22T10:34:39Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04706508007515433,"copyOk":true},{"repo":"tuyoogame/huatuo","owner":"tuyoogame","name":"huatuo","desc":"huatuo是一个特性完整、零成本、高性能、低内存的近乎完美的Unity全平台原生c#热更方案。 Huatuo is a fully featured, zero-cost, high-performance, low-memory solution for Unity's all-platform native c# hotfix","summaryCn":"Unity开发者的热更新救星，改代码不用等编译发布","reasonCn":"Huatuo是针对IL2CPP平台的Unity热更方案，用C++实现高性能热修复引擎。核心是通过方法注入技术实现不重新编译的热更新，相比ILRuntime等竞品，它在内存占用和运行性能上有明显优势，支持全平台原生C#热更，适合对热更性能敏感的游戏项目。","stars":2339,"starGrowth":0,"createdAt":"2022-01-11T03:50:55Z","silentRounds":24,"language":"C++","topics":["csharp","hot-fix","hot-update","hotfix","hotupdate","il2cpp","ilruntime","injectfix","lua","puerts","slua","tolua","unity","unlua","unreal","update","xlua"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["热更新","Unity","IL2CPP"],"domainKey":"热更新","pushedAt":"2023-02-16T03:17:00Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"hot-fix","source":"github","weight":0.7},{"name":"hot-update","source":"github","weight":0.7},{"name":"hotfix","source":"github","weight":0.7},{"name":"hotupdate","source":"github","weight":0.7},{"name":"il2cpp","source":"github","weight":0.7},{"name":"ilruntime","source":"github","weight":0.7},{"name":"injectfix","source":"github","weight":0.7},{"name":"lua","source":"github","weight":0.7},{"name":"puerts","source":"github","weight":0.7},{"name":"slua","source":"github","weight":0.7},{"name":"tolua","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unlua","source":"github","weight":0.7},{"name":"unreal","source":"github","weight":0.7},{"name":"update","source":"github","weight":0.7},{"name":"xlua","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.52,"source":"bigbro","url":"https://github.com/tuyoogame/huatuo","ts":"2023-02-16T03:17:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Scottcjn/rustchain-dialup","owner":"Scottcjn","name":"rustchain-dialup","desc":"Dial-up ISP + RustChain BBS + VoIP↔analog bridge for vintage hardware (Dreamcast, 386/486, 9600 baud modems) — mine RustChain over a phone line","summaryCn":"用90年代拨号互联网在古董电脑上挖Rust币的硬核实验","reasonCn":"这是一款面向复古计算爱好者的奇葩项目。它搭建了一个基于模拟电话线的拨号ISP环境，集成了RustChain BBS和VoIP转接桥，让Dreamcast或386/486电脑通过9600波特率调制解调器连接网络。最有趣的地方在于，你可以在这种极度受限的模拟信号环境中运行共识算法并挖矿，体验致敬早期互联网精神的极客乐趣，兼具技术挑战与怀旧情怀。","stars":17,"starGrowth":0,"createdAt":"2026-06-02T14:53:59Z","silentRounds":12,"language":"C","topics":["retrocomputing"],"aiDims":["非AI-好玩","学习资源"],"zone":"创意","funScore":1,"domainTags":["复古计算","拨号上网","RustChain","BBS"],"domainKey":"复古计算","pushedAt":"2026-09-07T18:11:07Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Scottcjn/rustchain-dialup","ts":"2026-09-07T18:11:07Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lucidrains/deep-daze","owner":"lucidrains","name":"deep-daze","desc":"","summaryCn":"用CLIP和Siren将文字描述转化为迷幻艺术图像的命令行工具。","reasonCn":"这个项目的核心原理是利用OpenAI的CLIP模型理解文本语义，配合Siren（神经隐式表示）技术将文本映射为图像。用户只需在命令行输入一段文字描述，比如「mist over green hills」（绿丘迷雾）或「a psychedelic experience on LSD」（LSD迷幻体验），就能生成对应的艺术化图像。","stars":4312,"starGrowth":0,"createdAt":"2021-01-17T06:00:50Z","silentRounds":5,"language":"Python","topics":["artificial-intelligence","deep-learning","transformers","siren","implicit-neural-representation","text-to-image","multi-modality"],"aiDims":["非AI-好玩"],"zone":"创意","funScore":0.5599999999999999,"domainTags":["文本生成图像","CLIP","神经网络渲染"],"domainKey":"文本生成图像","pushedAt":"2026-09-17T17:48:34.813Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"artificial-intelligence","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"siren","source":"github","weight":0.7},{"name":"implicit-neural-representation","source":"github","weight":0.7},{"name":"text-to-image","source":"github","weight":0.7},{"name":"multi-modality","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/lucidrains/deep-daze","ts":"2026-09-17T17:48:34.813Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"VectifyAI/PageIndex","owner":"VectifyAI","name":"PageIndex","desc":"📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG","summaryCn":"不切块不建向量库，让模型像人一样通读全文再作答的检索生成方案。","reasonCn":"这是一个反传统的RAG方案。它抛掉了向量数据库和文本切片这两样标配，改用推理模型直接通读文档结构来定位答案。装上SDK后，既能在本地用自己的大模型密钥完成索引、检索和对话，也能接入云端服务。对处理长文档和整批资料库的场景尤其合适，逻辑更接近人翻书找答案。","stars":35807,"starGrowth":8,"createdAt":"2025-04-01T10:53:54Z","silentRounds":0,"language":"Python","topics":["rag","vector-db","ai","llm","agents"],"aiDims":["RAG与知识","文档理解","AI搜索"],"zone":"AI","funScore":0.6,"domainTags":["RAG","文档索引","LLM","推理架构"],"domainKey":"RAG","pushedAt":"2026-09-21T09:27:51Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"vector-db","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.98,"source":"trending","url":"https://github.com/VectifyAI/PageIndex","ts":"2026-09-21T09:27:51Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02467026000502695,"copyOk":true},{"repo":"DietrichGebert/ponytail","owner":"DietrichGebert","name":"ponytail","desc":"Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.","summaryCn":"让 LLM 变得像人一样用一行代码干活，省时省力更安全。","reasonCn":"Ponytail 是一款专为 AI 编程助手设计的工具包，旨在解决大语言模型（LLM）在写代码时容易产生冗余、失控甚至安全问题。它能自动将复杂的自然语言指令压缩成极其精简、安全且有效的单行代码。安装后，它就像一个尽职的副驾驶，监控并优化 AI 的每一次输出，确保效率提升的同时不牺牲安全性。","stars":144386,"starGrowth":561,"createdAt":"2026-06-12T00:52:37Z","silentRounds":0,"language":"JavaScript","topics":["llm"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.8,"domainTags":["AI Agent","LLM","代码生成","开发工具","模型切换"],"domainKey":"AI Agent","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/DietrichGebert/ponytail","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.37398750571385037,"copyOk":true},{"repo":"bootdotdev/curriculum","owner":"bootdotdev","name":"curriculum","desc":"An open roadmap for Boot.dev's backend development curriculum","summaryCn":"全套后端学习路线，帮你从零到实战一站搞定","reasonCn":"curriculum 项目公开了 Boot.dev 官方后端开发课程的完整路线图，系统列出算法、数据结构、系统设计、数据库与 API 实践等关键章节，并提供学习资料、练习题和实战项目链接，帮助学习者循序渐进地从零掌握后端全栈技能，适合自学和教学使用。","stars":3919,"starGrowth":0,"createdAt":"2021-01-26T03:36:38Z","silentRounds":3,"language":"","topics":["algorithms"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["后端学习","课程路线","编程教育"],"domainKey":"后端学习","pushedAt":"2026-09-10T18:47:55Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"algorithms","source":"github","weight":0.7}],"aiScore":0.42,"source":"search","url":"https://github.com/bootdotdev/curriculum","ts":"2026-09-10T18:47:55Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chiraggude/awesome-laravel","owner":"chiraggude","name":"awesome-laravel","desc":"","summaryCn":"这份资源列表明确标注其灵感来源于 PHP 社区的资深项目 ziadoz/awesome-php","reasonCn":"列表的内容结构覆盖了软件开发的完整生命周期，从最基础的 Essentials 如 Laravel 官方文档和 Laracasts 学习平台，到具体的 Packages 推荐，甚至延伸至 Application Hosting 和 Deployment 部署环节。这种全链路视角非常实用，因为很多开发者往往忽略了非代码层面的问题，比如如何","stars":13104,"starGrowth":0,"createdAt":"2014-07-30T08:39:02Z","silentRounds":5,"language":"","topics":["awesome","laravel","php"],"aiDims":["非AI-实用","学习资源","开发者工具"],"zone":"资源","funScore":0,"domainTags":["Laravel","PHP框架","资源列表","开发工具链","生态导航"],"domainKey":"Laravel","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"laravel","source":"github","weight":0.7},{"name":"php","source":"github","weight":0.7}],"aiScore":0.38,"source":"tier","url":"https://github.com/chiraggude/awesome-laravel","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bettercap/bettercap","owner":"bettercap","name":"bettercap","desc":"The Swiss Army knife for 802.11, BLE, HID, CAN-bus, IPv4 and IPv6 networks reconnaissance and MITM attacks.","summaryCn":"网络安全瑞士军刀，WiFi蓝牙网络抓包攻击一站式搞定","reasonCn":"核心能力包括无线网络侦察、蓝牙设备扫描、ARP欺骗、DHCP劫持、HTTP中间人攻击、DNS欺骗等。它采用模块化架构，每个攻击模块都是独立的可插拔组件，用户可以根据需要组合使用。其强大的脚本引擎支持用Lua或JavaScript编写自定义攻击模块，扩展性极强。","stars":20018,"starGrowth":4,"createdAt":"2018-01-07T15:30:41Z","silentRounds":0,"language":"Go","topics":["security"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["渗透测试","网络侦察","MITM攻击","安全工具"],"domainKey":"渗透测试","pushedAt":"2026-08-13T13:47:49Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/bettercap/bettercap","ts":"2026-08-13T13:47:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"vortico/flama","owner":"vortico","name":"flama","desc":"The production framework for Predictive and Generative AI. Serve any model as an API in one line, with OpenAI/Anthropic/Ollama-compatible endpoints, a built-in chat UI, and native MCP.","summaryCn":"一行代码把AI模型变成API服务，带聊天界面还能接MCP协议","reasonCn":"技术上，它的核心亮点是「零代码适配」。你只需要把模型路径和类型告诉它，它自动帮你生成兼容主流协议的API端点。同时内置了一个Web聊天界面，可以直接在浏览器里测试模型。更前沿的是它原生支持MCP（Model Context Protocol）协议，这意味着它可以作为AI Agent的工具提供者，轻松","stars":297,"starGrowth":0,"createdAt":"2018-09-27T15:29:54Z","silentRounds":11,"language":"Python","topics":["web-framework"],"aiDims":["AI基础设施","模型部署","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI部署","模型推理","API服务","MCP协议"],"domainKey":"AI部署","pushedAt":"2026-09-23T18:07:43Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/vortico/flama","ts":"2026-09-23T18:07:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ai-collection/ai-collection","owner":"ai-collection","name":"ai-collection","desc":"The Generative AI Landscape - A Collection of Awesome Generative AI Applications","summaryCn":"集合了多种艺术风格的AI创作工具，激发你的创造力","reasonCn":"这个项目是一个AI艺术集，它包含了各种风格和类型的AI生成艺术作品。无论是油画、水彩还是3D模型，都可以在平台上找到灵感并尝试自己创作。用户可以自定义尺寸、颜色和样式，甚至还能将作品导出为JPEG或PNG格式。此外，平台还提供了一些教程和示例，帮助用户更好地理解和使用这些工","stars":9172,"starGrowth":5,"createdAt":"2023-01-03T06:02:03Z","silentRounds":0,"language":"","topics":["generative-art"],"aiDims":["AI应用","创意工具"],"zone":"创意","funScore":1,"domainTags":["生成艺术","AI创作","数字艺术"],"domainKey":"生成艺术","pushedAt":"2026-09-22T09:05:49Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"generative-art","source":"github","weight":0.7}],"aiScore":0.9,"source":"search","url":"https://github.com/ai-collection/ai-collection","ts":"2026-09-22T09:05:49Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.01663541212385521,"copyOk":false},{"repo":"amruthpillai/reactive-resume","owner":"amruthpillai","name":"reactive-resume","desc":"A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!","summaryCn":"完全自托管的简历生成器，隐私至上、模板精美","reasonCn":"技术上，这个项目采用React + TypeScript + Tailwind CSS构建，界面现代流畅，提供多种精心设计的简历模板供选择。它支持多种数据导入方式，可以从LinkedIn、其他简历格式导入数据，避免从零填起。导出格式支持PDF和Markdown，满足不同场景需求。","stars":43331,"starGrowth":30,"createdAt":"2020-03-25T14:08:22Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted","ai"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["简历生成","自托管","隐私保护","PDF导出"],"domainKey":"简历生成","pushedAt":"2026-09-10T10:48:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/amruthpillai/reactive-resume","ts":"2026-09-10T10:48:00Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09207703491726477,"copyOk":false},{"repo":"opa334/Dopamine","owner":"opa334","name":"Dopamine","desc":"Dopamine is a semi-untethered jailbreak for iOS 15 to 26(.0.1)","summaryCn":"iOS越狱神器，让你的iPhone跑起来像新机一样丝滑。","reasonCn":"它的核心功能是让你获得“系统级”的控制权。比如安装插件来美化锁屏、控制中心，或者像“CallBar”这样的实用工具来接管来电显示。它还支持很多Cydia源里的插件，基本上你想改的系统功能都能找到对应的工具。 安装过程并不复杂，前提是你的设备处于特定版本。","stars":6716,"starGrowth":3,"createdAt":"2023-01-31T21:28:44Z","silentRounds":0,"language":"C","topics":[],"aiDims":["游戏","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["iOS越狱","半越狱","iPhone","系统定制"],"domainKey":"iOS越狱","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/opa334/Dopamine","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aryabharat/what-where_for_Coding_Interview","owner":"aryabharat","name":"what-where_for_Coding_Interview","desc":"List of all the important links and resources to look before appearing for a Coding Interview/Job. These links are used by me during my preparation. There might be other links that could fit in so please feel free to contribute.open_hands","summaryCn":"汇总面试必备资源链接，帮你快速搭建刷题知识库","reasonCn":"准备技术面试时，总是找不到系统的学习材料？这份仓库把互联网上的优质面试资源全部收集起来，像一本可搜索的面试手册，让你省去逐个搜寻的时间。  \n核心亮点在于分类细致：按算法（数组、链表、动态规划等）、系统设计（高可用、缓存、数据库）和软技能（行为问题、简历写作）三个","stars":62,"starGrowth":0,"createdAt":"2020-04-16T19:35:43Z","silentRounds":10,"language":"","topics":["interview"],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["编程面试准备","算法刷题资源","系统设计学习"],"domainKey":"编程面试准备","pushedAt":"2026-09-12T06:50:44Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7}],"aiScore":0.4,"source":"search","url":"https://github.com/aryabharat/what-where_for_Coding_Interview","ts":"2026-09-12T06:50:44Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"every-app/open-seo","owner":"every-app","name":"open-seo","desc":"Open source alternative to Semrush and Ahrefs","summaryCn":"开源免费替代Ahrefs和Semrush的神器，功能强大得离谱。","reasonCn":"Open SEO 是一款基于 TypeScript 开发的全能型 SEO 分析平台，开源替代了传统的商业闭源工具。它具备全面的关键词排名追踪、网站审计、竞争对手分析等功能，支持多语言和多站点管理。项目代码结构清晰，模块化设计良好，适合开发者二次开发或集成到自己的项目中，为网站运营者提供了高性价比的替代方案。","stars":20031,"starGrowth":126,"createdAt":"2026-02-27T18:57:41Z","silentRounds":0,"language":"TypeScript","topics":["backlink-analysis","keyword-research","mcp","seo","seo-tools","site-audit","google-search-console-mcp","seo-agent","seo-skills","seo-audit","seo-automation"],"aiDims":["非AI-实用","开发工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["SEO","数据分析","网站审计"],"domainKey":"SEO","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backlink-analysis","source":"github","weight":0.7},{"name":"keyword-research","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"seo","source":"github","weight":0.7},{"name":"seo-tools","source":"github","weight":0.7},{"name":"site-audit","source":"github","weight":0.7},{"name":"google-search-console-mcp","source":"github","weight":0.7},{"name":"seo-agent","source":"github","weight":0.7},{"name":"seo-skills","source":"github","weight":0.7},{"name":"seo-audit","source":"github","weight":0.7},{"name":"seo-automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/every-app/open-seo","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.38264490040437316,"copyOk":false},{"repo":"SWE-agent/SWE-agent","owner":"SWE-agent","name":"SWE-agent","desc":"SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024] ","summaryCn":"AI程序员自动读issue、写代码、修bug的开源神器","reasonCn":"它的核心架构是一个智能Agent循环：首先用LLM理解issue描述，然后用shell命令探索代码库，定位问题所在，接着修改代码、测试、提交PR，整个流程自动化。它不是一个简单的代码补全工具，而是真正能自主完成一个完整的软件工程任务。","stars":20384,"starGrowth":2,"createdAt":"2024-04-02T04:09:47Z","silentRounds":0,"language":"Python","topics":["developer-tools"],"aiDims":["AI Agent"],"zone":"AI","funScore":0,"domainTags":["GitHub Issue 修复","代码库探索","自动化开发","Agent 框架","LLM 应用"],"domainKey":"GitHub Issue 修复","pushedAt":"2026-09-21T22:02:50Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/SWE-agent/SWE-agent","ts":"2026-09-21T22:02:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ankitects/anki","owner":"ankitects","name":"anki","desc":"Anki is a smart spaced repetition flashcard program","summaryCn":"科学记忆神器，用间隔重复帮你记住任何知识","reasonCn":"基于间隔重复算法的智能记忆卡片程序，帮助用户长期记忆语言词汇、医学知识、编程概念等。支持自定义卡片、富媒体内容和同步功能，跨平台可用。算法经过数十年认知科学验证，比传统死记硬背效率高得多。适合学生、语言学习者、备考人群和任何需要长期记忆的场景。","stars":31430,"starGrowth":23,"createdAt":"2012-12-21T08:03:31Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["非AI-实用","学习资源","效率工具"],"zone":"工具","funScore":0,"domainTags":["记忆","间隔重复","背单词"],"domainKey":"记忆","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.25,"source":"trending","url":"https://github.com/ankitects/anki","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07119535475660198,"copyOk":true},{"repo":"ravynsoft/ravynos","owner":"ravynsoft","name":"ravynos","desc":"An open-source OS project that aims to provide source and binary compatibility with macOS® and a similar user experience.","summaryCn":"一个能让Linux跑macOS软件的开源操作系统，想法挺疯狂","reasonCn":"ravynos是一个开源操作系统项目，目标是让Linux原生兼容macOS软件和框架。它通过兼容层实现macOS源码和二进制兼容，让用户在Linux上获得类似macOS的体验。项目用C语言开发，基于Linux内核，目前6875个stars。开发操作系统级别项目难度极高，需要处理大量系统调用和API兼容问题。","stars":7031,"starGrowth":1,"createdAt":"2021-01-31T23:17:36Z","silentRounds":3,"language":"C","topics":["desktop"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"创意","funScore":1,"domainTags":["操作系统","兼容层","开源项目","macOS","Linux"],"domainKey":"操作系统","pushedAt":"2026-09-06T21:02:41Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/ravynsoft/ravynos","ts":"2026-09-06T21:02:41Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AllThingsSmitty/css-protips","owner":"AllThingsSmitty","name":"css-protips","desc":"A collection of tips to help take your CSS skills pro. 🕹 ","summaryCn":"前端选手的CSS秘籍大全，3万星的大神级资源库","reasonCn":"该项目整理了超过300个CSS实用技巧，涵盖现代布局、渐变动画、响应式设计、Flexbox、Grid等核心知识点，每个技巧都配有可运行的代码示例。虽然纯前端资源与AI无关，但高质量教程对开发者极具实用价值，30263 star证明其社区认可度极高，是前端入门到进阶的必看清单。","stars":30283,"starGrowth":0,"createdAt":"2015-08-29T12:29:03Z","silentRounds":1,"language":"","topics":["frontend","cheatsheet"],"aiDims":["教程","学习资源","非AI-实用"],"zone":"资源","funScore":0.5,"domainTags":["CSS","前端开发","代码技巧","学习资源","Cheatsheet"],"domainKey":"CSS","pushedAt":"2026-09-23T01:27:44Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"cheatsheet","source":"github","weight":0.7}],"aiScore":0.22,"source":"search","url":"https://github.com/AllThingsSmitty/css-protips","ts":"2026-09-23T01:27:44Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"markbeep/vvzapi","owner":"markbeep","name":"vvzapi","desc":"Unofficial REST API and search for ETHZ VVZ","summaryCn":"ETHZ 课程表搜索的非官方 API，Python 写的","reasonCn":"这是一个针对苏黎世联邦理工学院（ETHZ）课程搜索系统的非官方 REST API 接口。它允许开发者通过编程方式查询课程信息、时间表和教授资料，无需手动在网页上操作。项目使用 Python 开发，代码简洁，易于集成到其他应用中。","stars":15,"starGrowth":0,"createdAt":"2025-10-13T15:38:44Z","silentRounds":10,"language":"Python","topics":["course"],"aiDims":["非AI-实用"],"zone":"资源","funScore":0,"domainTags":["ETHZ","课程表","API","RESTful","苏黎世联邦理工"],"domainKey":"ETHZ","pushedAt":"2026-09-23T15:28:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"course","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/markbeep/vvzapi","ts":"2026-09-23T15:28:26Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"airbytehq/airbyte","owner":"airbytehq","name":"airbyte","desc":"Open-source data movement for ELT pipelines and AI agents — from APIs, databases & files to warehouses, lakes, and AI applications. Both self-hosted and Cloud.","summaryCn":"免费的数据搬运工，让AI随时喝到最新鲜的数据奶","reasonCn":"开源ELT数据管道平台，预置400+连接器覆盖MySQL、PostgreSQL、API、S3等数据源，可将数据自动同步到数据仓库、数据湖或AI应用。支持自托管和云端部署，可视化配置让非技术人员也能搭建数据流，是RAG系统和AI Agent的数据底座。","stars":22117,"starGrowth":6,"createdAt":"2020-07-27T23:55:54Z","silentRounds":0,"language":"Python","topics":["self-hosted"],"aiDims":["AI基础设施","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["数据集成","ETL","数据管道","API同步","数据仓库"],"domainKey":"数据集成","pushedAt":"2026-09-23T21:37:55Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/airbytehq/airbyte","ts":"2026-09-23T21:37:55Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018813853596780755,"copyOk":false},{"repo":"solidjs/solid","owner":"solidjs","name":"solid","desc":"A declarative, efficient, and flexible JavaScript library for building user interfaces.","summaryCn":"很多前端开发者在选择框架时会纠结于性能与易用性的平衡，React、Vue 等虽功","reasonCn":"很多前端开发者在选择框架时会纠结于性能与易用性的平衡，React、Vue 等虽功能丰富，却在大规模更新时仍会产生虚拟 DOM 的 diff 开销。Solid 的核心思想是“编译时响应”，它在编译阶段把所有状态依赖解析为最小化的更新指令，运行时直接操作真实 DOM，几乎没有额外的抽象层，从而实","stars":36068,"starGrowth":2,"createdAt":"2018-04-24T16:36:27Z","silentRounds":0,"language":"TypeScript","topics":["javascript","performance","solid","proxies","declarative","fine-grained","jsx","reactive"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["前端框架","JavaScript","响应式"],"domainKey":"前端框架","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"javascript","source":"github","weight":0.7},{"name":"performance","source":"github","weight":0.7},{"name":"solid","source":"github","weight":0.7},{"name":"proxies","source":"github","weight":0.7},{"name":"declarative","source":"github","weight":0.7},{"name":"fine-grained","source":"github","weight":0.7},{"name":"jsx","source":"github","weight":0.7},{"name":"reactive","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/solidjs/solid","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tomlooman/ActionRoguelike","owner":"tomlooman","name":"ActionRoguelike","desc":"Co-op Action Roguelike in Unreal Engine C++","summaryCn":"虚幻引擎C++合作动作Roguelike，四人联机刷怪打BOSS","reasonCn":"这是一个基于虚幻引擎4开发的合作动作Roguelike游戏项目，源码完全公开。项目包含武器系统、角色技能树、敌人生成波次、战利品掉落、多人联机等完整游戏机制。采用C++与蓝图混合开发，代码结构清晰模块化，是学习UE4游戏开发的优质参考案例。","stars":4597,"starGrowth":2,"createdAt":"2020-09-07T16:52:50Z","silentRounds":0,"language":"C++","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["虚幻引擎","Roguelike","多人联机","游戏开发"],"domainKey":"虚幻引擎","pushedAt":"2026-09-06T03:04:30Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/tomlooman/ActionRoguelike","ts":"2026-09-06T03:04:30Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"StevenBlack/hosts","owner":"StevenBlack","name":"hosts","desc":"🔒 Consolidating and extending hosts files from several well-curated sources. Optionally pick extensions for porn, social media, and other categories.","summaryCn":"一键拦截全网垃圾广告追踪，让网络清净又安全的神器","reasonCn":"这个项目整合了多个权威黑名单源，自动合并生成hosts文件，支持一键屏蔽广告、恶意网站和社交追踪。用户可按需开启色情、社交等分类拦截，无需改代码就能彻底净化网络环境。虽然和AI无关，但作为基础网络工具非常实用。","stars":31093,"starGrowth":4,"createdAt":"2012-04-12T20:22:50Z","silentRounds":0,"language":"Python","topics":["security","privacy"],"aiDims":["安全工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["hosts 管理","广告拦截","网络隐私"],"domainKey":"hosts 管理","pushedAt":"2026-09-23T20:17:05Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/StevenBlack/hosts","ts":"2026-09-23T20:17:05Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Lightning-AI/LitServe","owner":"Lightning-AI","name":"LitServe","desc":"A minimal Python framework for building custom AI inference servers with full control over logic, batching, and scaling.","summaryCn":"想象一下你写好了一个超酷的AI模型，但要把它变成别人能通过网页调用的API接","reasonCn":"技术上，它构建在高性能后端之上，利用了自动批处理（Auto-batching）和多模型流水线并行技术。与单纯的FastAPI相比，它处理并发请求的效率极高，能动态控制硬件资源利用率，最大化GPU吞吐量，且支持自定义逻辑，无论是视觉模型还是大语言模型都能无缝适配。","stars":3943,"starGrowth":0,"createdAt":"2023-12-12T14:45:03Z","silentRounds":5,"language":"Python","topics":["api"],"aiDims":["推理引擎","模型部署","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["推理服务器","模型部署","高性能","框架"],"domainKey":"模型部署","pushedAt":"2026-09-22T07:34:21Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Lightning-AI/LitServe","ts":"2026-09-22T07:34:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"karakeep-app/karakeep","owner":"karakeep-app","name":"karakeep","desc":"A self-hostable bookmark-everything app (links, notes and images) with AI-based automatic tagging and full text search","summaryCn":"自家服务器托管的智能书签神器，AI自动打标让整理资料彻底躺平","reasonCn":"这款应用专为收藏党打造，核心亮点是用AI技术实现链接、笔记和图片的统一收纳。它不仅仅是简单的剪藏，更通过自动标签和全文搜索，把乱糟糟的信息流变成了可检索的知识库。技术栈现代，支持自部署，数据完全掌握在自己手里，是极客和效率追求者的福音。","stars":29218,"starGrowth":5,"createdAt":"2024-02-06T12:28:31Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted"],"aiDims":["AI应用","效率工具","AI搜索"],"zone":"工具","funScore":0,"domainTags":["书签管理","自托管","全文搜索"],"domainKey":"书签管理","pushedAt":"2026-09-22T19:52:20Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/karakeep-app/karakeep","ts":"2026-09-22T19:52:20Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015513382161681155,"copyOk":false},{"repo":"abougouffa/minemacs","owner":"abougouffa","name":"minemacs","desc":"MinEmacs: an Emacs configuration framework for daily use","summaryCn":"让 Emacs 编辑器开箱即用的日常配置框架，笔记写作样样精通","reasonCn":"这是一个专为日常使用设计的 Emacs 配置框架，整合了笔记管理、代码编辑、任务追踪等核心功能。基于 Spacemacs 理念简化而来，去除了复杂的双层按键体系，让新手能快速上手。内置 Org-mode 用于笔记和知识管理，支持 markdown 实时预览，配合内置终端和文件浏览器，打造统一的桌面工作环境。","stars":203,"starGrowth":0,"createdAt":"2022-09-17T13:27:16Z","silentRounds":14,"language":"Emacs Lisp","topics":["note-taking"],"aiDims":["效率工具"],"zone":"工具","funScore":0,"domainTags":["Emacs","配置","笔记","IDE","生产力"],"domainKey":"Emacs","pushedAt":"2026-09-18T14:30:51Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"Emacs Lisp","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/abougouffa/minemacs","ts":"2026-09-18T14:30:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JHUISI/charm","owner":"JHUISI","name":"charm","desc":"Charm: A Framework for Rapidly Prototyping Cryptosystems","summaryCn":"把加密协议当乐高拼，几段代码就能跑起来看效果","reasonCn":"Charm面向密码学研究者和课程实验，把常见加解密、密钥交换、签名认证、密钥派生与协议状态整理成可组合组件。它用Python降低从论文到可运行原型的成本，方便替换算法、调整消息顺序和状态转移，观察不同组合下的安全行为，适合做教学演示、协议对比与安全机制快速验证方案。","stars":649,"starGrowth":0,"createdAt":"2011-05-15T17:59:52Z","silentRounds":10,"language":"Python","topics":["cryptography"],"aiDims":["安全工具","开发者工具","学习资源"],"zone":"工具","funScore":0,"domainTags":["密码学","协议原型","加密库"],"domainKey":"密码学","pushedAt":"2026-09-10T04:07:47Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"cryptography","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/JHUISI/charm","ts":"2026-09-10T04:07:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"EbookFoundation/free-programming-books","owner":"EbookFoundation","name":"free-programming-books","desc":":books: Freely available programming books","summaryCn":"全网最全编程电子书仓库，涵盖所有主流编程语言和技术栈。","reasonCn":"这是一个开源的电子书集合仓库，收录了数千本免费的编程书籍，涵盖Python、Java、前端等多个领域。虽然不直接属于AI工具，但它是开发者学习AI基础知识的重要资源库，适合想系统提升编程能力的技术人员，社区活跃且持续更新。","stars":397464,"starGrowth":39,"createdAt":"2013-10-11T06:50:37Z","silentRounds":0,"language":"Python","topics":["education","books","list","resource","hacktoberfest"],"aiDims":["学习资源","开发者工具","效率工具"],"zone":"资源","funScore":0,"domainTags":["编程","电子书","知识库","学习资源"],"domainKey":"编程","pushedAt":"2026-08-07T04:07:05.660Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"education","source":"github","weight":0.7},{"name":"books","source":"github","weight":0.7},{"name":"list","source":"github","weight":0.7},{"name":"resource","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/EbookFoundation/free-programming-books","ts":"2026-08-07T04:07:05.660Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.11729436628222932,"copyOk":false},{"repo":"Kanaries/graphic-walker","owner":"Kanaries","name":"graphic-walker","desc":"An open source alternative to Tableau. Embeddable visual analytic","summaryCn":"开源可嵌入的 Tableau 替代品，拖拽生成交互式图表。","reasonCn":"Graphic Walker 用 TypeScript 构建，可在浏览器或嵌入式环境快速加载数据集，提供丰富的图表库与拖拽式配置界面，支持自定义 SQL 查询、实时过滤和多维度联动，生成的可视化作品可导出为图片或嵌入网页，满足企业内部数据分析与仪表盘需求。","stars":3251,"starGrowth":0,"createdAt":"2022-09-27T10:02:28Z","silentRounds":1,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["数据可视化","BI 工具","拖拽式分析","嵌入式","TypeScript"],"domainKey":"数据可视化","pushedAt":"2026-09-23T21:47:45Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Kanaries/graphic-walker","ts":"2026-09-23T21:47:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ZuodaoTech/everyone-can-use-english","owner":"ZuodaoTech","name":"everyone-can-use-english","desc":"人人都能用英语","summaryCn":"教你如何把英语当成母语一样用，告别哑巴英语","reasonCn":"这是一个开源的英语学习指南，旨在帮助中国人打破语言障碍，实现流利交流。项目通过系统化的课程设计，涵盖了从发音、词汇到听说读写的全方位训练。它强调实用性和场景化，通过大量真实对话示例和练习，让学习者在实践中掌握语言，而不是死记硬背语法规则。项目维护者结合自身经验，提供了一套可复制的英语学习路径。","stars":38203,"starGrowth":32,"createdAt":"2019-03-15T16:33:53Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["学习资源","非AI-实用","非AI-好玩"],"zone":"资源","funScore":0,"domainTags":["英语学习","语言学习","自学指南","口语训练"],"domainKey":"英语学习","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/ZuodaoTech/everyone-can-use-english","ts":"2026-09-22T04:07:30.369Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09851289165772321,"copyOk":false},{"repo":"kd2org/opodsync","owner":"kd2org","name":"opodsync","desc":"A minimalist GPodder server, using the same API, for self-hosting","summaryCn":"一个极简的 GPodder 服务器，让你能自托管播客订阅服务。","reasonCn":"它的主要功能是同步 GPodder 的订阅源。你可以把 GPodder 里的订阅列表导入到 Opodsync 中，然后通过网页界面来管理这些订阅。最棒的是，它支持 OPML 导入导出，这意味着你可以轻松地在不同的播客客户端之间迁移订阅列表。","stars":322,"starGrowth":0,"createdAt":"2022-08-04T00:49:34Z","silentRounds":8,"language":"PHP","topics":["podcast"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["播客同步","GPodder API","自托管","OPML管理","PHP应用"],"domainKey":"播客同步","pushedAt":"2026-09-11T23:30:03Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/kd2org/opodsync","ts":"2026-09-11T23:30:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hashicorp/vagrant","owner":"hashicorp","name":"vagrant","desc":"Vagrant is a tool for building and distributing development environments.","summaryCn":"用代码一键创建可复现的虚拟开发环境，告别'在我电脑上能跑'的尴尬","reasonCn":"Vagrant通过Vagrantfile配置文件声明式地管理虚拟机生命周期，底层对接VirtualBox、VMware、AWS等多种虚拟化平台，支持一键启动预配置的完整开发环境。团队只需共享配置文件，就能复现相同的工作空间，彻底消除环境差异带来的兼容性问题。","stars":27209,"starGrowth":0,"createdAt":"2010-01-21T08:34:27Z","silentRounds":4,"language":"Ruby","topics":["automation"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["虚拟机","开发环境","虚拟化","DevOps"],"domainKey":"虚拟机","pushedAt":"2026-09-21T07:05:09Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/hashicorp/vagrant","ts":"2026-09-21T07:05:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TailAdmin/tailadmin-free-tailwind-dashboard-template","owner":"TailAdmin","name":"tailadmin-free-tailwind-dashboard-template","desc":"Free and Open-source Tailwind CSS Dashboard Admin Template that comes with all essential dashboard UI components, pages and elements","summaryCn":"免费好用的Tailwind后台模板，快速搭建管理界面","reasonCn":"这是一个基于Tailwind CSS的免费开源后台管理模板，内置了丰富的仪表板组件、图表、表格和页面布局。适合需要快速搭建管理系统或数据可视化界面的前端开发者，无需从零写代码，直接拿来就能用，支持自定义扩展，大幅节省开发时间。","stars":2318,"starGrowth":2,"createdAt":"2023-01-24T11:09:51Z","silentRounds":0,"language":"HTML","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["后台管理模板","Tailwind CSS","仪表盘UI","前端模板"],"domainKey":"后台管理模板","pushedAt":"2026-09-15T03:58:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/TailAdmin/tailadmin-free-tailwind-dashboard-template","ts":"2026-09-15T03:58:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"neuron-core/neuron-ai","owner":"neuron-core","name":"neuron-ai","desc":"The Agentic Framework of the PHP ecosystem to build production-ready AI driven applications. Connect components (LLMs, Tools, vector DBs, memory) to agents that interact with your data and UI.","summaryCn":"PHP程序员的AI Agent神器，一行代码让网站学会思考","reasonCn":"PHP生态首个生产级Agent框架，将LLM、工具调用、向量数据库、记忆系统无缝集成。开发者通过可视化连接组件即可构建能自主调用API、查询知识库、管理会话的AI Agent，支持Swoole/ReactPHP等异步运行时，适合搭建智能客服、自动化工作流等场景。","stars":2103,"starGrowth":0,"createdAt":"2025-03-02T15:35:21Z","silentRounds":2,"language":"PHP","topics":["vector-db"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","PHP","向量数据库","模型调用"],"domainKey":"AI Agent","pushedAt":"2026-09-23T20:03:06Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/neuron-core/neuron-ai","ts":"2026-09-23T20:03:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"marc2332/freya","owner":"marc2332","name":"freya","desc":"Cross-platform native GUI library for 🦀 Rust","summaryCn":"用Rust语言打造的跨平台原生图形界面开发库，追求极致性能与流畅体验","reasonCn":"技术层面上，它巧妙地结合了Rust语言的强大内存安全特性和Skia这个顶级的2D图形渲染引擎。整个界面采用声明式语法设计，写起来有点像现代的前端框架，逻辑非常清晰。它能够直接编译成各个操作系统的原生机器码，省去了庞大的运行环境，内存占用极小。","stars":3168,"starGrowth":2,"createdAt":"2022-07-27T15:28:11Z","silentRounds":0,"language":"Rust","topics":["desktop"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Rust","GUI","跨平台"],"domainKey":"Rust","pushedAt":"2026-09-23T17:20:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/marc2332/freya","ts":"2026-09-23T17:20:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FooIbar/EhViewer","owner":"FooIbar","name":"EhViewer","desc":"","summaryCn":"FooIbar/EhViewer 是一款专注于高性能与现代化视觉体验的 Android 图片阅读","reasonCn":"在技术实现上，该项目被明确标榜为“高性能”分支，意味着它在加载更多图片、滚动流畅度等方面做了专门优化。项目使用了 Kotlin 语言开发，并依赖 Coil、Ktor 等现代 Android 生态库来确保网络请求和图片加载的高效性。","stars":4723,"starGrowth":2,"createdAt":"2022-11-17T19:24:13Z","silentRounds":0,"language":"Kotlin","topics":[],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["Android应用","Material Design","动态取色","图片阅读器"],"domainKey":"Android应用","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/FooIbar/EhViewer","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nubjs/nub","owner":"nubjs","name":"nub","desc":"The fast all-in-one Node.js toolkit","summaryCn":"用Rust重写的超快Node.js开发工具包","reasonCn":"这是一个用Rust重写的高性能Node.js开发工具包，号称所有功能合在一起都比传统npm工具快得多。集成了文件监控、任务运行、依赖管理等功能，一个工具替代多个。Rust底层带来极致性能，启动秒级完成，特别适合追求开发效率的开发者。\n","stars":4334,"starGrowth":2,"createdAt":"2026-06-03T18:36:21Z","silentRounds":0,"language":"Rust","topics":["javascript-runtime","node-version-manager","nodejs","package-manager","script-runner"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Node.js","工具链","Rust"],"domainKey":"Node.js","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"javascript-runtime","source":"github","weight":0.7},{"name":"node-version-manager","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"package-manager","source":"github","weight":0.7},{"name":"script-runner","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/nubjs/nub","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"prest/prest","owner":"prest","name":"prest","desc":"PostgreSQL ➕ REST, low-code, simplify and accelerate development, ⚡ instant, realtime, high-performance on any Postgres application, existing or new, MCP server","summaryCn":"PostgreSQL 的零代码 REST API 生成器，极速轻量。","reasonCn":"PREST 是一个用 Go 编写的低代码后端服务，只需连接 PostgreSQL 数据库，即可自动为数据表生成高性能的 RESTful API。它无需编写任何业务代码，通过简单的 SQL 配置即可实现数据的增删改查，非常适合快速原型开发和数据服务化场景。","stars":4619,"starGrowth":0,"createdAt":"2016-11-22T05:17:05Z","silentRounds":3,"language":"Go","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["PostgreSQL","REST API","后端开发","低代码"],"domainKey":"PostgreSQL","pushedAt":"2026-09-18T09:31:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/prest/prest","ts":"2026-09-18T09:31:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"niedev/RTranslator","owner":"niedev","name":"RTranslator","desc":"Open source real-time translation app for Android that runs locally","summaryCn":"实时翻译应用，支持Android本地运行","reasonCn":"该应用使用Java开发，提供实时翻译功能，用户可以直接在本地设备上使用，无需依赖云端服务。","stars":10449,"starGrowth":3,"createdAt":"2020-03-30T18:40:55Z","silentRounds":0,"language":"Java","topics":["transformers"],"aiDims":["AI Agent","AI-实用"],"zone":"工具","funScore":0,"domainTags":["Android","本地翻译","机器翻译","离线应用"],"domainKey":"Android","pushedAt":"2026-09-22T20:03:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI-实用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/niedev/RTranslator","ts":"2026-09-22T20:03:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jgauchia/IceNav-v3","owner":"jgauchia","name":"IceNav-v3","desc":"🛰️ ESP32 Based GPS Navigator with OSM offline maps.  Multi GNSS","summaryCn":"ESP32做的离线GPS导航仪，装个包就能走","reasonCn":"它的技术架构非常硬核。ESP32作为主控，配合高精度GPS模块，能够实时接收卫星信号并计算经纬度。地图数据通过SD卡存储在本地，支持瓦片地图格式，用户可以提前下载好区域的地图包。软件层面，它运行在ESP32的RTOS系统上，通过Web界面进行配置，支持自定义地图图层、轨迹记录和航点","stars":462,"starGrowth":0,"createdAt":"2022-03-27T15:20:08Z","silentRounds":4,"language":"C","topics":["arduino"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.7,"domainTags":["GPS导航","嵌入式","ESP32","户外装备"],"domainKey":"GPS导航","pushedAt":"2026-09-23T21:08:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/jgauchia/IceNav-v3","ts":"2026-09-23T21:08:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zealdocs/zeal","owner":"zealdocs","name":"zeal","desc":"Offline documentation browser. Your personal reference library, searchable in an instant.","summaryCn":"把成千上万种开发文档装进口袋，没网也能秒查，打工人福音","reasonCn":"Zeal 是一款本地离线文档浏览器，内置上百种编程语言和框架的 API 参考文档，支持即时全文搜索和代码高亮。采用 C++ 开发，运行轻量流畅，支持多标签页、自定义快捷键和插件扩展，跨 Windows/macOS/Linux 三平台无缝使用。","stars":12805,"starGrowth":0,"createdAt":"2013-01-20T01:50:30Z","silentRounds":1,"language":"C++","topics":["developer-tools"],"aiDims":["开发者工具","效率工具","学习资源"],"zone":"工具","funScore":0,"domainTags":["文档浏览","离线搜索","开发者工具","API参考"],"domainKey":"文档浏览","pushedAt":"2026-09-09T18:59:27Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/zealdocs/zeal","ts":"2026-09-09T18:59:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"SonicJs-Org/sonicjs","owner":"SonicJs-Org","name":"sonicjs","desc":"SonicJS - The edge-native headless CMS for Cloudflare Workers. Sub-100ms response times, zero cold starts, TypeScript-first. Built on D1, R2, and Hono.","summaryCn":"跑在Cloudflare边缘的极速无头CMS，告别冷启动","reasonCn":"SonicJS是一个专为Cloudflare Workers构建的边缘原生无头CMS，利用D1数据库、R2对象存储和Hono框架，实现亚100毫秒响应时间和零冷启动。全TypeScript开发，面向现代前端架构，适合需要超低延迟的Edge Computing场景。","stars":1735,"starGrowth":0,"createdAt":"2018-08-15T22:49:14Z","silentRounds":3,"language":"TypeScript","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["无头CMS","Cloudflare Workers","边缘计算","D1数据库"],"domainKey":"无头CMS","pushedAt":"2026-09-13T05:18:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/SonicJs-Org/sonicjs","ts":"2026-09-13T05:18:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bridgecrewio/checkov","owner":"bridgecrewio","name":"checkov","desc":"Prevent cloud misconfigurations and find vulnerabilities during build-time in infrastructure as code, container images and open source packages with Checkov by Bridgecrew.","summaryCn":"一键扫描云配置文件自动揪出安全漏洞和配置错误","reasonCn":"这是一个专为基础设施即代码设计的安全扫描工具，支持Terraform、Kubernetes、Docker、CloudFormation等多种 IaC 格式。它能在构建阶段检测云资源的安全配置错误，如S3存储桶公开访问、安全组开放过大等常见问题。项目由Bridgecrew公司维护，支持自定义策略和CI/CD集成，是云安全领域的实用工具。虽然不直接涉及AI，但AI系统的云基础设施安全配置同样需要这类工具来保障。","stars":9022,"starGrowth":3,"createdAt":"2019-11-27T08:55:14Z","silentRounds":0,"language":"Python","topics":["devops","kubernetes"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["云安全","IaC","基础设施","扫描工具"],"domainKey":"云安全","pushedAt":"2026-09-17T23:03:06Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/bridgecrewio/checkov","ts":"2026-09-17T23:03:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"invoke-ai/InvokeAI","owner":"invoke-ai","name":"InvokeAI","desc":"Invoke is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, and serves as the foundation for multiple commercial products.","summaryCn":"很多艺术家在尝试AI绘画时，总觉得界面太简陋、功能太分散。","reasonCn":"技术上，它在底层的扩散模型上做了大量工程优化，支持多种模型权重加载，特别是其独特的“Unified Canvas”（统一画布）功能，允许用户在无限的画布上进行扩展、补全和局部涂抹，这是它区别于其他GUI工具的最强利器，极大地提升了创作的连贯性和可控性。","stars":28271,"starGrowth":6,"createdAt":"2022-08-17T01:04:27Z","silentRounds":0,"language":"Python","topics":["generative-art"],"aiDims":["创意工具","模型部署"],"zone":"AI","funScore":0.7,"domainTags":["AI绘画","Stable Diffusion","图像生成","创意工具"],"domainKey":"图像生成","pushedAt":"2026-09-21T19:49:51Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"generative-art","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/invoke-ai/InvokeAI","ts":"2026-09-21T19:49:51Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.01863669484630894,"copyOk":false},{"repo":"Anil-matcha/Open-Generative-AI","owner":"Anil-matcha","name":"Open-Generative-AI","desc":"Unrestricted Open-source alternative to AI video platforms — Free AI image &amp; video generation studio with 600+ models (Flux, Midjourney, Kling, Sora, Veo). No content filters. Self-hosted, MIT licensed.","summaryCn":"500多个AI模型随便用，免费自托管的图像视频生成工厂","reasonCn":"这个项目把Flux、Midjourney、Kling、Sora等500多个主流AI模型打包成一个可自托管的平台，完全免费且无内容过滤。你可以像搭积木一样在本地部署AI图像和视频生成服务，隐私安全，不受第三方平台限制，是当前开源AI创作领域最全面的解决方案之一。","stars":29050,"starGrowth":26,"createdAt":"2023-05-09T14:12:37Z","silentRounds":0,"language":"JavaScript","topics":["ai-image-generation","ai-video-generation","generative-ai","muapi","text-to-video","creative-tools","image-to-video","javascript","kling-ai","open-source","sora-alternative","wan-video","ai-art-generator","midjourney-alternative","uncensored","lipsync","text-to-image","seedance2","fal-ai-alternative","invideo-alternative"],"aiDims":["多模态Agent","AI应用","模型部署"],"zone":"AI","funScore":1,"domainTags":["AI生成","图像视频","自托管"],"domainKey":"AI生成","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"ai-image-generation","source":"github","weight":0.7},{"name":"ai-video-generation","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"muapi","source":"github","weight":0.7},{"name":"text-to-video","source":"github","weight":0.7},{"name":"creative-tools","source":"github","weight":0.7},{"name":"image-to-video","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"kling-ai","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"sora-alternative","source":"github","weight":0.7},{"name":"wan-video","source":"github","weight":0.7},{"name":"ai-art-generator","source":"github","weight":0.7},{"name":"midjourney-alternative","source":"github","weight":0.7},{"name":"uncensored","source":"github","weight":0.7},{"name":"lipsync","source":"github","weight":0.7},{"name":"text-to-image","source":"github","weight":0.7},{"name":"seedance2","source":"github","weight":0.7},{"name":"fal-ai-alternative","source":"github","weight":0.7},{"name":"invideo-alternative","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Anil-matcha/Open-Generative-AI","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.08068502581755593,"copyOk":false},{"repo":"zilliztech/claude-context","owner":"zilliztech","name":"claude-context","desc":"Code search MCP for Claude Code. Make entire codebase the context for any coding agent.","summaryCn":"当你面对一个几十万行代码的大型项目时，Claude或ChatGPT往往显得‘笨拙’：你","reasonCn":"核心逻辑基于MCP（Model Context Protocol）标准。它是一个搜索工具，后端连接Zilliz Cloud（VectorDB），负责将整个仓库的代码切片、向量化并存储。当你让Claude Code工作时，这个工具会实时拦截你的问题，在向量数据库中进行语义搜索，把最相关的代码片段作为上下文注入到对话中","stars":12564,"starGrowth":1,"createdAt":"2025-06-06T02:12:47Z","silentRounds":1,"language":"TypeScript","topics":["vector-db"],"aiDims":[],"zone":"AI","funScore":0.6,"domainTags":["MCP","代码库检索","向量数据库","AI编程助手"],"domainKey":"向量数据库","pushedAt":"2026-07-14T12:00:38Z","tags":[{"name":"vector-db","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/zilliztech/claude-context","ts":"2026-07-14T12:00:38Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"openmobilemaps/maps-core","owner":"openmobilemaps","name":"maps-core","desc":"The lightweight and modern Map SDK for Android and iOS","summaryCn":"给App塞进谷歌地图能力却不给苹果税交钱的轻量SDK","reasonCn":"用C++打造的多平台地图引擎，让开发者在安卓和iOS上获得一致且轻量的地图体验。它剥离了传统庞大SDK的冗余，聚焦核心渲染与瓦片加载，支持自定义图层和离线数据，是独立应用摆脱重型地图依赖、实现高性能地理信息展示的理想底座。","stars":220,"starGrowth":0,"createdAt":"2021-02-18T07:43:04Z","silentRounds":8,"language":"C++","topics":["maps"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["地图SDK","移动开发","渲染引擎"],"domainKey":"地图SDK","pushedAt":"2026-09-11T12:03:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"maps","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/openmobilemaps/maps-core","ts":"2026-09-11T12:03:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opendataloader-project/opendataloader-pdf","owner":"opendataloader-project","name":"opendataloader-pdf","desc":"PDF Parser for AI-ready data. Automate PDF accessibility. Open-source.","summaryCn":"让AI能读懂PDF的神器，表格文字一键提取变数据","reasonCn":"这是一个专为AI和数据科学场景打造的PDF解析工具，用Java开发，能够将复杂的PDF文档智能提取为结构化数据。支持文本、表格、列表等多种内容形式，特别适合构建RAG系统时做数据预处理。相比传统解析库，它对表格和复杂排版的处理更加精准，且提供简单易用的API接口，几行代码即可完成文档解析。","stars":29346,"starGrowth":10,"createdAt":"2025-05-13T05:48:02Z","silentRounds":0,"language":"Java","topics":["rag"],"aiDims":["RAG与知识","文档理解","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["PDF解析","数据提取","RAG数据准备"],"domainKey":"PDF解析","pushedAt":"2026-09-23T05:13:27Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/opendataloader-project/opendataloader-pdf","ts":"2026-09-23T05:13:27Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03102228583111838,"copyOk":true},{"repo":"grandgaming9321-prog/reality-engine","owner":"grandgaming9321-prog","name":"reality-engine","desc":"Top Dynamic AI World Simulation & Storytelling Tools 2026","summaryCn":"想象一下你正在玩一个游戏，里面的NPC不是只会复读的复读机，而是能根据你的","reasonCn":"它的核心技术在于将动态叙事引擎与AI实时决策系统整合，允许创作者通过简单的指令定义世界的规则，剩下的复杂互动交给AI逻辑去填充。与传统的脚本式游戏开发完全不同，这种架构让故事的发展具有极高的不可预测性和新鲜感，每一轮游玩都是全新的体验。","stars":128,"starGrowth":0,"createdAt":"2026-06-28T21:06:56Z","silentRounds":18,"language":"HTML","topics":["procedural-generation"],"aiDims":["非AI-好玩","游戏","创意工具"],"zone":"创意","funScore":0,"domainTags":["程序化生成","AI","叙事引擎","游戏开发"],"domainKey":"程序化生成","pushedAt":"2026-09-23T20:06:30Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/grandgaming9321-prog/reality-engine","ts":"2026-09-23T20:06:30Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Hack-with-Github/Awesome-Hacking","owner":"Hack-with-Github","name":"Awesome-Hacking","desc":"A collection of various awesome lists for hackers, pentesters and security researchers","summaryCn":"渗透测试和安全研究者的资源宝库，涵盖工具、教程、CTF","reasonCn":"内容覆盖面很广：信息收集工具、漏洞扫描器、无线渗透、社工工具、CTF竞赛资源、安全写本、逆向工程等等。每个分类下都有精选的资源链接，包括开源工具、博客文章、在线课程、视频演示等。 使用方式很直接：打开仓库浏览分类，找到感兴趣的方向点进链接。","stars":120996,"starGrowth":36,"createdAt":"2016-03-30T15:47:10Z","silentRounds":0,"language":"","topics":["security"],"aiDims":["安全工具","学习资源"],"zone":"资源","funScore":0,"domainTags":["安全研究","渗透测试","资源清单"],"domainKey":"安全研究","pushedAt":"2026-07-26T03:29:02Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7}],"aiScore":0.4,"source":"search","url":"https://github.com/Hack-with-Github/Awesome-Hacking","ts":"2026-07-26T03:29:02Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.10889259149062779,"copyOk":false},{"repo":"iamkun/dayjs","owner":"iamkun","name":"dayjs","desc":"⏰ Day.js 2kB immutable date-time library alternative to Moment.js with the same modern API","summaryCn":"轻量级时间处理库，Moment.js的完美替代品","reasonCn":"Day.js 是一款极其轻量且现代化的日期处理库，体积仅有 2KB，却提供了与 Moment.js 完全一致的现代 API。它支持链式调用，能轻松处理日期的加减、格式化、验证等常见操作。核心优势在于零依赖，使用 Moment.js 的 API 即可无缝迁移，同时去掉了 Moment.js 沉重的体积负担，非常适合追求极致性能和加载速度的前端项目。","stars":48667,"starGrowth":0,"createdAt":"2018-04-10T09:26:44Z","silentRounds":3,"language":"JavaScript","topics":["time","date","date-formatting","datetime","moment","dayjs"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["日期处理","JavaScript","Moment.js替代","前端库"],"domainKey":"日期处理","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"time","source":"github","weight":0.7},{"name":"date","source":"github","weight":0.7},{"name":"date-formatting","source":"github","weight":0.7},{"name":"datetime","source":"github","weight":0.7},{"name":"moment","source":"github","weight":0.7},{"name":"dayjs","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/iamkun/dayjs","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"grafana/tempo","owner":"grafana","name":"tempo","desc":"Grafana Tempo is a high volume, minimal dependency distributed tracing backend.","summaryCn":"给分布式系统装上黑匣子，让你一眼看穿代码跑在哪一步卡住了","reasonCn":"Tempo 是 Grafana 旗下的分布式追踪后端系统，专注于高吞吐量和低延迟的链路追踪数据存储。它采用标签索引存储架构，能够快速查询海量追踪数据，并结合 Grafana UI 提供直观的拓扑图和性能分析，帮助开发人员快速定位系统性能瓶颈和故障点。","stars":5485,"starGrowth":0,"createdAt":"2020-01-24T18:05:02Z","silentRounds":2,"language":"Go","topics":["distributed-tracing","grafana"],"aiDims":["开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["分布式追踪","可观测性","日志","性能分析"],"domainKey":"分布式追踪","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"distributed-tracing","source":"github","weight":0.7},{"name":"grafana","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/grafana/tempo","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"shieldfy/API-Security-Checklist","owner":"shieldfy","name":"API-Security-Checklist","desc":"Checklist of the most important security countermeasures when designing, testing, and releasing your API","summaryCn":"做API开发时照着这份清单就能避免大部分安全问题","reasonCn":"这份清单由安全专家维护，全面覆盖了API设计、开发、测试和发布各环节的安全要点。从身份认证、权限控制到输入验证、错误处理，从速率限制到日志审计，每一条都是实战经验的结晶。star超过2.3万，是API安全领域最实用的开源参考文档之一，适合开发团队在代码审查和上线前对照检查。","stars":23323,"starGrowth":0,"createdAt":"2017-07-08T20:01:38Z","silentRounds":2,"language":"","topics":["security"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"资源","funScore":0,"domainTags":["API安全","安全检查清单","开发规范","代码审查"],"domainKey":"API安全","pushedAt":"2026-07-21T07:59:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7}],"aiScore":0.5,"source":"search","url":"https://github.com/shieldfy/API-Security-Checklist","ts":"2026-07-21T07:59:55Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"drl990114/MarkFlowy","owner":"drl990114","name":"MarkFlowy","desc":"AI-powered cross-platform editor for WYSIWYG Markdown and general-purpose text editing.","summaryCn":"对于经常写文章、做笔记或者搞创作的人来说，排版和灵感枯竭一直是两大痛点","reasonCn":"在技术架构上，它采用了现代化的跨平台前端技术栈构建，底层解析引擎对Markdown语法支持得极其丰富且实时响应。它深度整合了各大主流AI接口，通过流式传输技术让AI生成的文字像打字机一样实时蹦出来，同时在本地存储和缓存方面做了优化，确保海量笔记管理时依然丝般顺滑。","stars":2393,"starGrowth":1,"createdAt":"2022-05-25T13:03:26Z","silentRounds":5,"language":"TypeScript","topics":["editor"],"aiDims":["AI应用","AI写作","效率工具"],"zone":"工具","funScore":0,"domainTags":["Markdown编辑器","AI写作","跨平台"],"domainKey":"Markdown编辑器","pushedAt":"2026-09-22T15:56:03Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/drl990114/MarkFlowy","ts":"2026-09-22T15:56:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pdulvp/jellyfin-qnap","owner":"pdulvp","name":"jellyfin-qnap","desc":"Jellyfin server packaging for QNAP NAS","summaryCn":"QNAP NAS用户的福音，把你的黑胶变成私人媒体库。","reasonCn":"这是Jellyfin服务器的QNAP专用版本，让拥有QNAP存储设备的用户能轻松搭建自己的私有媒体服务器。Jellyfin是一个开源的流媒体解决方案，支持视频、音频和照片，支持Docker容器部署，实现了真正的去中心化和数据掌控。","stars":782,"starGrowth":0,"createdAt":"2021-12-24T14:23:39Z","silentRounds":10,"language":"Shell","topics":["jellyfin"],"aiDims":["实用工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Jellyfin","NAS","媒体服务器"],"domainKey":"Jellyfin","pushedAt":"2026-09-23T06:05:35Z","tags":[{"name":"实用工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/pdulvp/jellyfin-qnap","ts":"2026-09-23T06:05:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"testcontainers/testcontainers-python","owner":"testcontainers","name":"testcontainers-python","desc":"Testcontainers is a Python library that providing a friendly API to run Docker container. It is designed to create runtime environment to use during your automatic tests.","summaryCn":"测试代码时一键启动Docker环境，告别环境配置的噩梦","reasonCn":"核心技术上，它封装了Docker的API，支持主流数据库（PostgreSQL、MySQL、MongoDB）、消息队列（Redis、Kafka）、浏览器自动化（Selenium Grid）等常见服务。你只需要几行Python代码就能启动一个完整的测试环境，不用手动配Docker Compose文件，也不用在CI/CD里搞一堆复杂的shell脚","stars":2301,"starGrowth":1,"createdAt":"2017-03-22T08:22:35Z","silentRounds":3,"language":"Python","topics":["testing"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Docker","测试自动化","CI/CD","数据库测试"],"domainKey":"Docker","pushedAt":"2026-09-14T23:33:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/testcontainers/testcontainers-python","ts":"2026-09-14T23:33:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"marko-js/marko","owner":"marko-js","name":"marko","desc":"A declarative, HTML-based language that makes building web apps fun","summaryCn":"让写前端界面像搭积木一样爽快，声明式模板让代码既优雅又好维护的利器","reasonCn":"Marko 是 LinkedIn 开源的声明式 HTML 模板引擎，支持 HTML 优先的组件开发模式。它实现了细粒度 diffing 算法，只需更新变化的 DOM 部分而非整块重绘，极大提升了渲染性能。提供 TypeScript 支持和热重载，适合构建高性能、可维护的现代 Web 应用界面。","stars":14417,"starGrowth":0,"createdAt":"2014-01-07T23:58:21Z","silentRounds":3,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["前端模板","HTML模板","性能优化"],"domainKey":"前端模板","pushedAt":"2026-09-23T21:50:03Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/marko-js/marko","ts":"2026-09-23T21:50:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"markmead/hyperui","owner":"markmead","name":"hyperui","desc":"Free Tailwind CSS v4 components for your next project, designed to enhance your web development with the latest features and styles 🚀","summaryCn":"最新版Tailwind组件库，让前端页面开发效率翻倍","reasonCn":"HyperUI提供了大量基于Tailwind CSS v4的免费UI组件，专为现代网页设计打造。它解决了开发者在面对空白画布不知从何下手的痛点，提供了可直接复制使用的精美代码片段。无论是导航栏、卡片还是表单，都能快速集成，大幅提升前端开发效率，特别适合追求极简风格的开发者。","stars":12238,"starGrowth":0,"createdAt":"2021-11-12T17:14:11Z","silentRounds":3,"language":"HTML","topics":["frontend"],"aiDims":["效率工具","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Tailwind CSS","UI 组件","前端开发","组件库"],"domainKey":"Tailwind CSS","pushedAt":"2026-09-18T05:22:49Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/markmead/hyperui","ts":"2026-09-18T05:22:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alibaba/open-code-review","owner":"alibaba","name":"open-code-review","desc":"Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI &amp; Anthropic compatible.","summaryCn":"在大型组织里，代码审查往往是耗时且质量参差不齐的环节。","reasonCn":"在大型组织里，代码审查往往是耗时且质量参差不齐的环节。传统的人工 Review 受限于审查者经验，容易遗漏安全或性能隐患；而市面上的静态分析工具又缺乏上下文理解。Open‑Code‑Review 通过混合确定性流水线与 LLM 智能点评，既保证规则的精准可控，又利用大模型的语言理解能力提供","stars":39711,"starGrowth":393,"createdAt":"2026-05-18T02:16:36Z","silentRounds":0,"language":"Go","topics":["agent","code-review","code-review-assistant","harness","repository-level-context","agent-skills"],"aiDims":["代码助手","AI应用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["代码审查","CI/CD","静态分析"],"domainKey":"代码审查","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"code-review","source":"github","weight":0.7},{"name":"code-review-assistant","source":"github","weight":0.7},{"name":"harness","source":"github","weight":0.7},{"name":"repository-level-context","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.81,"source":"trending","url":"https://github.com/alibaba/open-code-review","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3956274080229659,"copyOk":false},{"repo":"NirDiamant/GenAI_Agents","owner":"NirDiamant","name":"GenAI_Agents","desc":"50+ tutorials and implementations for Generative AI Agent techniques, from basic conversational bots to complex multi-agent systems.","summaryCn":"50多个AI智能体实战教程，从入门到多Agent协作全覆盖","reasonCn":"这个项目收集了50多个生成式AI Agent的实战教程和完整代码实现，从最简单的对话机器人到复杂的多Agent协作系统都有覆盖。核心技术包括RAG检索增强、工具调用、任务规划、LangChain和AutoGen框架使用等，每个教程都配有可运行的Jupyter Notebook示例。","stars":24367,"starGrowth":6,"createdAt":"2024-09-09T20:10:19Z","silentRounds":0,"language":"Jupyter Notebook","topics":["rag","ml","llm","agents"],"aiDims":["AI Agent","教程","RAG与知识"],"zone":"资源","funScore":0.5,"domainTags":["AI Agent","教程","Jupyter Notebook"],"domainKey":"AI Agent","pushedAt":"2026-09-21T10:54:15Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/NirDiamant/GenAI_Agents","ts":"2026-09-21T10:54:15Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.018738703984897607,"copyOk":false},{"repo":"OpenZeppelin/openzeppelin-contracts","owner":"OpenZeppelin","name":"openzeppelin-contracts","desc":"OpenZeppelin Contracts is a library for secure smart contract development.","summaryCn":"Solidity 合约开发的护身符，开源界最靠谱的安全基建","reasonCn":"OpenZeppelin 是 Solidity 生态中最权威的智能合约安全库，提供经过社区审计的 ERC-20、ERC-721 等标准实现。内置可重入锁、访问控制、安全数学运算等核心防护机制，帮助开发者避免常见漏洞，是 Web3 安全基础设施的事实标准。","stars":27251,"starGrowth":3,"createdAt":"2016-08-01T20:54:54Z","silentRounds":0,"language":"Solidity","topics":["security"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["Solidity","智能合约","Web3安全"],"domainKey":"Solidity","pushedAt":"2026-09-23T16:30:48Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Solidity","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/OpenZeppelin/openzeppelin-contracts","ts":"2026-09-23T16:30:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"derbyjs/derby","owner":"derbyjs","name":"derby","desc":"","summaryCn":"Derby 解决的痛点是实时多用户应用开发中的“状态一致性”难题。","reasonCn":"与像 React 或 Vue 这样主要关注视图更新的框架不同，Derby 是一种面向数据驱动的架构。它使用双向绑定机制，模板中的变量会自动映射到数据模型，反之亦然。更重要的是，Derby 采用“同构”渲染策略，同一套模板代码既在服务器端执行生成初始 HTML 页面，又在客户端继续运行。","stars":4700,"starGrowth":0,"createdAt":"2011-04-09T04:29:16Z","silentRounds":6,"language":"TypeScript","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["MVC框架","实时同步","Node.js","数据持久化","离线应用"],"domainKey":"MVC框架","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/derbyjs/derby","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"open-cli-tools/concurrently","owner":"open-cli-tools","name":"concurrently","desc":"Run commands concurrently. Like `npm run watch-js & npm run watch-less` but better.","summaryCn":"同时跑多个命令，再也不用手动开多个终端窗口了","reasonCn":"concurrently 是一个命令行工具，允许你一次性并发执行多个命令。它让你不用再手动开多个终端窗口，也不用写复杂的 shell 脚本。你可以用一个命令同时启动前端开发服务器、后端 API 和实时编译工具，还能统一查看所有任务的输出和退出码，配置灵活，支持通过配置文件定义命令组。","stars":7853,"starGrowth":0,"createdAt":"2015-02-05T22:27:51Z","silentRounds":3,"language":"TypeScript","topics":["command-line"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行","并发","Node.js"],"domainKey":"命令行","pushedAt":"2026-09-15T02:35:07Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/open-cli-tools/concurrently","ts":"2026-09-15T02:35:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cocoindex-io/cocoindex","owner":"cocoindex-io","name":"cocoindex","desc":"Incremental engine for long horizon agents 🌟 Star if you like it!","summaryCn":"给AI装上“顺风耳”，让长时序智能体不再过目即忘，秒懂海量历史数据。","reasonCn":"这是一个专为长时序智能体设计的增量索引引擎，解决了AI在处理海量历史数据时容易遗忘或索引过慢的痛点。它采用Rust编写，能高效处理日志、文档、数据库等流式数据，并将其转化为向量索引，供后续Agent快速检索。支持断点续传和增量更新，是构建长期记忆系统的关键基础设施。","stars":11591,"starGrowth":5,"createdAt":"2025-03-03T23:03:09Z","silentRounds":0,"language":"Rust","topics":["rag"],"aiDims":["AI基础设施","AI Agent","向量数据库"],"zone":"AI","funScore":0.6,"domainTags":["向量索引","RAG","增量计算"],"domainKey":"RAG","pushedAt":"2026-09-22T17:04:40Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/cocoindex-io/cocoindex","ts":"2026-09-22T17:04:40Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0162941074971961,"copyOk":false},{"repo":"Panniantong/Agent-Reach","owner":"Panniantong","name":"Agent-Reach","desc":"Give your AI agent eyes to see the entire internet. Read &amp; search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.","summaryCn":"给AI装上眼睛，一个命令行能刷遍全网平台不用花一分钱","reasonCn":"超热门AI Agent工具，让AI代理获得互联网浏览能力。支持Twitter、Reddit、YouTube、GitHub、Bilibili、小红书等六大平台的搜索和内容读取，纯命令行操作零API费用。底层基于Playwright等自动化技术实现，是构建联网AI Agent的实用基础设施。","stars":84792,"starGrowth":342,"createdAt":"2026-02-24T02:10:24Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","网页自动化","数据抓取","社交媒体监控","浏览器自动化"],"domainKey":"AI Agent","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/Panniantong/Agent-Reach","ts":"2026-09-17T04:13:51.053Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.37452023775827903,"copyOk":false},{"repo":"glanceapp/glance","owner":"glanceapp","name":"glance","desc":"A self-hosted dashboard that puts all your feeds in one place","summaryCn":"一个把所有新闻源和视频都汇总在一块的炫酷仪表盘","reasonCn":"Glance是一款轻量级自托管信息仪表盘，支持RSS订阅、新闻聚合、YouTube视频、天气、日历等多种数据源的整合展示。采用Go语言开发，Docker一键部署，前端基于Vue.js实现。它解决信息分散、需频繁切换多个平台查看更新的问题，可自定义布局主题和刷新间隔。与商业RSS服务相比数据完全自控，隐私安全有保障。","stars":37163,"starGrowth":8,"createdAt":"2024-04-27T18:55:38Z","silentRounds":0,"language":"Go","topics":["self-hosted","docker"],"aiDims":["非AI-实用","可视化","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["信息聚合","自托管","RSS阅读"],"domainKey":"信息聚合","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/glanceapp/glance","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02464580362188198,"copyOk":true},{"repo":"jjxxs/websocket-ts","owner":"jjxxs","name":"websocket-ts","desc":"A WebSocket for browsers with optional auto-reconnect and message buffering written in TypeScript.","summaryCn":"自动重连带缓冲的WebSocket，前端必备","reasonCn":"技术核心：基于TypeScript实现，支持浏览器和Node.js环境。特性包括自动重连（可配置延迟）、消息缓冲（断网时缓存发送）、错误恢复和类型安全接口。相比原生WebSocket，它提供更高级的抽象，简化开发流程。性能上，利用现代JS引擎优化，开销小。","stars":158,"starGrowth":0,"createdAt":"2020-09-11T12:54:04Z","silentRounds":24,"language":"TypeScript","topics":["browser","buffer","client-websocket","reconnecting","typescript","websocket"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WebSocket","TypeScript","前端网络库"],"domainKey":"WebSocket","pushedAt":"2026-07-11T16:30:34Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"browser","source":"github","weight":0.7},{"name":"buffer","source":"github","weight":0.7},{"name":"client-websocket","source":"github","weight":0.7},{"name":"reconnecting","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"websocket","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"bigbro","url":"https://github.com/jjxxs/websocket-ts","ts":"2026-07-11T16:30:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Jakubantalik/Libraries","owner":"Jakubantalik","name":"Libraries","desc":"Collection of effects for React: Border beam, Liquid Gooey","summaryCn":"给 React 组件加个流光边框，瞬间提升 UI 高级感","reasonCn":"这是一个专为 React 打造的流光边框动画库，利用 CSS border-image 和 SVG 蒙版实现无限循环的渐变扫描效果。它不仅解决了传统边框单调的问题，还能通过配置实现多种形态的呼吸与流动，非常适合追求极致视觉体验的前端项目。","stars":3629,"starGrowth":14,"createdAt":"2026-04-12T17:27:34Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["创意工具","可视化","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["React","UI特效","CSS动画"],"domainKey":"React","pushedAt":"2026-08-17T15:33:31Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/Jakubantalik/Libraries","ts":"2026-08-17T15:33:31Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.06428812344998622,"copyOk":false},{"repo":"NirDiamant/agents-towards-production","owner":"NirDiamant","name":"agents-towards-production","desc":"End-to-end, code-first tutorials for building production-grade GenAI agents. From prototype to enterprise deployment.","summaryCn":"手把手教你把AI Agent从原型变成生产级应用","reasonCn":"这个仓库提供端到端的代码教程，从基础的RAG问答系统到复杂的Agent工作流，覆盖生产环境所需的全部关键环节。每个教程都有完整的代码实现，涵盖LangChain、LlamaIndex等主流框架，还讲解了部署策略、性能优化和监控方案。star超过2.1万，是学习AI Agent生产化最热门的中文友好教程库之一。","stars":21487,"starGrowth":5,"createdAt":"2025-06-16T17:33:44Z","silentRounds":0,"language":"Jupyter Notebook","topics":["rag","agents"],"aiDims":["AI Agent","RAG与知识","教程"],"zone":"资源","funScore":0,"domainTags":["AI Agent","RAG","教程"],"domainKey":"AI Agent","pushedAt":"2026-09-21T10:54:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/NirDiamant/agents-towards-production","ts":"2026-09-21T10:54:17Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.015698096523479314,"copyOk":true},{"repo":"Unstructured-IO/unstructured","owner":"Unstructured-IO","name":"unstructured","desc":"Convert documents to structured data effortlessly. Unstructured is open-source ETL solution for transforming complex documents into clean, structured formats for language models.  Visit our website to learn more about our enterprise grade Platform product for production grade workflows, partitioning, enrichments, chunking and embedding.","summaryCn":"它使用自然语言处理技术来识别和提取文档中的关键词和信息，然后通过算法将","reasonCn":"它使用自然语言处理技术来识别和提取文档中的关键词和信息，然后通过算法将这些信息转化为结构化的数据格式，如JSON或CSV。用户无需手动编写代码，就能快速获得所需的结构化数据。非常适合需要将非结构化文本转化为结构化数据的用户，如研究人员、开发人员等。","stars":15469,"starGrowth":5,"createdAt":"2022-09-26T21:53:41Z","silentRounds":0,"language":"HTML","topics":["ml"],"aiDims":["AI基础设施"],"zone":"AI","funScore":0,"domainTags":["文档处理","ETL","数据提取","大模型","NLP"],"domainKey":"文档处理","pushedAt":"2026-09-22T04:15:13Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Unstructured-IO/unstructured","ts":"2026-09-22T04:15:13Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.01596968129808003,"copyOk":false},{"repo":"Canop/broot","owner":"Canop","name":"broot","desc":"A new way to see and navigate directory trees : https://dystroy.org/broot","summaryCn":"告别cd和ls！让目录导航变成一种享受的炫酷工具","reasonCn":"一款基于Rust打造的新一代目录浏览和导航工具，革命性地解决了传统命令行cd、ls效率低下的问题。支持交互式目录树遍历、实时文件预览、vi风格快捷键导航、递归搜索和多种自定义主题，性能卓越。安装便捷，Mac和Linux用户一条命令即可上手，彻底改变开发者日常文件管理体验。","stars":12956,"starGrowth":1,"createdAt":"2018-11-15T20:09:19Z","silentRounds":2,"language":"Rust","topics":["command-line"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["目录树","命令行工具","Rust","文件管理","终端"],"domainKey":"目录树","pushedAt":"2026-09-23T18:48:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/Canop/broot","ts":"2026-09-23T18:48:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rive-app/rive-flutter","owner":"rive-app","name":"rive-flutter","desc":"Flutter runtime for Rive","summaryCn":"Flutter动画运行时，让设计师的交互式动画在App里动起来","reasonCn":"rive-flutter是Rive动画的Flutter运行时，让设计师创作的交互式动画能在Flutter应用中运行。Rive支持状态机驱动动画，可以根据用户交互实时改变动画表现。相比Lottie，它支持更复杂的交互逻辑和实时参数控制。适合需要高质量动画效果的Flutter应用开发，让设计师和开发者协作更顺畅。","stars":1515,"starGrowth":0,"createdAt":"2019-12-03T02:48:44Z","silentRounds":3,"language":"Dart","topics":["design","game-development"],"aiDims":["创意工具","可视化"],"zone":"工具","funScore":1,"domainTags":["Flutter","交互动画","Rive"],"domainKey":"Flutter","pushedAt":"2026-09-23T20:44:20Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/rive-app/rive-flutter","ts":"2026-09-23T20:44:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"0x5e/wechat-deleted-friends","owner":"0x5e","name":"wechat-deleted-friends","desc":"","summaryCn":"曾通过新建群组接口探测微信好友删除状态的经典Python小工具。","reasonCn":"这是一个早期利用微信网页版接口来检测被删好友的Python脚本工具。它的核心原理是通过尝试将好友拉入新群组，如果遇到阻碍则说明对方已将自己删除。由于微信接口策略的调整及安全限制，该项目目前已经无法正常使用，作者建议后来者仅作代码参考，但它在当年曾引发了广泛关注与讨论。","stars":4773,"starGrowth":0,"createdAt":"2016-01-02T01:28:59Z","silentRounds":4,"language":"Python","topics":["wechat","python"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["微信","好友清理","Python脚本"],"domainKey":"微信","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"wechat","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/0x5e/wechat-deleted-friends","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"huggingface/transformers.js","owner":"huggingface","name":"transformers.js","desc":"State-of-the-art Machine Learning for the web. Run 🤗 Transformers directly in your browser, with no need for a server!","summaryCn":"把最强大模型装进浏览器，本地运行 AI 不再依赖服务器","reasonCn":"Transformers.js 是 Hugging Face 推出的开源项目，它能在浏览器中直接运行机器学习模型，无需后端服务器支持。基于 WebGPU 技术，它让 AI 应用可以在客户端实现实时推理，保护数据隐私的同时提供了极致的响应速度。","stars":16309,"starGrowth":0,"createdAt":"2023-02-13T13:51:45Z","silentRounds":2,"language":"JavaScript","topics":["transformers"],"aiDims":["模型部署","推理引擎","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["WebGPU","机器学习","浏览器推理","前端AI","模型转换"],"domainKey":"WebGPU","pushedAt":"2026-09-23T12:31:57Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/huggingface/transformers.js","ts":"2026-09-23T12:31:57Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"harvard-edge/cs249r_book","owner":"harvard-edge","name":"cs249r_book","desc":"Machine Learning Systems: Foundations, Scaling, Agentic AI, and Physical AI (Vols I–IV) • Harvard CS249r | https://mlsysbook.ai","summaryCn":"哈佛明星课程教材，带你从零搭建工业级机器学习系统","reasonCn":"这是哈佛CS249r课程的完整教材，系统讲解机器学习系统的核心架构与工程实践。从数据流水线、模型训练到在线服务部署，覆盖分布式训练、特征存储、特征漂移检测等关键主题。代码示例采用Python实现，适合想深入理解ML系统底层原理的开发者。","stars":28416,"starGrowth":6,"createdAt":"2023-09-06T19:31:06Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["学习资源","教程","模型与训练"],"zone":"资源","funScore":0,"domainTags":["机器学习系统","课程教材","ML工程","分布式训练","模型部署"],"domainKey":"机器学习系统","pushedAt":"2026-09-23T18:37:48Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/harvard-edge/cs249r_book","ts":"2026-09-23T18:37:48Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.018633445945945945,"copyOk":false},{"repo":"NaoTu/DesktopNaotu","owner":"NaoTu","name":"DesktopNaotu","desc":"","summaryCn":"对于习惯在脱机环境下深度思考的用户，桌面版脑图提供了一个无需浏览器即可","reasonCn":"项目的核心亮点在于其稳定性与轻量化，它兼容 Windows、Mac OS X 和 Linux 等多种操作系统，确保了不同平台下的一致性体验。自动保存功能的加入，进一步降低了数据丢失的风险。虽然在 AI 时代，新的 MindMap 工具层出不穷，但这种强调本地原生体验的软件，依然是很多追求效率的个人","stars":5087,"starGrowth":0,"createdAt":"2016-11-08T01:32:30Z","silentRounds":3,"language":"TypeScript","topics":["desktop-app","mindmap","naotu","mind","inaotu"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["思维导图","桌面应用","本地协作"],"domainKey":"思维导图","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop-app","source":"github","weight":0.7},{"name":"mindmap","source":"github","weight":0.7},{"name":"naotu","source":"github","weight":0.7},{"name":"mind","source":"github","weight":0.7},{"name":"inaotu","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/NaoTu/DesktopNaotu","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nektos/act","owner":"nektos","name":"act","desc":"Run your GitHub Actions locally 🚀","summaryCn":"本地模拟运行GitHub Actions，CI配置开发测试两不误","reasonCn":"开源工具，让开发者能在本地复现GitHub Actions的完整执行流程，无需提交代码到服务器即可验证配置是否正确。支持多种Linux镜像，解决CI调试痛点，特别适合开发自定义Action或排查流水线问题的场景，大幅提升开发效率。","stars":72082,"starGrowth":6,"createdAt":"2019-01-02T19:53:43Z","silentRounds":0,"language":"Go","topics":["github-actions","golang","ci","devops"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["GitHub Actions本地运行","CI/CD调试工具","Docker容器模拟"],"domainKey":"GitHub Actions本地运行","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"github-actions","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"ci","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/nektos/act","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01824971560167587,"copyOk":false},{"repo":"metalbear-co/mirrord","owner":"metalbear-co","name":"mirrord","desc":"Run any process, on your machine or in an AI agent's environment, as if it were a pod in your Kubernetes cluster: real env vars, DNS, network, traffic.","summaryCn":"本地代码直接连K8s集群，调试开发从此不再煎熬","reasonCn":"mirrord的技术实现非常巧妙。它通过注入代理的方式，拦截你本地进程的网络请求，将这些请求透明地转发到Kubernetes集群中对应的服务。这意味着你的本地代码认为自己是在本地运行，但实际上访问的是集群里的数据库、缓存、消息队列等服务。","stars":5336,"starGrowth":0,"createdAt":"2022-02-01T16:42:30Z","silentRounds":2,"language":"Rust","topics":["backend","devops","developer-tools"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","开发调试","容器化"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T19:49:50Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/metalbear-co/mirrord","ts":"2026-09-23T19:49:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JimLiu/baoyu-design","owner":"JimLiu","name":"baoyu-design","desc":"Run Claude Design locally as an Agent Skill — Cursor, Claude Code & more. Produce polished UI mockups, prototypes, decks & wireframes as self-contained HTML, without claude.ai/design. Best with Opus 4.8.","summaryCn":"让AI设计师在你的代码编辑器里直接跑，秒出专业UI稿","reasonCn":"Baoyu Design是一个基于Claude Design能力的本地Agent工具，可在Cursor、Claude Code等开发环境中直接运行，无需访问claude.ai网站。它能生成专业的UI原型、设计稿、演示文稿和线框图，输出为自包含的HTML文件，支持高保真视觉呈现。项目特别适合使用Claude Opus 4模型的开发者，能将AI设计能力无缝集成到日常开发流程中。","stars":4144,"starGrowth":5,"createdAt":"2026-06-07T01:16:18Z","silentRounds":0,"language":"JavaScript","topics":["design"],"aiDims":["AI应用","工具调用","创意工具"],"zone":"AI","funScore":0.6,"domainTags":["Agent Skill","UI原型","Claude API"],"domainKey":"Agent Skill","pushedAt":"2026-09-23T05:34:49Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/JimLiu/baoyu-design","ts":"2026-09-23T05:34:49Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02234362934362934,"copyOk":false},{"repo":"xang1234/stock-screener","owner":"xang1234","name":"stock-screener","desc":"Stock scanner with multiple fundamental and technical criteria. Features 80+ filters, AI chatbot (Groq/DeepSeek/Gemini),   theme discovery, and StockBee-style breadth indicators.","summaryCn":"集成了AI聊天与80多项多维筛选指标的股票扫描神器。","reasonCn":"技术架构上，该项目采用Python构建，内置了超过80种精细的筛选过滤机制。它不仅支持传统的量化指标扫描，还创新性地接入了Groq、DeepSeek以及Gemini等主流大模型，让AI作为你的私人投研分析师。此外，它还独具匠心地实现了题材自动发现功能，并引入了对短线交易者极具参考价值的St","stars":317,"starGrowth":2,"createdAt":"2026-01-12T15:06:06Z","silentRounds":0,"language":"Python","topics":["finance"],"aiDims":["AI应用","非AI-实用"],"zone":"AI","funScore":0.7,"domainTags":["股票筛选","AI分析","金融"],"domainKey":"股票筛选","pushedAt":"2026-09-22T15:17:12Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/xang1234/stock-screener","ts":"2026-09-22T15:17:12Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wordpress-mobile/WordPress-Android","owner":"wordpress-mobile","name":"WordPress-Android","desc":"WordPress for Android","summaryCn":"WordPress官方安卓App，开源免费，管理博客网站超方便。","reasonCn":"WordPress-Android是WordPress官方的移动端应用开源代码库，使用Kotlin开发，支持完整的CMS功能，包括撰写文章、管理媒体库、查看统计数据等。作为全球最流行的博客平台的官方客户端，它的代码质量高、架构规范，是学习Android UI开发和Java/Kotlin混合项目的绝佳素材，同时也方便有定制需求的企业进行二次开发。","stars":3153,"starGrowth":0,"createdAt":"2013-04-08T20:52:40Z","silentRounds":24,"language":"Kotlin","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Android 开发","移动应用","CMS"],"domainKey":"Android 开发","pushedAt":"2026-09-23T18:41:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/wordpress-mobile/WordPress-Android","ts":"2026-09-23T18:41:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ultrasecurity/Storm-Breaker","owner":"ultrasecurity","name":"Storm-Breaker","desc":"","summaryCn":"该项目在实现层面展示了现代Web技术与设备交互的边界，通过将原本复杂的命令","reasonCn":"工具集成了地理位置追踪、实时Webcam捕捉及音频监控等高级功能，这些功能在授权的安全审计场景中极具参考意义。特别值得注意的是其日志管理系统，用户不仅可以即时查看监听记录，还支持随时清空日志并下载归档，确保了测试过程的可追踪性。","stars":5995,"starGrowth":1,"createdAt":"2020-12-17T07:21:14Z","silentRounds":3,"language":"HTML","topics":["hack","hacking-tool","social-engineering","social-engineering-attacks","social-engeneering-toolkit"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网络安全","设备监控","Web技术","地理定位"],"domainKey":"网络安全","pushedAt":"2026-09-16T18:07:39.903Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"hack","source":"github","weight":0.7},{"name":"hacking-tool","source":"github","weight":0.7},{"name":"social-engineering","source":"github","weight":0.7},{"name":"social-engineering-attacks","source":"github","weight":0.7},{"name":"social-engeneering-toolkit","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/ultrasecurity/Storm-Breaker","ts":"2026-09-16T18:07:39.903Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nolimits4web/swiper","owner":"nolimits4web","name":"swiper","desc":"Most modern mobile touch slider with hardware accelerated transitions","summaryCn":"现代移动端触摸轮播组件，支持硬件加速过渡效果","reasonCn":"Swiper是最流行的移动端触摸滑动组件库，支持多种布局模式和过渡动画效果。采用原生JavaScript开发，不依赖任何框架，可轻松集成到Vue、React、Angular等项目。提供硬件加速的流畅动画，适合展示图片轮播、内容滑块等场景。文档完善，社区活跃，是前端开发者的实用工具箱之一。","stars":41902,"starGrowth":0,"createdAt":"2012-03-14T19:11:45Z","silentRounds":6,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["触摸交互","轮播组件","前端开发","移动端UI"],"domainKey":"触摸交互","pushedAt":"2026-09-14T20:08:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/nolimits4web/swiper","ts":"2026-09-14T20:08:32Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zhitongblog/solomd","owner":"zhitongblog","name":"solomd","desc":"A markdown editor — and the bridge to your LLM. Local-first, MIT, ~15 MB. Bundled MCP server lets Claude Code / Codex / Cursor drive your vault directly. 14 AI providers BYOK.","summaryCn":"不仅是轻量笔记软件，更是能让AI帮你写文档的编辑器","reasonCn":"它的核心亮点是支持 MCP（Model Context Protocol）协议。这意味着像 Claude Code 或 Cursor 这样的 AI 工具可以通过这个协议，直接‘读写’你的本地笔记文件。你不需要手动上传文档，AI可以直接在你的笔记库里搜索相关内容并据此写作。","stars":1099,"starGrowth":1,"createdAt":"2026-04-08T03:52:45Z","silentRounds":1,"language":"TypeScript","topics":["note-taking"],"aiDims":["AI写作","代码助手","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["Markdown编辑器","本地笔记","AI集成"],"domainKey":"Markdown编辑器","pushedAt":"2026-09-23T14:20:57Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/zhitongblog/solomd","ts":"2026-09-23T14:20:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"neomjs/neo","owner":"neomjs","name":"neo","desc":"Neo.mjs is a self-evolving software organism: a professional end-to-end AI engineering team whose cross-model swarm inhabits live apps via Neural Link, Active Hybrid GraphRAG, DreamService, and self-healing loops.","summaryCn":"一个会自我进化的AI软件工程团队，多人智能体协作写代码的先锋。","reasonCn":"这个项目概念非常超前，它不只是单点AI，而是构建了一个“软件组织”。它利用多模态模型通过GraphRAG等技术，让多个AI Agent像真正的工程师一样分工协作，在活的应用上共同完成任务。虽然概念很火，但作为实战工具还处于较早期阶段。","stars":3282,"starGrowth":0,"createdAt":"2019-11-10T23:56:02Z","silentRounds":3,"language":"JavaScript","topics":["frontend","ai-agent"],"aiDims":["AI Agent","多模态Agent","代码助手"],"zone":"AI","funScore":0.7,"domainTags":["AI Agent","多智能体协作","代码生成"],"domainKey":"AI Agent","pushedAt":"2026-09-23T17:10:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/neomjs/neo","ts":"2026-09-23T17:10:42Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"adonisjs/core","owner":"adonisjs","name":"core","desc":"AdonisJS is a TypeScript-first web framework for building web apps and API servers. It comes with support for testing, modern tooling, an ecosystem of official packages, and more.","summaryCn":"TypeScript优先的企业级Web框架，开箱即用功能齐全","reasonCn":"技术上，AdonisJS提供了一套完整的应用开发工具链：内置HTTP服务器和路由器、数据库ORM和迁移系统、认证和授权中间件、模板引擎和验证器、队列和任务调度等。它的架构设计遵循MVC模式，代码结构清晰，易于维护和团队协作。","stars":19132,"starGrowth":2,"createdAt":"2015-08-15T17:03:54Z","silentRounds":0,"language":"TypeScript","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Web框架","TypeScript","Node.js","API开发","MVC"],"domainKey":"Web框架","pushedAt":"2026-09-23T14:03:40Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/adonisjs/core","ts":"2026-09-23T14:03:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"m1k1o/neko","owner":"m1k1o","name":"neko","desc":"A self hosted virtual browser that runs in docker and uses WebRTC.","summaryCn":"自托管虚拟浏览器，在Docker里开个云电脑浏览器","reasonCn":"Neko是一个自托管的虚拟浏览器，运行在Docker容器中，通过WebRTC技术让浏览器界面流式传输到你的本地设备。你可以用它来安全地访问不信任的网站、进行自动化测试或作为远程桌面使用，完全掌控数据，不依赖第三方云服务。","stars":22369,"starGrowth":0,"createdAt":"2020-03-14T20:46:07Z","silentRounds":1,"language":"Go","topics":["self-hosted","docker"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Docker","虚拟浏览器","WebRTC"],"domainKey":"Docker","pushedAt":"2026-09-10T20:43:39Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/m1k1o/neko","ts":"2026-09-10T20:43:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"backend-br/desafios","owner":"backend-br","name":"desafios","desc":"Aqui você encontrará uma coleção de desafios projetados para aprimorar suas habilidades de programação com foco em backend.","summaryCn":"程序员后端刷题练手的好帮手，实战项目提技术","reasonCn":"项目中的题目设计比较贴近实际工作场景，不是那种纯刷LeetCode的纯算法题。你会遇到需要设计RESTful API、处理数据库查询、实现缓存逻辑、做身份验证等各种真实开发中会遇到的问题。这种题目能让你在解决过程中自然地学到后端开发的完整知识体系。","stars":2703,"starGrowth":1,"createdAt":"2018-02-06T14:21:40Z","silentRounds":8,"language":"","topics":["backend"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["后端开发","编程练习","面试题库"],"domainKey":"后端开发","pushedAt":"2026-09-17T19:33:41Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7}],"aiScore":0.3,"source":"search","url":"https://github.com/backend-br/desafios","ts":"2026-09-17T19:33:41Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"0xsline/OpenChatCut","owner":"0xsline","name":"OpenChatCut","desc":"Open-source, local-first conversational AI video editor with a professional multi-track timeline, Agent Skills, MCP integration, and Remotion rendering.","summaryCn":"把聊天记录直接变成精美短视频，动动嘴就能让AI帮你剪辑片子。","reasonCn":"从技术架构来看，它采用了本地优先的隐私保护设计，保证用户的素材安全。核心集成了强大的多轨道时间轴，支持复杂的视频层叠与音频混音。同时，它引入了智能代理（Agent）技能和MCP集成，能够智能理解用户的剪辑意图，并无缝对接Remotion引擎进行高性能的视频渲染和输出，技术栈非","stars":1975,"starGrowth":15,"createdAt":"2026-07-15T02:49:42Z","silentRounds":0,"language":"TypeScript","topics":["video-editor"],"aiDims":["AI应用","创意工具"],"zone":"创意","funScore":1,"domainTags":["视频剪辑","AI视频生成","Remotion","多轨道时间轴","自动化创作"],"domainKey":"视频剪辑","pushedAt":"2026-09-21T11:37:58Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/0xsline/OpenChatCut","ts":"2026-09-21T11:37:58Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.10167721518987341,"copyOk":false},{"repo":"conorbronsdon/avoid-ai-writing","owner":"conorbronsdon","name":"avoid-ai-writing","desc":"Skill that audits and rewrites content to remove AI writing patterns. Use it with your favorite agents including Claude Code, OpenClaw, Codex, and Hermes.","summaryCn":"专治AI写作痕迹神器让ChatGPT生成的文章读起来像人写的","reasonCn":"这个工具专门检测并去除AI写作的套路痕迹。用ChatGPT写过文章的人都知道那种首先其次最后的人味缺失感。它可以接入Claude Code、Codex等AI助手，在内容生成时自动审计并改写，去除明显AI模式，让你的文章回归自然表达。","stars":4672,"starGrowth":17,"createdAt":"2026-03-06T07:47:40Z","silentRounds":0,"language":"JavaScript","topics":["writing"],"aiDims":["AI写作","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI写作","文本分析","技能包"],"domainKey":"AI写作","pushedAt":"2026-09-23T18:17:24Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/conorbronsdon/avoid-ai-writing","ts":"2026-09-23T18:17:24Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.07429931506849315,"copyOk":false},{"repo":"gleam-wisp/wisp","owner":"gleam-wisp","name":"wisp","desc":"🧚 A practical web framework for Gleam","summaryCn":"Gleam语言的Web开发框架，小众但设计优雅","reasonCn":"Wisp是一个为Gleam语言量身打造的Web框架，Gleam是编译到Erlang和JavaScript的新型函数式语言。Wisp提供路由、中间件、HTTP处理等Web开发核心能力，语法风格类似Elixir的Phoenix但更安全。适合喜欢探索新语言的开发者尝试。","stars":1482,"starGrowth":0,"createdAt":"2023-08-04T19:41:07Z","silentRounds":3,"language":"Gleam","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web框架","函数式编程","后端开发"],"domainKey":"Web框架","pushedAt":"2026-09-14T14:59:19Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Gleam","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/gleam-wisp/wisp","ts":"2026-09-14T14:59:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"marcelscruz/dev-resources","owner":"marcelscruz","name":"dev-resources","desc":"A collaborative list of resources for developers","summaryCn":"程序员必备的宝藏资源库，涵盖前端后端各类开发技能","reasonCn":"这是一个由社区维护的开发者资源清单，汇集了大量学习材料、工具推荐和设计资源。内容覆盖HTML、CSS、JavaScript、框架及后端技术，采用Markdown格式易于阅读。虽非AI项目，却是提升编码效率、拓展技术视野的实用导航站，适合初学者与资深开发者收藏备查。","stars":1341,"starGrowth":2,"createdAt":"2021-01-25T19:34:25Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["开发者工具","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["开发者资源","编程学习","技术导航"],"domainKey":"开发者资源","pushedAt":"2026-09-23T11:05:37Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/marcelscruz/dev-resources","ts":"2026-09-23T11:05:37Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zouyee/zig-learning","owner":"zouyee","name":"zig-learning","desc":"A bunch of links to blog posts, articles, videos, etc for learning Zig","summaryCn":"Zig 语言新手入门的精选学习资源导航集合","reasonCn":"这是一个精心整理的 Zig 编程语言学习链接集合，涵盖了从官方文档、博客文章到视频教程的各种资源。它帮助开发者快速构建知识体系，避免在浩瀚的信息海洋中迷失方向。虽然不涉及复杂算法，但作为入门拐杖非常实用，适合想尝试新一代系统编程语言的学习者。无需运行复杂代码，浏览即可获取价值，是低成本探索新语言的高效途径，协助你迈出 Zig 学习的第一步。","stars":204,"starGrowth":0,"createdAt":"2024-11-27T03:21:33Z","silentRounds":8,"language":"","topics":["learning"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["Zig语言","编程教程","学习导航","开发者资源"],"domainKey":"Zig语言","pushedAt":"2026-09-14T11:30:37Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7}],"aiScore":0.35,"source":"search","url":"https://github.com/zouyee/zig-learning","ts":"2026-09-14T11:30:37Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"oceanbase/oceanbase","owner":"oceanbase","name":"oceanbase","desc":"The Fastest Distributed Database for Transactional, Analytical, and  AI Workloads.","summaryCn":"国产分布式数据库巨头，兼顾交易分析与AI向量检索的全能选手","reasonCn":"OceanBase是蚂蚁集团出品的分布式关系型数据库，以极高性能和强一致性著称。近期增强了对AI工作负载的支持，内置向量检索能力，支持HTAP混合负载。对于需要处理海量数据且兼顾传统事务与AI向量化查询的场景，它是极具竞争力的国产开源选择。","stars":10285,"starGrowth":0,"createdAt":"2021-05-31T14:36:17Z","silentRounds":1,"language":"C++","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","大语言模型"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","HTAP","向量检索","国产数据库"],"domainKey":"分布式数据库","pushedAt":"2026-09-23T01:00:38Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/oceanbase/oceanbase","ts":"2026-09-23T01:00:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"asgeirtj/system_prompts_leaks","owner":"asgeirtj","name":"system_prompts_leaks","desc":"Extracted system prompts from Anthropic - Claude Fable 5.1, Opus 5, Claude Design, Claude Code. OpenAI - ChatGPT GPT-6-Astra, Codex. Google - Gemini 3.8 Flash, 3.1 Pro, Antigravity. xAI - Grok, Grok Bot, Cursor, Kimi and more! Updated regularly.","summaryCn":"你有没有好奇过：ChatGPT回答问题的方式，Claude写出那么有条理的文字，背后","reasonCn":"收集的内容相当全面，包括Anthropic的Claude系列、OpenAI的ChatGPT和Codex、Google的Gemini系列，以及xAI的Grok等。每条提示词都标注了模型版本和来源，方便对比不同模型的'性格'差异。 使用方法很简单，clone仓库直接查看即可。","stars":68087,"starGrowth":48,"createdAt":"2025-05-03T02:43:56Z","silentRounds":0,"language":"JavaScript","topics":["llm","ai","chatbot"],"aiDims":["提示工程","学习资源","安全工具"],"zone":"资源","funScore":1,"domainTags":["系统提示词","提示工程","大模型内部机制"],"domainKey":"系统提示词","pushedAt":"2026-09-22T20:10:09Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/asgeirtj/system_prompts_leaks","ts":"2026-09-22T20:10:09Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.14611494117819848,"copyOk":false},{"repo":"javalite/javalite","owner":"javalite","name":"javalite","desc":"JavaLite is a cohesive collection of frameworks designed from ground up to add pleasure back to your daily life","summaryCn":"让写Java代码像搭积木一样简单的轻量级框架","reasonCn":"Javalite是一套从零开始设计的Java轻量级框架集合，旨在为开发者提供愉悦的Web开发体验。它包含轻量级的ORM、安全框架和Web框架，核心设计理念是零配置、极简依赖和高度可扩展性，让你告别繁琐的XML配置和臃肿的Spring容器，专注于业务逻辑本身。","stars":870,"starGrowth":0,"createdAt":"2013-05-31T07:33:11Z","silentRounds":25,"language":"Java","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Java Web框架","ORM","极简开发"],"domainKey":"Java Web框架","pushedAt":"2026-09-03T08:07:53Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/javalite/javalite","ts":"2026-09-03T08:07:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"go-goyave/goyave","owner":"go-goyave","name":"goyave","desc":"🍐 The enterprise REST API framework","summaryCn":"写后端API像搭乐高一样简单的Go框架。","reasonCn":"Goyave是一个现代的Go语言REST API框架，旨在提供类型安全和清晰的代码结构。它内置了强大的验证功能，自动处理HTTP请求和响应，支持依赖注入和中间件机制。框架设计注重可维护性和可扩展性，内置了错误处理和日志记录，能显著提升开发效率。","stars":1778,"starGrowth":0,"createdAt":"2019-10-21T09:44:34Z","silentRounds":8,"language":"Go","topics":["backend","web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Go语言","REST API","后端框架","依赖注入","数据验证"],"domainKey":"Go语言","pushedAt":"2026-09-23T14:35:06Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/go-goyave/goyave","ts":"2026-09-23T14:35:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"visgl/react-map-gl","owner":"visgl","name":"react-map-gl","desc":"React friendly API wrapper around MapboxGL JS","summaryCn":"React写地图太简单，拖代码进去就能跑出高大上的交互地图界面","reasonCn":"这是visgl家族的核心成员，专为React打造的地图开发库。底层基于MapboxGL JS，封装成React组件后，状态管理和渲染逻辑都变得极其顺手。无论是做数据大屏、地理信息系统还是LBS应用，都能快速集成，还支持自定义图层和3D效果，是前端地图开发的强力工具。","stars":8505,"starGrowth":0,"createdAt":"2015-10-08T16:56:26Z","silentRounds":3,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["React","地图开发","WebGL","数据可视化","LBS"],"domainKey":"React","pushedAt":"2026-09-03T03:30:14Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/visgl/react-map-gl","ts":"2026-09-03T03:30:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Kong/kong","owner":"Kong","name":"kong","desc":"🦍 The API and AI Gateway","summaryCn":"既能管API流量又能调AI模型的中转站，企业级AI网关的开源之选","reasonCn":"老牌API网关Kong加入了AI网关功能，支持请求转发到OpenAI等大模型服务，内置限流、认证、日志等治理能力。Lua编写的核心插件系统高度可定制，配合Kong Gateway可轻松搭建企业级AI应用架构，处理并发调用、成本管理、敏感词过滤不在话下。","stars":44175,"starGrowth":8,"createdAt":"2014-11-17T23:56:08Z","silentRounds":0,"language":"Lua","topics":["devops","ai"],"aiDims":["AI基础设施","工具调用","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["API网关","流量治理","企业中间件","Lua插件"],"domainKey":"API网关","pushedAt":"2026-09-23T13:39:08Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Lua","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/Kong/kong","ts":"2026-09-23T13:39:08Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024543293718166385,"copyOk":false},{"repo":"53AI/53AIHub","owner":"53AI","name":"53AIHub","desc":"53AI Hub is an open-source AI portal and knowledge base for managing enterprise knowledge, AI agents, prompts, and AI tools, seamlessly integrating with Coze, Dify, FastGPT, RAGFlow. 一个AI知识库与Agent门户","summaryCn":"53AI Hub是一个开源的AI知识库和AI代理门户，集成了多种AI工具和服务。它支持与Coze、Dify等AI平台无缝集成，为开发者提供一站式管理企业知识、AI工具的平台。","reasonCn":"53AI Hub是一个集成了多种AI工具和服务的知识库和AI代理门户。它支持与Coze、Dify等AI平台无缝集成，为开发者提供了一站式管理企业知识、AI工具的功能。","stars":4658,"starGrowth":0,"createdAt":"2025-04-24T11:24:11Z","silentRounds":9,"language":"Go","topics":["rag"],"aiDims":["AI基础设施","AI Agent"],"zone":"AI","funScore":0,"domainTags":["AI知识库","Agent门户","RAG","提示词管理","Coze集成"],"domainKey":"RAG","pushedAt":"2026-09-22T09:39:38Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/53AI/53AIHub","ts":"2026-09-22T09:39:38Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"trezor/blockbook","owner":"trezor","name":"blockbook","desc":":blue_book: Trezor address/account balance backend ","summaryCn":"自己搭个加密货币查询服务，支持比特币以太坊等主流币种","reasonCn":"Blockbook是Trezor硬件钱包的官方后端，用Go编写。通过索引区块链数据，提供地址余额查询和交易记录检索API。支持比特币、以太坊、莱特币等数十种币，可对接不同节点。部署后可替代第三方服务，实现数据隐私可控，适合需要自建加密资产查询的开发者使用。","stars":795,"starGrowth":0,"createdAt":"2017-08-28T15:51:55Z","silentRounds":24,"language":"Go","topics":["backend"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["区块链","比特币","以太坊","索引服务","自托管"],"domainKey":"区块链","pushedAt":"2026-09-23T12:56:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/trezor/blockbook","ts":"2026-09-23T12:56:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"osmdroid/osmdroid","owner":"osmdroid","name":"osmdroid","desc":"","summaryCn":"这是一个在Android移动开发领域具有悠久历史的开源地图类库，其诞生时间甚至","reasonCn":"在十六年的发展历程中，该项目吸引了超过七十五位全球贡献者的加入，共同塑造了其作为Android早期最知名开源地图库的地位。尽管随着移动生态的变迁与维护成本的不断增加，项目维护团队最终选择将其归档，不再发布官方的新版本或进行常规更新，但其沉淀的技术方案对整个开源地图领域","stars":3078,"starGrowth":0,"createdAt":"2014-03-06T19:03:48Z","silentRounds":4,"language":"Java","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Android","地图开发","开源地图","MapView"],"domainKey":"Android","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/osmdroid/osmdroid","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"renpy/renpy","owner":"renpy","name":"renpy","desc":"The Ren'Py Visual Novel Engine","summaryCn":"用Python写代码就能做视觉小说，零门槛创作互动故事","reasonCn":"Ren'Py是基于Python和Pygame的开源视觉小说引擎，内置完整的剧情脚本系统、角色立绘管理、分支对话树、存档加载、字幕选项和背景音乐控制。代码用Python编写，对程序员极其友好，无需图形编辑器就能快速搭建互动叙事作品。相比商业引擎如Visual Novel Maker，它完全免费开源，可深度定制，支持跨平台发布到Windows、macOS、Linux、Android和iOS，社区活跃且文档完善。","stars":6851,"starGrowth":0,"createdAt":"2012-06-28T01:21:22Z","silentRounds":1,"language":"Ren'Py","topics":["game"],"aiDims":["游戏","创意工具","非AI-好玩"],"zone":"创意","funScore":0.5,"domainTags":["视觉小说","互动叙事","Ren'Py","Python脚本","游戏引擎"],"domainKey":"视觉小说","pushedAt":"2026-08-08T05:54:36Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Ren'Py","source":"language","weight":0.5}],"aiScore":0.48,"source":"search","url":"https://github.com/renpy/renpy","ts":"2026-08-08T05:54:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"YILS-LIN/short-video-factory","owner":"YILS-LIN","name":"short-video-factory","desc":"一键生成产品营销与泛内容短视频，AI批量自动剪辑，高颜值跨平台桌面端工具 One click generation of product marketing and general content short videos, AI batch automatic cliping, beautiful cross platform desktop tool","summaryCn":"AI一键批量剪短视频，营销内容生产终于不用加班了","reasonCn":"支持AI批量自动剪辑短视频的桌面端工具，上传素材即可自动生成带字幕、配乐、转场效果的产品营销视频，跨平台支持Windows和Mac。内置多种模板和风格，一键导出适配抖音、B站、小红书等平台的尺寸格式，极大提升内容生产效率。","stars":5474,"starGrowth":5,"createdAt":"2025-07-09T05:55:44Z","silentRounds":0,"language":"TypeScript","topics":["desktop","automation"],"aiDims":["AI应用","多模态Agent","工具调用"],"zone":"AI","funScore":1,"domainTags":["短视频","AI 剪辑","视频制作","批量处理"],"domainKey":"短视频","pushedAt":"2026-09-22T06:34:09Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/YILS-LIN/short-video-factory","ts":"2026-09-22T06:34:09Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.017740226525392765,"copyOk":true},{"repo":"Q00/ai-agent-engineering-101","owner":"Q00","name":"ai-agent-engineering-101","desc":"AI Agent Engineering 101 — SeoulTech 2026 Fall. Lecture notes and PR-based assignment submission.","summaryCn":"首尔科技大学2026秋季的AI Agent工程课程，配套作业和实战项目","reasonCn":"这是一套完整的AI Agent工程教学材料，涵盖从基础理论到实战开发的全部流程。通过 Lecture Notes 和基于 PR 的作业提交方式，学生可以系统学习如何构建、调试和优化 AI Agent，是深入学习 Agent Engineering 的优质资源。","stars":39,"starGrowth":0,"createdAt":"2026-09-01T06:45:53Z","silentRounds":3,"language":"HTML","topics":["course"],"aiDims":["AI Agent","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["AI Agent","课程","软件工程"],"domainKey":"AI Agent","pushedAt":"2026-09-22T09:56:37Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"course","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Q00/ai-agent-engineering-101","ts":"2026-09-22T09:56:37Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rlaope/estudy","owner":"rlaope","name":"estudy","desc":"컴퓨터 공학 학문에서 쌓아온 나의 견문(見聞)과 견해(見解).","summaryCn":"计算机科班生的核心课程笔记与代码合集，操作系统网络分布式全都有，备考复习必备。","reasonCn":"内容涵盖操作系统（进程管理、内存管理、文件系统）、计算机网络（TCP/IP协议栈、HTTP、DNS）、分布式系统（一致性协议、分布式存储）等计算机核心课程。每个主题下面都有概念梳理、课后习题解答和配套的代码实现。全部用Markdown书写，结构清晰，适合在线阅读。","stars":1089,"starGrowth":0,"createdAt":"2022-06-27T03:24:36Z","silentRounds":3,"language":"","topics":["backend"],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["计算机科学","操作系统","计算机网络","分布式系统","笔记"],"domainKey":"计算机科学","pushedAt":"2026-09-23T11:14:47Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7}],"aiScore":0.2,"source":"search","url":"https://github.com/rlaope/estudy","ts":"2026-09-23T11:14:47Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"oracle/visualvm","owner":"oracle","name":"visualvm","desc":"","summaryCn":"集成 JDK 命令行工具与轻量级性能分析功能。","reasonCn":"项目支持对 Java 进程进行实时监控和性能分析，例如查看堆转储、分析内存泄漏、检查线程死锁等。这些功能对于排查生产环境中的性能问题非常有帮助。VisualVM 的界面友好，操作直观，即使是初学者也能快速上手。对于有一定经验的开发者来说，它还能通过插件系统扩展更多功能，例如连","stars":3267,"starGrowth":1,"createdAt":"2016-09-12T14:44:30Z","silentRounds":4,"language":"Java","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["JDK工具","性能分析","堆转储","系统监控","Java开发"],"domainKey":"JDK工具","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/oracle/visualvm","ts":"2026-09-19T18:31:48.187Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"donnemartin/system-design-primer","owner":"donnemartin","name":"system-design-primer","desc":"Learn how to design large-scale systems. Prep for the system design interview. Includes Anki flashcards.","summaryCn":"大厂系统设计面试必看的开源圣经，帮你搞定高并发架构难题","reasonCn":"这是一个由社区维护的高质量开源学习库，专门帮助工程师准备大型系统的架构设计面试。内容涵盖负载均衡、数据库分片、缓存策略等核心概念，并配有详细的图示和代码示例。虽然不涉及AI算法开发，但它是进入AI工程化领域的必备理论基础，适合想从事AI系统架构的同学阅读。","stars":371314,"starGrowth":88,"createdAt":"2017-02-26T16:15:28Z","silentRounds":0,"language":"Python","topics":["design","interview"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["系统设计","面试准备","分布式系统","后端架构"],"domainKey":"系统设计","pushedAt":"2026-09-15T01:10:09Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"interview","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.25,"source":"trending","url":"https://github.com/donnemartin/system-design-primer","ts":"2026-09-15T01:10:09Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.26471098854338915,"copyOk":true},{"repo":"trufflesecurity/trufflehog","owner":"trufflesecurity","name":"trufflehog","desc":"Find, verify, and analyze leaked credentials","summaryCn":"代码库里藏着的密钥“照妖镜”，扫描一遍，泄露的密码Token全现形。","reasonCn":"TruffleHog 是个专注于检测 Git 历史中硬编码凭据的开源工具，支持扫描本地仓库、GitHub 及云存储。它内置多种高熵检测器和验证逻辑，能发现 AWS Key、JWT、API Token 等各类敏感信息，帮助团队在代码上线前堵住安全漏洞，是 DevSecOps 流程中的必备“体检仪”。","stars":28025,"starGrowth":9,"createdAt":"2016-12-31T05:08:12Z","silentRounds":0,"language":"Go","topics":["security"],"aiDims":["安全工具","开发者工具","实用工具"],"zone":"工具","funScore":0,"domainTags":["DevSecOps","凭证检测","Git扫描"],"domainKey":"DevSecOps","pushedAt":"2026-09-23T20:27:46Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"实用工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/trufflesecurity/trufflehog","ts":"2026-09-23T20:27:46Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02796342551293488,"copyOk":false},{"repo":"vibheksoni/stealth-browser-mcp","owner":"vibheksoni","name":"stealth-browser-mcp","desc":"","summaryCn":"这是一个旨在打破大模型联网壁垒的底层基础设施项目，专门为支持 MCP（Mode","reasonCn":"在核心亮点方面，该项目最大的优势在于能够从容应对现代网站严苛的安全防护机制。无论是令人头疼的 Cloudflare 挑战验证、各类动态反爬虫检测，还是必须输入账号密码的登录墙，它都能凭借真实的浏览器环境和高度隐蔽的操作策略平稳绕过。","stars":2057,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.2,"domainTags":["浏览器自动化","AI代理","爬虫代理","MCP协议"],"domainKey":"浏览器自动化","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"tier","url":"https://github.com/vibheksoni/stealth-browser-mcp","ts":"2026-09-23T18:23:13.861Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dokku/dokku","owner":"dokku","name":"dokku","desc":"A docker-powered PaaS that helps you build and manage the lifecycle of applications","summaryCn":"你自己的Heroku，一行命令托管所有应用","reasonCn":"dokku的核心思路是把Docker容器的复杂性全部隐藏起来。你只需要像用Heroku一样，git push到服务器，应用就自动部署。它基于Docker，用Shell脚本驱动，非常轻量，安装完占用资源极少。支持Node.js、Python、Ruby、Go、PHP等多种语言，还能一键挂载PostgreSQL、MySQL、Redis等数据库","stars":32151,"starGrowth":1,"createdAt":"2013-06-08T10:26:57Z","silentRounds":3,"language":"Shell","topics":["devops","self-hosted"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["PaaS","Docker","自托管","DevOps","部署"],"domainKey":"PaaS","pushedAt":"2026-09-23T08:28:27Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/dokku/dokku","ts":"2026-09-23T08:28:27Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EasyTier/EasyTier","owner":"EasyTier","name":"EasyTier","desc":"A simple, decentralized mesh VPN with WireGuard support.","summaryCn":"像搭积木一样，用 WireGuard 搭建分布式 VPN 网络工具。","reasonCn":"EasyTier 是一个基于 WireGuard 协议的分布式虚拟专用网络（VPN）解决方案，旨在提供简单、去中心化的网络连接。它支持节点自动发现和路由，允许用户在公网环境中像在局域网一样通信，适合远程办公、开发环境互通等场景。","stars":13769,"starGrowth":9,"createdAt":"2023-09-29T17:03:53Z","silentRounds":0,"language":"Rust","topics":["nat-traversal","p2p","rust","tailscale","vpn","zerotier"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["Mesh VPN","WireGuard","去中心化网络","P2P路由","远程协作"],"domainKey":"Mesh VPN","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"nat-traversal","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"tailscale","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"zerotier","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/EasyTier/EasyTier","ts":"2026-08-14T01:23:41.092Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.028960926719442226,"copyOk":false},{"repo":"sv222/devops-interview-questions","owner":"sv222","name":"devops-interview-questions","desc":"This repository contains a collection of answers to common DevOps interview questions. The answers cover a wide range of DevOps topics.","summaryCn":"在竞争激烈的技术招聘市场，DevOps 已成为许多企业必备的岗位技能，但面试题","reasonCn":"在竞争激烈的技术招聘市场，DevOps 已成为许多企业必备的岗位技能，但面试题目往往涉及广泛的工具链和概念，让求职者备感压力。这个仓库把常见的面试题目按主题系统化，提供了简洁而完整的答案，帮助你在短时间内梳理核心要点。","stars":40,"starGrowth":0,"createdAt":"2023-10-28T23:40:37Z","silentRounds":11,"language":"","topics":["interview"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["DevOps","面试题","Kubernetes","CI/CD"],"domainKey":"DevOps","pushedAt":"2026-09-08T12:24:58Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7}],"aiScore":0.41,"source":"search","url":"https://github.com/sv222/devops-interview-questions","ts":"2026-09-08T12:24:58Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ai-boost/awesome-prompts","owner":"ai-boost","name":"awesome-prompts","desc":"Curated list of chatgpt prompts from the top-rated GPTs in the GPTs Store. Prompt Engineering, prompt attack & prompt protect. Advanced Prompt Engineering papers.","summaryCn":"精选自GPT商店最顶尖的提示词宝库，带你玩转高阶提示词。 (29字)","reasonCn":"技术层面上，它不仅仅是简单的文本堆砌，而是一个系统化的提示词工程知识库。里面详细剖析了各种复杂的逻辑控制结构、角色设定技巧以及少样本提示方法。更厉害的是，它还专门收录了防范恶意提示词攻击的安全策略，教你如何防止自己的AI被坏人三言两语就套出底裤，同时收录了多篇行","stars":8926,"starGrowth":2,"createdAt":"2024-01-19T15:15:10Z","silentRounds":0,"language":"","topics":["awesome"],"aiDims":["提示工程","学习资源"],"zone":"资源","funScore":0.7,"domainTags":["提示词工程","GPTs商店","提示词攻击","资源清单","大模型应用"],"domainKey":"提示词工程","pushedAt":"2026-09-23T08:27:28Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.82,"source":"search","url":"https://github.com/ai-boost/awesome-prompts","ts":"2026-09-23T08:27:28Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"resux1338/OSCP-Cheatsheet-v2","owner":"resux1338","name":"OSCP-Cheatsheet-v2","desc":"Quick reference sheet for OSCP exam.","summaryCn":"OSCP认证通关速查表，黑客攻防必备工具清单","reasonCn":"它的内容非常全面，从信息收集（如Nmap、Google Dorks）、漏洞扫描（如Nessus、Burp Suite）到漏洞利用（如Metasploit、Exploit-DB），再到提权（如Linux、Windows提权技巧），应有尽有。它还包含了一些常用的Python脚本和PowerShell命令，方便你在考试中快速编写工具。","stars":25,"starGrowth":1,"createdAt":"2026-05-31T21:23:44Z","silentRounds":5,"language":"PowerShell","topics":["cheatsheet"],"aiDims":["安全工具","非AI-实用","学习资源"],"zone":"资源","funScore":0.5,"domainTags":["渗透测试","OSCP","速查表","安全工具"],"domainKey":"渗透测试","pushedAt":"2026-09-23T12:03:41Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"PowerShell","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/resux1338/OSCP-Cheatsheet-v2","ts":"2026-09-23T12:03:41Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"appium/appium","owner":"appium","name":"appium","desc":"Cross-platform automation framework for all kinds of apps, built on top of the W3C WebDriver protocol","summaryCn":"跨平台自动化测试神器，一套代码搞定手机和网页的自动化操作","reasonCn":"Appium是基于W3C WebDriver协议的开源自动化测试框架，支持iOS、Android和Web应用的全平台测试。通过JSON Wire协议与设备通信，可使用JavaScript、Python、Java等多种语言编写测试脚本。内置丰富的元素定位策略和智能等待机制，能高效处理复杂交互场景，是目前移动端自动化测试领域最主流的解决方案。","stars":21996,"starGrowth":0,"createdAt":"2013-01-09T22:49:50Z","silentRounds":2,"language":"TypeScript","topics":["automation"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["移动端测试","自动化测试","WebDriver","跨平台","Appium"],"domainKey":"移动端测试","pushedAt":"2026-09-23T09:25:19Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/appium/appium","ts":"2026-09-23T09:25:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"c3lang/c3c","owner":"c3lang","name":"c3c","desc":"","summaryCn":"与 C 语言实现全 ABI 兼容的特性，意味着你可以在任何现有的 C 项目中直接引","reasonCn":"作为一种通用编程语言，C3 的工具链非常完备。官方提供了针对 Windows、Linux、MacOS 和 OpenBSD 的预编译二进制文件，开发者可以像使用现代编译器一样快速上手。如果你想在浏览器中直接体验，项目组还提供了在线运行环境，让你在不安装任何复杂工具的情况下，就能先睹为快地尝试它","stars":5831,"starGrowth":1,"createdAt":"2019-07-31T21:24:37Z","silentRounds":2,"language":"C3","topics":["language","compiler","c3"],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["系统编程","编译器","C语言演进"],"domainKey":"系统编程","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"language","source":"github","weight":0.7},{"name":"compiler","source":"github","weight":0.7},{"name":"c3","source":"github","weight":0.7},{"name":"C3","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/c3lang/c3c","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"julienXX/terminal-notifier","owner":"julienXX","name":"terminal-notifier","desc":"Send User Notifications on macOS from the command-line.","summaryCn":"在终端里敲命令就能弹出macOS桌面通知的神器","reasonCn":"一个轻量级macOS命令行通知工具，用Objective-C编写，直接调用系统通知中心API。安装后可在shell脚本、后台任务完成后弹出系统通知，无需GUI，支持分组、音效、自定义标题等多种参数配置，适合终端重度用户提升效率。","stars":7336,"starGrowth":1,"createdAt":"2012-07-15T15:48:59Z","silentRounds":1,"language":"Objective-C","topics":["command-line"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["macOS","命令行工具","通知系统","Shell脚本"],"domainKey":"macOS","pushedAt":"2026-08-30T13:06:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Objective-C","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/julienXX/terminal-notifier","ts":"2026-08-30T13:06:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"girafe-ai/ml-course","owner":"girafe-ai","name":"ml-course","desc":"Open Machine Learning course","summaryCn":"手把手带你从零基础入门机器学习的顶级开源实战课程","reasonCn":"从技术架构来看，整个课程仓库完全基于 Jupyter Notebook 构建，所有的代码、图表和文字解释都完美融合在同一个文档中。它不仅覆盖了线性回归、决策树等传统的机器学习算法，还深入到了神经网络、计算机视觉和自然语言处理等前沿深度学习领域。","stars":3563,"starGrowth":1,"createdAt":"2019-02-01T16:20:39Z","silentRounds":2,"language":"Jupyter Notebook","topics":["course"],"aiDims":["模型与训练","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["机器学习","课程","Jupyter Notebook"],"domainKey":"机器学习","pushedAt":"2026-09-22T12:53:11Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"course","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/girafe-ai/ml-course","ts":"2026-09-22T12:53:11Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"codecoradev/uteke","owner":"codecoradev","name":"uteke","desc":"🧠 The Brain for Your AI — Local-first memory engine for AI agents. Store, recall, and search memories with semantic embeddings. Single Rust binary, zero config, fully offline.","summaryCn":"当你的AI应用开始变多，模型本身就会变得“健忘”，这时候你需要一个靠谱的记","reasonCn":"技术架构上，Uteke选用了Rust语言开发，确保了内存安全和惊人的处理速度。它不仅是一个简单的数据库，更是一个专为语义搜索优化过的嵌入引擎。它支持高效的向量相似度匹配，即使在数据量巨大的情况下，也能在毫秒级时间内完成查询。","stars":263,"starGrowth":0,"createdAt":"2026-05-29T00:27:31Z","silentRounds":2,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","RAG与知识"],"zone":"AI","funScore":0.6,"domainTags":["向量数据库","语义搜索","AI记忆","Rust"],"domainKey":"向量数据库","pushedAt":"2026-09-23T01:30:41Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/codecoradev/uteke","ts":"2026-09-23T01:30:41Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PetoiCamp/OpenCatEsp32-Quadruped-Robot","owner":"PetoiCamp","name":"OpenCatEsp32-Quadruped-Robot","desc":"An ESP32-based open source quadruped robot pet framework for developing Boston Dynamics-style four-legged robots that are perfect for STEM, coding & robotics education, IoT robotics applications, AI-enhanced robotics application services, research, and DIY robotics kit development.","summaryCn":"OpenCatEsp32 让你用 ESP32 开发四足机器人，模仿 Boston Dynamics 的四足运","reasonCn":"OpenCatEsp32 让你用 ESP32 开发四足机器人，模仿 Boston Dynamics 的四足运动。它提供一套完整的硬件连接图、驱动代码与运动学模型，适合 STEM 教育、机器人竞赛或 IoT 原型。\\n\\n项目核心是 ESP32 的双核心和低功耗特性，配合 PWM 舵机驱动与 IMU 惯性测量，代码实现了闭环控制、","stars":420,"starGrowth":3,"createdAt":"2021-02-19T09:25:42Z","silentRounds":0,"language":"C++","topics":["arduino"],"aiDims":["非AI-好玩"],"zone":"工具","funScore":0.7,"domainTags":["四足机器人","ESP32","STEM教育","运动学"],"domainKey":"四足机器人","pushedAt":"2026-09-08T04:54:40Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/PetoiCamp/OpenCatEsp32-Quadruped-Robot","ts":"2026-09-08T04:54:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NexusForever/NexusForever","owner":"NexusForever","name":"NexusForever","desc":"An emulator for the defunct MMORPG WildStar which supports build 16042 (Final Patch)","summaryCn":"WildStar 作为曾经风靡的第一人称 MMORPG，官方在 2017 年停止运营后，玩家","reasonCn":"WildStar 作为曾经风靡的第一人称 MMORPG，官方在 2017 年停止运营后，玩家只能靠记忆与旧文件重现。NexusForever 就是为这些怀旧玩家量身打造的完整仿真平台。它在 Windows/Linux/macOS 上跑得通，内部实现了官方服务器协议，允许玩家在本机上搭建多服并进行跨服游戏，甚至还可以","stars":669,"starGrowth":0,"createdAt":"2018-09-13T02:53:39Z","silentRounds":15,"language":"C#","topics":["emulator"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["MMORPG","游戏服务器","逆向工程","怀旧"],"domainKey":"MMORPG","pushedAt":"2026-08-31T00:33:36Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/NexusForever/NexusForever","ts":"2026-08-31T00:33:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FortAwesome/react-fontawesome","owner":"FortAwesome","name":"react-fontawesome","desc":"Official React Component for Font Awesome Icons","summaryCn":"Font Awesome图标在React里一键搞定，前端开发必备神器","reasonCn":"这是Font Awesome官方出的React组件库，让开发者能在React项目中轻松引入和使用Font Awesome图标。支持TypeScript类型定义，内置超过16000个图标，涵盖solid、regular、brands等多种风格。通过npm安装后，只需在代码中导入对应图标组件，设置size、color等属性即可调用，还能自定义CSS样式，非常适合需要大量图标的Web应用开发。","stars":3736,"starGrowth":0,"createdAt":"2017-09-06T15:06:27Z","silentRounds":23,"language":"TypeScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["React图标组件","Font Awesome","SVG图标","前端UI库","TypeScript类型"],"domainKey":"React图标组件","pushedAt":"2026-09-13T15:30:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/FortAwesome/react-fontawesome","ts":"2026-09-13T15:30:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xPlaygrounds/rig","owner":"0xPlaygrounds","name":"rig","desc":"⚙️🦀 Build modular and scalable LLM Applications in Rust","summaryCn":"用Rust构建模块化LLM应用，让AI集成更灵活高效","reasonCn":"Rig是一个用Rust编写的框架，专门用于构建模块化和可扩展的大语言模型应用。它提供了统一的接口来连接各种LLM后端、工具调用和记忆系统，开发者可以像搭积木一样组合不同组件。相比Python方案，Rust版本在性能和内存安全上有显著优势。适合需要高性能AI应用的开发团队，支持本地部署和云端API等多种接入方式。","stars":8699,"starGrowth":2,"createdAt":"2024-06-05T13:42:28Z","silentRounds":0,"language":"Rust","topics":["llm-model"],"aiDims":["大语言模型","AI基础设施","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["LLM框架","Agent运行时","Rust","RAG检索","函数调用"],"domainKey":"LLM框架","pushedAt":"2026-09-23T21:19:15Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/0xPlaygrounds/rig","ts":"2026-09-23T21:19:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"evidence-dev/evidence","owner":"evidence-dev","name":"evidence","desc":"Business intelligence as code: build fast, interactive data visualizations in SQL and markdown","summaryCn":"把SQL查出来的数据直接变成漂亮的网页图表，像搭积木一样简单","reasonCn":"Evidence是一款让开发者能用SQL和Markdown直接构建交互式数据可视化仪表盘的工具。它允许你编写SQL查询来获取数据，然后配合Markdown编写描述，自动生成美观且响应式的网页。项目支持多种图表类型，并能轻松嵌入现有网站，极大简化了BI开发流程。","stars":6956,"starGrowth":0,"createdAt":"2021-05-27T03:50:19Z","silentRounds":1,"language":"JavaScript","topics":["data-visualization","data-science"],"aiDims":["可视化","非AI-实用","AI应用"],"zone":"工具","funScore":0,"domainTags":["BI","数据可视化","SQL","Markdown","仪表盘"],"domainKey":"BI","pushedAt":"2026-09-23T18:49:43Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/evidence-dev/evidence","ts":"2026-09-23T18:49:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Loyalsoldier/clash-rules","owner":"Loyalsoldier","name":"clash-rules","desc":"🦄️ 🎃 👻 Clash Premium 规则集(RULE-SET)，兼容 ClashX Pro、Clash for Windows 等基于 Clash Premium 内核的客户端。","summaryCn":"一个开源的Clash规则集，帮你自动管理广告过滤和节点分流","reasonCn":"这个项目收集了全球主流广告拦截规则和域名分类列表，自动整合成Clash可直接使用的规则格式。支持广告过滤、分流中国大陆与境外流量，无需手动配置。规则持续更新，覆盖AdGuard、EasyList等知名规则源，稳定性高，适合需要网络分流和广告拦截的用户使用。","stars":28518,"starGrowth":5,"createdAt":"2020-07-30T23:53:49Z","silentRounds":0,"language":"","topics":["adblock","adguard","anticensorship","chinalist","clash","dnsmasq","easylist","geosite","gfw","gfwlist","proxy","routing","shadowsocks","ss","ssr","surge","v2ray"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Clash","代理规则","广告过滤","网络分流","DNS"],"domainKey":"Clash","pushedAt":"2026-08-18T22:52:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"adblock","source":"github","weight":0.7},{"name":"adguard","source":"github","weight":0.7},{"name":"anticensorship","source":"github","weight":0.7},{"name":"chinalist","source":"github","weight":0.7},{"name":"clash","source":"github","weight":0.7},{"name":"dnsmasq","source":"github","weight":0.7},{"name":"easylist","source":"github","weight":0.7},{"name":"geosite","source":"github","weight":0.7},{"name":"gfw","source":"github","weight":0.7},{"name":"gfwlist","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"routing","source":"github","weight":0.7},{"name":"shadowsocks","source":"github","weight":0.7},{"name":"ss","source":"github","weight":0.7},{"name":"ssr","source":"github","weight":0.7},{"name":"surge","source":"github","weight":0.7},{"name":"v2ray","source":"github","weight":0.7}],"aiScore":0.45,"source":"bigbro","url":"https://github.com/Loyalsoldier/clash-rules","ts":"2026-08-18T22:52:09Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015525983589312013,"copyOk":true},{"repo":"shadow1ng/fscan","owner":"shadow1ng","name":"fscan","desc":"一款内网综合扫描工具，方便一键自动化、全方位漏扫扫描。(An intranet comprehensive scanning tool, enabling one-click automated, all-round vulnerability scanning)","summaryCn":"想象一下，你拿到一个内网网段，想看看里面有哪些“门户大开”的服务。","reasonCn":"它的核心技术亮点在于“一键”和“全”。它不需要你手动配置复杂的参数，下载下来就能用，支持几十种常见漏洞检测，覆盖面非常广。相比传统的Nmap或者专门针对特定漏洞的扫描器，fscan 的优势在于“综合”，它能在一个流程里完成从端口扫描到漏洞验证的全过程，这对于快速评估一个内网的","stars":14583,"starGrowth":0,"createdAt":"2020-11-13T16:35:20Z","silentRounds":2,"language":"Go","topics":["security"],"aiDims":["安全工具"],"zone":"工具","funScore":0.5,"domainTags":["内网渗透","漏洞扫描","安全工具","自动化","端口探测"],"domainKey":"内网渗透","pushedAt":"2026-09-17T04:19:08Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/shadow1ng/fscan","ts":"2026-09-17T04:19:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"milanm/DevOps-Roadmap","owner":"milanm","name":"DevOps-Roadmap","desc":"DevOps Roadmap for 2026. with learning resources","summaryCn":"超全DevOps学习路线图，从入门到进阶一网打尽，新手福音","reasonCn":"这张路线图详细梳理了2026年DevOps工程师需要掌握的技术栈，涵盖Linux、Git、CI/CD、容器化、云原生等核心领域。它不仅列出工具名称，还贴心地附上了配套的学习资源链接。对于想系统转行或进阶的开发者来说，这是份极其实用的避坑指南，帮你理清杂乱的技术脉络，不再迷茫。","stars":20596,"starGrowth":6,"createdAt":"2022-10-23T09:19:46Z","silentRounds":0,"language":"","topics":["devops"],"aiDims":["开发者工具","学习资源","效率工具"],"zone":"资源","funScore":0,"domainTags":["DevOps","学习路线图","技术栈"],"domainKey":"DevOps","pushedAt":"2026-07-14T15:33:20Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7}],"aiScore":0.45,"source":"search","url":"https://github.com/milanm/DevOps-Roadmap","ts":"2026-07-14T15:33:20Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.01887395610798213,"copyOk":false},{"repo":"lightgbm-org/LightGBM","owner":"lightgbm-org","name":"LightGBM","desc":"A fast, distributed, high performance gradient boosting (GBT, GBDT, GBRT, GBM or MART) framework based on decision tree algorithms, used for ranking, classification and many other machine learning tasks.","summaryCn":"当你的老板丢给你一个几百万行、几十个特征的Excel表格，让你预测下个月的用","reasonCn":"其核心技术灵魂在于“基于梯度的单边采样（GOSS）”和“互斥特征捆绑（EFB）”。简单来说，它跳过了那些无关紧要的数据点，只关注对决策至关重要的部分，同时把相关联的特征合并处理，从而在保证模型精度的前提下，实现了几何级数的训练速度提升。","stars":18797,"starGrowth":1,"createdAt":"2016-08-05T05:45:50Z","silentRounds":1,"language":"C++","topics":["ml"],"aiDims":["AI基础设施","模型与训练"],"zone":"AI","funScore":0,"domainTags":["梯度提升","机器学习","分类回归"],"domainKey":"梯度提升","pushedAt":"2026-09-23T14:57:58Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/lightgbm-org/LightGBM","ts":"2026-09-23T14:57:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mickael-kerjean/filestash","owner":"mickael-kerjean","name":"filestash","desc":"📁 Universal File Storage Client","summaryCn":"一个长得像SaaS的文件管理器，把本地硬盘和云盘串成一个统一入口","reasonCn":"Filestash是一款基于Go语言开发的全能文件管理前端，它通过Web界面集成了多种存储协议，包括本地文件系统、S3、Google Drive、Dropbox、Nextcloud等，支持SSH、FTP、SFTP等连接方式。项目采用模块化架构，支持插件扩展，拥有现代化的响应式UI设计，可以跨平台运行。它解决了用户需要在不同云服务之间频繁切换的痛点，提供统一的操作体验，同时支持通过WebDAV、WebDAV-S等协议挂载到本地，实现文件互通。","stars":14707,"starGrowth":1,"createdAt":"2017-06-13T05:34:14Z","silentRounds":1,"language":"Go","topics":["git"],"aiDims":["非AI-实用","效率工具","开发工具"],"zone":"工具","funScore":0,"domainTags":["文件管理","云存储聚合","WebDAV","自托管","多协议存储"],"domainKey":"文件管理","pushedAt":"2026-09-23T14:39:53Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/mickael-kerjean/filestash","ts":"2026-09-23T14:39:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alisadeghiaghili/Deutsch","owner":"alisadeghiaghili","name":"Deutsch","desc":"Personal German (Deutsch) language learning notes — vocabulary, grammar, conjugation, and cases from A1 level","summaryCn":"如果你正准备自学德语，或者在课堂上需要一本结构化的教材，Deutsch 项目提","reasonCn":"如果你正准备自学德语，或者在课堂上需要一本结构化的教材，Deutsch 项目提供了一个完整的、可搜索的知识库。传统教材往往是纸质或大型 PDF，检索不便，而这里把所有内容拆分成独立的 Markdown 文件，并用 Python 脚本生成索引，打开任意章节都能立即定位到需要的词汇或语法点。","stars":11,"starGrowth":0,"createdAt":"2024-07-09T14:17:59Z","silentRounds":9,"language":"Python","topics":["cheatsheet"],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["德语学习","词汇语法","Markdown笔记"],"domainKey":"德语学习","pushedAt":"2026-09-10T22:15:10Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/alisadeghiaghili/Deutsch","ts":"2026-09-10T22:15:10Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"attify/firmware-analysis-toolkit","owner":"attify","name":"firmware-analysis-toolkit","desc":"Toolkit to emulate firmware and analyse it for security vulnerabilities ","summaryCn":"Rust 写的固件仿真分析套件，找出 IoT 设备安全漏洞","reasonCn":"通过模拟硬件和解析二进制，快速定位固件中的安全缺陷。采用 Rust 高性能，支持多种协议和设备，易于集成到 CI 流程，帮助安全研究员和厂商预防漏洞。支持插件化分析模块，允许自定义脚本以识别特定漏洞模式。已集成多种流行固件格式，如 UEFI、Zephyr 和 ESP32，用户可在本地或云端快速扫描。","stars":1603,"starGrowth":0,"createdAt":"2016-10-24T02:50:15Z","silentRounds":3,"language":"Rust","topics":["iot"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["固件分析","IoT安全","Rust","二进制解析"],"domainKey":"固件分析","pushedAt":"2026-09-08T17:54:03Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"iot","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/attify/firmware-analysis-toolkit","ts":"2026-09-08T17:54:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opentiny/tiny-engine","owner":"opentiny","name":"tiny-engine","desc":"TinyEngine is a low-code engine based on which you can build or develop low-code platforms in different domains/TinyEngine是一个低代码引擎，基于这个引擎可以构建或者开发出不同领域的低代码平台","summaryCn":"阿里开源的低代码引擎，可视化搭建页面，让普通人也能轻松造系统","reasonCn":"TinyEngine 是蚂蚁集团开源的高扩展性低代码引擎，支持可视化拖拽生成前端页面。虽然不直接包含 AI，但它代表了未来人机协作的一种形态——让非程序员也能通过自然语言或图形界面快速构建应用，非常适合想了解低代码或搭建内部工具的场景。","stars":2789,"starGrowth":1,"createdAt":"2023-09-18T02:34:36Z","silentRounds":10,"language":"Vue","topics":["frontend"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["低代码引擎","前端开发","可视化搭建"],"domainKey":"低代码引擎","pushedAt":"2026-08-28T02:35:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/opentiny/tiny-engine","ts":"2026-08-28T02:35:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zi-yue-1129/DATAGEN","owner":"zi-yue-1129","name":"DATAGEN","desc":"DATAGEN: AI-driven multi-agent research assistant automating hypothesis generation, data analysis, and report writing. ","summaryCn":"想象一下，如果你需要一个博士级别的助手，24小时不间断地帮你做实验设计、","reasonCn":"它的核心架构采用多智能体协作模式。系统会自动分配角色，比如“研究员”负责提出假设，“数据分析师”负责处理数据，“报告撰写人”负责总结结论。这些智能体之间通过精心设计的提示词和工具调用进行交互，就像一个真实的研究团队在讨论工作。","stars":1804,"starGrowth":0,"createdAt":"2024-07-23T05:50:15Z","silentRounds":5,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","AI写作","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["多智能体","科研自动化","数据分析"],"domainKey":"多智能体","pushedAt":"2026-09-21T20:37:16Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/zi-yue-1129/DATAGEN","ts":"2026-09-21T20:37:16Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Fosowl/agenticSeek","owner":"Fosowl","name":"agenticSeek","desc":"Fully Local Manus AI. No APIs, No $200 monthly bills. Enjoy an autonomous agent that thinks, browses the web, and code for the sole cost of electricity.","summaryCn":"完全在本地运行的Manus平替，不用花昂贵API费用的全能开源智能体","reasonCn":"在技术架构上，它巧妙地绕过了对各大商业闭源API的依赖，深度整合了本地大模型运行框架与网页自动化抓取工具。它内部的智能循环机制能够让大模型自主拆解复杂任务，比如你让它去全网搜集某个行业的最新研报并写出分析代码，它就会自己打开浏览器、搜索、筛选网页、提取关键信息并在","stars":27273,"starGrowth":1,"createdAt":"2025-02-19T18:03:06Z","silentRounds":2,"language":"Python","topics":["llm","agents"],"aiDims":["AI Agent","AI搜索","大语言模型"],"zone":"AI","funScore":1,"domainTags":["本地运行","自主智能体","AI Agent"],"domainKey":"AI Agent","pushedAt":"2026-09-21T19:29:58Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Fosowl/agenticSeek","ts":"2026-09-21T19:29:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"airbnb/javascript","owner":"airbnb","name":"javascript","desc":"JavaScript Style Guide","summaryCn":"Airbnb团队出品的JS代码规范指南，让代码写得更漂亮。","reasonCn":"这是 Airbnb 团队维护的 JavaScript 编码规范指南，涵盖了从变量命名到异步编程的最佳实践。它不仅是一份风格指南，更是一套帮助团队提升代码可读性、可维护性和协作效率的工程标准。通过统一的编码风格，减少代码审查中的争论，降低技术债务。","stars":148261,"starGrowth":5,"createdAt":"2012-11-01T23:13:50Z","silentRounds":0,"language":"JavaScript","topics":["javascript","eslint","naming-conventions","arrow-functions","style-guide","style-linter","es6","es2015","linting","styleguide","es2016","es2017","es2018","tc39"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["JavaScript","代码规范","最佳实践","指南"],"domainKey":"JavaScript","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"javascript","source":"github","weight":0.7},{"name":"eslint","source":"github","weight":0.7},{"name":"naming-conventions","source":"github","weight":0.7},{"name":"arrow-functions","source":"github","weight":0.7},{"name":"style-guide","source":"github","weight":0.7},{"name":"style-linter","source":"github","weight":0.7},{"name":"es6","source":"github","weight":0.7},{"name":"es2015","source":"github","weight":0.7},{"name":"linting","source":"github","weight":0.7},{"name":"styleguide","source":"github","weight":0.7},{"name":"es2016","source":"github","weight":0.7},{"name":"es2017","source":"github","weight":0.7},{"name":"es2018","source":"github","weight":0.7},{"name":"tc39","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/airbnb/javascript","ts":"2026-08-16T00:54:51.023Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015101172931519415,"copyOk":false},{"repo":"chubin/wttr.in","owner":"chubin","name":"wttr.in","desc":":partly_sunny: The right way to check the weather","summaryCn":"一行命令就能看天气，比手机App还文艺。","reasonCn":"它的核心亮点在于“零配置”和“极简美学”。不需要注册账号，没有广告，不需要安装任何库，只要在浏览器里打开 `wttr.in` 或者在终端里敲 `curl wttr.in`，立马就能看到当前天气。它支持全球2000多个城市，还能自动识别你的IP地址。","stars":30577,"starGrowth":2,"createdAt":"2015-12-26T21:03:43Z","silentRounds":0,"language":"Go","topics":["ascii-art"],"aiDims":["非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["命令行工具","天气查询","ASCII艺术","极简设计"],"domainKey":"命令行工具","pushedAt":"2026-09-07T20:12:34Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"ascii-art","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/chubin/wttr.in","ts":"2026-09-07T20:12:34Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenListTeam/OpenList","owner":"OpenListTeam","name":"OpenList","desc":"A new AList Fork to Anti Trust Crisis","summaryCn":"网盘聚合管理神器，几十个云盘统一挂载访问，文件管理从未如此丝滑","reasonCn":"ate协作……账号多、入口杂、切换烦，OpenList就是来解决这个混乱局面的——它把所有网盘聚合到一个Web界面里，像操作本地文件夹一样管理云端文件。 技术上基于Go语言开发，轻量高效，单二进制文件即可运行。核心原理是调用各网盘的API将远程存储挂载为本地文件系统路径，支持WebDAV协","stars":24735,"starGrowth":5,"createdAt":"2025-06-11T23:35:02Z","silentRounds":0,"language":"Go","topics":["alist","aliyunpan","baidupan","openlist"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["网盘聚合","文件管理","WebDAV","云存储","自托管"],"domainKey":"网盘聚合","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"alist","source":"github","weight":0.7},{"name":"aliyunpan","source":"github","weight":0.7},{"name":"baidupan","source":"github","weight":0.7},{"name":"openlist","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/OpenListTeam/OpenList","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015606428138265615,"copyOk":false},{"repo":"vbcherepanov/total-agent-memory","owner":"vbcherepanov","name":"total-agent-memory","desc":"Persistent memory for Claude Code & Codex CLI. Auto-extracted knowledge graph, multi-representation embeddings, 3D WebGL visualization. LongMemEval R@5=97.45%. Self-hosted, Ollama-optional","summaryCn":"给代码助手装上持久记忆与3D知识图谱的黑科技工具。","reasonCn":"技术架构上，它能够自动从交互中提炼核心知识，构建出结构化的知识图谱，并结合多表征嵌入技术进行深度检索。它在严苛的LongMemEval基准测试中拿下了惊人的高分。同时，它还内置了酷炫的3D WebGL可视化界面，让你可以直观地用图形化方式看到AI脑子里到底记了你代码库的哪些关系网。","stars":69,"starGrowth":1,"createdAt":"2026-02-16T08:14:09Z","silentRounds":14,"language":"Python","topics":["knowledge-management"],"aiDims":["AI Agent","RAG与知识","知识图谱"],"zone":"AI","funScore":1,"domainTags":["知识图谱","AI 记忆","代码助手","可视化"],"domainKey":"知识图谱","pushedAt":"2026-09-21T18:08:31Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/vbcherepanov/total-agent-memory","ts":"2026-09-21T18:08:31Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pytorch/executorch","owner":"pytorch","name":"executorch","desc":"On-device AI across mobile, embedded and edge for PyTorch","summaryCn":"把AI模型装进手机和物联网设备里，随时随地跑模型不联网","reasonCn":"PyTorch官方的端侧AI推理框架，专为移动端、嵌入式设备和边缘计算场景设计。支持模型量化压缩、硬件加速后端（如Qualcomm NNAPI、ARM Ethos），能将PyTorch模型导出为高效格式并在本地离线运行，实现真正的设备端智能。","stars":5045,"starGrowth":1,"createdAt":"2022-02-25T17:58:31Z","silentRounds":2,"language":"Python","topics":["mobile"],"aiDims":["模型部署","AI基础设施","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["端侧推理","移动端","模型部署"],"domainKey":"模型部署","pushedAt":"2026-09-23T21:06:13Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/pytorch/executorch","ts":"2026-09-23T21:06:13Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"egametang/ET","owner":"egametang","name":"ET","desc":"Unity3D Client And C# Server Framework","summaryCn":"国产C#游戏框架，基于Actor模型做MMO网游，性能强易上手","reasonCn":"ET是一套基于Actor模型的Unity游戏框架，用C#编写，专门为大型多人在线游戏设计。它采用成熟的ECS架构思想，支持高并发处理，代码结构清晰。目前GitHub接近一万star，在国内Unity开发者社区口碑不错，是学习游戏服务器架构的优质开源项目。","stars":9916,"starGrowth":1,"createdAt":"2010-06-27T15:50:52Z","silentRounds":1,"language":"C#","topics":["actor","c-sharp","dotnet","game","game-engine","meta","metaverse","mmo","mmorpg","server","unity","unity3d","web3"],"aiDims":["游戏"],"zone":"工具","funScore":0,"domainTags":["游戏开发","游戏引擎","服务器架构"],"domainKey":"游戏开发","pushedAt":"2026-08-26T13:25:42Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"actor","source":"github","weight":0.7},{"name":"c-sharp","source":"github","weight":0.7},{"name":"dotnet","source":"github","weight":0.7},{"name":"game","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"meta","source":"github","weight":0.7},{"name":"metaverse","source":"github","weight":0.7},{"name":"mmo","source":"github","weight":0.7},{"name":"mmorpg","source":"github","weight":0.7},{"name":"server","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"web3","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"bigbro","url":"https://github.com/egametang/ET","ts":"2026-08-26T13:25:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jonashackt/spring-boot-vuejs","owner":"jonashackt","name":"spring-boot-vuejs","desc":"Example project showing how to build a Spring Boot App providing a GUI with Vue.js","summaryCn":"前后端分离开发模板，一键搭建企业级全栈应用","reasonCn":"这是一个Spring Boot后端加Vue.js前端的完整示例项目，帮你快速搭建前后端分离的企业级应用。Spring Boot负责业务逻辑和API，Vue.js负责用户界面，两者配合默契。无需从零开始配置，直接clone下来就能跑起来，学习全栈开发非常友好。","stars":2123,"starGrowth":0,"createdAt":"2017-09-11T09:42:52Z","silentRounds":24,"language":"Java","topics":["backend","frontend"],"aiDims":["学习资源","教程","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["Spring Boot","Vue.js","全栈开发","企业应用","示例项目"],"domainKey":"Spring Boot","pushedAt":"2026-09-20T00:00:16Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/jonashackt/spring-boot-vuejs","ts":"2026-09-20T00:00:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"exceptionless/Exceptionless","owner":"exceptionless","name":"Exceptionless","desc":"Exceptionless application","summaryCn":"让程序员不再漏掉线上bug的自动报错提醒小助手","reasonCn":"这是一个强大的应用故障监控平台，专门用来收集、分类和追踪应用程序中抛出的异常。支持几十种编程语言的Agent接入，能自动识别错误签名，避免重复报警。提供Web界面查看错误趋势、用户影响面以及代码库映射，帮助团队快速定位并修复生产环境的问题，提升软件稳定性。","stars":2456,"starGrowth":0,"createdAt":"2014-02-07T00:48:37Z","silentRounds":8,"language":"C#","topics":["monitoring"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["异常监控","错误追踪","自托管","APM"],"domainKey":"异常监控","pushedAt":"2026-09-13T04:50:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/exceptionless/Exceptionless","ts":"2026-09-13T04:50:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AI4Finance-Foundation/FinGPT","owner":"AI4Finance-Foundation","name":"FinGPT","desc":"FinGPT: Open-Source Financial Large Language Models!  Revolutionize 🔥    We release the trained model on HuggingFace.","summaryCn":"很多大模型对金融专业术语的理解往往流于表面，而FinGPT的目标就是做一个“金","reasonCn":"技术上，它采用了轻量化的微调策略，允许用户在消费级显卡上通过极低的成本定制自己的金融垂直模型。它不仅包含预训练模型，还配套了专门的金融数据清洗工具，这解决了金融AI研发中最大的痛点——高质量数据的缺失。与商业化的闭源金融模型相比，FinGPT提供了完全透明的模型架构与训","stars":21275,"starGrowth":2,"createdAt":"2023-02-11T20:21:34Z","silentRounds":0,"language":"Jupyter Notebook","topics":["ml"],"aiDims":["大语言模型","微调","AI应用"],"zone":"AI","funScore":0,"domainTags":["金融大模型","LLM微调","量化交易","AI金融"],"domainKey":"金融大模型","pushedAt":"2026-09-23T01:59:08Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/AI4Finance-Foundation/FinGPT","ts":"2026-09-23T01:59:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Barca0412/Introduction-to-Quantitative-Finance","owner":"Barca0412","name":"Introduction-to-Quantitative-Finance","desc":"AI+金融（量化）：1.多因子股票量化框架开源教程 2.学界和业界的经典资料收录 3.AI + 金融的相关工作，包括LLM, Agent, benchmark(evaluation), etc.","summaryCn":"把最前沿的人工智能技术和量化金融结合起来的开源保姆级硬核教程。","reasonCn":"深度融合人工智能与量化金融的开源学习教程，不仅系统讲解过多因子股票量化框架的搭建，还全面收录了学界与业界的经典文献。特别前瞻性地探讨了大语言模型、智能体以及模型评测在金融领域的创新应用，是跨界探索AI+金融方向的宝贵实战指南。","stars":1765,"starGrowth":0,"createdAt":"2023-08-04T04:10:47Z","silentRounds":1,"language":"Python","topics":["finance","trading"],"aiDims":["学习资源","AI Agent","大语言模型"],"zone":"资源","funScore":0,"domainTags":["量化金融","多因子模型","LLM应用"],"domainKey":"量化金融","pushedAt":"2026-09-22T23:45:23Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/Barca0412/Introduction-to-Quantitative-Finance","ts":"2026-09-22T23:45:23Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ggez/ggez","owner":"ggez","name":"ggez","desc":"Rust library to create a Good Game Easily","summaryCn":"用Rust写2D游戏的极简神器，几分钟就能让你做出能玩的像素游戏","reasonCn":"ggez是Rust生态中最成熟的2D游戏开发框架之一，提供完整的图形渲染、音频播放、输入处理和游戏循环机制。它基于OpenGL实现，支持Windows、macOS、Linux和Web平台。与Unity、Unreal等重型引擎不同，ggez轻量灵活，代码量极小，适合从零搭建游戏逻辑，且Rust的内存安全特性让游戏运行稳定不易崩溃。","stars":4694,"starGrowth":1,"createdAt":"2016-01-30T14:29:51Z","silentRounds":3,"language":"Rust","topics":["game"],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Rust","游戏引擎","2D游戏开发","图形渲染","开源框架"],"domainKey":"Rust","pushedAt":"2026-08-24T21:20:24Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/ggez/ggez","ts":"2026-08-24T21:20:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"badafans/better-cloudflare-ip","owner":"badafans","name":"better-cloudflare-ip","desc":"","summaryCn":"在使用 Cloudflare 加速的网站时，用户常常会发现偶尔出现加载缓慢或连接超","reasonCn":"从技术实现上看，该程序运行后会自动遍历 Cloudflare 的公共 IP 列表，并对每个 IP 进行连接测试，记录关键的的网络质量数据。最终，它会生成一个排名列表，将丢包率低、延迟小的优质 IP 地址排在最前。用户可以根据这些结果，手动修改本机的 hosts 文件，将特定域名指向最优 IP，","stars":3462,"starGrowth":3,"createdAt":"2020-07-27T04:07:46Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0.2,"domainTags":["网络测速","Cloudflare IP","丢包率检测","Go语言"],"domainKey":"网络测速","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/badafans/better-cloudflare-ip","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"needleworm/bhban_rpa","owner":"needleworm","name":"bhban_rpa","desc":"<6개월 치 업무를 하루 만에 끝내는 업무 자동화(생능출판사, 2020)>의 예제 코드입니다. 파이썬을 한 번도 배워본 적 없는 분들을 위한 예제이며, 엑셀부터 디자인, 매크로, 크롤링까지 업무 자동화와 관련된 다양한 분야 예제가 제공됩니다.","summaryCn":"办公党救星：Python自动化脚本搞定Excel爬虫，下班不再加班","reasonCn":"这是韩国畅销书配套开源代码，针对零编程基础用户打造。涵盖Excel读写、网页爬虫、Word和PPT自动化处理、PDF提取、数据库操作、UI自动点击、定时任务调度等近20个真实场景。每个例子都配有简洁注释，直接复制粘贴就能跑起来，是职场人从\"手动搬砖\"转型\"自动执行\"的最佳入门项目。","stars":1153,"starGrowth":0,"createdAt":"2020-02-11T08:49:41Z","silentRounds":22,"language":"Python","topics":["design"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["办公自动化","Python","Excel处理"],"domainKey":"办公自动化","pushedAt":"2026-08-26T05:28:03Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/needleworm/bhban_rpa","ts":"2026-08-26T05:28:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Wei-Shaw/sub2api","owner":"Wei-Shaw","name":"sub2api","desc":"Sub2API 一站式开源中转服务，让 Claude、Openai 、Gemini、Grok订阅统一接入，支持拼车共享，更高效分摊成本，原生工具无缝使用。","summaryCn":"一接口同时调用Claude、GPT等多模型，拼车分摊降低使用成本","reasonCn":"这是一个API聚合中转服务，能同时对接Claude、OpenAI、Gemini、Grok等主流AI平台。用户只需一个API地址和密钥，就能调用多家模型，还支持多人拼车分摊订阅费用，大幅降低使用成本，特别适合需要多模型对比和选型的开发者与团队。","stars":42409,"starGrowth":86,"createdAt":"2025-12-18T02:26:18Z","silentRounds":0,"language":"Go","topics":["2api","antigravity2api","claude","claude-code","codex","crs","gemini","cc2api","crs2"],"aiDims":["AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["API 聚合","多模型","中转服务","成本优化"],"domainKey":"API 聚合","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"2api","source":"github","weight":0.7},{"name":"antigravity2api","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"crs","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"cc2api","source":"github","weight":0.7},{"name":"crs2","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/Wei-Shaw/sub2api","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.31690033719257704,"copyOk":false},{"repo":"gradio-app/gradio","owner":"gradio-app","name":"gradio","desc":"Build and share delightful machine learning apps, all in Python. 🌟 Star to support our work!","summaryCn":"把模型代码变成可分享的Web应用，让非技术人员也能一键体验你的AI","reasonCn":"Gradio是Hugging Face推出的Python库，让你用几行代码就把机器学习模型包装成美观的Web界面，支持文本、图像、音频等多种输入输出类型，还能一键生成分享链接发给任何人体验。它内置流式传输技术，大模型输出可以逐字显示，体验非常流畅，是目前AI社区最流行的模型展示工具之一。","stars":43600,"starGrowth":3,"createdAt":"2018-12-19T08:24:04Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","模型部署"],"zone":"工具","funScore":0,"domainTags":["机器学习","模型部署","Web界面","Python","AI演示"],"domainKey":"机器学习","pushedAt":"2026-09-23T19:58:21Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/gradio-app/gradio","ts":"2026-09-23T19:58:21Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"shenwei356/rush","owner":"shenwei356","name":"rush","desc":"A cross-platform command-line tool for executing jobs in parallel","summaryCn":"帮你一键并行跑任务，省时省力的终端神器来","reasonCn":"rush 是用 Go 写的跨平台 CLI，支持多线程/多进程并行执行，可通过配置文件指定依赖与执行顺序。它不需要额外容器，直接把本地脚本、命令或可执行文件串起来，适合 bioinformatics、数据处理等流水线。使用时只需写一个 yaml 或 json，rush 自动调度，错误时可重试或并行终止。","stars":1134,"starGrowth":0,"createdAt":"2017-01-03T07:01:03Z","silentRounds":5,"language":"Go","topics":["bioinformatics"],"aiDims":["工具调用","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["并行任务","YAML配置","生物信息学","数据流水线"],"domainKey":"并行任务","pushedAt":"2026-09-21T16:36:46Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"bioinformatics","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.22,"source":"search","url":"https://github.com/shenwei356/rush","ts":"2026-09-21T16:36:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wix/Detox","owner":"wix","name":"Detox","desc":"Gray box end-to-end testing and automation framework for mobile apps","summaryCn":"用代码给App做体检，自动化测试再也不用人工点点点了","reasonCn":"Detox是Wix开源的灰盒端到端测试框架，专为React Native和原生移动应用设计。它通过直接在App内注入JavaScript，精准定位UI元素并执行自动化操作，避免了传统黑盒测试的盲点。支持iOS和Android双平台，测试脚本用JavaScript编写，可与CI/CD流水线无缝集成，显著提升移动端应用的质量保障效率。","stars":12027,"starGrowth":0,"createdAt":"2016-06-15T11:58:01Z","silentRounds":11,"language":"JavaScript","topics":["automation"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["移动端测试","端到端测试","自动化测试"],"domainKey":"移动端测试","pushedAt":"2026-09-07T14:48:39Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/wix/Detox","ts":"2026-09-07T14:48:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bytecodealliance/wasmtime","owner":"bytecodealliance","name":"wasmtime","desc":"A lightweight WebAssembly runtime that is fast, secure, and standards-compliant","summaryCn":"让你的网页应用秒变超快，秒开不卡顿的神器","reasonCn":"Wasmtime是一个高性能WebAssembly运行时，用Rust编写，兼容标准规范。它通过即时编译技术，将WebAssembly字节码高效翻译为机器码，在浏览器和服务器端都能极速执行。支持多种语言编译输出，让Web开发体验更接近原生应用。","stars":18650,"starGrowth":0,"createdAt":"2017-08-29T14:01:55Z","silentRounds":2,"language":"Rust","topics":["wasm","jit","sandbox","cranelift","wasi","wasmtime","webassembly","runtime","rust","aot","standalone","lightweight","portable"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["WebAssembly","运行时","Rust","JIT编译"],"domainKey":"WebAssembly","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"wasm","source":"github","weight":0.7},{"name":"jit","source":"github","weight":0.7},{"name":"sandbox","source":"github","weight":0.7},{"name":"cranelift","source":"github","weight":0.7},{"name":"wasi","source":"github","weight":0.7},{"name":"wasmtime","source":"github","weight":0.7},{"name":"webassembly","source":"github","weight":0.7},{"name":"runtime","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"aot","source":"github","weight":0.7},{"name":"standalone","source":"github","weight":0.7},{"name":"lightweight","source":"github","weight":0.7},{"name":"portable","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/bytecodealliance/wasmtime","ts":"2026-09-12T22:00:03.438Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"streamlit/streamlit","owner":"streamlit","name":"streamlit","desc":"Streamlit — A faster way to build and share data apps.","summaryCn":"几行Python代码就能做出漂亮的数据可视化大屏","reasonCn":"Streamlit是Python生态中最受欢迎的数据应用开发框架，用纯Python就能快速构建交互式数据面板和AI应用界面。无需任何前端知识，内置丰富的UI组件，适合数据科学家和AI开发者快速原型验证和模型演示。star数45000+，是AI应用开发必备工具。","stars":45822,"starGrowth":6,"createdAt":"2019-08-24T00:14:52Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Python","数据可视化","Web框架"],"domainKey":"Python","pushedAt":"2026-09-23T21:43:01Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/streamlit/streamlit","ts":"2026-09-23T21:43:01Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01839282440748985,"copyOk":true},{"repo":"weaviate/weaviate","owner":"weaviate","name":"weaviate","desc":"Weaviate is an open-source vector database that stores both objects and vectors, allowing for the combination of vector search with structured filtering with the fault tolerance and scalability of a cloud-native database​.","summaryCn":"开源向量数据库，存储和搜索Embedding，支持混合向量加过滤查询","reasonCn":"它的独特之处在于同时支持向量搜索和结构化过滤。比如你可以搜\"最近6个月发布的关于Python的向量\"，向量相似度匹配内容相关性，结构化字段过滤时间范围和语言类型。还支持多租户、GraphQL API、内置向量生成器——可以直接调OpenAI或Hugging Face模型把文本转成向量存进去。","stars":16839,"starGrowth":2,"createdAt":"2016-03-30T15:03:17Z","silentRounds":0,"language":"Go","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["向量数据库","RAG","Embedding","AI搜索","向量检索"],"domainKey":"向量数据库","pushedAt":"2026-09-23T21:25:42Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/weaviate/weaviate","ts":"2026-09-23T21:25:42Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"glzr-io/glazewm","owner":"glzr-io","name":"glazewm","desc":"GlazeWM is a tiling window manager for macOS and Windows inspired by i3wm.","summaryCn":"macOS和Windows都能用的平铺窗口管理器，告别拖拽窗口","reasonCn":"GlazeWM是macOS和Windows平台的平铺窗口管理器，灵感来自Linux的i3wm。用Rust编写，支持键盘快捷键和鼠标拖拽，自动将窗口按网格排列，大幅提升多任务效率。配置文件简单，学习曲线平缓，适合追求高效桌面体验的开发者。","stars":12787,"starGrowth":2,"createdAt":"2020-06-28T10:28:26Z","silentRounds":0,"language":"Rust","topics":["i3wm","window-manager","windows","rust","tiling-window-manager","macos"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["窗口管理","平铺","Rust","桌面环境"],"domainKey":"窗口管理","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"i3wm","source":"github","weight":0.7},{"name":"window-manager","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"tiling-window-manager","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/glzr-io/glazewm","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gerardog/gsudo","owner":"gerardog","name":"gsudo","desc":"Sudo for Windows","summaryCn":"Windows版sudo，不用切换窗口直接提权跑命令","reasonCn":"gsudo是Windows平台上的权限提升工具，让你在当前PowerShell或CMD窗口里直接以管理员身份执行命令，无需弹出UAC确认框或切换到管理员终端。支持别名功能、多Session并发管理、以及复杂的命令行参数传递，对习惯Linux sudo的开发者来说体验接近原生生地，还能和现有终端工作流无缝集成。","stars":6056,"starGrowth":1,"createdAt":"2019-11-07T14:10:34Z","silentRounds":3,"language":"C#","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Windows","提权","命令行","运维"],"domainKey":"Windows","pushedAt":"2026-08-12T04:39:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/gerardog/gsudo","ts":"2026-08-12T04:39:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"foundation/foundation-emails","owner":"foundation","name":"foundation-emails","desc":"","summaryCn":"该框架的设计核心是跨平台兼容性，这是邮件营销和自动化通知领域最头疼的问","reasonCn":"上手流程非常现代，通过命令行工具启动项目后，开发者可以使用配套的模板堆栈进行快速开发，并实时通过浏览器同步查看修改结果。这种开发模式模拟了现代Web开发的敏捷流程，极大地提升了UI设计的迭代速度。其提供的测试组件覆盖了市面上主流的邮件客户端，确保最终产出在发布前就已","stars":7794,"starGrowth":0,"createdAt":"2013-09-18T20:13:27Z","silentRounds":7,"language":"HTML","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["邮件开发","响应式设计","邮件模板","前端开发"],"domainKey":"邮件开发","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/foundation/foundation-emails","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mikedaley/web-a2e","owner":"mikedaley","name":"web-a2e","desc":"Cycle-accurate Apple //e and Apple II Plus emulator in the browser — WebAssembly core, WebGL CRT rendering, signal-accurate composite video, no frameworks","summaryCn":"在浏览器里完整复刻 Apple //e，像真机一样玩复古游戏","reasonCn":"web-a2e 是一款基于 WebAssembly 的 Apple //e 与 Apple II Plus 循环精确模拟器，核心采用 C++ 编写的 CPU 与硬件模型，渲染采用 WebGL 实现 CRT 显示，能够在浏览器中完整运行 6502 汇编代码、磁盘映像以及信号级别的视频合成，提供了无框架、即点即用的复古计算体验。","stars":20,"starGrowth":0,"createdAt":"2026-01-05T08:55:58Z","silentRounds":10,"language":"C++","topics":["retrocomputing"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["WebAssembly","复古计算","模拟器","CRT渲染"],"domainKey":"WebAssembly","pushedAt":"2026-09-23T13:35:21Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.79,"source":"search","url":"https://github.com/mikedaley/web-a2e","ts":"2026-09-23T13:35:21Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zb614433612/CodeCraft","owner":"zb614433612","name":"CodeCraft","desc":"基于 DeepSeek 大语言模型的 AI 编程助手 Agent —— 自然语言驱动的终端编码助手。支持自动读写文件、代码搜索、命令执行、Git 管理，复杂任务智能拆解与多 Agent 并行执行。","summaryCn":"像个全能码农助手，帮你搞定代码读写Git等繁琐活儿","reasonCn":"在技术架构上，它采用了智能任务拆解和多Agent并行执行机制，这是它区别于普通代码补全工具的核心优势。当接到一个复杂的编程任务时，它不会盲目地生成代码，而是会像人类专家一样，先规划任务步骤，然后分配给不同的Agent并行处理——比如一个Agent负责代码生成，另一个负责语法检查","stars":10,"starGrowth":0,"createdAt":"2026-05-22T16:00:33Z","silentRounds":22,"language":"Java","topics":["llm-model"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["编程助手","Agent","自然语言编程"],"domainKey":"Agent","pushedAt":"2026-09-22T03:34:24Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/zb614433612/CodeCraft","ts":"2026-09-22T03:34:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"focus-creative-games/leanclr","owner":"focus-creative-games","name":"leanclr","desc":"LeanCLR is a production-oriented CLR implementation. Its core goal is to provide high ECMA-335 compatibility, low integration complexity, and strong cross-platform capability, so developers can truly achieve “Write C#, Run Anywhere.”","summaryCn":"能把.NET程序跑到各种平台上的轻量运行时","reasonCn":"LeanCLR是一个面向生产的ECMA-335兼容运行时实现，专为跨平台部署设计，目标是低集成复杂度和高兼容性。它让你无需安装完整.NET框架就能在Linux、macOS、Windows甚至嵌入式设备上运行.NET程序，对游戏开发等场景尤其有价值。","stars":673,"starGrowth":1,"createdAt":"2025-12-23T01:26:17Z","silentRounds":5,"language":"C++","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":[".NET","运行时","跨平台","游戏开发","ECMA-335"],"domainKey":".NET","pushedAt":"2026-07-23T04:03:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"bigbro","url":"https://github.com/focus-creative-games/leanclr","ts":"2026-07-23T04:03:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mr-highball/wfc","owner":"mr-highball","name":"wfc","desc":"wave function collapse algorithm","summaryCn":"用波函数坍缩算法在屏幕上画出一片片梦幻般的随机像素图案","reasonCn":"核心亮点在于它对WFC算法的经典实现。WFC最初被用于解决数独等约束满足问题，后来被发现能生成极为逼真的纹理。这个项目用Pascal这种古老但高效的语言编写，剥离了现代框架的复杂封装，直接展示算法中“观测”、“坍缩”和“消除”这三个关键步骤是如何在网格上运算的。","stars":12,"starGrowth":0,"createdAt":"2021-01-10T17:50:28Z","silentRounds":12,"language":"Pascal","topics":["procedural-generation"],"aiDims":["非AI-好玩","创意工具","学习资源"],"zone":"创意","funScore":1,"domainTags":["波函数坍缩","程序化生成","像素艺术","算法演示"],"domainKey":"波函数坍缩","pushedAt":"2026-09-07T02:18:17Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Pascal","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/mr-highball/wfc","ts":"2026-09-07T02:18:17Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DioxusLabs/blitz","owner":"DioxusLabs","name":"blitz","desc":"A radically modular HTML/CSS rendering engine","summaryCn":"基于Rust的高性能HTML与CSS渲染引擎","reasonCn":"这是一个用Rust重写的高性能HTML和CSS渲染引擎，专为现代Web应用设计。核心优势是渲染速度极快，内存安全，支持增量更新和精细的DOM diffing。相比传统JavaScript渲染方案，它在复杂页面和大数据量场景下表现优异，适合对性能敏感的前端项目。\n","stars":4174,"starGrowth":0,"createdAt":"2022-02-28T04:42:02Z","silentRounds":1,"language":"Rust","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["渲染引擎","WebAssembly","前端","Rust"],"domainKey":"渲染引擎","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/DioxusLabs/blitz","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"risingwavelabs/risingwave","owner":"risingwavelabs","name":"risingwave","desc":"Event streaming platform for agentic AI. Continuously ingest, transform, and serve event streams in real time, at scale.","summaryCn":"让你的AI助手拥有实时感知能力的云原生数据库","reasonCn":"技术上，它采用了云原生架构，实现了计算与存储的完全分离。最核心的竞争力在于其强大的物化视图能力，支持标准的 SQL 语法，这意味着开发者不需要学习复杂的流处理框架（如 Flink），就能用简单的 SQL 语句处理每秒百万级的事件流。","stars":9341,"starGrowth":3,"createdAt":"2022-01-28T12:58:03Z","silentRounds":0,"language":"Rust","topics":["database"],"aiDims":["AI基础设施","向量数据库"],"zone":"工具","funScore":0,"domainTags":["流处理数据库","实时计算","物化视图","Kafka","Rust"],"domainKey":"流处理数据库","pushedAt":"2026-09-23T18:29:04Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/risingwavelabs/risingwave","ts":"2026-09-23T18:29:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NirDiamant/Agent_Memory_Techniques","owner":"NirDiamant","name":"Agent_Memory_Techniques","desc":"Agent memory for LLMs: 30 runnable Jupyter notebooks covering conversation buffers, vector stores, knowledge graphs, episodic and semantic memory, MemGPT, Mem0, Letta, Zep, Graphiti, LoCoMo benchmarks, and production patterns.","summaryCn":"让AI记住每次对话，30个教程教你实现Agent持久记忆","reasonCn":"这个仓库包含了30个可运行的Jupyter笔记本，从基础的对话缓存到高级的向量存储、知识图谱记忆，再到MemGPT、Mem0、Letta等知名Agent框架的持久化记忆实现，还附带LoCoMo基准测试。涵盖情景记忆、语义记忆等AI Agent核心能力，是系统学习LLM记忆技术的一站式资源。","stars":1073,"starGrowth":4,"createdAt":"2026-05-05T21:08:23Z","silentRounds":0,"language":"Jupyter Notebook","topics":["vector-db"],"aiDims":["AI Agent","RAG与知识","教程"],"zone":"资源","funScore":0,"domainTags":["Agent记忆","向量数据库","Jupyter","教程","知识图谱"],"domainKey":"Agent记忆","pushedAt":"2026-09-21T10:54:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/NirDiamant/Agent_Memory_Techniques","ts":"2026-09-21T10:54:20Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bengarrett/retrotxtgo","owner":"bengarrett","name":"retrotxtgo","desc":"Handle legacy, code page text files in a modern terminal.","summaryCn":"现代终端里读老代码页文本，告别乱码的历史包袱","reasonCn":"这是一款针对现代终端的复古文本处理工具，专门解决Windows老系统代码页（如CP437）文件在UTF-8环境下显示乱码的问题。它通过内置映射表将遗留字符正确渲染，保留了旧式计算机的视觉风格，是复古计算爱好者的实用小工具。","stars":12,"starGrowth":0,"createdAt":"2019-11-17T07:57:07Z","silentRounds":9,"language":"HTML","topics":["retrocomputing"],"aiDims":["非AI-好玩","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["终端工具","代码页转换","复古计算"],"domainKey":"终端工具","pushedAt":"2026-09-12T10:55:00Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/bengarrett/retrotxtgo","ts":"2026-09-12T10:55:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tauri-apps/tauri-action","owner":"tauri-apps","name":"tauri-action","desc":"Build your Web application as a Tauri binary for macOS, Linux and Windows","summaryCn":"一键把网页应用打包成原生桌面应用，支持三大操作系统","reasonCn":"tauri-action是Tauri官方的GitHub Actions集成，帮助开发者在CI/CD流水线中自动构建跨平台桌面应用。支持macOS、Linux和Windows三大平台，只需配置YAML文件即可自动化打包流程，让Web前端技术轻松打造原生体验的桌面软件。","stars":1613,"starGrowth":0,"createdAt":"2020-07-07T21:41:23Z","silentRounds":22,"language":"TypeScript","topics":["github-actions","hacktoberfest"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["GitHub Actions","CI/CD","桌面打包","跨平台构建","自动化部署"],"domainKey":"GitHub Actions","pushedAt":"2026-08-16T04:32:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"github-actions","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"bigbro","url":"https://github.com/tauri-apps/tauri-action","ts":"2026-08-16T04:32:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jupyter/docker-stacks","owner":"jupyter","name":"docker-stacks","desc":"","summaryCn":"该项目提供了一系列预构建的 Jupyter 应用镜像，支持多种开发环境开箱即用，","reasonCn":"在核心亮点方面，该栈不仅兼容了多种不同的应用前端，还能无缝对接 JupyterHub，支持团队协作与资源共享。对于需要在大规模计算集群上部署作业的团队来说，其标准化程度极高，配合现有的 DevOps 工作流，可以实现自动化的环境部署与清理。","stars":8467,"starGrowth":0,"createdAt":"2015-07-15T15:30:00Z","silentRounds":6,"language":"Python","topics":["notebook","jupyter","docker","jupyterhub","jupyter-notebook","jupyterlab","python","ipython","ipython-notebook","jupyter-notebooks"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Jupyter","Docker","数据科学","开发环境"],"domainKey":"Jupyter","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"notebook","source":"github","weight":0.7},{"name":"jupyter","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"jupyterhub","source":"github","weight":0.7},{"name":"jupyter-notebook","source":"github","weight":0.7},{"name":"jupyterlab","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"ipython","source":"github","weight":0.7},{"name":"ipython-notebook","source":"github","weight":0.7},{"name":"jupyter-notebooks","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/jupyter/docker-stacks","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jangtrinh/design-os-3d-blender","owner":"jangtrinh","name":"design-os-3d-blender","desc":"AI-agent operating system for Blender 5.2: skills, verified bpy knowledge base, AGENT_OK/AGENT_FAIL execution contract, production gate for 3D-printable parts, robot-arm demo","summaryCn":"给Blender装上AI大脑，自动建模验证并输出可打印件。","reasonCn":"针对Blender 5.2构建的AI代理操作系统，内置经过验证的bpy知识库。通过AGENT_OK/FAIL执行契约规范操作，集成生产门控机制，自动检查模型是否具备3D打印可行性。项目还包含机械臂演示场景，展示AI在复杂场景中的任务规划能力。","stars":79,"starGrowth":0,"createdAt":"2026-09-06T00:18:05Z","silentRounds":3,"language":"Python","topics":["3d-printing"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["AI代理","Blender","3D建模","3D打印","bpy"],"domainKey":"AI代理","pushedAt":"2026-09-22T09:26:31Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/jangtrinh/design-os-3d-blender","ts":"2026-09-22T09:26:31Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"treeverse/dvc","owner":"treeverse","name":"dvc","desc":"🦉 Data Versioning and ML Experiments","summaryCn":"机器学习项目的数据版本管理器，像Git一样管理数据","reasonCn":"简单来说，DVC就像一个给数据版本控制的Git，它把你的数据集、模型和实验配置都纳入版本管理，但你不需要把几百GB的训练数据塞进代码仓库。它用轻量级的.dvc文件记录引用，真正的数据存在云存储里，比如AWS S3、Google Cloud Storage或者甚至你自己的服务器。","stars":15882,"starGrowth":1,"createdAt":"2017-03-04T08:16:33Z","silentRounds":1,"language":"Python","topics":["developer-tools","ml"],"aiDims":["AI基础设施","模型与训练","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据版本控制","机器学习","Git集成","实验追踪","模型管理"],"domainKey":"数据版本控制","pushedAt":"2026-09-21T18:46:11Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/treeverse/dvc","ts":"2026-09-21T18:46:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ThreeMammals/Ocelot","owner":"ThreeMammals","name":"Ocelot","desc":"","summaryCn":"Ocelot是一款专为.NET开发者设计的API网关，核心目标是解决微服务架构下前端","reasonCn":"该项目的核心亮点在于其配置驱动架构与中间件的结合。所有路由规则、标头操纵、服务发现等操作均通过JSON配置定义，但底层执行依赖于ASP.NET Core的管道系统。这意味着当请求到达一个特殊的\"请求构建器\"中间件时，原始HttpRequest对象会被转换并发送给下游服务，且这是管道中的最后","stars":8714,"starGrowth":0,"createdAt":"2016-06-29T20:35:24Z","silentRounds":7,"language":"C#","topics":["ocelot","dotnet","aspnetcore","gateway","microservices","reverse-proxy","api-gateway"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["API网关","微服务","ASP.NET Core","HTTP路由"],"domainKey":"API网关","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ocelot","source":"github","weight":0.7},{"name":"dotnet","source":"github","weight":0.7},{"name":"aspnetcore","source":"github","weight":0.7},{"name":"gateway","source":"github","weight":0.7},{"name":"microservices","source":"github","weight":0.7},{"name":"reverse-proxy","source":"github","weight":0.7},{"name":"api-gateway","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.32,"source":"tier","url":"https://github.com/ThreeMammals/Ocelot","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kuzzleio/kuzzle","owner":"kuzzleio","name":"kuzzle","desc":"Open-source Back-end, self-hostable & ready to use - Real-time, storage, advanced search - Web, Apps, Mobile, IoT -","summaryCn":"自建后端平台，实时数据加搜索，手机网页物联网全搞定","reasonCn":"Kuzzle是一款自托管开源后端平台，内置Elasticsearch搜索引擎和Redis存储引擎。支持WebSocket实时通信、消息队列、身份认证、权限管理等核心功能。可用于快速搭建移动端、Web端、IoT设备的统一后端，无需从零开发API。","stars":1673,"starGrowth":0,"createdAt":"2015-08-20T08:26:57Z","silentRounds":24,"language":"JavaScript","topics":["backend"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["后端","实时通信","物联网","搜索引擎"],"domainKey":"后端","pushedAt":"2026-08-11T09:08:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/kuzzleio/kuzzle","ts":"2026-08-11T09:08:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rizinorg/cutter","owner":"rizinorg","name":"cutter","desc":"Free and Open Source Reverse Engineering Platform powered by rizin","summaryCn":"图形化逆向分析神器，让二进制代码不再高深莫测","reasonCn":"基于Rizin逆向引擎打造的开源图形化平台，支持反汇编、二进制结构解析、调试和符号分析。提供可视化交叉引用图、字符串提取、十六进制编辑等丰富功能，集成数十种插件扩展，可处理PE、ELF、mach-o等多种可执行文件格式，无需命令行的复杂操作即可完成专业级逆向分析。","stars":19785,"starGrowth":4,"createdAt":"2017-09-25T09:50:51Z","silentRounds":0,"language":"C++","topics":["security"],"aiDims":["安全工具"],"zone":"工具","funScore":0.6,"domainTags":["逆向工程","二进制分析","安全研究"],"domainKey":"逆向工程","pushedAt":"2026-09-11T12:44:05Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/rizinorg/cutter","ts":"2026-09-11T12:44:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"siyuan-note/siyuan","owner":"siyuan-note","name":"siyuan","desc":"A privacy-first, self-hosted, fully open source personal knowledge management software, written in typescript and golang.","summaryCn":"本地搭建的个人知识库，AI帮你自动整理连接","reasonCn":"思源笔记是一款隐私优先的自托管个人知识管理系统，采用TypeScript和Go开发，数据完全存储在本地，支持端到端加密同步。核心特性包括块级编辑、双向链接、标签体系和插件系统。通过AI插件可集成本地大模型，实现智能总结、知识问答、自动关联等能力。项目star数超过4.5万，是中文社区最活跃的开源知识管理工具之一，文档完善，插件生态丰富。","stars":46467,"starGrowth":6,"createdAt":"2020-08-30T09:21:35Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"工具","funScore":0,"domainTags":["知识管理","双向链接","块级编辑","自托管笔记","端到端加密"],"domainKey":"知识管理","pushedAt":"2026-09-23T17:38:38Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/siyuan-note/siyuan","ts":"2026-09-23T17:38:38Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.018387371683129962,"copyOk":false},{"repo":"langchain4j/langchain4j","owner":"langchain4j","name":"langchain4j","desc":"LangChain4j is an idiomatic, open-source Java library for building LLM-powered applications on the JVM. It offers a unified API over popular LLM providers and vector stores, and makes implementing tool calling (including MCP support), agents and RAG easy. It integrates seamlessly with enterprise Java frameworks like Quarkus and Spring Boot.","summaryCn":"Java开发者的AI应用脚手架，链式调用搞定大模型","reasonCn":"技术上，它的核心设计是链式调用（Chain）模式，可以把多个LLM调用串起来形成复杂流程。内置了多种LLM提供商的适配器，覆盖OpenAI、Claude、Gemini、通义千问等主流模型。支持RAG（检索增强生成），对接Pinecone、Milvus、Chroma等向量数据库。","stars":13146,"starGrowth":4,"createdAt":"2023-06-20T15:30:29Z","silentRounds":0,"language":"Java","topics":["vector-db"],"aiDims":["AI基础设施","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["Java","大模型","RAG","链式调用"],"domainKey":"RAG","pushedAt":"2026-09-22T08:24:45Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/langchain4j/langchain4j","ts":"2026-09-22T08:24:45Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"taye/interact.js","owner":"taye","name":"interact.js","desc":"JavaScript drag and drop, resizing and multi-touch gestures with inertia and snapping for modern browsers (and also IE9+)","summaryCn":"浏览器拖拽神器，支持触摸惯性吸附，做交互超简单","reasonCn":"这是一个专注于现代浏览器的高性能拖拽与缩放库，不仅支持鼠标，还完美适配移动端多点触控。它内置了惯性滑动和智能吸附功能，能极大提升网页交互体验。对于开发者而言，它代码轻量、无依赖，且兼容IE9以上，是构建复杂UI拖拽布局的实用利器，尤其适合需要手势交互的Web应用开发场景。","stars":12929,"starGrowth":0,"createdAt":"2012-10-21T10:05:19Z","silentRounds":8,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["拖拽交互","前端库","手势处理"],"domainKey":"拖拽交互","pushedAt":"2026-08-01T04:04:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/taye/interact.js","ts":"2026-08-01T04:04:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Critlist/protoHack","owner":"Critlist","name":"protoHack","desc":"Restoring Jay Fenlason's original 1981—1982 Hack to run on modern Linux.","summaryCn":"把上古时代的远古开源游戏硬核移植到现代Linux系统运行","reasonCn":"这是一场跨越四十年的计算机考古，通过修复古老的源码让一九八一年至八二年的经典游戏在今天发光。开发者重构了底层的驱动和输入输出逻辑，让现代终端能够直接渲染当年的文字冒险与角色扮演游戏。对于复古计算和游戏史爱好者来说，这是一个极为难得且充满情怀的开源实验项目，完美展现了代码的跨时代生命力。","stars":14,"starGrowth":0,"createdAt":"2026-02-01T09:38:18Z","silentRounds":15,"language":"C","topics":["roguelike","retrocomputing"],"aiDims":["非AI-好玩","开发者工具"],"zone":"创意","funScore":0,"domainTags":["Roguelike","复古游戏","移植","C语言"],"domainKey":"Roguelike","pushedAt":"2026-08-13T00:03:24Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"roguelike","source":"github","weight":0.7},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Critlist/protoHack","ts":"2026-08-13T00:03:24Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"trivago/prettier-plugin-sort-imports","owner":"trivago","name":"prettier-plugin-sort-imports","desc":"","summaryCn":"这是一款集成在Prettier代码格式化工具中的插件，它专门用于解决开发过程中","reasonCn":"它的核心优势在于高度的可配置性，你完全可以根据团队的编码规范，制定出一套属于自己的导入排序逻辑。使用方法也非常简单，只需在你的Prettier配置文件中添加相应的order排序规则即可。如果是开发Vue单文件组件，它还能配合相关插件顺利工作，不会因为文件格式的特殊性而产生兼容","stars":3952,"starGrowth":1,"createdAt":"2020-07-20T08:44:42Z","silentRounds":3,"language":"TypeScript","topics":["prettier","prettierrc","typescript","import","sort","plugin","javascript","opensource","hacktoberfest"],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["前端开发","代码规范","自动格式化"],"domainKey":"前端开发","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"prettier","source":"github","weight":0.7},{"name":"prettierrc","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"import","source":"github","weight":0.7},{"name":"sort","source":"github","weight":0.7},{"name":"plugin","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"opensource","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/trivago/prettier-plugin-sort-imports","ts":"2026-09-18T20:05:54.376Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bobeff/open-source-games","owner":"bobeff","name":"open-source-games","desc":"A list of open source games.","summaryCn":"硬核开源游戏大本营，拒绝氪金也能玩通关。","reasonCn":"开源游戏往往能带来意想不到的惊喜。有些游戏在玩法上极具创意，比如《Aseprite》虽然是绘图工具，但被玩家开发出了无数种玩法；有些则在技术上令人惊叹，比如使用WebAssembly在浏览器里跑3D大作。这里没有强制付费，代码完全公开，你可以随意修改、研究，甚至把喜欢的游戏魔改成自","stars":15323,"starGrowth":29,"createdAt":"2021-09-15T21:35:29Z","silentRounds":0,"language":"Python","topics":["awesome-list","games","open-source"],"aiDims":["游戏"],"zone":"创意","funScore":0,"domainTags":["开源游戏","游戏列表","游戏库","游戏资源","独立游戏"],"domainKey":"开源游戏","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"awesome-list","source":"github","weight":0.7},{"name":"games","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/bobeff/open-source-games","ts":"2026-09-19T04:31:19.347Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.09267773934608105,"copyOk":false},{"repo":"feder-cr/AIHawk","owner":"feder-cr","name":"AIHawk","desc":"Born as an AI web agent that applied to jobs in bulk. Becoming a general one: an agent you point at the web and tell what to do.","summaryCn":"对着网页说句话就自动帮你办事，批量投简历的AI打工人。","reasonCn":"AIHawk 是一个基于大语言模型的网页自动化 Agent，最初专为批量投递求职而设计。它能将自然语言指令转化为完整的浏览器操作序列，自动填写表单、点击按钮、处理验证码。核心优势在于具备页面理解能力，可解析 DOM 结构、定位交互元素并自主决策下一步动作，告别传统自动化脚本的硬编码限制，让网页操作真正通过对话驱动。","stars":31626,"starGrowth":5,"createdAt":"2024-08-04T12:15:06Z","silentRounds":0,"language":"Python","topics":["automation","ai-agent"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["网页自动化","AI Agent","求职自动化","浏览器自动化","表单填写"],"domainKey":"AI Agent","pushedAt":"2026-09-21T22:36:38Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/feder-cr/AIHawk","ts":"2026-09-21T22:36:38Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01547429330297856,"copyOk":false},{"repo":"BeiDouMS/BeiDou-Server","owner":"BeiDouMS","name":"BeiDou-Server","desc":"Global MapleStory Server BeiDou(冒险岛GMS服务端北斗)","summaryCn":"开源冒险岛服务端，自己搭建怀旧游戏服务器，和朋友一起重温经典","reasonCn":"这是冒险岛全球服（GMS）的开源服务端项目，用JavaScript编写，支持完整的游戏服务器功能。你可以下载下来自建服务器，和朋友一起重温经典怀旧版本。项目star618，在怀旧游戏服务器领域有一定社区基础，适合喜欢冒险岛的老玩家和技术爱好者尝试搭建私服。","stars":722,"starGrowth":0,"createdAt":"2024-08-03T20:36:18Z","silentRounds":1,"language":"JavaScript","topics":["maplestory","gms083"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0,"domainTags":["冒险岛","游戏服务器","MMORPG","怀旧"],"domainKey":"冒险岛","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"maplestory","source":"github","weight":0.7},{"name":"gms083","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/BeiDouMS/BeiDou-Server","ts":"2026-08-11T09:45:50.026Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"johnlui/Learn-Laravel-5","owner":"johnlui","name":"Learn-Laravel-5","desc":"","summaryCn":"这套教程自2017版以来在中文开发者社区中口碑颇佳，其核心亮点在于将 Larav","reasonCn":"项目本身是一个包含代码示例的资源库，要求开发环境具备 PHP 7.0.0 或更高版本。这意味着它虽然使用的是 Laravel 5 版本，但其教学理念聚焦于框架基础结构的理解，如对路由、控制器、视图引擎以及服务容器的配置。读者下载代码后，需运行 composer update 命令来同步依赖，这一过程","stars":3455,"starGrowth":0,"createdAt":"2015-03-07T07:56:52Z","silentRounds":4,"language":"PHP","topics":["laravel","php"],"aiDims":["教程","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["Laravel教程","PHP入门","代码示例"],"domainKey":"Laravel教程","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"laravel","source":"github","weight":0.7},{"name":"php","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/johnlui/Learn-Laravel-5","ts":"2026-09-18T22:32:16.581Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mubeng/mubeng","owner":"mubeng","name":"mubeng","desc":"An incredibly fast proxy checker &amp; IP rotator with ease.","summaryCn":"批量测代理、自动换IP，爬虫开发者的必备瑞士军刀","reasonCn":"基于Go语言编写的高性能代理检测与IP轮换工具，支持HTTP、HTTPS、SOCKS5等多种代理协议，采用异步并发架构可同时检测数千个代理IP，内置JSON格式输出、批量导入导出和API接口，适合数据采集、爬虫项目和自动化测试场景快速集成使用。","stars":2713,"starGrowth":0,"createdAt":"2021-01-16T00:33:07Z","silentRounds":3,"language":"Go","topics":["proxy","proxy-server","proxy-list","proxy-checker","http-proxy","https-proxy","socks5","socks5-proxy","ip-rate-limit","ip-rotation","proxy-rotator","proxy-rotation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["代理检测","IP轮换","爬虫工具","性能测试","网络工具"],"domainKey":"代理检测","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"proxy","source":"github","weight":0.7},{"name":"proxy-server","source":"github","weight":0.7},{"name":"proxy-list","source":"github","weight":0.7},{"name":"proxy-checker","source":"github","weight":0.7},{"name":"http-proxy","source":"github","weight":0.7},{"name":"https-proxy","source":"github","weight":0.7},{"name":"socks5","source":"github","weight":0.7},{"name":"socks5-proxy","source":"github","weight":0.7},{"name":"ip-rate-limit","source":"github","weight":0.7},{"name":"ip-rotation","source":"github","weight":0.7},{"name":"proxy-rotator","source":"github","weight":0.7},{"name":"proxy-rotation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/mubeng/mubeng","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"RSSNext/Folo","owner":"RSSNext","name":"Folo","desc":"🧡 Folo is the AI RSS Reader","summaryCn":"AI 驱动的 RSS 阅读器，帮你把长文摘要得明明白白。","reasonCn":"它的工作流程非常智能：首先从各种网站和博客拉取 RSS 订阅源，然后通过 AI 算法分析文章内容，提取核心观点和关键信息。你不需要阅读全文，只需要看它的摘要就能知道这篇文章讲了什么。这极大地节省了时间，让你能专注于真正有价值的内容。","stars":39004,"starGrowth":0,"createdAt":"2024-04-08T03:23:35Z","silentRounds":1,"language":"TypeScript","topics":["ai"],"aiDims":["AI搜索","AI应用","信息聚合"],"zone":"AI","funScore":0,"domainTags":["AI摘要","RSS阅读器","信息聚合"],"domainKey":"AI摘要","pushedAt":"2026-09-22T08:43:31Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"信息聚合","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/RSSNext/Folo","ts":"2026-09-22T08:43:31Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Open-Source-Legal/OpenContracts","owner":"Open-Source-Legal","name":"OpenContracts","desc":"The open document intelligence platform for builders and hackers - DMS for the agentic world","summaryCn":"开源智能合同平台，用AI解析文档，给黑客-builder的神器","reasonCn":"这是一个面向开发者的开放文档智能平台，特别适合Agent时代。它支持文档管理系统，能帮你解析和处理各种合同、法律文件。虽然名字带Legal，但本质是一个通用的文档智能基础设施工具。对于想给AI Agent加文档处理能力的开发者来说，这是个很棒的起点，能把向量数据库和文档解析能力结合在一起。","stars":1487,"starGrowth":1,"createdAt":"2022-10-24T04:08:13Z","silentRounds":2,"language":"Python","topics":["vector-db"],"aiDims":["AI应用","文档理解","向量数据库"],"zone":"AI","funScore":0,"domainTags":["文档智能","RAG","Agent"],"domainKey":"RAG","pushedAt":"2026-09-23T16:33:16Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/Open-Source-Legal/OpenContracts","ts":"2026-09-23T16:33:16Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"react/create-react-app","owner":"react","name":"create-react-app","desc":"Set up a modern web app by running one command.","summaryCn":"一键生成React项目模板，小白也能秒上手的建站神器","reasonCn":"Create React App 是由 Facebook 官方维护的 React 项目脚手架工具，零配置开箱即用，内置 Babel、Webpack 等核心依赖，一条命令即可创建完整可运行的前端项目。支持 TypeScript、CSS Modules、热更新，社区超过十万星，是 React 初学者入门的首选工具。","stars":103254,"starGrowth":0,"createdAt":"2016-07-17T14:55:11Z","silentRounds":1,"language":"JavaScript","topics":["react","zero-configuration","build-tools"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["React","脚手架","前端开发","Node.js","构建工具"],"domainKey":"React","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"react","source":"github","weight":0.7},{"name":"zero-configuration","source":"github","weight":0.7},{"name":"build-tools","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/react/create-react-app","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"a-peirogon/cal-annas","owner":"a-peirogon","name":"cal-annas","desc":"Calibre store plugin for Anna's Archive/Libgen, world's largest shadow library. Real-time uptime data via SLUM.","summaryCn":"对于经常使用 Calibre 管理电子书的人来说，寻找可靠的大型书源往往耗时又容","reasonCn":"对于经常使用 Calibre 管理电子书的人来说，寻找可靠的大型书源往往耗时又容易遇到失效链接。cal-annas 就是为此而生的插件，它把 Calibre 的书源搜索与 Anna's Archive/Libgen 的海量图书数据库结合，并通过 SLUM（Status of Library Usability Monitor）实时上报书源可用性。","stars":39,"starGrowth":0,"createdAt":"2026-02-02T14:14:05Z","silentRounds":11,"language":"Python","topics":["ebook"],"aiDims":["非AI-实用"],"zone":"资源","funScore":0.7,"domainTags":["电子书","Calibre插件","资源获取"],"domainKey":"电子书","pushedAt":"2026-09-07T23:55:28Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.39,"source":"search","url":"https://github.com/a-peirogon/cal-annas","ts":"2026-09-07T23:55:28Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"neondatabase/neon","owner":"neondatabase","name":"neon","desc":"Neon: Serverless Postgres. We separated storage and compute to offer autoscaling, code-like database branching, and scale to zero.","summaryCn":"把数据库变成代码一样的分支管理，用多少付多少的神器","reasonCn":"Neon把传统Postgres的存储和计算彻底拆分，让你能像操作Git代码一样对数据库进行分支、合并和回滚。它支持自动扩缩容到零，不用的时候自动停机，按实际使用量计费。这种架构对AI应用特别友好——每个用户、每个项目都能拥有独立隔离的数据库实例，开发环境也能随意fork出来独立测试，再也不用担心污染共享数据库。","stars":23123,"starGrowth":2,"createdAt":"2021-03-26T14:20:47Z","silentRounds":0,"language":"Rust","topics":["database"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["PostgreSQL","Serverless","数据库"],"domainKey":"PostgreSQL","pushedAt":"2026-08-31T11:03:14Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/neondatabase/neon","ts":"2026-08-31T11:03:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jeecgboot/jimureport","owner":"jeecgboot","name":"jimureport","desc":"一款真正的AI报表诞生了：一句话生成报表和数据大屏，支持对话式AI修改报表；类 Excel 在线报表设计，兼容国产信创 数据源。 对着AI说句话，报表就出来了——表格、图表、分组汇总全部自动配好，轻松应对各类复杂报表场景。想做大屏？同样一句话搞定。","summaryCn":"对着AI说句话，报表大屏自动生成，告别繁琐配置","reasonCn":"这是一个基于JeecgBoot的低代码报表引擎，核心亮点是引入了AI对话能力。用户只需自然语言描述需求，系统就能自动分析意图，智能匹配数据源并生成复杂的表格或可视化大屏。它兼容国产信创环境，提供类Excel的在线设计器，支持多维度分组汇总和多种图表类型。对于开发人员，它极大降低了报表开发门槛；对于业务人员，它让数据分析不再依赖技术团队，真正实现业务自助化，显著提升工作效率。","stars":8499,"starGrowth":2,"createdAt":"2019-12-05T12:25:59Z","silentRounds":0,"language":"Java","topics":["data-visualization"],"aiDims":["AI应用","可视化","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI报表","数据可视化","低代码","大模型应用","企业BI"],"domainKey":"AI报表","pushedAt":"2026-09-23T14:34:20Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/jeecgboot/jimureport","ts":"2026-09-23T14:34:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/BotFramework-WebChat","owner":"microsoft","name":"BotFramework-WebChat","desc":"A highly-customizable web-based client for Azure Bot Services.","summaryCn":"微软官方的网页聊天机器人客户端，可高度定制","reasonCn":"从技术实现来看，它采用了React框架开发，这意味着它拥有现代化的组件架构和极高的灵活性。最强大的地方在于它的“高度可定制性”，你可以通过修改React组件的props，把原本普通的对话框变成完全符合你品牌调性的界面，甚至可以集成到任何现有的Web应用中。","stars":1783,"starGrowth":0,"createdAt":"2016-07-07T23:16:57Z","silentRounds":13,"language":"HTML","topics":["chatbot"],"aiDims":["AI应用","工具调用","开发工具"],"zone":"工具","funScore":0,"domainTags":["聊天机器人","Web组件","React","对话界面","微软生态"],"domainKey":"聊天机器人","pushedAt":"2026-09-23T10:04:32Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/microsoft/BotFramework-WebChat","ts":"2026-09-23T10:04:32Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"tw-in-js/twind","owner":"tw-in-js","name":"twind","desc":"","summaryCn":"Twind的核心逻辑是将Tailwind极其丰富的原子类系统，直接带入到了运行时环境","reasonCn":"这一工具的亮点在于其“按需编译”的机制，它能根据实际使用的类名，在运行时动态生成所对应的CSS，从而保证最终生成的样式表不仅精简，而且完全符合业务需求。如果你是Tailwind的忠实用户，但又厌倦了复杂的PostCSS配置，Twind会让你感到豁然开朗。","stars":3948,"starGrowth":0,"createdAt":"2020-12-05T17:59:21Z","silentRounds":3,"language":"JavaScript","topics":["tailwindcss","tailwind-in-js","css-in-js","tailwind","css"],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["CSS框架","TailwindCSS","运行时编译"],"domainKey":"CSS框架","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"tailwindcss","source":"github","weight":0.7},{"name":"tailwind-in-js","source":"github","weight":0.7},{"name":"css-in-js","source":"github","weight":0.7},{"name":"tailwind","source":"github","weight":0.7},{"name":"css","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/tw-in-js/twind","ts":"2026-09-18T20:05:54.376Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"persiyanov/herdr-reviewr","owner":"persiyanov","name":"herdr-reviewr","desc":"A code review + file viewer sidebar for herdr. Comment on a diff and send back to agent. Inspect diffs, files, and a PR state.","summaryCn":"herdr侧边打开代码审查窗口，直接在浏览器写评注，省去切换工具。","reasonCn":"herdr‑reviewr 是为 herdr 开发的侧栏插件，提供文件浏览、diff 展示与即时评论功能，支持将评注发送回后端代理，帮助团队在同一界面完成代码审查、查看文件变更和 PR 状态，无需跳转至外部 IDE 或审查平台。插件用 Rust 编写，直接嵌入 herdr，支持多语言高亮和自定义快捷键。","stars":753,"starGrowth":4,"createdAt":"2026-06-26T12:50:56Z","silentRounds":0,"language":"Rust","topics":["code-review","herdr","herdr-plugin","rust","tui"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["代码审查","IDE插件","Rust","开发工具"],"domainKey":"代码审查","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"code-review","source":"github","weight":0.7},{"name":"herdr","source":"github","weight":0.7},{"name":"herdr-plugin","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"tui","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/persiyanov/herdr-reviewr","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yo35/rpb-chessboard","owner":"yo35","name":"rpb-chessboard","desc":"A WordPress plugin to display chess games and diagrams (using the standard FEN and PGN notations).","summaryCn":"想象一下，你是一个热爱国际象棋的博主，想要在文章里分享一局精彩的对局，","reasonCn":"技术上，它通过前端脚本解析标准的PGN（Portable Game Notation）和FEN（Forsyth-Edwards Notation）格式数据，并利用轻量级的CSS/JS逻辑在浏览器端进行渲染。相比于截图或使用外部图片服务，这种方式不仅加载速度快，而且棋盘是矢量渲染的，在不同屏幕尺寸下都不会失真，同时支持","stars":38,"starGrowth":0,"createdAt":"2013-10-26T11:26:47Z","silentRounds":15,"language":"PHP","topics":["chess"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0,"domainTags":["国际象棋","WordPress插件","棋谱展示","FEN解析","互动棋盘"],"domainKey":"国际象棋","pushedAt":"2026-09-05T09:04:01Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/yo35/rpb-chessboard","ts":"2026-09-05T09:04:01Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"userfrosting/UserFrosting","owner":"userfrosting","name":"UserFrosting","desc":"[READ ONLY] Modern PHP user login and management framework","summaryCn":"老牌PHP安全登录框架，虽然停更但代码规范堪称教科书","reasonCn":"这是一个非常老牌的PHP用户认证与管理系统框架，拥有1600多星。专注于用户注册、登录、权限管理等核心功能，代码结构严谨，遵循现代PHP标准。虽然项目已转为只读模式，但作为学习PHP安全编码和构建用户系统的参考资料，它依然具有很高的实用价值和历史意义。","stars":1667,"starGrowth":0,"createdAt":"2014-03-26T18:05:59Z","silentRounds":24,"language":"PHP","topics":["web-framework"],"aiDims":["非AI-好玩","游戏","安全工具"],"zone":"工具","funScore":0,"domainTags":["PHP框架","用户认证","ACL","安全"],"domainKey":"PHP框架","pushedAt":"2026-09-19T02:43:41Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/userfrosting/UserFrosting","ts":"2026-09-19T02:43:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"larksuite/cli","owner":"larksuite","name":"cli","desc":"The official Lark/飞书 CLI tool, maintained by the larksuite team — built for humans and AI Agents. Covers core business domains including Messenger, Docs, Base, Sheets, Calendar, Mail, Tasks, Meetings, and more, with 200+ commands and 20+ AI Agent Skills.","summaryCn":"飞书官方命令行工具，AI Agent操作飞书的利器","reasonCn":"这个项目是飞书官方团队维护的CLI工具，设计目标是让AI Agent和开发者能方便地调用飞书的API能力。它覆盖了飞书的核心业务域，包括即时通讯、文档、多维表格、电子表格、日历、邮件、任务、会议等功能。通过一个统一的命令行接口，你可以用自然语言描述操作，或者用脚本自动化执行","stars":17402,"starGrowth":18,"createdAt":"2026-03-25T07:13:30Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI Agent","效率工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["飞书","CLI","办公自动化","企业协作"],"domainKey":"飞书","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/larksuite/cli","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.06852370991840018,"copyOk":false},{"repo":"nextcloud/nextcloudpi","owner":"nextcloud","name":"nextcloudpi","desc":"📦 Build code for NextcloudPi: Raspberry Pi, Odroid, Rock64, curl installer...","summaryCn":"在树莓派上“一键”装 Nextcloud，省心省力的私有云轻松","reasonCn":"NextcloudPi 将 Nextcloud 与 Raspbian/Debian 系统打包，提供一键安装脚本和自动更新。它支持多种单板机（Raspberry Pi、Odroid、Rock64 等），并可通过 curl 安装到任何 Linux。无需手动配置数据库或 Web 服务器，直接即可使用 Web 界面管理文件、联系人和日历，支持同步和加密。它还内置了自动更新工具，可按需检查并安装安全补丁，防止被攻击；集成了 Let’s Encrypt SSL，可轻松启用 HTTPS；支持多用户、共享文件夹、加密云盘，还可以通过 WebDAV 或 Sync Client 进行同步。安装后只需访问 192.168.1.100:80，即可登录管理面板。","stars":2940,"starGrowth":0,"createdAt":"2017-02-12T19:11:16Z","silentRounds":9,"language":"Shell","topics":["raspberry-pi"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["私有云","树莓派","文件同步","自托管"],"domainKey":"私有云","pushedAt":"2026-09-07T20:54:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/nextcloud/nextcloudpi","ts":"2026-09-07T20:54:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KitionAI/kition","owner":"KitionAI","name":"kition","desc":"Kition brings Markdown, DataTable, WhiteBoard, a tool-using AI agent, browser research, and visual workflows into one desktop workspace. ","summaryCn":"想象一下你正在进行一项复杂的项目调研，平时你需要打开浏览器搜集资料，打","reasonCn":"技术架构上，它采用了现代化的 TypeScript 技术栈，通过轻量化的桌面容器将多种功能组件解耦，但又通过统一的状态管理保证了数据的一致性。最亮眼的地方在于它的工具链调用机制（Tool-using AI Agent），能够根据用户的指令动态调用浏览器接口和数据处理引擎，这种交互方式远比传统","stars":338,"starGrowth":3,"createdAt":"2026-07-25T04:41:35Z","silentRounds":0,"language":"TypeScript","topics":["note-taking"],"aiDims":["AI Agent","效率工具","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["知识管理","AI Agent","桌面应用","浏览器自动化","Markdown笔记"],"domainKey":"AI Agent","pushedAt":"2026-09-21T23:30:23Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/KitionAI/kition","ts":"2026-09-21T23:30:23Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"github/gh-aw","owner":"github","name":"gh-aw","desc":"GitHub Agentic Workflows","summaryCn":"让AI自动帮你搞定GitHub上那些重复无聊的操作","reasonCn":"基于Go语言打造的GitHub AI工作流引擎，专为开发者日常操作深度优化。支持多步骤任务编排，自动调用GitHub API完成issue管理、PR审查等操作。接入大模型后无需手动配置即可自动执行工作流，大幅减少人工干预，提升团队开发效率。","stars":5171,"starGrowth":6,"silentRounds":0,"language":"Go","topics":["gh-extension","actions","cai","ci","claude-code","codex","copilot","github-actions"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["GitHub自动化","AI工作流","Agent","API集成","代码审查"],"domainKey":"Agent","pushedAt":"2026-09-23T07:35:20.006Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"gh-extension","source":"github","weight":0.7},{"name":"actions","source":"github","weight":0.7},{"name":"cai","source":"github","weight":0.7},{"name":"ci","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"github-actions","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/github/gh-aw","ts":"2026-09-23T07:35:20.006Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02148095146006575,"copyOk":false},{"repo":"HariSekhon/DevOps-Bash-tools","owner":"HariSekhon","name":"DevOps-Bash-tools","desc":"1200+ DevOps Bash Scripts - AWS, GCP, Kubernetes, Docker, CI/CD, APIs, SQL, PostgreSQL, MySQL, Hive, Impala, Kafka, Hadoop, Jenkins, GitHub, GitLab, BitBucket, Azure DevOps, TeamCity, Spotify, MP3, LDAP, Code/Build Linting, pkg mgmt for Linux, Mac, Python, Perl, Ruby, NodeJS, Golang, Advanced dotfiles: .bashrc, .vimrc, .gitconfig, .screenrc, tmux..","summaryCn":"海量DevOps脚本库，一键解决云、容器、数据库等运维难题。","reasonCn":"这是一个包含1200多个Bash脚本的集合，涵盖AWS、GCP、Kubernetes、Docker、CI/CD、APIs、SQL、PostgreSQL、MySQL、Kafka、Hadoop、Jenkins、GitHub等数十种工具和平台。每个脚本都经过实战验证，可直接复制使用，大幅减少重复造轮子的时间。适合DevOps工程师、云运维和自动化测试开发者。","stars":8415,"starGrowth":1,"createdAt":"2016-01-17T12:44:26Z","silentRounds":2,"language":"Shell","topics":["devops","git"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["DevOps","Bash脚本","云运维","Kubernetes","CI/CD"],"domainKey":"DevOps","pushedAt":"2026-09-15T17:40:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/HariSekhon/DevOps-Bash-tools","ts":"2026-09-15T17:40:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kyrolabs/awesome-langchain","owner":"kyrolabs","name":"awesome-langchain","desc":"😎 Awesome list of tools and projects with the awesome LangChain framework","summaryCn":"如果你正准备利用AI技术开发自己的应用程序，但面对LangChain浩如烟海的文档","reasonCn":"这个项目的核心亮点在于其极高的更新频率和覆盖度。它不仅仅是一个简单的链接清单，而是通过科学的分类方式，将分散在GitHub、博客、论文中的优质AI资源聚合起来。它涵盖了向量检索、模型微调、提示词工程等所有热门板块，是很多核心AI基础设施项目的摇篮，对于理解现代LLM应用栈的","stars":9544,"starGrowth":1,"createdAt":"2023-02-28T20:25:33Z","silentRounds":1,"language":"","topics":["awesome"],"aiDims":["学习资源","AI基础设施","RAG与知识"],"zone":"资源","funScore":0,"domainTags":["LangChain","Awesome列表","AI资源导航","开发者工具","提示工程"],"domainKey":"LangChain","pushedAt":"2026-09-22T03:04:20Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.88,"source":"search","url":"https://github.com/kyrolabs/awesome-langchain","ts":"2026-09-22T03:04:20Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kurikomi-labs/komi-store","owner":"kurikomi-labs","name":"komi-store","desc":"🩵 A free, open-source app store for developers' releases on GitHub, Codeberg & Forgejo — browse, discover, and install apps with one click. Formerly GitHub Store.","summaryCn":"GitHub开源应用的百宝箱，一键发现安装各种神器","reasonCn":"这是一个专门收集GitHub、Codeberg、Forgejo等平台上开源应用的桌面应用商店。不像传统软件分发平台，它直接从代码托管平台抓取开源发布版本，帮你发现各种开发者工具、实用软件和创意项目。支持Kotlin编写，跨平台运行，界面简洁现代，让搜索和安装开源软件变得像刷App Store一样简单，无需手动下载压缩包。","stars":18690,"starGrowth":13,"createdAt":"2025-11-21T13:03:23Z","silentRounds":0,"language":"Kotlin","topics":["desktop"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["应用商店","开源分发","桌面软件"],"domainKey":"应用商店","pushedAt":"2026-07-29T17:44:50Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/kurikomi-labs/komi-store","ts":"2026-07-29T17:44:50Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04930401284109149,"copyOk":true},{"repo":"quickjs-ng/quickjs","owner":"quickjs-ng","name":"quickjs","desc":"QuickJS, the Next Generation: a mighty JavaScript engine","summaryCn":"用C写的神级JS引擎，轻量快速，嵌入式应用必备利器","reasonCn":"这是用C语言编写的轻量级JavaScript引擎，专为嵌入式场景优化。相比V8引擎，它启动更快、内存占用更小，可轻松嵌入到各种应用程序中。支持标准ES2023语法，兼容Node.js模块系统，还内置了WebAssembly支持。项目持续活跃迭代，适合需要内置脚本解释器的开发者使用。","stars":3780,"starGrowth":4,"createdAt":"2023-11-01T18:49:43Z","silentRounds":0,"language":"C","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript引擎","嵌入式","C语言"],"domainKey":"JavaScript引擎","pushedAt":"2026-08-16T20:55:12Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/quickjs-ng/quickjs","ts":"2026-08-16T20:55:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MaterialDesignInXAML/MaterialDesignInXamlToolkit","owner":"MaterialDesignInXAML","name":"MaterialDesignInXamlToolkit","desc":"Google's Material Design in XAML & WPF, for C# & VB.Net. ","summaryCn":"C#开发者秒变UI设计师，一键套用Google风界面模板","reasonCn":"专为WPF和XAML打造的Material Design组件库，内置上百个现成控件，包括按钮、卡片、表单、导航等，全部遵循Google Material Design规范。支持深色与浅色主题自由切换，可通过NuGet一键安装，无需手写大量XAML样式代码，即可快速构建现代化Windows桌面应用界面。","stars":16259,"starGrowth":0,"createdAt":"2015-02-07T23:30:48Z","silentRounds":5,"language":"C#","topics":["design"],"aiDims":["可视化","非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WPF","XAML","UI组件","Material Design",".NET"],"domainKey":"WPF","pushedAt":"2026-09-11T04:53:14Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/MaterialDesignInXAML/MaterialDesignInXamlToolkit","ts":"2026-09-11T04:53:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mozilla-ai/document-to-podcast","owner":"mozilla-ai","name":"document-to-podcast","desc":"Blueprint by Mozilla.ai for generating podcasts from documents using local AI","summaryCn":"Document to Podcast是一款功能强大的转换工具，可以将各种类型的文本内容转","reasonCn":"Document to Podcast是一款功能强大的转换工具，可以将各种类型的文本内容转换为可听的播客。用户只需上传文本文件，软件就会自动处理并生成相应的音频文件。这款工具支持多种格式的文本文件，且转换后的音频文件音质清晰，能够满足用户对于播客内容的需求。","stars":183,"starGrowth":0,"createdAt":"2024-11-14T11:33:38Z","silentRounds":19,"language":"Python","topics":["podcast"],"aiDims":["非AI-好玩"],"zone":"AI","funScore":0,"domainTags":["播客生成","本地AI","文本转音频"],"domainKey":"播客生成","pushedAt":"2026-09-21T17:03:40Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/mozilla-ai/document-to-podcast","ts":"2026-09-21T17:03:40Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"atuinsh/atuin","owner":"atuinsh","name":"atuin","desc":"✨ Making your shell magical","summaryCn":"给命令行装上记忆大脑，搜索历史命令像谷歌一样快","reasonCn":"这是一个用Rust编写的命令行历史记录管理工具，它能自动同步你的Shell历史记录到云端数据库，并支持强大的全文检索和模糊搜索功能。它解决了传统Shell历史记录分散、无法跨设备同步、搜索困难等痛点，通过加密同步和智能压缩技术，让你在任何终端都能快速找回过去敲过的任何命令。","stars":31786,"starGrowth":6,"createdAt":"2020-10-04T23:01:58Z","silentRounds":0,"language":"Rust","topics":["shell","rust","zsh","history","fish","bash"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["Shell","命令历史","同步工具","终端增强"],"domainKey":"Shell","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"shell","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"zsh","source":"github","weight":0.7},{"name":"history","source":"github","weight":0.7},{"name":"fish","source":"github","weight":0.7},{"name":"bash","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/atuinsh/atuin","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.018566287044610833,"copyOk":false},{"repo":"libsdl-org/SDL","owner":"libsdl-org","name":"SDL","desc":"Simple DirectMedia Layer","summaryCn":"跨平台游戏开发神器，Windows Mac手机都能跑","reasonCn":"SDL是老牌跨平台多媒体开发库，支持Windows、Mac、Linux、iOS、Android等平台。提供音频、视频、输入、网络等基础功能，是许多著名游戏的底层支撑。用C语言编写，性能优秀，社区成熟稳定，适合想做游戏或创意项目的开发者。","stars":16647,"starGrowth":4,"createdAt":"2021-01-15T19:55:54Z","silentRounds":0,"language":"C","topics":["sdl","sdl2","sdl3"],"aiDims":["游戏","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["跨平台多媒体","游戏开发","C语言"],"domainKey":"跨平台多媒体","pushedAt":"2026-08-19T01:02:53Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"sdl","source":"github","weight":0.7},{"name":"sdl2","source":"github","weight":0.7},{"name":"sdl3","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/libsdl-org/SDL","ts":"2026-08-19T01:02:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"popcorntime/popcorntime","owner":"popcorntime","name":"popcorntime","desc":"Popcorn Time™ puts everything in one place. Your favorite platforms, your shows, your movies-ready when you are.","summaryCn":"一个应用聚合所有流媒体资源的经典视频神器","reasonCn":"基于Electron开发的桌面应用，把电影、剧集和流媒体平台整合到一个界面里。用户无需切换不同APP，一个搜索就能跨平台观看。技术栈TypeScript，支持多平台。因版权原因已停止维护，GitHub仓库被删，但仍有粉丝镜像版本可用。","stars":10685,"starGrowth":0,"createdAt":"2016-02-16T11:07:20Z","silentRounds":4,"language":"TypeScript","topics":["desktop"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["流媒体","P2P","视频播放","Electron"],"domainKey":"流媒体","pushedAt":"2026-09-04T03:18:38Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/popcorntime/popcorntime","ts":"2026-09-04T03:18:38Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deven96/ahnlich","owner":"deven96","name":"ahnlich","desc":"Suite of tools containing an in-memory vector datastore and AI proxy","summaryCn":"用Rust写的超快内存向量数据库加AI代理，专为极速检索设计。","reasonCn":"技术上，它完全用Rust编写，充分利用了Rust在内存安全和并发性能上的绝对优势。由于采用纯内存架构，它的读写延迟达到了微秒级别，远超传统基于磁盘或臃肿架构的数据库。同时它内置了AI代理模块，能够很好地作为大模型与后端数据存储之间的桥梁，简化了AI应用的架构复杂度。","stars":245,"starGrowth":0,"createdAt":"2024-05-01T05:15:29Z","silentRounds":16,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["向量数据库","Rust","内存检索","AI代理"],"domainKey":"向量数据库","pushedAt":"2026-09-22T11:49:33Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/deven96/ahnlich","ts":"2026-09-22T11:49:33Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"pimalaya/himalaya","owner":"pimalaya","name":"himalaya","desc":"CLI to manage emails","summaryCn":"在终端里收发邮件的神器，让你告别网页版和笨重客户端","reasonCn":"这是一个用Rust编写的开源命令行邮件客户端，支持通过IMAP/SMTP协议收发和管理邮件。你可以在终端里直接写邮件、查看收件箱、搜索历史邮件、批量删除垃圾邮件，界面清爽现代，配置灵活，支持Gmail、Outlook等多种邮箱。","stars":7304,"starGrowth":5,"createdAt":"2020-12-23T23:56:25Z","silentRounds":0,"language":"Rust","topics":["cli","rust","email","himalaya","imap","gpg","notmuch","pgp","smtp","pimalaya","sendmail","client","terminal"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行客户端","邮件管理","IMAP协议","SMTP协议","终端UI"],"domainKey":"命令行客户端","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"cli","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"email","source":"github","weight":0.7},{"name":"himalaya","source":"github","weight":0.7},{"name":"imap","source":"github","weight":0.7},{"name":"gpg","source":"github","weight":0.7},{"name":"notmuch","source":"github","weight":0.7},{"name":"pgp","source":"github","weight":0.7},{"name":"smtp","source":"github","weight":0.7},{"name":"pimalaya","source":"github","weight":0.7},{"name":"sendmail","source":"github","weight":0.7},{"name":"client","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/pimalaya/himalaya","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01705366922234392,"copyOk":false},{"repo":"cloudflare/agents","owner":"cloudflare","name":"agents","desc":"Build and deploy AI Agents on Cloudflare ","summaryCn":"你以前写个AI聊天机器人，还得自己维护云服务器，处理并发压力，配置负载均","reasonCn":"架构上，它采用了事件驱动的模式，能够自动在全球最靠近用户的数据中心触发AI处理逻辑。它最大的技术亮点在于和Cloudflare Workers AI的深度绑定，你不需要关心模型背后的GPU集群在哪，直接调用API就能使用全球各种顶尖大模型。","stars":5646,"starGrowth":16,"createdAt":"2025-01-29T23:14:04Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","AI基础设施","模型部署"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","边缘计算","Workers AI","TypeScript"],"domainKey":"AI Agent","pushedAt":"2026-09-19T22:52:04Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/cloudflare/agents","ts":"2026-09-19T22:52:04Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.05650159404888416,"copyOk":false},{"repo":"apache/shardingsphere-elasticjob","owner":"apache","name":"shardingsphere-elasticjob","desc":"Distributed scheduled job","summaryCn":"基于Java的分布式任务调度框架，帮你轻松管理定时作业。","reasonCn":"ElasticJob 是 Apache ShardingSphere 生态的分布式调度子项目，基于 Java 实现，提供弹性伸缩的定时任务执行、错过任务补偿、分片并行和分布式锁等特性，支持与 Spring、Spring Boot、Dubbo 等主流框架无缝集成，适用于微服务或大数据场景的批量作业管理。","stars":8205,"starGrowth":0,"createdAt":"2015-09-11T11:28:52Z","silentRounds":9,"language":"Java","topics":["database"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["分布式调度","定时任务","Java"],"domainKey":"分布式调度","pushedAt":"2026-09-11T01:49:24Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/apache/shardingsphere-elasticjob","ts":"2026-09-11T01:49:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chartist-js/chartist","owner":"chartist-js","name":"chartist","desc":"","summaryCn":"基于 SVG 的轻量响应式图表库，主打设计师友好与极简配置。","reasonCn":"Chartist 是一个用 SVG 技术构建的简单且响应式的前端图表库。它避开了体积庞大、技术选型争议大以及配置复杂的传统图表库陷阱，强调在设计自由度与代码简洁性之间取得平衡。通过直观的 API 设计，它让开发者能够轻松集成各种数据可视化图表，同时确保最终呈现效果贴合设计稿，适合追求轻量与灵活性的 Web 应用开发场景。","stars":13390,"starGrowth":0,"createdAt":"2014-03-09T22:40:07Z","silentRounds":8,"language":"TypeScript","topics":["chartist","responsive-charts","charting-library","javascript","svg"],"aiDims":["开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["数据可视化","SVG","前端图表","响应式设计"],"domainKey":"数据可视化","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"chartist","source":"github","weight":0.7},{"name":"responsive-charts","source":"github","weight":0.7},{"name":"charting-library","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"svg","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"search","url":"https://github.com/chartist-js/chartist","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"spf13/cobra","owner":"spf13","name":"cobra","desc":"A Commander for modern Go CLI interactions","summaryCn":"Go语言写命令行工具的神级框架，K8s和GitHub CLI都在用它","reasonCn":"Cobra是Go生态中最流行的命令行框架，专为构建功能完整的CLI应用而生。它原生支持子命令、标志参数、命令别名、自动帮助生成、配置文件读取、自动补全等功能。设计哲学强调清晰的结构和可扩展性，开发者只需定义命令树，框架自动处理参数解析、错误提示和文档输出。被Kubernetes、GitHub CLI、Hugo等顶级项目采用，是Go CLI开发的事实标准。","stars":44631,"starGrowth":4,"createdAt":"2013-09-03T20:40:26Z","silentRounds":0,"language":"Go","topics":["cobra","cobra-library","cobra-generator","posix-compliant-flags","command-cobra","cli-app","command-line","commandline","command","cli","go","golang","golang-library","golang-application","subcommands","posix"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Go语言","CLI框架","命令行","参数解析","开发工具"],"domainKey":"Go语言","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"cobra","source":"github","weight":0.7},{"name":"cobra-library","source":"github","weight":0.7},{"name":"cobra-generator","source":"github","weight":0.7},{"name":"posix-compliant-flags","source":"github","weight":0.7},{"name":"command-cobra","source":"github","weight":0.7},{"name":"cli-app","source":"github","weight":0.7},{"name":"command-line","source":"github","weight":0.7},{"name":"commandline","source":"github","weight":0.7},{"name":"command","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"golang-library","source":"github","weight":0.7},{"name":"golang-application","source":"github","weight":0.7},{"name":"subcommands","source":"github","weight":0.7},{"name":"posix","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/spf13/cobra","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Azgaar/Fantasy-Map-Generator","owner":"Azgaar","name":"Fantasy-Map-Generator","desc":"Web application generating interactive and highly customizable maps","summaryCn":"在线生成神级奇幻地图，脑洞大开的地理模拟器。","reasonCn":"核心技术在于程序化生成算法，它使用 Perlin Noise（柏林噪声）和 Diamond-Square 算法来模拟自然界的随机性。它不仅仅是简单的噪点叠加，而是通过多层噪声叠加来模拟山脉的起伏、河流的流向以及植被的分布。它内置了一套完善的地图符号库，能自动给地图添加山脉符号、河流方向、城","stars":6017,"starGrowth":0,"createdAt":"2017-03-16T20:59:50Z","silentRounds":2,"language":"HTML","topics":["procedural-generation"],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["程序化生成","地图生成","奇幻","地理模拟"],"domainKey":"程序化生成","pushedAt":"2026-09-23T21:37:05Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/Azgaar/Fantasy-Map-Generator","ts":"2026-09-23T21:37:05Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"siderolabs/talos","owner":"siderolabs","name":"talos","desc":"Talos Linux is a modern Linux distribution built for Kubernetes.","summaryCn":"传统的 Linux 发行版（如 Ubuntu、CentOS）功能太臃肿，装个 K8s 集群需要配","reasonCn":"它的技术架构非常独特，采用了分片设计。整个系统被拆分成很多微小的分片，每个分片只负责一个特定的功能（比如网络、存储、日志）。当你安装时，只需要下载几个小的镜像，然后告诉它你想用哪些分片，剩下的就自动完成了。","stars":11219,"starGrowth":4,"createdAt":"2017-11-03T23:19:12Z","silentRounds":0,"language":"Go","topics":["kubernetes"],"aiDims":["AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","Linux发行版","容器编排"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T19:09:30Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/siderolabs/talos","ts":"2026-09-23T19:09:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EFForg/rayhunter","owner":"EFForg","name":"rayhunter","desc":"Rust tool to detect cell site simulators on an orbic mobile hotspot","summaryCn":"专治各种伪基站和钓鱼WiFi，守护你的通信安全","reasonCn":"rayhunter是一款用Rust编写的安全检测工具，专门用于识别和检测移动网络中的基站模拟器。伪基站是攻击者在公共场所搭建的虚假信号塔，可以拦截你的手机通信、窃取个人信息甚至定位你的位置。这个工具能实时监控你手机连接到的基站信息，一旦发现可疑的异常基站信号就会发出警告，帮你及时发现潜在的网络攻击。","stars":5873,"starGrowth":3,"createdAt":"2023-11-08T23:25:16Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["网络安全","伪基站检测","Rust工具"],"domainKey":"网络安全","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/EFForg/rayhunter","ts":"2026-08-17T00:55:40.962Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Facepunch/Facepunch.Steamworks","owner":"Facepunch","name":"Facepunch.Steamworks","desc":"Another fucking c# Steamworks implementation","summaryCn":"给C#游戏开发者写的Steam API神器，写游戏必备。","reasonCn":"在技术实现上，它采用了底层原生库调用和C#互操作相结合的方式，保证了性能的同时，提供了非常符合C#语言习惯的调用接口。它支持异步操作，避免了阻塞主线程，这对于游戏开发来说非常重要。它还处理了Steam的验证、登录、心跳等复杂的后台逻辑，你只需要关注业务代码就行。","stars":3764,"starGrowth":0,"createdAt":"2016-07-06T16:38:16Z","silentRounds":3,"language":"C#","topics":["game"],"aiDims":["游戏","非AI-好玩","开发工具"],"zone":"工具","funScore":0,"domainTags":["Steam","游戏开发","C#","SDK","游戏库"],"domainKey":"Steam","pushedAt":"2026-09-15T16:48:11Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/Facepunch/Facepunch.Steamworks","ts":"2026-09-15T16:48:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"golang-migrate/migrate","owner":"golang-migrate","name":"migrate","desc":"Database migrations. CLI and Golang library.","summaryCn":"数据库版本管理利器，Go语言写的迁移工具，帮你搞定表结构变更","reasonCn":"golang-migrate 是一款数据库迁移管理工具，支持 SQL 和 Go 两种定义方式。提供命令行接口，可轻松管理数据库版本的向前和向后迁移。兼容 PostgreSQL、MySQL、SQLite、MongoDB 等多种数据库。在 CI/CD 流程中非常实用，帮助开发团队安全地同步数据库结构变更，避免手动执行 SQL 脚本带来的风险。","stars":18935,"starGrowth":0,"createdAt":"2018-01-19T09:30:58Z","silentRounds":1,"language":"Go","topics":["database"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["数据库迁移","Go语言","CI/CD","版本管理"],"domainKey":"数据库迁移","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/golang-migrate/migrate","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/agents-cli","owner":"google","name":"agents-cli","desc":"The CLI and skills that turn any coding assistant into an expert at creating, evaluating, and deploying AI agents on Google Cloud.","summaryCn":"如果你在开发AI Agent时被各种复杂的云配置、接口适配和反复的测试流程搞得","reasonCn":"技术上，它完美兼容Google Cloud生态，支持智能体定义、环境配置以及自动化评估管线。它最大的亮点在于把原本分散的开发环节（如本地调试、云端测试、生产环境发布）通过一套CLI命令串联起来。这种规范化的开发工作流，确保了无论你是开发简单的单节点助手，还是复杂的分布式智能体","stars":5980,"starGrowth":1,"createdAt":"2026-04-08T17:09:55Z","silentRounds":1,"language":"Python","topics":["agents"],"aiDims":["AI Agent","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["Agent开发","CLI工具","Google Cloud","评估管线","部署工具"],"domainKey":"Agent开发","pushedAt":"2026-09-22T12:15:00Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/google/agents-cli","ts":"2026-09-22T12:15:00Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"triton-inference-server/server","owner":"triton-inference-server","name":"server","desc":"The Triton Inference Server provides an optimized cloud and edge inferencing solution. ","summaryCn":"模型训练完只是第一步，真正难的是让别人稳定调用它。","reasonCn":"技术上看，它的强项是工程化。它不把某个模型框架绑死，而是提供一套推理服务架构，让你把模型放进服务里，再暴露给业务调用。动态批处理、负载均衡、多后端支持、指标采集等能力，都是为了在高并发生产环境里跑得更稳。","stars":11002,"starGrowth":0,"createdAt":"2018-10-04T21:10:30Z","silentRounds":2,"language":"Python","topics":["ml"],"aiDims":["推理引擎","模型部署","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["推理引擎","模型部署","深度学习"],"domainKey":"推理引擎","pushedAt":"2026-09-23T17:05:39Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.79,"source":"search","url":"https://github.com/triton-inference-server/server","ts":"2026-09-23T17:05:39Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"github/git-sizer","owner":"github","name":"git-sizer","desc":"Compute various size metrics for a Git repository, flagging those that might cause problems","summaryCn":"给臃肿的代码仓库做CT，揪出偷吃硬盘空间的元凶","reasonCn":"用Go语言编写的Git仓库分析工具，通过解析历史提交计算文件体积及增长趋势。自动标记二进制文件大对象、未压缩资源及冗余依赖，生成美观HTML报告。支持GitHub API集成，帮助团队定位仓库膨胀原因，优化CI构建速度与克隆效率，是精简代码库的利器。","stars":4076,"starGrowth":0,"createdAt":"2018-01-28T04:38:06Z","silentRounds":8,"language":"Go","topics":["git"],"aiDims":["开发者工具","非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["Git仓库分析","代码库瘦身","二进制文件检测","CI/CD优化"],"domainKey":"Git仓库分析","pushedAt":"2026-09-10T21:27:35Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/github/git-sizer","ts":"2026-09-10T21:27:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Liniyous/ElaWidgetTools","owner":"Liniyous","name":"ElaWidgetTools","desc":"","summaryCn":"基于Qt开发的精美FluentUI风格桌面端组件与工具库。","reasonCn":"对于长期使用C++进行桌面客户端开发的程序员来说，原生界面往往显得陈旧不够现代。该项目带来了一套优雅的流利设计风格组件库，完美契合现代操作系统的视觉审美。它不仅封装了大量常用的界面元素，还集成了许多实用的功能模块，让开发者无需从零开始就能打造出颜值极高的精美桌面应用程序。","stars":2503,"starGrowth":1,"createdAt":"2024-06-09T01:23:28Z","silentRounds":1,"language":"C++","topics":[],"aiDims":["开发者工具","可视化"],"zone":"工具","funScore":0.5,"domainTags":["Qt组件库","FluentUI","C++界面开发","桌面端UI"],"domainKey":"Qt组件库","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/Liniyous/ElaWidgetTools","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"medusajs/medusa","owner":"medusajs","name":"medusa","desc":"The world's most flexible commerce platform for agents and developers","summaryCn":"开源电商引擎，帮你搭店铺管库存做支付，灵活得像搭积木一样","reasonCn":"基于Headless架构的开源电商平台，支持商品管理、订单处理、支付集成等完整电商功能。采用模块化设计，开发者可自由组合插件，支持多仓库多店铺。技术栈为TypeScript/Next.js，配合PostgreSQL数据库，适合中大型电商项目定制开发，社区活跃且文档完善。","stars":36431,"starGrowth":11,"createdAt":"2020-01-18T13:39:04Z","silentRounds":0,"language":"TypeScript","topics":["nodejs","medusa","commerce","e-commerce","react","ecommerce","javascript","typescript","framework","ai-coding","ai-tools"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["电商平台","Headless架构","TypeScript","订单管理","支付集成"],"domainKey":"电商平台","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"nodejs","source":"github","weight":0.7},{"name":"medusa","source":"github","weight":0.7},{"name":"commerce","source":"github","weight":0.7},{"name":"e-commerce","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"ecommerce","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"ai-coding","source":"github","weight":0.7},{"name":"ai-tools","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/medusajs/medusa","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03390582196481019,"copyOk":false},{"repo":"forjd/better-writing","owner":"forjd","name":"better-writing","desc":"An agent skill for prose that sounds clear, specific, and human.","summaryCn":"让AI写出来的文章不再像假大空的机器人，变得像真人写的一样自然流畅。","reasonCn":"技术上，它并不是一个独立运行的大模型，而是一套高级的提示词和智能体行为技能。它通过特殊的约束条件和思维链引导，强制AI在生成内容时放弃那些套话和陈词滥调，转而使用更具体、更具细节的动词和形容词，精准捕捉人类写作时的呼吸感和节奏感，从而大幅提升文本的可读性。","stars":36,"starGrowth":0,"createdAt":"2026-05-27T09:28:26Z","silentRounds":1,"language":"Python","topics":["writing"],"aiDims":["AI写作","提示工程"],"zone":"AI","funScore":0,"domainTags":["写作","提示词","AI","Agent","文风"],"domainKey":"Agent","pushedAt":"2026-09-22T22:05:51Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/forjd/better-writing","ts":"2026-09-22T22:05:51Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"volcengine/OpenViking","owner":"volcengine","name":"OpenViking","desc":"Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.","summaryCn":"给AI大脑装上持久记忆，告别每次对话都像失忆","reasonCn":"这是一个专为AI智能体设计的自演进上下文数据库，旨在统一管理智能体的记忆、知识库和技能。它通过向量数据库存储非结构化数据，支持RAG技术检索知识，并具备动态知识图谱能力，让智能体能像人类一样积累经验、持续学习。","stars":38459,"starGrowth":123,"createdAt":"2026-01-05T07:11:17Z","silentRounds":0,"language":"Python","topics":["context-database","agentic-rag","agent-memory","self-evolving","agent-plugins","dsh-plugin"],"aiDims":["AI Agent","向量数据库","知识图谱"],"zone":"AI","funScore":0.6,"domainTags":["AI智能体","向量数据库","RAG","知识图谱","Python"],"domainKey":"向量数据库","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"context-database","source":"github","weight":0.7},{"name":"agentic-rag","source":"github","weight":0.7},{"name":"agent-memory","source":"github","weight":0.7},{"name":"self-evolving","source":"github","weight":0.7},{"name":"agent-plugins","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/volcengine/OpenViking","ts":"2026-09-17T04:13:51.053Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.37151355989495305,"copyOk":false},{"repo":"pingdotgg/t3code","owner":"pingdotgg","name":"t3code","desc":"","summaryCn":"T3 Stack 代码脚手架，一键生成全栈项目","reasonCn":"t3code 是基于 T3 Stack 生态的代码脚手架工具，集成 Next.js、tRPC、Prisma、Tailwind CSS 等热门技术栈。开发者只需运行一条命令，即可自动生成完整的全栈应用结构，省去手动配置框架、编写 boilerplate 代码的时间成本，特别适合追求开发效率的前端和全栈开发者。","stars":23329,"starGrowth":48,"createdAt":"2026-02-08T23:46:22Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["脚手架","全栈开发","TypeScript"],"domainKey":"脚手架","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"trending","url":"https://github.com/pingdotgg/t3code","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.18020708988812206,"copyOk":false},{"repo":"jarun/buku","owner":"jarun","name":"buku","desc":":bookmark: Personal mini-web in text","summaryCn":"在终端管理大量书签的神器，Markdown格式超方便","reasonCn":"这是一个功能强大的命令行书签管理器，采用纯文本Markdown格式存储所有书签数据，提供全文搜索和标签分类功能。无需打开浏览器就能在终端快速检索和管理成百上千的书签链接。支持从Chrome、Firefox、Safari等主流浏览器一键导入现有书签数据。还可通过SSH协议将书签同步到远程服务器、VPS或NAS设备，实现多设备统一管理。对技术用户来说是提升效率的绝佳工具。","stars":7206,"starGrowth":0,"createdAt":"2015-11-01T19:53:18Z","silentRounds":1,"language":"Python","topics":["command-line"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["书签管理","命令行工具","Markdown","终端应用"],"domainKey":"书签管理","pushedAt":"2026-09-15T05:47:34Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/jarun/buku","ts":"2026-09-15T05:47:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LaravelRUS/SleepingOwlAdmin","owner":"LaravelRUS","name":"SleepingOwlAdmin","desc":"🦉 Administrative interface builder for Laravel (Laravel admin)","summaryCn":"一键生成Laravel后台管理界面，告别重复造轮子","reasonCn":"LaravelRUS/SleepingOwlAdmin是一款专为Laravel框架设计的后台管理面板构建工具。它通过简洁的PHP配置代码即可生成包含表格、表单、导航等常用组件的管理界面，支持数据模型自动映射、权限管理和响应式布局。开发者无需手动编写大量HTML和JavaScript代码，大幅缩短后台系统开发周期，适合中小企业快速搭建内部管理平台。","stars":804,"starGrowth":0,"createdAt":"2016-02-20T07:04:49Z","silentRounds":24,"language":"PHP","topics":["backend"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Laravel框架","后台管理","PHP开发"],"domainKey":"Laravel框架","pushedAt":"2026-09-10T16:55:30Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/LaravelRUS/SleepingOwlAdmin","ts":"2026-09-10T16:55:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Mirrowel/LLM-API-Key-Proxy","owner":"Mirrowel","name":"LLM-API-Key-Proxy","desc":"Universal LLM Gateway: One API, every LLM. OpenAI/Anthropic-compatible endpoints with multi-provider translation and intelligent load-balancing.","summaryCn":"一个API网关把几十家大模型服务统一管理，一行配置搞定多模型接入","reasonCn":"LLM-API-Key-Proxy是一款开源的大语言模型API网关工具，核心功能是帮你把多个LLM服务商的API统一到一个接口下管理。支持OpenAI、Anthropic等主流服务商，自动处理协议转换和密钥管理。内置智能负载均衡，可以按QPS、延迟等指标自动分配请求。配置简单，只需修改一个配置文件就能切换不同的模型服务商，非常适合需要频繁更换模型或做多模型路由的场景。","stars":555,"starGrowth":1,"createdAt":"2025-06-10T17:07:26Z","silentRounds":4,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["API 网关","负载均衡","LLM 管理","协议转换"],"domainKey":"API 网关","pushedAt":"2026-09-23T21:06:19Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/Mirrowel/LLM-API-Key-Proxy","ts":"2026-09-23T21:06:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"cursor/plugins","owner":"cursor","name":"plugins","desc":"Cursor plugin specification and official plugins","summaryCn":"给AI编程助手装上各种插件，让它干活更顺手","reasonCn":"这是Cursor IDE官方维护的插件规范和插件仓库。Cursor是AI驱动的代码编辑器，这个项目定义了如何为Cursor开发插件，允许用户扩展其AI能力，比如自定义代码生成、自动化任务或集成特定工具。官方插件库提供现成插件，极大增强了IDE的智能水平。","stars":8411,"starGrowth":42,"createdAt":"2026-01-23T01:57:42Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["代码助手","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","IDE插件","扩展规范","代码自动化","开发工具链"],"domainKey":"AI编程助手","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/cursor/plugins","ts":"2026-09-17T04:13:51.053Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.16917645939840684,"copyOk":false},{"repo":"jandelgado/jled","owner":"jandelgado","name":"jled","desc":"Non-blocking LED controlling library for Arduino and friends.","summaryCn":"Arduino LED控制的神器，非阻塞式编程，再也不用担心程序卡死。","reasonCn":"它的技术实现非常巧妙，利用了Arduino框架中的`yield()`函数。这意味着在控制LED时，程序会主动让出CPU给其他任务，等LED状态变化后再继续执行。这就像在餐厅吃饭，你不用一直盯着厨师做菜（阻塞），而是可以一边看菜单一边聊天（非阻塞），厨师做好菜了自然会叫你。","stars":348,"starGrowth":0,"createdAt":"2017-06-18T18:35:16Z","silentRounds":11,"language":"C++","topics":["arduino"],"aiDims":["非AI-实用","开发工具"],"zone":"工具","funScore":0,"domainTags":["Arduino","嵌入式开发","库"],"domainKey":"Arduino","pushedAt":"2026-09-04T11:33:45Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/jandelgado/jled","ts":"2026-09-04T11:33:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wine-mirror/wine","owner":"wine-mirror","name":"wine","desc":"","summaryCn":"Wine 项目最核心的技术细节在于其名为 Winelib 的组件，该组件负责将 Windo","reasonCn":"该项目的兼容性范围极广，文档明确列出了包括 DOS、Windows 3.x、Win32 和 Win64 在内的所有主要 Windows 可执行文件架构。这意味着无论是二十年前遗留下来的特定行业软件，还是现代的 64 位商业应用，都有可能在 Wine 环境下找到对应的支持方案，用户无需为不同架构准备不同的运行","stars":4198,"starGrowth":0,"createdAt":"2014-08-07T07:02:05Z","silentRounds":2,"language":"C","topics":[],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.26,"domainTags":["Windows兼容","Linux系统","API实现","二进制加载","Wine","开源软件"],"domainKey":"Windows兼容","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/wine-mirror/wine","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ktbyers/netmiko","owner":"ktbyers","name":"netmiko","desc":"","summaryCn":"网络设备的运维一直以来都是IT领域中较痛苦的任务之一，由于不同厂商的CLI指","reasonCn":"不仅在数据获取方面表现出色，Netmiko同样能够稳健地处理配置变更请求。它对错误处理和命令响应的超时控制进行了深度封装，使得在批量配置多台设备时，脚本的稳定性和容错率得到了大幅提升。通过该库，原本需要手动登录几十台设备进行维护的操作，现在可以通过简单的Python逻辑在短","stars":4291,"starGrowth":0,"createdAt":"2014-11-28T21:42:52Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["网络自动化","CLI连接","多厂商设备"],"domainKey":"网络自动化","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/ktbyers/netmiko","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Pr1nted/Open-Doctrines","owner":"Pr1nted","name":"Open-Doctrines","desc":"A free, open-source grand strategy game — play in a browser, on desktop, or in a Discord call. 1641 provinces, six historical scenarios, a map editor and a WebAssembly mod SDK for thirteen languages.","summaryCn":"浏览器也能玩的大战略游戏，自定义地图超自由","reasonCn":"游戏的地图编辑器功能非常强大，允许玩家自定义地图和势力。你可以创建自己的地图，添加省份、城市、山脉等地理特征，然后设置不同的资源分布和地形加成。这使得游戏的可玩性大大提高，你可以根据自己的喜好创建各种不同的游戏环境。","stars":19,"starGrowth":0,"createdAt":"2026-07-21T13:06:59Z","silentRounds":8,"language":"C++","topics":["raylib"],"aiDims":["游戏","Raylib"],"zone":"创意","funScore":0.8,"domainTags":["大战略游戏","浏览器游戏","Raylib","WebAssembly","策略模拟"],"domainKey":"大战略游戏","pushedAt":"2026-09-21T17:17:05Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"Raylib","source":"llm","weight":1},{"name":"raylib","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Pr1nted/Open-Doctrines","ts":"2026-09-21T17:17:05Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"vasu-devs/JustHireMe","owner":"vasu-devs","name":"JustHireMe","desc":"Local-first AI job intelligence workbench for scraping roles, ranking fit, and generating tailored application materials.","summaryCn":"本地运行的AI求职助手，自动扒岗位并生成简历","reasonCn":"这是一个本地优先的AI求职工作台，核心能力在于自动抓取招聘网站职位、利用向量数据库评估人岗匹配度，并一键生成定制化求职材料。它将繁琐的求职流程自动化，让候选人能高效应对海量岗位，是职场人的实用效率神器。","stars":2252,"starGrowth":1,"createdAt":"2026-02-28T08:37:49Z","silentRounds":6,"language":"Python","topics":["vector-db"],"aiDims":["AI应用","工具调用","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI求职","简历生成","向量检索"],"domainKey":"AI求职","pushedAt":"2026-09-23T10:06:14Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/vasu-devs/JustHireMe","ts":"2026-09-23T10:06:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"grafana/loki","owner":"grafana","name":"loki","desc":"Like Prometheus, but for logs.","summaryCn":"像Prometheus一样，但专门收集日志的轻量级方案","reasonCn":"技术上看，Loki采用标签（label）+流（stream）的架构，每个日志流由一组标签唯一标识，比如kubernetes_pod_name、namespace等。日志以压缩格式存储在对象存储（如S3、GCS）上，查询语言叫LogQL，语法和Prometheus的PromQL非常相似，学习成本低。","stars":28934,"starGrowth":3,"createdAt":"2018-04-16T09:22:48Z","silentRounds":0,"language":"Go","topics":["loki","grafana","prometheus","logging","cloudnative","hacktoberfest"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["日志聚合","可观测性","日志系统","Grafana","存储"],"domainKey":"日志聚合","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"loki","source":"github","weight":0.7},{"name":"grafana","source":"github","weight":0.7},{"name":"prometheus","source":"github","weight":0.7},{"name":"logging","source":"github","weight":0.7},{"name":"cloudnative","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/grafana/loki","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nilbuild/developer-roadmap","owner":"nilbuild","name":"developer-roadmap","desc":"Interactive roadmaps, guides and other educational content to help developers grow in their careers.","summaryCn":"开发者职业成长地图，AI时代技能升级必看指南","reasonCn":"GitHub上Star超过36万的开源项目，收录了前端、后端、DevOps、区块链、数据库管理员等三十多个技术方向的完整学习路线。采用交互式地图形式，每个阶段标注了需要掌握的核心技能、推荐工具和免费学习资源，帮助程序员系统规划职业路径。","stars":367920,"starGrowth":47,"createdAt":"2017-03-15T13:45:52Z","silentRounds":0,"language":"TypeScript","topics":["angular-roadmap","backend-roadmap","blockchain-roadmap","computer-science","dba-roadmap","developer-roadmap","devops-roadmap","frontend-roadmap","go-roadmap","java-roadmap","javascript-roadmap","nodejs-roadmap","python-roadmap","qa-roadmap","react-roadmap","roadmap","software-architect-roadmap","vue-roadmap"],"aiDims":["学习资源","教程","可视化"],"zone":"资源","funScore":0,"domainTags":["学习路线","开发者成长","技术指南"],"domainKey":"学习路线","pushedAt":"2026-08-18T11:56:04Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"angular-roadmap","source":"github","weight":0.7},{"name":"backend-roadmap","source":"github","weight":0.7},{"name":"blockchain-roadmap","source":"github","weight":0.7},{"name":"computer-science","source":"github","weight":0.7},{"name":"dba-roadmap","source":"github","weight":0.7},{"name":"developer-roadmap","source":"github","weight":0.7},{"name":"devops-roadmap","source":"github","weight":0.7},{"name":"frontend-roadmap","source":"github","weight":0.7},{"name":"go-roadmap","source":"github","weight":0.7},{"name":"java-roadmap","source":"github","weight":0.7},{"name":"javascript-roadmap","source":"github","weight":0.7},{"name":"nodejs-roadmap","source":"github","weight":0.7},{"name":"python-roadmap","source":"github","weight":0.7},{"name":"qa-roadmap","source":"github","weight":0.7},{"name":"react-roadmap","source":"github","weight":0.7},{"name":"roadmap","source":"github","weight":0.7},{"name":"software-architect-roadmap","source":"github","weight":0.7},{"name":"vue-roadmap","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.32,"source":"bigbro","url":"https://github.com/nilbuild/developer-roadmap","ts":"2026-08-18T11:56:04Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1413832354859752,"copyOk":false},{"repo":"apptension/saas-boilerplate","owner":"apptension","name":"saas-boilerplate","desc":"SaaS Boilerplate - Open Source and free SaaS stack that lets you build SaaS products faster in React, Django and AWS. Focus on essential business logic instead of coding repeatable features!","summaryCn":"开箱即用的SaaS开发模板，省去重复搭建基础功能的时间","reasonCn":"技术上，这个项目整合了React做前端、Django做后端、AWS做云服务的基础架构。内置功能包括：基于OAuth的用户认证、Stripe订阅支付集成、邮件模板系统、角色权限管理、API限流、错误监控等。它不是完整的商业产品，而是一个工程模板，帮你跳过前期繁琐的基础设施搭建。","stars":2999,"starGrowth":0,"createdAt":"2020-04-14T09:31:26Z","silentRounds":5,"language":"TypeScript","topics":["backend","frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["SaaS","全栈开发","React"],"domainKey":"SaaS","pushedAt":"2026-09-02T11:37:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/apptension/saas-boilerplate","ts":"2026-09-02T11:37:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"leon-ai/leon","owner":"leon-ai","name":"leon","desc":"🧠 Leon is your open-source personal assistant.","summaryCn":"你的开源个人AI助手，能听懂语音、执行任务、控制设备","reasonCn":"Leon支持多种交互方式：自然语言对话、语音命令、手机App界面。它能理解你的意图并执行相应操作，比如设置提醒、搜索信息、控制智能家居设备、播放音乐、发送消息等。技术上它使用本地NLP引擎处理自然语言理解，支持多种语言，包括中文。","stars":17539,"starGrowth":2,"createdAt":"2019-02-10T12:25:12Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","chatbot"],"aiDims":["AI Agent","AI应用"],"zone":"工具","funScore":0,"domainTags":["AI Agent","智能助手","语音交互","自动化"],"domainKey":"AI Agent","pushedAt":"2026-09-22T07:51:55Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/leon-ai/leon","ts":"2026-09-22T07:51:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hmjz100/LinkSwift","owner":"hmjz100","name":"LinkSwift","desc":"一个基于 JavaScript 的网盘文件下载地址获取工具。基于【网盘直链下载助手】修改 ，支持 百度网盘 / 阿里云盘 / 中国移动云盘 / 天翼云盘 / 迅雷云盘 / 夸克网盘 / UC网盘 / 123云盘 八大网盘","summaryCn":"一键获取八大网盘直链，解决下载慢到崩溃的痛点","reasonCn":"这是一个纯JavaScript编写的网盘文件下载地址提取工具，修改自知名开源项目网盘直链下载助手。它通过解析网页接口，自动获取百度网盘、阿里云盘、夸克网盘等八大主流网盘的直链，支持批量下载和断点续传，有效解决网盘限速问题。","stars":20741,"starGrowth":34,"createdAt":"2022-08-10T11:59:44Z","silentRounds":0,"language":"JavaScript","topics":["aliyun-drive","baidu","baidu-netdisk","quark-netdisk","tianyi-netdisk","xunlei-netdisk","yidong-netdisk","tampermonkey","tampermonkey-script","tampermonkey-userscript","userscript","baidunetdisk","baiduyun","123pan","uc-netdisk","motrix","aria2","aliyunpan","baidupan","guangya-netdisk"],"aiDims":["非AI-实用","开发工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["网盘","直链提取","JavaScript","下载工具","批量下载"],"domainKey":"网盘","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"aliyun-drive","source":"github","weight":0.7},{"name":"baidu","source":"github","weight":0.7},{"name":"baidu-netdisk","source":"github","weight":0.7},{"name":"quark-netdisk","source":"github","weight":0.7},{"name":"tianyi-netdisk","source":"github","weight":0.7},{"name":"xunlei-netdisk","source":"github","weight":0.7},{"name":"yidong-netdisk","source":"github","weight":0.7},{"name":"tampermonkey","source":"github","weight":0.7},{"name":"tampermonkey-script","source":"github","weight":0.7},{"name":"tampermonkey-userscript","source":"github","weight":0.7},{"name":"userscript","source":"github","weight":0.7},{"name":"baidunetdisk","source":"github","weight":0.7},{"name":"baiduyun","source":"github","weight":0.7},{"name":"123pan","source":"github","weight":0.7},{"name":"uc-netdisk","source":"github","weight":0.7},{"name":"motrix","source":"github","weight":0.7},{"name":"aria2","source":"github","weight":0.7},{"name":"aliyunpan","source":"github","weight":0.7},{"name":"baidupan","source":"github","weight":0.7},{"name":"guangya-netdisk","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/hmjz100/LinkSwift","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.10691779567041126,"copyOk":false},{"repo":"Ismaestro/angular-example-app","owner":"Ismaestro","name":"angular-example-app","desc":"Angular Example App is a beginner-friendly, production-ready web application built with Angular 21. It serves as a real-world example showcasing core Angular features such as CRUD operations, authentication, i18n, lazy loading, and signals.","summaryCn":"一个手把手教你写Angular的全栈示例，CRUD和登录功能全都有","reasonCn":"这是Angular官方风格的完整示例应用，涵盖用户认证、CRUD操作、状态管理、响应式表单等核心功能。基于Angular 21最新版本，采用最新的控制流语法如@if @for，代码结构清晰规范。包含服务端交互、错误处理、路由守卫等完整功能，非常适合前端初学者通过这个项目快速掌握Angular开发，也可以作为团队内部开发规范的参考模板使用。","stars":2388,"starGrowth":0,"createdAt":"2016-11-22T17:10:18Z","silentRounds":25,"language":"TypeScript","topics":["frontend"],"aiDims":["非AI-实用","学习资源"],"zone":"资源","funScore":0,"domainTags":["Angular","前端开发","示例项目","教程","TypeScript"],"domainKey":"Angular","pushedAt":"2026-09-18T16:14:45Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/Ismaestro/angular-example-app","ts":"2026-09-18T16:14:45Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"google/dawn","owner":"google","name":"dawn","desc":"Native WebGPU implementation. Mirror of https://dawn.googlesource.com/dawn. File bugs here: https://crbug.com/dawn/new","summaryCn":"浏览器里跑WebGPU的原生实现，图形渲染性能直接起飞","reasonCn":"Dawn是Google主导的WebGPU原生实现，将WebGPU API绑定到Vulkan、Metal、DirectX等底层图形接口，让浏览器里的3D渲染、计算任务能充分利用GPU硬件加速。它是Chromium的核心组件，驱动Chrome和Edge的WebGPU支持，支持Compute Shader、渲染管线、多GPU设备管理等高级特性。","stars":1141,"starGrowth":1,"createdAt":"2023-10-04T16:54:23Z","silentRounds":1,"language":"C++","topics":["dawn","emdawnwebgpu","webgpu"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["WebGPU","图形渲染","C++","Vulkan"],"domainKey":"WebGPU","pushedAt":"2026-08-19T00:15:44Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"dawn","source":"github","weight":0.7},{"name":"emdawnwebgpu","source":"github","weight":0.7},{"name":"webgpu","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/google/dawn","ts":"2026-08-19T00:15:44Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"NorthwoodsSoftware/GoJS","owner":"NorthwoodsSoftware","name":"GoJS","desc":"JavaScript diagramming library for interactive flowcharts, org charts, design tools, planning tools, visual languages.","summaryCn":"JavaScript图表库，画流程图拓扑图轻松搞定","reasonCn":"纯JavaScript开发的交互式图表库，支持流程图、组织结构图、网络拓扑图、甘特图等多种图表类型。基于HTML5 Canvas渲染，无需额外插件，API设计简洁直观。虽然与AI无直接关联，但可为AI应用提供可视化配置面板和结果展示界面。","stars":8483,"starGrowth":1,"createdAt":"2014-04-02T18:18:40Z","silentRounds":3,"language":"JavaScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["图表库","流程图","可视化","JavaScript","Canvas"],"domainKey":"图表库","pushedAt":"2026-09-14T19:39:55Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/NorthwoodsSoftware/GoJS","ts":"2026-09-14T19:39:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xorbitsai/inference","owner":"xorbitsai","name":"inference","desc":"Swap GPT for any LLM by changing a single line of code. Xinference lets you run open-source, speech, and multimodal models on cloud, on-prem, or your laptop — all through one unified, production-ready inference API.","summaryCn":"一行代码切换大模型，云端本地都能跑，AI开发者的万能钥匙","reasonCn":"Xinference是一个统一的LLM推理引擎，支持开源大模型、语音和多模态模型。一行代码即可替换GPT接口，提供OpenAI兼容的RESTful API，支持多种推理后端如vLLM、TensorRT等。可在云端、本地或笔记本上部署，生产就绪，适合快速搭建AI应用原型。","stars":9588,"starGrowth":1,"createdAt":"2023-06-14T07:05:04Z","silentRounds":1,"language":"Python","topics":["transformers"],"aiDims":["推理引擎","大语言模型","多模态"],"zone":"AI","funScore":0,"domainTags":["大模型推理","LLM部署","模型切换"],"domainKey":"大模型推理","pushedAt":"2026-09-23T16:35:27Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/xorbitsai/inference","ts":"2026-09-23T16:35:27Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"crynta/terax-ai","owner":"crynta","name":"terax-ai","desc":"Lightweight (7MB) Terminal-first AI-native dev workspace","summaryCn":"只有7MB大小的终端AI编程工作空间，让命令行也能拥有大模型大脑","reasonCn":"它在技术上最大的亮点就是极致的“轻量与专注”。采用TypeScript编写，不臃肿，去掉了所有华而不实的图形界面开销。通过深度集成的Agent机制，它能够理解当前项目的上下文结构，不仅是简单的问答，还能直接在终端环境中替你执行复杂的开发任务，响应速度极快，对老旧电脑或者云端服务","stars":9231,"starGrowth":4,"createdAt":"2026-04-21T18:33:12Z","silentRounds":0,"language":"TypeScript","topics":["agents"],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"AI","funScore":0.7,"domainTags":["AI编程","终端工具","Agent","TypeScript","开发环境"],"domainKey":"Agent","pushedAt":"2026-09-21T21:53:23Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/crynta/terax-ai","ts":"2026-09-21T21:53:23Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"JimLiu/baoyu-skills","owner":"JimLiu","name":"baoyu-skills","desc":"","summaryCn":"号称能帮你通关所有游戏的辅助工具，实测效果如何？","reasonCn":"技术上，它采用计算机视觉技术，通过截图分析游戏画面，提取关键特征点。它支持多种游戏引擎，比如Unity、Cocos等，能够识别2D和3D游戏中的物体。除了基础的动作自动化，它还支持脚本录制和回放，你可以自己编写规则，比如“遇到红色敌人就放技能A”。","stars":26102,"starGrowth":9,"createdAt":"2026-01-13T03:04:56Z","silentRounds":0,"language":"TypeScript","topics":["agent-skills","claude-skills","codex-skills","openclaw-skills"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":1,"domainTags":["游戏自动化","计算机视觉","游戏辅助","Unity","脚本录制"],"domainKey":"游戏自动化","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"claude-skills","source":"github","weight":0.7},{"name":"codex-skills","source":"github","weight":0.7},{"name":"openclaw-skills","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/JimLiu/baoyu-skills","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03364128419278216,"copyOk":false},{"repo":"soxoj/maigret","owner":"soxoj","name":"maigret","desc":"🕵️‍♂️ Collect a dossier on a person by username from 3000+ sites","summaryCn":"一键全网搜集某人所有社交账号的黑科技，侦探必备神器","reasonCn":"Maigret 是一个用 Python 编写的开源工具，它能够通过用户名在 3000 多个网站和社交平台上收集该用户的所有公开档案。它支持多种数据源和输出格式，非常适合用于人肉搜索、背景调查或安全研究，虽然功能强大但操作非常简单，命令行界面即可使用。","stars":37901,"starGrowth":8,"createdAt":"2020-06-27T14:16:42Z","silentRounds":0,"language":"Python","topics":["osint","social-network","identification","socmint","sherlock","investigation","python","python3","open-source","cybersecurity","scraping","osint-python","redteam","osint-framework","cli","reconnaissance","infosec","pentesting","information-gathering","username"],"aiDims":["非AI-实用","开源项目"],"zone":"工具","funScore":0.7,"domainTags":["用户名搜索","OSINT","信息安全","爬虫","数字取证"],"domainKey":"用户名搜索","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开源项目","source":"llm","weight":1},{"name":"osint","source":"github","weight":0.7},{"name":"social-network","source":"github","weight":0.7},{"name":"identification","source":"github","weight":0.7},{"name":"socmint","source":"github","weight":0.7},{"name":"sherlock","source":"github","weight":0.7},{"name":"investigation","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"python3","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"scraping","source":"github","weight":0.7},{"name":"osint-python","source":"github","weight":0.7},{"name":"redteam","source":"github","weight":0.7},{"name":"osint-framework","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"reconnaissance","source":"github","weight":0.7},{"name":"infosec","source":"github","weight":0.7},{"name":"pentesting","source":"github","weight":0.7},{"name":"information-gathering","source":"github","weight":0.7},{"name":"username","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/soxoj/maigret","ts":"2026-08-14T01:23:41.092Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024633228674705153,"copyOk":false},{"repo":"adaltas/node-csv","owner":"adaltas","name":"node-csv","desc":"","summaryCn":"在Node.js后端开发中，处理逗号分隔值（CSV）文件是一项基础但繁琐的任务，","reasonCn":"该库的核心价值在于其全面的功能覆盖和对边缘情况的稳健处理。它不仅能够将CSV文本转换为JavaScript数组或对象，还支持反向操作，将数据记录序列化为标准的CSV文本。更重要的是，它提供了处理复杂场景的能力，如自定义分隔符、处理引号内的特殊字符以及流式处理大文件，避免了因一","stars":4287,"starGrowth":0,"createdAt":"2010-09-25T14:20:57Z","silentRounds":4,"language":"JavaScript","topics":["csv","parser","nodejs","streaming"],"aiDims":["非AI-实用","开发者工具","数据处理"],"zone":"工具","funScore":0.2,"domainTags":["CSV解析","Node.js","数据转换","流式处理"],"domainKey":"CSV解析","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"数据处理","source":"llm","weight":1},{"name":"csv","source":"github","weight":0.7},{"name":"parser","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"streaming","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/adaltas/node-csv","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JohnSnowLabs/spark-nlp","owner":"JohnSnowLabs","name":"spark-nlp","desc":"State of the Art Natural Language Processing","summaryCn":"专为企业级海量数据设计的高性能自然语言处理和AI分析平台","reasonCn":"基于Apache Spark构建的企业级自然语言处理框架，完美融合了传统NLP与最前沿的Transformer大模型。它支持分布式集群并行计算，能够以极高的吞吐量处理TB级别的文本数据，广泛应用于医疗、金融等高合规要求领域的文本挖掘。","stars":4159,"starGrowth":0,"createdAt":"2017-09-24T19:36:44Z","silentRounds":21,"language":"Scala","topics":["transformers"],"aiDims":["AI基础设施","大语言模型"],"zone":"工具","funScore":0,"domainTags":["NLP","Spark","Transformer","分布式计算","企业级文本分析"],"domainKey":"NLP","pushedAt":"2026-09-23T21:47:22Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Scala","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/JohnSnowLabs/spark-nlp","ts":"2026-09-23T21:47:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PaddlePaddle/Paddle","owner":"PaddlePaddle","name":"Paddle","desc":"PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）","summaryCn":"百度出品深度学习框架，国产AI训练首选，支持国产芯片灵活部署。","reasonCn":"PaddlePaddle（飞桨）是百度自主研发的开源深度学习平台，国内用户量最大的AI框架。它提供了从底层张量计算到上层应用开发的全套工具链，支持分布式训练、模型压缩和端侧部署。其最大优势在于对国产硬件（如华为昇腾、寒武纪）的深度适配，以及丰富的中文文档和教程生态，是国内开发者学习和生产落地的重要选择。","stars":24102,"starGrowth":0,"createdAt":"2016-08-15T06:59:08Z","silentRounds":2,"language":"C++","topics":["ml"],"aiDims":["AI基础设施","模型与训练","开发者工具"],"zone":"AI","funScore":0,"domainTags":["深度学习","机器学习","神经网络","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-23T04:02:51Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/PaddlePaddle/Paddle","ts":"2026-09-23T04:02:51Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"brightdata/cli","owner":"brightdata","name":"cli","desc":"Official Bright Data CLI - scrape, search, and extract structured web data directly from your terminal.","summaryCn":"一行命令从互联网上抓取你想要的数据，不用再写爬虫代码","reasonCn":"Bright Data官方推出的命令行工具，无需编写代码即可从终端直接抓取、搜索和提取结构化网页数据。底层调用Bright Data成熟的代理网络和爬虫基础设施，支持防屏蔽、反爬虫绕过等能力，适合需要批量采集数据的开发者和数据工程师，省去了自己维护爬虫架构的繁琐。","stars":6304,"starGrowth":0,"createdAt":"2026-03-07T10:25:57Z","silentRounds":24,"language":"TypeScript","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["数据采集","爬虫","命令行工具","网页抓取","数据提取"],"domainKey":"数据采集","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/brightdata/cli","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"akuity/kargo","owner":"akuity","name":"kargo","desc":"Application lifecycle orchestration","summaryCn":"K8s应用的全自动管家，从代码到上线一键搞定不操心","reasonCn":"Kargo是专为Kubernetes设计的GitOps应用交付平台，通过可视化界面管理应用全生命周期。支持自动同步上游变更、触发部署、回滚操作，内置多环境策略管理。相比Argo CD和Flux，它更强调开箱即用的工作流体验，无需复杂配置即可实现持续交付。","stars":3673,"starGrowth":1,"createdAt":"2022-07-14T22:16:06Z","silentRounds":1,"language":"Go","topics":["argocd","gitops","k8s","kubernetes","cd","delivery","hacktoberfest","promotions"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","GitOps","持续交付","容器部署","工作流编排"],"domainKey":"Kubernetes","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"argocd","source":"github","weight":0.7},{"name":"gitops","source":"github","weight":0.7},{"name":"k8s","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"cd","source":"github","weight":0.7},{"name":"delivery","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"promotions","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/akuity/kargo","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/RD-Agent","owner":"microsoft","name":"RD-Agent","desc":"Research and development (R&D) is crucial for the enhancement of industrial productivity, especially in the AI era, where the core aspects of R&D are mainly focused on data and models. We are committed to automating these high-value generic R&D processes through R&D-Agent, which lets AI drive data-driven AI. 🔗https://aka.ms/RD-Agent-Tech-Report","summaryCn":"AI驱动的工业研发自动化，让机器自己搞科研","reasonCn":"微软研究院开源的RD-Agent，专注于工业研发场景的自动化。它通过AI代理自动执行数据准备、模型训练、超参调优等繁琐任务，特别适用于材料科学、药物研发等领域。这是一个将AI从“聊天”推向“实干”的重要项目，展示了AI在垂直领域的强大生产力。","stars":14716,"starGrowth":4,"createdAt":"2024-04-03T03:39:33Z","silentRounds":0,"language":"Python","topics":["automation","data-science"],"aiDims":["AI基础设施","RAG与知识","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["AI研发","自动化","材料科学","药物研发","Agent"],"domainKey":"Agent","pushedAt":"2026-09-23T11:05:33Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/microsoft/RD-Agent","ts":"2026-09-23T11:05:33Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"KyleBanks/scene-ref-attribute","owner":"KyleBanks","name":"scene-ref-attribute","desc":"Unity C# attribute for serializing component and interface references within the scene or prefab during OnValidate.","summaryCn":"Unity场景引用自动序列化的超实用小插件","reasonCn":"技术上，它利用了Unity的OnValidate回调机制，在Inspector字段变化时自动扫描场景中符合要求的组件，通过接口或类型匹配找到对应的引用并完成序列化。核心思路是用C# Attribute语法糖封装了一层自动化逻辑，开发者只需在字段上加上自定义特性标记即可，不需要写任何额外的初始化代","stars":577,"starGrowth":0,"createdAt":"2023-02-15T17:45:32Z","silentRounds":24,"language":"C#","topics":[],"aiDims":["开发者工具","非AI-实用","游戏"],"zone":"工具","funScore":0,"domainTags":["Unity3D","C#语言","游戏开发","组件序列化","编辑器扩展"],"domainKey":"Unity3D","pushedAt":"2026-08-01T04:38:14Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/KyleBanks/scene-ref-attribute","ts":"2026-08-01T04:38:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sam-vdp/bepuphysics1int","owner":"sam-vdp","name":"bepuphysics1int","desc":"Pure C# deterministic fixed-point 3D real time physics simulation library","summaryCn":"纯C#写的确定性物理引擎，游戏里的物理模拟结果每次都一模一样","reasonCn":"BepuPhysics1Int是一个纯C#实现的确定性定点3D物理引擎，专为游戏和仿真场景设计。采用定点数运算替代浮点数，确保物理模拟在任何平台上都能产生完全相同的结果，这对于需要同步物理状态的网络游戏或回放系统非常重要。","stars":547,"starGrowth":0,"createdAt":"2018-02-09T12:55:15Z","silentRounds":24,"language":"C#","topics":[],"aiDims":["非AI-好玩","游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["物理引擎","确定性模拟","定点数运算","游戏开发","C#库"],"domainKey":"物理引擎","pushedAt":"2019-02-03T05:08:00Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/sam-vdp/bepuphysics1int","ts":"2019-02-03T05:08:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"antvis/G6","owner":"antvis","name":"G6","desc":"♾ A Graph Visualization Framework in JavaScript.","summaryCn":"专门画复杂关系图的 JS 库，节点连线再多也能清晰展示不混乱","reasonCn":"这是蚂蚁集团开源的一个专注于图可视化的 JavaScript 引擎，能够高效渲染复杂的网络关系图。它支持多种布局算法，如力导向、环形、分层等，能自动排列节点，让数据间的联系一目了然。特别适合用于社交网络分析、知识图谱和流程图展示，让数据关系不再晦涩难懂。","stars":12304,"starGrowth":3,"createdAt":"2017-02-13T10:00:43Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["图可视化","数据可视化","JavaScript"],"domainKey":"图可视化","pushedAt":"2026-09-23T10:35:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/antvis/G6","ts":"2026-09-23T10:35:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"givanz/Vvveb","owner":"givanz","name":"Vvveb","desc":"Powerful and easy to use cms to build websites, blogs or ecommerce stores.","summaryCn":"不用写一行代码也能搭出企业级网站，这开源CMS简直是懒人建站神器。","reasonCn":"Vvveb 是一款功能强大的开源 CMS，采用拖拽式可视化设计，支持博客、电商和常规企业官网搭建。无需深厚的编程基础，通过直观的页面构建器即可快速生成响应式网站，极大降低了建站门槛，非常适合追求效率的独立开发者和内容创作者。","stars":1192,"starGrowth":0,"createdAt":"2018-05-29T11:33:24Z","silentRounds":3,"language":"HTML","topics":["backend"],"aiDims":["非AI-实用","效率工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["内容管理系统","可视化建站","拖拽编辑器","电子商务","PHP应用"],"domainKey":"内容管理系统","pushedAt":"2026-08-05T19:00:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/givanz/Vvveb","ts":"2026-08-05T19:00:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vrothberg/vgrep","owner":"vrothberg","name":"vgrep","desc":"a user-friendly pager for grep","summaryCn":"很多人使用 grep 来定位日志中的关键字，但搜索完后往往只能把结果倒在屏幕","reasonCn":"很多人使用 grep 来定位日志中的关键字，但搜索完后往往只能把结果倒在屏幕上，想要进一步浏览就得借助 less，操作繁琐。vgrep 把这两个步骤合并成一个工具，让你在终端里像翻书一样逐页查看匹配行，省去切换命令的麻烦。","stars":707,"starGrowth":0,"createdAt":"2015-03-27T13:35:15Z","silentRounds":11,"language":"Go","topics":["frontend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["命令行工具","日志分析","终端搜索"],"domainKey":"命令行工具","pushedAt":"2026-09-18T08:37:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/vrothberg/vgrep","ts":"2026-09-18T08:37:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tldraw/tldraw","owner":"tldraw","name":"tldraw","desc":"Build infinite canvas apps in React with the tldraw SDK. World's best, top-most agent recommended #1 five star SDK.","summaryCn":"一个让你脑洞无限的数字画布，手残也能画出专业级设计稿","reasonCn":"基于React的无限画布框架，内置丰富的绘图、框选、连线、便签等工具，支持多人实时协作。导出功能完善，可输出PNG、SVG或嵌入网页。SDK高度可定制，能快速搭建白板、流程图、思维导图等创意工具应用，无需从零写Canvas逻辑。","stars":50521,"starGrowth":13,"createdAt":"2021-05-09T11:48:37Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["Canvas","Drawing","Collaboration","React","Whiteboard"],"domainKey":"Canvas","pushedAt":"2026-09-23T19:02:58Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/tldraw/tldraw","ts":"2026-09-23T19:02:58Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.039771956216226916,"copyOk":false},{"repo":"1038lab/ComfyUI-RMBG","owner":"1038lab","name":"ComfyUI-RMBG","desc":"","summaryCn":"在图形处理和 AI 视觉创作领域，如何干净利落地把主体从复杂的背景中分离出","reasonCn":"该插件最大的亮点在于其庞大且持续更新的模型生态。它不仅包含了像 RMBG-2.0 这样经典的抠图模型，还融入了 BiRefNet、SAM2 等能够处理极复杂边缘的深度学习架构。通过引入诸如 Lucida 这样的微调模型，它甚至能够完美应对透明物体、伪装物体、毛发纤维以及发光特效等传统抠图工具","stars":2123,"starGrowth":17,"createdAt":"2024-11-14T17:48:27Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","多模态Agent","创意工具"],"zone":"AI","funScore":0.5,"domainTags":["ComfyUI","background removal","segmentation","BiRefNet","SAM2"],"domainKey":"ComfyUI","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/1038lab/ComfyUI-RMBG","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0750226095148375,"copyOk":false},{"repo":"lscambo13/ElegantFin","owner":"lscambo13","name":"ElegantFin","desc":"A Jellyfin theme inspired from Jellyseerr. This theme improves the overall look and experience with various little fixes to the UI/UX.","summaryCn":"让你的Jellyfin影音库颜值飙升的精美主题，开箱即用超省心","reasonCn":"一个专为Jellyfin媒体服务器设计的精美主题，灵感来自Jellyseerr的UI风格。通过CSS自定义大幅改善Jellyfin的界面外观和用户体验，包括现代化卡片布局、平滑过渡动画、暗色主题优化等细节。安装只需把文件放到Jellyfin的themes目录并启用，无需修改核心代码，升级不受影响。","stars":2169,"starGrowth":1,"createdAt":"2024-12-01T10:55:03Z","silentRounds":1,"language":"CSS","topics":["design","jellyfin"],"aiDims":["非AI-好玩","可视化","非AI-实用"],"zone":"创意","funScore":0.7,"domainTags":["Jellyfin主题","CSS设计","UI美化","媒体服务器","暗色模式"],"domainKey":"Jellyfin主题","pushedAt":"2026-09-09T11:16:14Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"jellyfin","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/lscambo13/ElegantFin","ts":"2026-09-09T11:16:14Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nikolaydubina/go-recipes","owner":"nikolaydubina","name":"go-recipes","desc":"🦩 Tools for Go projects","summaryCn":"Go语言开发者的效率神器，一堆实用小工具合集让你写代码更快更稳","reasonCn":"这个项目汇总了大量Go语言开发中的实用工具，包括性能分析、可视化、代码生成等功能。虽不直接涉及AI领域，但作为开发者工具集非常实用，适合Go开发者提升编码效率。项目由知名Go专家维护，质量有保障，是学习Go最佳实践的好资源。","stars":4762,"starGrowth":1,"createdAt":"2021-06-26T12:57:05Z","silentRounds":3,"language":"Go","topics":["data-visualization","developer-tools"],"aiDims":["开发者工具","非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["Go语言","开发工具","性能分析","代码生成","可视化"],"domainKey":"Go语言","pushedAt":"2026-09-09T09:58:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/nikolaydubina/go-recipes","ts":"2026-09-09T09:58:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"esengine/ccesengine","owner":"esengine","name":"ccesengine","desc":"Cross-platform Game Engine | Based on Cocos Creator","summaryCn":"用C++做游戏也能很简单，一键发布到网页和手机的跨平台引擎","reasonCn":"基于Cocos Creator深度开发的跨平台游戏引擎，使用C++编写核心逻辑，支持WebGL渲染。提供TypeScript接口，让开发者可以用熟悉的Web技术栈快速开发游戏，并能一键发布到浏览器、Android和iOS平台，大幅降低跨平台游戏开发门槛。","stars":83,"starGrowth":0,"createdAt":"2025-12-30T14:02:54Z","silentRounds":24,"language":"C++","topics":["cocos","cross-platform","game-development","game-engine","typescript","webgl"],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"创意","funScore":0,"domainTags":["游戏引擎","Cocos Creator","游戏开发","跨平台"],"domainKey":"游戏引擎","pushedAt":"2026-01-19T07:50:06Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"cocos","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/esengine/ccesengine","ts":"2026-01-19T07:50:06Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"shap/shap","owner":"shap","name":"shap","desc":"A game theoretic approach to explain the output of any machine learning model.","summaryCn":"让AI模型不再黑盒，每一步决策都能看懂且可解释","reasonCn":"基于博弈论的可解释性框架，计算每个特征对模型输出的贡献值，支持树模型、神经网络等多种模型类型。提供Jupyter Notebook集成和多种可视化方案，帮助数据科学家理解模型行为，让黑盒模型变得透明可信，是模型审计和debug的必备工具。","stars":25773,"starGrowth":1,"createdAt":"2016-11-22T19:17:08Z","silentRounds":2,"language":"Jupyter Notebook","topics":["ml"],"aiDims":["模型与训练","可视化","推理引擎"],"zone":"工具","funScore":0,"domainTags":["机器学习","模型可解释性","Shapley值","数据科学","可视化"],"domainKey":"机器学习","pushedAt":"2026-09-21T18:23:28Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/shap/shap","ts":"2026-09-21T18:23:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Tencent/ncnn","owner":"Tencent","name":"ncnn","desc":"ncnn is a high-performance neural network inference framework optimized for the mobile platform","summaryCn":"让手机端也能流畅跑深度学习模型的性能怪兽","reasonCn":"技术上，ncnn 极其纯粹。它放弃了所有沉重的第三方库，完全用 C++ 重新实现，这意味着它没有冗余，运行速度极快。它最厉害的地方是对 Vulkan 的支持，这让它能跨平台地调用各种手机 GPU 的算力。同时，它支持模型量化和精简，能把模型体积压缩得很小，而性能损失却微乎其微，在移动","stars":23861,"starGrowth":3,"createdAt":"2017-06-30T10:55:37Z","silentRounds":0,"language":"C++","topics":["deep-learning"],"aiDims":["推理引擎","模型部署","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["移动端推理","神经网络","Vulkan"],"domainKey":"移动端推理","pushedAt":"2026-09-23T02:19:30Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Tencent/ncnn","ts":"2026-09-23T02:19:30Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jacksondunstan/NativeCollections","owner":"jacksondunstan","name":"NativeCollections","desc":"Native Collection Types for Unity","summaryCn":"Unity游戏开发的性能神器，告别内存卡顿和GC烦恼","reasonCn":"Unity的Jobs System催生了对高性能集合的需求。NativeCollections提供了NativeArray、NativeList等原生集合类型，直接在非托管内存中分配，完全绕开C#垃圾回收机制。配合Job并行计算框架，可以实现超大数量级的数据处理，对游戏性能敏感的项目来说几乎是必装神器。","stars":421,"starGrowth":0,"createdAt":"2018-08-06T04:44:24Z","silentRounds":24,"language":"C#","topics":["csharp","unity","unity-scripts","unity3d","unity3d-development","unity3d-script"],"aiDims":["游戏","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Unity","性能优化","C#","Native Memory"],"domainKey":"Unity","pushedAt":"2023-08-26T16:58:48Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity-scripts","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"unity3d-development","source":"github","weight":0.7},{"name":"unity3d-script","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.42,"source":"bigbro","url":"https://github.com/jacksondunstan/NativeCollections","ts":"2023-08-26T16:58:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"prisma/prisma","owner":"prisma","name":"prisma","desc":"Next-generation ORM for Node.js & TypeScript | PostgreSQL, MySQL, MariaDB, SQL Server, SQLite, MongoDB and CockroachDB","summaryCn":"编程搭桥神器，让代码与数据库对话不再头疼","reasonCn":"Prisma是Node.js和TypeScript开发者的数据库 ORM 工具，支持PostgreSQL等多种主流数据库。它通过类型安全的查询构建器，让开发者用代码优雅地操作数据，极大减少了写SQL的繁琐，提升开发效率，特别适合现代Web应用后端开发。","stars":47660,"starGrowth":5,"createdAt":"2019-06-20T13:33:47Z","silentRounds":0,"language":"TypeScript","topics":["database"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["ORM","数据库","TypeScript"],"domainKey":"ORM","pushedAt":"2026-08-18T20:08:47Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/prisma/prisma","ts":"2026-08-18T20:08:47Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015314729332773815,"copyOk":false},{"repo":"michidk/Unity-Script-Collection","owner":"michidk","name":"Unity-Script-Collection","desc":"A maintained collection of useful & free unity scripts / library's / plugins and extensions","summaryCn":"Unity必备脚本全家桶，粒子特效寻路物理都有现成的","reasonCn":"从技术角度看，这个脚本库覆盖了游戏开发最常见的几个方向：粒子系统方面有多达几十种预设效果，从烟花爆炸到雨水模拟都有；寻路系统整合了A*算法的优化版本，支持多层导航网格；物理部分有弹簧、绳索、车辆模拟等实用组件；还有各种材质Shader、UI动画、音效管理工具。","stars":6372,"starGrowth":0,"createdAt":"2016-09-16T16:11:23Z","silentRounds":2,"language":"","topics":["csharp","hacktoberfest","material","particles","pathfinding","physics","script","scripts","shader","shaders","steamworks","unity","unity-asset","unity-editor","unity-scripts","unity2d","unity3d","unity3d-plugin","unity3d-script"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Unity游戏开发","粒子系统","寻路算法","物理模拟","脚本组件"],"domainKey":"Unity游戏开发","pushedAt":"2026-09-09T13:31:40Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"material","source":"github","weight":0.7},{"name":"particles","source":"github","weight":0.7},{"name":"pathfinding","source":"github","weight":0.7},{"name":"physics","source":"github","weight":0.7},{"name":"script","source":"github","weight":0.7},{"name":"scripts","source":"github","weight":0.7},{"name":"shader","source":"github","weight":0.7},{"name":"shaders","source":"github","weight":0.7},{"name":"steamworks","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity-asset","source":"github","weight":0.7},{"name":"unity-editor","source":"github","weight":0.7},{"name":"unity-scripts","source":"github","weight":0.7},{"name":"unity2d","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"unity3d-plugin","source":"github","weight":0.7},{"name":"unity3d-script","source":"github","weight":0.7}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/michidk/Unity-Script-Collection","ts":"2026-09-09T13:31:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"i-am-bee/beeai-framework","owner":"i-am-bee","name":"beeai-framework","desc":"Build production-ready AI agents in both Python and Typescript.","summaryCn":"你是否曾因为团队习惯不同，在Python和TypeScript之间纠结，导致AI Agent的","reasonCn":"技术上，它抛弃了臃肿的冗余设计，追求的是一种“高性能、低配置”的开发体验。它拥有强大的工具调用抽象，能轻松连接任何API或本地逻辑。同时，它内置了多种Agent编排策略，无论是简单的单步响应，还是复杂的任务链式处理，它都能通过简单的配置项实现，让你从繁杂的底层代码中解脱","stars":3410,"starGrowth":2,"createdAt":"2024-08-23T14:03:06Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","Python","TypeScript","智能体","开发框架"],"domainKey":"AI Agent","pushedAt":"2026-09-21T07:10:34Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/i-am-bee/beeai-framework","ts":"2026-09-21T07:10:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/maka","owner":"apache","name":"maka","desc":"Apache Maka (Incubating) is a high-performance agent workspace that keeps a complete record of everything it did.","summaryCn":"把所有AI代理的操作都自动记录，像个人助理的全程日志本。","reasonCn":"Maka提供高性能的代理工作区，基于TypeScript实现，能够实时捕获每一次指令、文件修改、网络请求等细节，并以结构化日志保存，便于调试、审计和复现。它支持插件扩展，可对接各类AI模型和工具链，实现跨语言、跨平台的自动化工作流。同时提供可视化面板，帮助用户直观查看历史记录并回滚到任意节点，极大提升团队协作效率。","stars":5615,"starGrowth":1,"createdAt":"2026-05-27T15:46:05Z","silentRounds":1,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","操作审计","日志系统","可视化回放","工作流追踪"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:21:25Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"trending","url":"https://github.com/apache/maka","ts":"2026-09-23T21:21:25Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sundowndev/phoneinfoga","owner":"sundowndev","name":"phoneinfoga","desc":"Information gathering framework for phone numbers","summaryCn":"号码侦探神器，开源OSINT工具扒出电话背后信息","reasonCn":"技术亮点包括：支持全球130+个国家的号码验证；通过Google搜索、Viber、Truecaller等数据源获取关联信息；批量扫描模式适合大规模测试；输出格式支持JSON、Markdown、纯文本等多种选项。框架模块化设计，可以轻松扩展新的信息源插件。","stars":17947,"starGrowth":7,"createdAt":"2018-10-25T09:19:47Z","silentRounds":0,"language":"Go","topics":["information-gathering","phone-number","osint","open-source-intelligence","reconnaissance","footprint","phone","investigation","reputation","agregation"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["OSINT","信息收集","安全工具"],"domainKey":"OSINT","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"information-gathering","source":"github","weight":0.7},{"name":"phone-number","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"open-source-intelligence","source":"github","weight":0.7},{"name":"reconnaissance","source":"github","weight":0.7},{"name":"footprint","source":"github","weight":0.7},{"name":"phone","source":"github","weight":0.7},{"name":"investigation","source":"github","weight":0.7},{"name":"reputation","source":"github","weight":0.7},{"name":"agregation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/sundowndev/phoneinfoga","ts":"2026-09-07T02:44:20.540Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.022170111996433945,"copyOk":false},{"repo":"imthenachoman/How-To-Secure-A-Linux-Server","owner":"imthenachoman","name":"How-To-Secure-A-Linux-Server","desc":"An evolving how-to guide for securing a Linux server.","summaryCn":"运维老炮总结的Linux服务器安全终极指南","reasonCn":"技术上，它不只是理论，而是实打实的配置命令和操作步骤。从防火墙iptables/ufw设置、SSH端口和密钥加固、文件权限最小化、自动安全更新，到入侵检测、日志审计、禁用无用服务，每个环节都有具体方案。相比零散的安全博客，它提供了系统化防护思路。","stars":31459,"starGrowth":4,"createdAt":"2019-02-09T06:44:37Z","silentRounds":0,"language":"","topics":["security"],"aiDims":["安全工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["Linux安全","服务器运维","系统安全","网络安全"],"domainKey":"Linux安全","pushedAt":"2026-09-07T20:43:13Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7}],"aiScore":0.35,"source":"search","url":"https://github.com/imthenachoman/How-To-Secure-A-Linux-Server","ts":"2026-09-07T20:43:13Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"XIU2/CloudflareSpeedTest","owner":"XIU2","name":"CloudflareSpeedTest","desc":"🌩「自选优选 IP」测试 Cloudflare CDN 延迟和速度，获取最快 IP ！当然也支持其他 CDN / 多个解析 IP 的网站 ~","summaryCn":"免费测速找到最快CF IP，科学上网冲浪提速神器","reasonCn":"这是一款用Go编写的开源网络测速工具，专门用于批量测试Cloudflare CDN各IP的延迟和传输速度。用户只需提供IP列表或域名，程序会自动并发检测所有候选IP，按速度排序输出最优结果。支持自定义超时、并发数等参数，也兼容其他CDN服务。项目近3万star，社区活跃，持续迭代，是网络爱好者提升访问体验的实用利器。","stars":29123,"starGrowth":6,"createdAt":"2020-08-30T14:24:52Z","silentRounds":0,"language":"Go","topics":["cloudflare","cdn","speedtest","golang","go"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["网络测速","Cloudflare","Go语言"],"domainKey":"网络测速","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"cloudflare","source":"github","weight":0.7},{"name":"cdn","source":"github","weight":0.7},{"name":"speedtest","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/XIU2/CloudflareSpeedTest","ts":"2026-09-07T02:44:20.540Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018618068193524018,"copyOk":false},{"repo":"RediSearch/RediSearch","owner":"RediSearch","name":"RediSearch","desc":"A query and indexing engine for Redis, providing secondary indexing, full-text search, vector similarity search and aggregations.","summaryCn":"Redis的智能搜索插件，全文向量都精通，让数据库也能做语义检索","reasonCn":"技术上，RediSearch提供了三大核心能力：一是全文搜索，支持分词、模糊匹配、短语搜索；二是向量相似性搜索，支持HNSW等算法，可以存embedding做语义检索；三是聚合框架，支持分组、过滤、排序等操作。它作为Redis模块运行，不需要独立部署，数据存储在Redis内存中，读取速度极快。","stars":6241,"starGrowth":0,"createdAt":"2016-05-05T13:25:43Z","silentRounds":4,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI搜索"],"zone":"工具","funScore":0,"domainTags":["Redis","向量搜索","全文检索","HNSW"],"domainKey":"Redis","pushedAt":"2026-09-23T21:49:09Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/RediSearch/RediSearch","ts":"2026-09-23T21:49:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vitali87/code-graph-rag","owner":"vitali87","name":"code-graph-rag","desc":"The ultimate RAG for your monorepo. Query, understand, and edit multi-language codebases with the power of AI and knowledge graphs","summaryCn":"用AI和知识图谱给你的多语言代码库装上超级大脑","reasonCn":"该项目利用知识图谱技术将代码库的调用关系、文件依赖、函数结构等信息组织成可视化图谱，结合RAG检索增强生成，让开发者能用自然语言查询和编辑复杂的monorepo。支持Python、TypeScript、Go等多种语言，能自动解析代码结构并建立实体间关联，回答代码理解类问题的准确率远高于传统搜索引擎。","stars":5169,"starGrowth":3,"createdAt":"2025-06-16T14:31:57Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["RAG与知识","知识图谱","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["代码理解","知识图谱","RAG","多语言代码库","Monorepo"],"domainKey":"知识图谱","pushedAt":"2026-09-23T08:33:14Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/vitali87/code-graph-rag","ts":"2026-09-23T08:33:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"phoenixframework/phoenix","owner":"phoenixframework","name":"phoenix","desc":"Peace of mind from prototype to production","summaryCn":"Elixir写的Web框架，让你轻松构建高并发应用，稳如老狗","reasonCn":"Phoenix是Elixir语言的明星Web框架，依托BEAM虚拟机实现超强并发处理能力。它提供脚手架工具快速搭建项目，内置LiveView支持实时交互无需写JavaScript。采用函数式编程，代码简洁易维护，非常适合构建高可用、低延迟的生产级应用。","stars":23168,"starGrowth":1,"createdAt":"2014-01-20T14:14:11Z","silentRounds":3,"language":"Elixir","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Web框架","Elixir","BEAM虚拟机","高并发"],"domainKey":"Web框架","pushedAt":"2026-09-17T21:41:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Elixir","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/phoenixframework/phoenix","ts":"2026-09-17T21:41:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Bitwise-01/Instagram-","owner":"Bitwise-01","name":"Instagram-","desc":"","summaryCn":"作为一款明确标注用于账号暴力破解的自动化工具，它体现了极其硬核的底层网","reasonCn":"该工具的运行环境需要Python 3.9版本，并强制要求用户准备一份格式化的代理IP池。在完成依赖项安装后，程序会将上传的代理保存至内部数据库，以备调用。安装步骤包括创建隔离的虚拟环境、安装必备的Pipenv工具，以及通过 pip 安装项目依赖，整个过程对于熟悉命令行操作的开发者而言","stars":5139,"starGrowth":2,"createdAt":"2017-01-29T12:34:06Z","silentRounds":0,"language":"Python","topics":["instagram","hack","bruteforce"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0.2,"domainTags":["Instagram","自动化脚本","暴力破解"],"domainKey":"Instagram","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"instagram","source":"github","weight":0.7},{"name":"hack","source":"github","weight":0.7},{"name":"bruteforce","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/Bitwise-01/Instagram-","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MetaCubeX/ClashMetaForAndroid","owner":"MetaCubeX","name":"ClashMetaForAndroid","desc":"A rule-based tunnel for Android.","summaryCn":"安卓神器级代理客户端，规则驱动网络 tunnel 全家桶","reasonCn":"基于Clash内核的Android代理客户端，采用规则驱动的智能隧道技术。支持自定义规则集、DNS配置、多节点管理、后台保活等功能。Kotlin开发，界面现代简洁，提供Meta增强版特性。适合安卓用户管理代理连接，支持订阅配置自动更新，规则匹配精准高效。","stars":46578,"starGrowth":23,"createdAt":"2022-06-07T06:38:00Z","silentRounds":0,"language":"Kotlin","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网络代理","Android应用","规则引擎"],"domainKey":"网络代理","pushedAt":"2026-08-17T23:52:44Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/MetaCubeX/ClashMetaForAndroid","ts":"2026-08-17T23:52:44Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07048138606208941,"copyOk":false},{"repo":"nextlevelbuilder/ui-ux-pro-max-skill","owner":"nextlevelbuilder","name":"ui-ux-pro-max-skill","desc":"An AI SKILL that provide design intelligence for building professional UI/UX multiple platforms","summaryCn":"给程序员装个设计总监，AI帮你搞定专业UI界面","reasonCn":"这是专为Claude等AI打造的UI/UX设计提示词增强Skill。它内置海量设计规范、组件库和最佳实践，让AI生成的界面告别简陋。开发者只需对话，即可获得媲美专业设计师的界面方案，大幅降低前端开发门槛，提升UI质量与效率。","stars":129876,"starGrowth":242,"createdAt":"2025-11-30T11:36:31Z","silentRounds":0,"language":"Python","topics":["command-line"],"aiDims":["提示工程","AI应用","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["UI设计","提示词工程","AI编程","设计规范"],"domainKey":"UI设计","pushedAt":"2026-09-21T13:01:10Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/nextlevelbuilder/ui-ux-pro-max-skill","ts":"2026-09-21T13:01:10Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3667079368012566,"copyOk":true},{"repo":"pixeltable/pixeltable","owner":"pixeltable","name":"pixeltable","desc":"Unified multimodal backend for AI data apps","summaryCn":"一个让多模态数据自动向量化检索的AI数据底座","reasonCn":"pixeltable是一个面向AI数据应用的统一多模态后端，支持图像、文本、音频等数据的声明式管理。内置向量化能力，可对接OpenAI、Anthropic等主流模型，通过简单API完成数据加载、搜索和RAG构建，让开发者专注业务逻辑而非数据基建。","stars":1624,"starGrowth":0,"createdAt":"2023-05-10T18:03:02Z","silentRounds":25,"language":"Python","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","多模态Agent"],"zone":"工具","funScore":0,"domainTags":["Python","Vector Database","Multimodal","Data Pipeline","RAG"],"domainKey":"Python","pushedAt":"2026-09-23T21:26:44Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/pixeltable/pixeltable","ts":"2026-09-23T21:26:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deckarep/golang-set","owner":"deckarep","name":"golang-set","desc":"","summaryCn":"Go语言通用的集合数据结构实现，提供类似哈希表的高效增删查功能。","reasonCn":"Go语言标准库中缺少通用的泛型集合（Set）实现，该项目填补了这一空白。它提供了类型无关的集合数据结构，支持快速添加、删除和元素检查操作，性能经过优化以满足大规模数据场景需求。新版本中移除了对MongoDB BSON驱动的外部依赖，减少了依赖树复杂度，同时增加了IsDisjoint等实用方法。适合在Go项目中需要高效去重或成员资格判断场景下使用，是替代map[string]bool实现集合语义的标准选择。","stars":4748,"starGrowth":2,"createdAt":"2013-07-03T21:52:01Z","silentRounds":0,"language":"Go","topics":["go","set","threadsafe","datastructures","concurrency","golang","generics"],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Go语言","数据结构","集合库","开源组件"],"domainKey":"Go语言","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"set","source":"github","weight":0.7},{"name":"threadsafe","source":"github","weight":0.7},{"name":"datastructures","source":"github","weight":0.7},{"name":"concurrency","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"generics","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/deckarep/golang-set","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ikuaitu/vue-fabric-editor","owner":"ikuaitu","name":"vue-fabric-editor","desc":"快图设计-基于fabric.js和Vue的开源图片编辑器，可自定义字体、素材、设计模板。fabric.js and Vue based image editor, can customize fonts, materials, design templates.","summaryCn":"拖拽就能做海报banner的在线设计神器","reasonCn":"基于Vue 3和fabric.js打造的可视化编辑器，支持拖拽添加文字、图片、图形等元素，内置丰富设计模板和素材库。可自定义字体、颜色、布局，导出PNG/SVG等格式。项目架构清晰，组件化设计，适合快速集成到项目中作为嵌入式设计工具使用。","stars":7975,"starGrowth":3,"createdAt":"2022-09-03T16:43:33Z","silentRounds":0,"language":"Vue","topics":["design"],"aiDims":["创意工具","可视化"],"zone":"工具","funScore":0,"domainTags":["图片编辑","海报设计","fabric.js","Vue","可视化编辑器"],"domainKey":"图片编辑","pushedAt":"2026-07-21T09:03:00Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/ikuaitu/vue-fabric-editor","ts":"2026-07-21T09:03:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"github/spec-kit","owner":"github","name":"spec-kit","desc":"💫 Toolkit to help you get started with Spec-Driven Development","summaryCn":"用自然语言写需求规格，AI帮你自动生成代码和文档的神兵利器","reasonCn":"spec-kit是GitHub官方推出的规范驱动开发工具包，帮助你用自然语言描述需求规格，然后自动生成代码、文档和测试。它支持多种编程语言和框架，提供模板系统和代码生成引擎。相比传统开发方式，spec-driven开发能让非技术人员也能参与需求定义，减少沟通成本，提高开发效率。它的核心理念是把需求文档变成可执行的规范。","stars":138405,"starGrowth":97,"createdAt":"2025-08-21T22:54:31Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["代码助手","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["规范驱动开发","代码生成","需求管理","AI辅助开发"],"domainKey":"规范驱动开发","pushedAt":"2026-09-23T21:39:44Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/github/spec-kit","ts":"2026-09-23T21:39:44Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.29310252519778907,"copyOk":true},{"repo":"google/osv-scanner","owner":"google","name":"osv-scanner","desc":"Vulnerability scanner written in Go which uses the data provided by https://osv.dev","summaryCn":"一键扫描项目依赖，找出存在安全漏洞的包，让代码远离黑客攻击。","reasonCn":"Go语言编写的开源漏洞扫描器，从OSV全球漏洞数据库中获取数据，自动检测项目依赖包是否存在已知安全问题。支持扫描Go、Python、npm、Maven等多种语言的依赖文件，可生成JSON或HTML格式的详细报告，方便集成到CI/CD流水线中自动化检测。","stars":11074,"starGrowth":2,"createdAt":"2022-11-14T01:05:20Z","silentRounds":0,"language":"Go","topics":["scanner","security-audit","security-tools","vulnerability-scanner"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["漏洞扫描","安全审计","依赖管理"],"domainKey":"漏洞扫描","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"scanner","source":"github","weight":0.7},{"name":"security-audit","source":"github","weight":0.7},{"name":"security-tools","source":"github","weight":0.7},{"name":"vulnerability-scanner","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/google/osv-scanner","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"MobiFlight/MobiFlight-Connector","owner":"MobiFlight","name":"MobiFlight-Connector","desc":"MobiFlight is an open source project that allows you to create your own home cockpit for your favorite flight simulator in a flexible, affordable and extremely user-friendly way.","summaryCn":"把Arduino变成硬核飞行模拟舱，低成本DIY神器","reasonCn":"其核心是一个C#编写的通用接口程序，运行在PC上，通过串口与Arduino通信。它具备极其强大的“映射”功能：你可以定义“当引擎启动信号灯亮起时，点亮我的LED”或“当襟翼收起时，敲响我的步进电机”。它支持多种硬件协议，包括RS485、GPIO和模拟输出，甚至能控制复杂的数码管显示高度和速","stars":357,"starGrowth":0,"createdAt":"2021-05-28T14:38:10Z","silentRounds":10,"language":"C#","topics":["arduino"],"aiDims":["非AI-好玩","效率工具","开发者工具"],"zone":"创意","funScore":1,"domainTags":["模拟飞行","Arduino","硬件映射"],"domainKey":"模拟飞行","pushedAt":"2026-09-23T13:32:30Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/MobiFlight/MobiFlight-Connector","ts":"2026-09-23T13:32:30Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"javierbyte/visual-center","owner":"javierbyte","name":"visual-center","desc":"Tool to better align logos in the center of a container.","summaryCn":"专治Logo居中强迫症的在线工具，一键搞定排版。","reasonCn":"这是一个纯前端运行的Logo对齐工具，旨在解决网页设计中Logo视觉重心不一致的常见痛点。它支持水平垂直居中、边缘对齐等多种对齐模式，并能实时预览调整后的效果。用户无需安装任何软件，直接在浏览器中拖拽文件即可使用，极大提升了设计效率。","stars":1450,"starGrowth":0,"createdAt":"2016-05-12T06:51:39Z","silentRounds":24,"language":"JavaScript","topics":["design","frontend"],"aiDims":["非AI-实用","创意工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Logo居中","前端工具","对齐工具","浏览器端","设计"],"domainKey":"Logo居中","pushedAt":"2026-08-23T21:24:48Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/javierbyte/visual-center","ts":"2026-08-23T21:24:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"phaserjs/phaser","owner":"phaserjs","name":"phaser","desc":"Phaser is a fun, free and fast 2D game framework for making HTML5 games for desktop and mobile web browsers, supporting Canvas and WebGL rendering.","summaryCn":"用JavaScript写2D游戏的万能框架，浏览器就能跑","reasonCn":"Phaser是业界最流行的2D HTML5游戏框架，基于JavaScript开发。支持Canvas和WebGL双渲染模式，内置物理引擎、动画系统、输入处理、音频管理等全套功能。跨平台支持桌面和移动端浏览器，也兼容Discord活动、Facebook小游戏等平台。","stars":40347,"starGrowth":1,"createdAt":"2013-04-12T12:27:51Z","silentRounds":2,"language":"JavaScript","topics":["canvas","discord-activities","facebook-instant-games","game-development","game-frameworks","gamedev","html5-game-development","javascript","phaser","phaser-development","phaserjs","webgl","youtube-playables"],"aiDims":["游戏","创意工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["游戏开发","HTML5","2D游戏","JavaScript"],"domainKey":"游戏开发","pushedAt":"2026-08-21T06:56:13Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"canvas","source":"github","weight":0.7},{"name":"discord-activities","source":"github","weight":0.7},{"name":"facebook-instant-games","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"game-frameworks","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"html5-game-development","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"phaser","source":"github","weight":0.7},{"name":"phaser-development","source":"github","weight":0.7},{"name":"phaserjs","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"youtube-playables","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"bigbro","url":"https://github.com/phaserjs/phaser","ts":"2026-08-21T06:56:13Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"facebookarchive/Tweaks","owner":"facebookarchive","name":"Tweaks","desc":"","summaryCn":"Tweaks是Facebook为其iOS应用团队开发并开源的一套参数微调系统，旨在解决应","reasonCn":"该工具的核心工作流非常贴合日常开发习惯。开发者首先定义好可调整的参数及其范围，然后在运行中的应用中通过特定手势或UI元素打开调节面板。在这里，可以拖动滑块或输入具体数值来改变参数，并立即看到界面或行为的反馈。","stars":4701,"starGrowth":0,"createdAt":"2014-03-14T02:58:52Z","silentRounds":5,"language":"Objective-C","topics":[],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["iOS开发","参数调试","应用优化","动画调参"],"domainKey":"iOS开发","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Objective-C","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/facebookarchive/Tweaks","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"openai-php/client","owner":"openai-php","name":"client","desc":"⚡️ OpenAI PHP is a supercharged community-maintained PHP API client that allows you to interact with OpenAI API.","summaryCn":"有了它，PHP程序员也能轻松接入GPT大模型","reasonCn":"技术上，它封装了OpenAI的所有主流API——Chat Completions、Images、Audio、Fine-tuning、Embeddings。支持PHP 8.0+，充分利用了现代PHP的类型系统，IDE补全体验很好。核心亮点包括：函数调用（让模型根据你的函数定义生成调用参数，实现Agent效果）、流式输出（逐字实时返回，适合","stars":5834,"starGrowth":2,"createdAt":"2022-09-25T22:08:39Z","silentRounds":0,"language":"PHP","topics":["api"],"aiDims":["AI应用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["OpenAI API","PHP","AI集成","函数调用","SDK"],"domainKey":"OpenAI API","pushedAt":"2026-09-23T15:15:40Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/openai-php/client","ts":"2026-09-23T15:15:40Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lutzroeder/netron","owner":"lutzroeder","name":"netron","desc":"Visualizer for neural network, deep learning and machine learning models","summaryCn":"神经网络模型可视化神器，架构图一眼看透细节","reasonCn":"Netron 是 AI 领域必备的可视化工具，支持打开各种主流格式的机器学习模型文件。无论是 PyTorch、TensorFlow、ONNX 还是 CoreML，都能生成清晰易懂的架构图。它可以深入查看每一层的参数和结构，对于调试模型和理解复杂神经网络架构非常有帮助，界面简洁且跨平台。","stars":33508,"starGrowth":3,"createdAt":"2010-12-26T12:53:43Z","silentRounds":0,"language":"JavaScript","topics":["ml","ai"],"aiDims":["AI应用","可视化"],"zone":"工具","funScore":0,"domainTags":["模型可视化","神经网络","深度学习","工具","架构图"],"domainKey":"模型可视化","pushedAt":"2026-09-22T12:03:21Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/lutzroeder/netron","ts":"2026-09-22T12:03:21Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dzharii/awesome-typescript","owner":"dzharii","name":"awesome-typescript","desc":"","summaryCn":"一个沉淀十一年的 TypeScript 资源精选指南，涵盖生态全貌。","reasonCn":"项目涵盖了广泛的资源维度，包括基础教程、高级类型技巧以及各种实用的第三方库。它解决了开发者在面对海量碎片化信息时难以筛选高质量学习路径的问题，通过人工策展的方式，将那些经过时间检验的坚实素材汇总在一起，让后来的学习者不必重复经历早期的摸索阶段。","stars":5115,"starGrowth":0,"createdAt":"2015-06-06T22:00:48Z","silentRounds":5,"language":"","topics":[],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0.2,"domainTags":["TypeScript","编程资源","技术清单"],"domainKey":"TypeScript","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1}],"aiScore":0.26,"source":"search","url":"https://github.com/dzharii/awesome-typescript","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/cameraview","owner":"google","name":"cameraview","desc":"","summaryCn":"曾用于简化 Android 应用相机功能集成的开发库（现已废弃）。","reasonCn":"这曾是一个旨在帮助安卓开发者轻松集成相机功能的开源辅助库，能够通过简单的布局配置实现预览、闪光灯和自动对焦。它在底层根据不同的系统 API 版本自动切换调用策略，从而简化了早期安卓相机开发的适配难度。不过需要注意的是，该项目目前已经正式宣布废弃并不再更新，官方建议广大开发者迁移到更新的 Jetpack CameraX 框架。","stars":4702,"starGrowth":0,"createdAt":"2016-08-04T05:22:14Z","silentRounds":7,"language":"Java","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Android开发","相机集成","UI布局","自动对焦","API适配"],"domainKey":"Android开发","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/google/cameraview","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"nicholas-fedor/watchtower","owner":"nicholas-fedor","name":"watchtower","desc":"Automate Docker container image updates","summaryCn":"你的Docker容器有了自动续费功能，镜像更新不用你操心","reasonCn":"Watchtower是一个Docker容器自动化更新工具，能监控指定容器使用的镜像版本，一旦发现Docker Hub或其他镜像仓库有更新，就会自动拉取最新镜像并重建容器，支持多种触发方式和通知回调，大幅降低容器化服务的手动维护成本。","stars":4482,"starGrowth":2,"createdAt":"2025-01-16T23:29:36Z","silentRounds":0,"language":"Go","topics":["community-project","docker","go","golang"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Docker","容器管理","自动化","DevOps"],"domainKey":"Docker","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"community-project","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/nicholas-fedor/watchtower","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"getmaxun/maxun","owner":"getmaxun","name":"maxun","desc":"🔥 The open-source no-code platform for web scraping, crawling, search and AI data extraction • Turn websites into structured APIs in minutes 🔥 ","summaryCn":"不用写一行代码就能抓取网页数据，AI帮你自动填表导数据，懒人神器","reasonCn":"它是个开源的无代码爬虫平台，能在浏览器里录制你的操作，自动识别按钮和输入框。更厉害的是加上了AI能力，能从复杂网页里自动抽取结构化的JSON数据。把网页变成API只需几分钟，不用写正则或选择器，对小白特别友好。","stars":17539,"starGrowth":6,"createdAt":"2023-10-23T21:40:19Z","silentRounds":0,"language":"TypeScript","topics":["automation","agents"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.8,"domainTags":["网页爬虫","数据提取","无代码"],"domainKey":"网页爬虫","pushedAt":"2026-09-21T17:18:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/getmaxun/maxun","ts":"2026-09-21T17:18:32Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.019026284280745764,"copyOk":true},{"repo":"Desert-Ant-Labs/demo-clipper","owner":"Desert-Ant-Labs","name":"demo-clipper","desc":"Example macOS app and CLI to generate short clips, fully on device.","summaryCn":"苹果电脑上随时截取并生成精彩短视频的轻量级桌面命令行神器","reasonCn":"利用原生macOS框架与命令行工具实现的本地视频剪辑小应用，能够快速高效地将长视频或播客音频切片导出。整个处理过程完全在设备本地离线运行，无需上传云端服务器，兼顾了极高的操作速度与隐私安全性，非常适合内容创作者日常使用。","stars":24,"starGrowth":0,"createdAt":"2026-08-09T08:02:09Z","silentRounds":2,"language":"Swift","topics":["podcast"],"aiDims":["效率工具","非AI-实用","创意工具"],"zone":"工具","funScore":0.6,"domainTags":["视频剪辑","本地处理","macOS","CLI工具"],"domainKey":"视频剪辑","pushedAt":"2026-09-23T12:28:50Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/Desert-Ant-Labs/demo-clipper","ts":"2026-09-23T12:28:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GreptimeTeam/greptimedb","owner":"GreptimeTeam","name":"greptimedb","desc":"The open-source observability database. One columnar engine for metrics, logs, and traces, on object storage.","summaryCn":"一个数据库同时存指标日志追踪，基于对象存储，可观测性神器","reasonCn":"GreptimeDB是开源的可观测性数据库，一套引擎同时处理metrics、logs、traces三种数据，基于对象存储实现低成本高扩展。Rust编写性能出色，兼容Prometheus/InfluxDB/MySQL协议，是监控系统、APM、日志平台的优秀存储底座。","stars":6703,"starGrowth":3,"createdAt":"2022-04-11T03:46:48Z","silentRounds":0,"language":"Rust","topics":["cloud-native","rust","metrics","logs","observability","traces","promql","observability-database","greptimedb","object-storage","opentelemetry","prometheus","ai-agents","distributed-database","high-cardinality","self-hosted","wide-events"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["时序数据库","可观测性","对象存储"],"domainKey":"时序数据库","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cloud-native","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"metrics","source":"github","weight":0.7},{"name":"logs","source":"github","weight":0.7},{"name":"observability","source":"github","weight":0.7},{"name":"traces","source":"github","weight":0.7},{"name":"promql","source":"github","weight":0.7},{"name":"observability-database","source":"github","weight":0.7},{"name":"greptimedb","source":"github","weight":0.7},{"name":"object-storage","source":"github","weight":0.7},{"name":"opentelemetry","source":"github","weight":0.7},{"name":"prometheus","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"distributed-database","source":"github","weight":0.7},{"name":"high-cardinality","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"wide-events","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/GreptimeTeam/greptimedb","ts":"2026-09-07T02:44:20.540Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"a-anjos/python-opencv","owner":"a-anjos","name":"python-opencv","desc":"Cheat sheet for OpenCV 4.x (Python)","summaryCn":"OpenCV速查表，程序员写代码时的救命稻草","reasonCn":"在技术架构上，该项目纯粹由LaTeX源码构成，没有任何动态网页或脚本依赖。这意味着它不仅加载速度极快，而且可以生成任意格式的文档，无论是PDF还是HTML，都能完美呈现。它梳理了OpenCV Python接口中最常用的几百个函数，从基础的图像读取、灰度转换，到复杂的边缘检测、特征匹配，","stars":58,"starGrowth":0,"createdAt":"2019-12-13T14:57:46Z","silentRounds":11,"language":"TeX","topics":["cheatsheet"],"aiDims":["开发者工具","效率工具","学习资源"],"zone":"资源","funScore":0,"domainTags":["OpenCV","计算机视觉","速查表","Python","图像处理"],"domainKey":"OpenCV","pushedAt":"2026-08-14T13:29:05Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"TeX","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/a-anjos/python-opencv","ts":"2026-08-14T13:29:05Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"infinitered/reactotron","owner":"infinitered","name":"reactotron","desc":"A desktop app for inspecting your React JS and React Native projects. macOS, Linux, and Windows.","summaryCn":"React开发者的桌面调试神器，帮你一眼看穿代码问题。","reasonCn":"Reactotron是一款专为React和React Native设计的桌面调试客户端，支持macOS、Linux和Windows三大平台。它允许开发者实时监控Redux状态、查看网络请求、调试JavaScript错误以及监控设备性能，提供类似Chrome DevTools的直观界面，让调试过程更加高效便捷。","stars":15590,"starGrowth":0,"createdAt":"2016-04-15T21:58:32Z","silentRounds":11,"language":"TypeScript","topics":["desktop"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["React","React Native","调试工具","状态管理","前端开发"],"domainKey":"React","pushedAt":"2026-08-13T22:50:44Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/infinitered/reactotron","ts":"2026-08-13T22:50:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"genkit-ai/genkit","owner":"genkit-ai","name":"genkit","desc":"Open-source framework for building agentic apps in JavaScript, Go, Dart, and Python, built and used in production by Google","summaryCn":"谷歌亲儿子框架，让写AI Agent应用像搭积木一样简单顺手","reasonCn":"由Google开源的AI应用开发框架，支持多语言（TS/Go/Dart/Python）。专为构建Agentic AI应用设计，内置了流程编排、工具调用、记忆管理和LLM集成功能。对于前端或后端开发者来说，这是快速构建生产级AI功能的高效工具。","stars":6454,"starGrowth":2,"createdAt":"2024-04-29T22:04:42Z","silentRounds":0,"language":"TypeScript","topics":["vector-db","rag","agents"],"aiDims":["AI基础设施","AI Agent","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","应用框架","流程编排","多语言","工具调用"],"domainKey":"AI Agent","pushedAt":"2026-09-23T21:19:47Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/genkit-ai/genkit","ts":"2026-09-23T21:19:47Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/doris","owner":"apache","name":"doris","desc":"Apache Doris is a real-time analytics and hybrid search database for AI agents.","summaryCn":"让AI Agent秒查海量数据瞬间做决策的数据库神器","reasonCn":"Apache Doris是Apache顶级实时分析数据库，专为AI Agent打造。支持SQL查询和向量检索，毫秒级响应海量数据。兼容MySQL协议，部署简单。让AI快速获取知识做决策，而非只会聊天。社区活跃star 1.5万+，是AI基础设施明星项目。","stars":15986,"starGrowth":4,"createdAt":"2017-08-10T12:13:30Z","silentRounds":0,"language":"Java","topics":["database"],"aiDims":["AI基础设施","AI Agent","AI搜索"],"zone":"工具","funScore":0,"domainTags":["OLAP","向量检索","实时分析","分布式数据库"],"domainKey":"OLAP","pushedAt":"2026-09-23T16:03:18Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/apache/doris","ts":"2026-09-23T16:03:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"spine/spine","owner":"spine","name":"spine","desc":"Lightweight MVC library for building JavaScript applications","summaryCn":"轻量级JavaScript MVC框架快速构建网页应用","reasonCn":"这是一个轻量级的JavaScript MVC框架，帮助开发者用Model-View-Controller模式构建前端应用。核心特点是小巧易用，依赖少，学习成本低。虽然社区相对小众，但在简单单页应用和教学场景中仍有其价值，适合喜欢简洁架构的开发者。\n","stars":3694,"starGrowth":0,"createdAt":"2011-03-29T06:34:32Z","silentRounds":25,"language":"JavaScript","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript","前端框架","MVC架构"],"domainKey":"JavaScript","pushedAt":"2020-04-04T23:20:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"bigbro","url":"https://github.com/spine/spine","ts":"2020-04-04T23:20:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"GameServerManagers/LinuxGSM","owner":"GameServerManagers","name":"LinuxGSM","desc":"The command-line tool for quick, simple deployment and management of Linux dedicated game servers.","summaryCn":"游戏服务器管理神器，一行命令秒部署《我的世界》《CS》等联机大厅","reasonCn":"LinuxGSM 是一个强大的命令行工具，能让任何人一键部署和管理各种热门游戏的专用服务器。虽然和 AI 无关，但它极致实用且好玩，特别适合想和朋友开黑、搭建独立游戏世界的玩家，让硬核的服务器运维变得像点外卖一样简单。","stars":4904,"starGrowth":0,"createdAt":"2013-08-09T12:37:43Z","silentRounds":7,"language":"Shell","topics":["game"],"aiDims":["非AI-好玩","游戏","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["Linux","游戏服务器","运维工具"],"domainKey":"Linux","pushedAt":"2026-08-07T19:24:41Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/GameServerManagers/LinuxGSM","ts":"2026-08-07T19:24:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/guava","owner":"google","name":"guava","desc":"Google core libraries for Java","summaryCn":"Java开发者的瑞士军刀，什么都有就是不用写","reasonCn":"Guava 是 Google 出品的 Java 核心工具库，提供了集合操作、缓存、并发工具、字符串处理、函数式编程等常用功能，是 Java 开发者最常用的第三方库之一。它填补了 Java 标准库的空白，让代码更简洁、更安全、更高效。几乎所有企业级 Java 项目都会依赖它。虽然不涉及 AI，但作为基础工具，它是构建任何 AI Java 应用的前提。","stars":51906,"starGrowth":0,"createdAt":"2014-05-29T16:23:17Z","silentRounds":5,"language":"Java","topics":["guava","java"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Java","工具库","缓存","函数式编程"],"domainKey":"Java","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"guava","source":"github","weight":0.7},{"name":"java","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/google/guava","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"tradesdontlie/tradingview-mcp","owner":"tradesdontlie","name":"tradingview-mcp","desc":"AI-assisted TradingView chart analysis — connect Claude Code to your TradingView Desktop for personal workflow automation","summaryCn":"把Claude变成你的量化交易助手，自动分析图表找买卖点。","reasonCn":"它的核心机制是MCP（Model Context Protocol），让Claude能“看懂”你电脑上的图表。你只需要告诉它“帮我分析一下AAPL最近30天的走势，找找看有没有双底形态”，它就会自动去TradingView抓取数据，然后分析出来，甚至能给出买卖建议。","stars":6620,"starGrowth":12,"createdAt":"2026-03-29T14:39:25Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["AI Agent","工具调用","效率工具"],"zone":"AI","funScore":0,"domainTags":["量化交易","技术分析","MCP协议","TradingView","Claude集成","金融科技"],"domainKey":"量化交易","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/tradesdontlie/tradingview-mcp","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.049725679758308154,"copyOk":false},{"repo":"linorobot/linorobot2","owner":"linorobot","name":"linorobot2","desc":"Autonomous mobile robots (2WD, 4WD, Mecanum Drive)","summaryCn":"如果说人工智能是AI的灵魂，那机器人就是它的身体。","reasonCn":"它的核心优势在于对ROS 2框架的完美兼容。它不仅支持复杂的麦轮运动控制，还集成了激光雷达测距和深度相机视觉接口。与市面上昂贵的商用底盘相比，它更加透明，每一行代码都向开发者开放，方便你自定义避障逻辑或是接入更高级的视觉模型。","stars":1026,"starGrowth":0,"createdAt":"2021-08-03T19:54:54Z","silentRounds":3,"language":"Python","topics":["robotics"],"aiDims":["非AI-好玩","AI应用"],"zone":"工具","funScore":0,"domainTags":["机器人","ROS","嵌入式"],"domainKey":"机器人","pushedAt":"2026-09-23T17:44:02Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/linorobot/linorobot2","ts":"2026-09-23T17:44:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rasadov/EcommerceAPI","owner":"rasadov","name":"EcommerceAPI","desc":"Modular e-commerce backend with a GraphQL gateway and gRPC microservices for accounts, products, orders, payments, and recommendations.","summaryCn":"一个用Go写的模块化电商后端，GraphQL+gRPC架构很现代","reasonCn":"这个项目是一个基于Go语言的模块化电商后端系统，采用GraphQL作为统一查询入口，后端拆分了多个gRPC微服务来分别处理账户、商品、订单、支付和推荐等核心业务。技术上采用了现代化的微服务架构设计，服务间通信效率高，代码结构清晰，对于想学习电商系统架构或需要快速搭建电商后端的开发者来说，是一个很好的参考项目。","stars":697,"starGrowth":0,"createdAt":"2025-01-10T13:02:51Z","silentRounds":24,"language":"Go","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["电商后端","GraphQL","gRPC","微服务","Go"],"domainKey":"电商后端","pushedAt":"2026-07-12T17:48:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/rasadov/EcommerceAPI","ts":"2026-07-12T17:48:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ovexro/dockpanel","owner":"ovexro","name":"dockpanel","desc":"Modern server management panel built with Rust and React. Sites, databases, Docker apps, Git deploy, mail, DNS, monitoring, backups, and security — all in one panel.","summaryCn":"Rust打造的现代化服务器管理面板，功能一站式","reasonCn":"dockpanel是一款基于Rust和React构建的现代化服务器管理面板，集成了网站托管、数据库管理、Docker应用部署、Git自动部署、邮件服务、DNS配置、监控告警、备份恢复和安全防护等功能。一个面板搞定所有运维需求，适合中小团队和个人开发者。","stars":968,"starGrowth":0,"createdAt":"2026-03-23T19:35:12Z","silentRounds":3,"language":"Rust","topics":["control-panel","devops","docker","hosting-panel","react","rust","self-hosted","server-management","vps","web-hosting"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Rust","服务器管理","Docker","运维工具"],"domainKey":"Rust","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"control-panel","source":"github","weight":0.7},{"name":"devops","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"hosting-panel","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"server-management","source":"github","weight":0.7},{"name":"vps","source":"github","weight":0.7},{"name":"web-hosting","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/ovexro/dockpanel","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tsrpc/tsrpc","owner":"tsrpc","name":"tsrpc","desc":"A TypeScript RPC framework, with runtime type checking and serialization, support both HTTP and WebSocket. It is very suitable for website / APP / games, and absolutely comfortable to full-stack TypeScript developers.","summaryCn":"写后端再也不用手写接口文档了，前端后端一套TS代码搞定通信","reasonCn":"基于TypeScript接口类型，自动生成序列化代码和接口文档。支持HTTP和WebSocket两种传输协议，前端定义接口，后端自动获得完整类型提示。消除了前后端联调的痛苦，特别适合游戏和全栈项目，让开发者告别手写API文档的繁琐。","stars":1935,"starGrowth":0,"createdAt":"2017-09-26T03:22:12Z","silentRounds":24,"language":"TypeScript","topics":["ajax","axios","backend-framework","express","framework","full-stack","full-stack-developer","full-stack-development","grpc","nestjs","node","nodejs","protobuf","rpc","runtime-type-checking","serialization-algorithm","server-framework","typescript","typescript-rpc-framework","typescript-serialization"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["RPC框架","TypeScript","全栈开发","WebSocket","API通信"],"domainKey":"RPC框架","pushedAt":"2026-02-26T19:16:11Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ajax","source":"github","weight":0.7},{"name":"axios","source":"github","weight":0.7},{"name":"backend-framework","source":"github","weight":0.7},{"name":"express","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"full-stack","source":"github","weight":0.7},{"name":"full-stack-developer","source":"github","weight":0.7},{"name":"full-stack-development","source":"github","weight":0.7},{"name":"grpc","source":"github","weight":0.7},{"name":"nestjs","source":"github","weight":0.7},{"name":"node","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"protobuf","source":"github","weight":0.7},{"name":"rpc","source":"github","weight":0.7},{"name":"runtime-type-checking","source":"github","weight":0.7},{"name":"serialization-algorithm","source":"github","weight":0.7},{"name":"server-framework","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"typescript-rpc-framework","source":"github","weight":0.7},{"name":"typescript-serialization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.4,"source":"bigbro","url":"https://github.com/tsrpc/tsrpc","ts":"2026-02-26T19:16:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Dicklesworthstone/destructive_command_guard","owner":"Dicklesworthstone","name":"destructive_command_guard","desc":"The Destructive Command Guard (dcg) is for blocking dangerous git and shell commands from being executed by agents.","summaryCn":"给AI Agent装的刹车系统，防止它乱删代码库闯祸","reasonCn":"这是一个专门保护 AI Agent 执行安全的小工具，基于 Rust 编写，性能优秀。它可以拦截并阻止危险的 Git 和 Shell 命令，比如 rm -rf、git push --force 这类可能摧毁代码库的致命操作。通过配置文件定义规则，灵活控制哪些命令允许执行、哪些需要警告或拒绝。","stars":6043,"starGrowth":5,"createdAt":"2026-01-07T22:27:34Z","silentRounds":0,"language":"Rust","topics":["developer-tools"],"aiDims":["AI Agent","安全工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["AI安全","命令拦截","Agent防护","Rust工具"],"domainKey":"AI安全","pushedAt":"2026-09-23T18:34:53Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/Dicklesworthstone/destructive_command_guard","ts":"2026-09-23T18:34:53Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02097865298692702,"copyOk":false},{"repo":"hendurhance/ui-ux","owner":"hendurhance","name":"ui-ux","desc":"📚 This guide is designed to help you learn UI/UX design, and is divided into three levels: Beginner, Intermediate, and Expert. It includes learning resource, guides and tools that cover all aspects of designing user interfaces and user experiences.","summaryCn":"想成为顶尖设计师却不知道从何下手？这份指南就像一位贴心的导师，从零开始","reasonCn":"想成为顶尖设计师却不知道从何下手？这份指南就像一位贴心的导师，从零开始带你领略UI/UX设计的奥秘。它不仅适合初学者快速入门，也能帮助资深设计师查漏补缺，巩固知识体系。内容涵盖了设计理论、用户研究、交互原则、视觉表现等核心模块，层层递进，逻辑清晰。","stars":1437,"starGrowth":1,"createdAt":"2021-06-28T13:32:57Z","silentRounds":4,"language":"","topics":["design"],"aiDims":["学习资源","教程","设计"],"zone":"资源","funScore":0,"domainTags":["UI设计","用户体验","设计理论"],"domainKey":"UI设计","pushedAt":"2026-08-14T00:18:35Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"设计","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7}],"aiScore":0.4,"source":"search","url":"https://github.com/hendurhance/ui-ux","ts":"2026-08-14T00:18:35Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"borski/travel-hacking-toolkit","owner":"borski","name":"travel-hacking-toolkit","desc":"AI-powered travel hacking and search with cash, points, miles, and award flights. Drop-in skills and MCP servers for Claude, Codex, and OpenCode.","summaryCn":"用AI帮你薅羊毛，把里程点数变成机票的神奇工具","reasonCn":"这是一个专为旅行达人和积分党打造的AI辅助工具箱，通过MCP协议连接Claude等大模型。它能自动搜索航线、分析里程价值、计算最佳兑换方案，还能直接把搜索结果喂给AI Agent进行决策。支持现金、里程、积分混合支付，让AI帮你像老司机一样“飞”遍全球。","stars":678,"starGrowth":2,"createdAt":"2026-03-21T06:40:36Z","silentRounds":0,"language":"Python","topics":["travel"],"aiDims":["AI Agent","工具调用","AI搜索"],"zone":"AI","funScore":0.8,"domainTags":["里程兑换","旅行","AI"],"domainKey":"里程兑换","pushedAt":"2026-09-21T11:42:24Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"travel","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/borski/travel-hacking-toolkit","ts":"2026-09-21T11:42:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"github/docs","owner":"github","name":"docs","desc":"The open-source repo for docs.github.com","summaryCn":"GitHub官方文档的开源仓库，教你怎么玩转GitHub这个代码平台","reasonCn":"这是GitHub官方文档网站的开源代码仓库，把docs.github.com的全部内容都开源了。它用Astro框架搭建，支持本地开发调试，通过Markdown编写文档内容，集成GitHub Actions自动部署。文档涵盖GitHub平台的功能指南、API参考、Actions工作流教程、Codespaces和GitHub Desktop等内容。项目结构规范，是学习技术文档写作和静态网站构建的绝佳参考案例。","stars":20872,"starGrowth":4,"createdAt":"2019-05-31T15:57:39Z","silentRounds":0,"language":"TypeScript","topics":["docs","works-with-codespaces"],"aiDims":["非AI-实用","教程"],"zone":"资源","funScore":0,"domainTags":["文档","静态网站","GitHub"],"domainKey":"文档","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"docs","source":"github","weight":0.7},{"name":"works-with-codespaces","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/github/docs","ts":"2026-09-05T02:38:58.868Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"BastiTee/tolino-notes","owner":"BastiTee","name":"tolino-notes","desc":"Convert Tolino notes into useful formats","summaryCn":"批量导出Tolino标注，自动生成TXT、CSV或Markdown。","reasonCn":"tolino-notes 基于 Python 实现，读取 Tolino 设备导出的 annot.zip 包，解析其中的 xml 结构，提取高亮文字、个人笔记和书签位置。随后根据用户配置的模板，可一次性生成纯文本、CSV 表格或 Markdown 文档，支持批量处理多本书籍，方便后续在文献管理软件或写作时引用。","stars":11,"starGrowth":0,"createdAt":"2023-08-19T14:37:45Z","silentRounds":12,"language":"Python","topics":["ebook"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["电子书标注","数据导出","文件转换"],"domainKey":"电子书标注","pushedAt":"2026-09-06T17:50:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/BastiTee/tolino-notes","ts":"2026-09-06T17:50:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SigNoz/signoz","owner":"SigNoz","name":"signoz","desc":"SigNoz is an open-source, OpenTelemetry-native observability platform for your team and their AI agents. Get logs, metrics, and traces in one tool with features like APM, distributed tracing, log management, infra monitoring, etc. Combined with SigNoz MCP and a native AI teammate (in SigNoz Cloud) it helps you build more resilient apps.","summaryCn":"给AI agent装个行车记录仪，日志指标追踪全搞定","reasonCn":"它基于OpenTelemetry打造，这是一个开源可观测性标准协议。SigNoz把日志、指标、追踪三种数据统一收集存储，用OpenTelemetry协议接入后，不需要写一堆埋点代码，自动采集。内置APM功能，能可视化调用链，追踪每个请求的完整路径。","stars":32170,"starGrowth":4,"createdAt":"2021-01-03T12:44:36Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted","monitoring"],"aiDims":["AI基础设施","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["可观测性","APM","分布式追踪","OpenTelemetry","日志监控"],"domainKey":"可观测性","pushedAt":"2026-09-23T20:23:27Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/SigNoz/signoz","ts":"2026-09-23T20:23:27Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"datawhalechina/zero-to-sglang","owner":"datawhalechina","name":"zero-to-sglang","desc":"面向大模型开发者的 SGLang 系统化开源教程：从推理基础与环境搭建开始，逐步学习模型部署、结构化生成、服务开发和性能优化，   结合实战案例带你从 0 到 1 掌握 SGLang，构建高性能 LLM 推理应用","summaryCn":"零起点 SGLang 教程，让你轻松搭建高速 LLM 推理服务","reasonCn":"SGLang 是一款高性能的低层大模型推理框架，支持多 GPU 并行、混合精度和自定义算子。该教程从环境搭建、基础语法到模型部署、微调、结构化生成，配合实战案例，让读者在 2 周内能从零跑通 70B 模型，并通过 API 实时推理。通过逐步拆解每个推理步骤，读者将了解张量并行、梯度检查点和推理缓存的原理。最终，你可以将 SGLang 与 Flask 或 FastAPI 对接，快速上线 Web 服务。","stars":1232,"starGrowth":11,"createdAt":"2026-08-27T02:24:12Z","silentRounds":0,"language":"Python","topics":["tutorial"],"aiDims":["大语言模型","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["LLM教程","推理优化","模型部署","开源课程","SGLang"],"domainKey":"LLM教程","pushedAt":"2026-09-20T06:34:56Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/datawhalechina/zero-to-sglang","ts":"2026-09-20T06:34:56Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.08967857142857144,"copyOk":false},{"repo":"CommunityToolkit/Maui","owner":"CommunityToolkit","name":"Maui","desc":"The .NET MAUI Community Toolkit is a community-created library that contains .NET MAUI Extensions, Advanced UI/UX Controls, and Behaviors to help make your life as a .NET MAUI developer easier","summaryCn":"微软MAUI开发的必备工具箱，控件行为全都有","reasonCn":"这是.NET MAUI跨平台移动应用开发的实用工具库，由社区贡献者维护。提供大量高级UI控件、行为Behaviors和特效组件，涵盖动画、手势识别、弹出层、输入验证等常用功能。无需从零造轮子，直接拿来就能用，显著提升开发效率，适合所有.NET MAUI开发者。","stars":2703,"starGrowth":0,"createdAt":"2021-06-02T19:02:20Z","silentRounds":5,"language":"C#","topics":["mobile","desktop"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["移动开发",".NET MAUI","UI 控件"],"domainKey":"移动开发","pushedAt":"2026-09-17T06:55:15Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/CommunityToolkit/Maui","ts":"2026-09-17T06:55:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ynqa/jnv","owner":"ynqa","name":"jnv","desc":"Interactive JSON filter using jq","summaryCn":"命令行里的JSON放大镜，随手一敲就能交互式筛选数据","reasonCn":"jnv 是一款用 Rust 打造的交互式 JSON 过滤工具，底层基于 jq，但提供了更友好的交互体验。它支持自动补全、语法高亮、历史命令记录，还能与 shell 管道无缝配合。运行后直接在终端里逐层浏览 JSON 结构，实时过滤数据，再也不用反复拼接复杂命令了。","stars":6116,"starGrowth":0,"createdAt":"2024-03-18T13:33:43Z","silentRounds":3,"language":"Rust","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["JSON","命令行","数据过滤"],"domainKey":"JSON","pushedAt":"2026-09-17T02:40:15Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/ynqa/jnv","ts":"2026-09-17T02:40:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"simple-icons/simple-icons","owner":"simple-icons","name":"simple-icons","desc":"SVG icons for popular brands","summaryCn":"几千个品牌官方SVG图标，前端开发者和设计师的免费素材库","reasonCn":"这个项目的设计理念是简洁和标准化。所有图标都来自各品牌的官方设计指南，颜色、比例、风格都经过严格审核，确保品牌标识的准确性。图标以SVG格式提供，支持无损缩放，可以随意调整大小和颜色而不失真。与PNG等其他格式相比，SVG文件更小、更清晰，在现代网页中几乎是唯一推荐使用","stars":25897,"starGrowth":5,"createdAt":"2012-11-16T11:09:13Z","silentRounds":0,"language":"JavaScript","topics":["design"],"aiDims":["非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["SVG图标","品牌标识","前端素材"],"domainKey":"SVG图标","pushedAt":"2026-09-20T04:25:37Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/simple-icons/simple-icons","ts":"2026-09-20T04:25:37Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015579217670000386,"copyOk":true},{"repo":"mosaico-labs/mosaico","owner":"mosaico-labs","name":"mosaico","desc":"Mosaico - The data platform for Physical AI","summaryCn":"机器人通常很呆板，给它设定什么路径就走什么路。","reasonCn":"该项目的核心亮点在于其架构的灵活性，通过一套高效的数据管道，将机器人的各种传感器信息（RGB、深度、点云）高效同步给AI模型。它支持多种主流的视觉语言模型（VLM）接入，能够把自然语言指令转化为机器人的动作指令，在处理多任务协同方面表现卓越，与传统的ROS（机器人操作系统","stars":1051,"starGrowth":1,"createdAt":"2025-12-09T13:28:50Z","silentRounds":3,"language":"Python","topics":["robotics"],"aiDims":["多模态Agent","机器人"],"zone":"AI","funScore":0,"domainTags":["机器人","ROS","多模态","传感器","物理AI"],"domainKey":"多模态","pushedAt":"2026-09-23T12:48:06Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"机器人","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/mosaico-labs/mosaico","ts":"2026-09-23T12:48:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nextcloud/server","owner":"nextcloud","name":"server","desc":"☁️ Nextcloud server, a safe home for all your data","summaryCn":"把数据牢牢攥在自己手里的自建云盘，安全可靠还能装各种好玩应用","reasonCn":"Nextcloud 是开源的私有云存储方案，让你把数据从 Google Drive 和 Dropbox 手中夺回来。支持文件同步、文档协作、日历 contacts、视频会议等。企业级加密、权限管理完善，兼容各种对象存储后端，生态插件丰富，还能作为 AI 应用的数据存储与知识库底座，支持 Docker 一键部署，上手门槛低。","stars":36889,"starGrowth":5,"createdAt":"2016-06-02T07:44:14Z","silentRounds":0,"language":"PHP","topics":["design"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["私有云","文件同步","协作平台"],"domainKey":"私有云","pushedAt":"2026-09-23T20:25:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/nextcloud/server","ts":"2026-09-23T20:25:34Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015406625281249153,"copyOk":false},{"repo":"TQDuysuke/kinx_lab_photoframe","owner":"TQDuysuke","name":"kinx_lab_photoframe","desc":"Công cụ web tự động gắn thông số EXIF và thương hiệu cá nhân lên ảnh","summaryCn":"一键给照片加EXIF信息和个人品牌标记，省时省力的在线小工具","reasonCn":"kinx_lab_photoframe 是基于 JavaScript 的网页工具，能够批量读取本地图片，自动写入或覆盖 EXIF 元数据，并叠加自定义品牌水印。用户只需上传照片和标志，即可生成统一标识的作品，省去手动编辑的繁琐，适合摄影师或电商快速加品牌。","stars":11,"starGrowth":0,"createdAt":"2026-03-07T15:17:26Z","silentRounds":11,"language":"JavaScript","topics":["photography"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["图片处理","EXIF","水印"],"domainKey":"图片处理","pushedAt":"2026-08-25T18:58:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"photography","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"search","url":"https://github.com/TQDuysuke/kinx_lab_photoframe","ts":"2026-08-25T18:58:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Mugen87/yuka","owner":"Mugen87","name":"yuka","desc":"JavaScript library for developing Game AI.","summaryCn":"给游戏角色装上大脑，让它们像人一样绕开障碍物","reasonCn":"在技术层面，Yuka基于ECS（实体-组件-系统）架构构建，这是一种非常高效的游戏开发模式。它内置了行为树（Behavior Trees）和GOAP（目标指向行动计划）两种主流的AI规划算法。行为树擅长处理确定性的步骤，比如“先检测血量，如果低就逃跑”；而GOAP则适合处理复杂的目标达成，比如“","stars":1369,"starGrowth":0,"createdAt":"2017-07-07T21:19:47Z","silentRounds":8,"language":"JavaScript","topics":["game-development"],"aiDims":["游戏","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["游戏开发","AI","行为树"],"domainKey":"游戏开发","pushedAt":"2026-09-05T07:10:18Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/Mugen87/yuka","ts":"2026-09-05T07:10:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"roots/bedrock","owner":"roots","name":"bedrock","desc":"","summaryCn":"Bedrock是一个将现代PHP工程理念注入WordPress的样板项目，它的核心目标是将","reasonCn":"该项目的设计哲学深受\"十二因子应用\"（Twelve-Factor App）方法论的启发。它重新组织了WordPress的默认目录结构，将应用代码、配置文件和运行时数据严格分离。例如，网站配置不再硬编码在config.php中，而是通过.env环境变量文件进行管理，使得本地开发环境与生产服务器环境的切换","stars":6573,"starGrowth":0,"createdAt":"2013-11-02T21:10:54Z","silentRounds":6,"language":"PHP","topics":["composer","wordpress","bedrock","wordpress-boilerplate","php","wordpress-starter","wp-composer","wordpress-composer","wp-packages"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["WordPress","样板项目","PHP开发","工程化"],"domainKey":"WordPress","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"composer","source":"github","weight":0.7},{"name":"wordpress","source":"github","weight":0.7},{"name":"bedrock","source":"github","weight":0.7},{"name":"wordpress-boilerplate","source":"github","weight":0.7},{"name":"php","source":"github","weight":0.7},{"name":"wordpress-starter","source":"github","weight":0.7},{"name":"wp-composer","source":"github","weight":0.7},{"name":"wordpress-composer","source":"github","weight":0.7},{"name":"wp-packages","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/roots/bedrock","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xCUB3/wBlock","owner":"0xCUB3","name":"wBlock","desc":"The next-generation ad blocker for Safari. Free and open source on macOS, iOS, iPadOS, and visionOS, with 750,000 rules, userscripts, userstyles, and an element zapper.","summaryCn":"给 Safari 浏览器装上终极盾牌，屏蔽广告又保护隐私。","reasonCn":"这是款专为苹果全家桶设计的下一代广告拦截器，完全开源免费。它内置75万条规则库，支持脚本和样式注入，甚至能像 CSS 选择器一样用“元素抓取”功能精准屏蔽页面上的广告位。底层基于开源的 uBlock Origin 规则，不仅拦截广告还能有效防止追踪，保护你的上网隐私。","stars":2929,"starGrowth":5,"createdAt":"2024-07-18T12:49:38Z","silentRounds":0,"language":"Swift","topics":["privacy"],"aiDims":["安全工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["广告拦截","Safari扩展","隐私保护","开源软件","macOS"],"domainKey":"广告拦截","pushedAt":"2026-09-23T20:32:53Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/0xCUB3/wBlock","ts":"2026-09-23T20:32:53Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02012120177534995,"copyOk":false},{"repo":"remotion-dev/remotion","owner":"remotion-dev","name":"remotion","desc":"🎥 Make videos programmatically with React","summaryCn":"用 React 写代码，像做 PPT 一样生成炫酷视频。","reasonCn":"Remotion 是一个基于 React 的视频生成框架，允许开发者使用代码控制视频的每一帧。它提供了强大的合成器、动画库和媒体处理能力，支持 React 组件直接作为视频内容，并能导出为 MP4 或 GIF 等格式，彻底改变了传统视频剪辑的工作流。","stars":60053,"starGrowth":53,"createdAt":"2020-06-23T19:49:10Z","silentRounds":0,"language":"TypeScript","topics":["react","video","javascript"],"aiDims":["创意工具","非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["视频生成","React","创意编程","动态图形","视频合成"],"domainKey":"视频生成","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"react","source":"github","weight":0.7},{"name":"video","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/remotion-dev/remotion","ts":"2026-09-17T04:13:51.053Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1616476612325779,"copyOk":true},{"repo":"rust-lang/rust","owner":"rust-lang","name":"rust","desc":"Empowering everyone to build reliable and efficient software.","summaryCn":"建造软件基石，让程序快如闪电且安全不掉坑","reasonCn":"Rust是一门注重内存安全与执行性能的系统级编程语言，由Mozilla开发。它采用所有权机制防止内存泄漏和悬空指针，无需垃圾回收即可实现零成本抽象。广泛应用于操作系统、嵌入式设备、区块链、云计算及高性能网络服务，也是众多AI基础设施和推理引擎的首选语言。","stars":119046,"starGrowth":11,"createdAt":"2010-06-16T20:39:03Z","silentRounds":0,"language":"Rust","topics":["rust","compiler","language"],"aiDims":["开发者工具","非AI-实用","学习资源"],"zone":"工具","funScore":0,"domainTags":["系统编程语言","编译器","内存安全","工具链"],"domainKey":"系统编程语言","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"rust","source":"github","weight":0.7},{"name":"compiler","source":"github","weight":0.7},{"name":"language","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.25,"source":"trending","url":"https://github.com/rust-lang/rust","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03327720376997127,"copyOk":false},{"repo":"gzuidhof/coi-serviceworker","owner":"gzuidhof","name":"coi-serviceworker","desc":"Cross-origin isolation (COOP and COEP) through a service worker for situations in which you can't control the headers (e.g. GH pages)","summaryCn":"不用改服务器配置就能开启跨域隔离的实用小工具","reasonCn":"技术上，coi-serviceworker的思路很巧妙：它用一个Service Worker拦截所有请求，动态添加所需的响应头，而不需要服务端做任何配置。核心就是一个十几行的JS文件，注册后自动为所有同源请求注入Cross-Origin-Opener-Policy和Cross-Origin-Embedder-Policy头。","stars":579,"starGrowth":0,"createdAt":"2021-08-06T00:56:16Z","silentRounds":10,"language":"JavaScript","topics":[],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["Service Worker","跨域隔离","WebGPU","静态托管"],"domainKey":"Service Worker","pushedAt":"2023-12-09T15:47:30Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"bigbro","url":"https://github.com/gzuidhof/coi-serviceworker","ts":"2023-12-09T15:47:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sjai013/unity-gameplay-ability-system","owner":"sjai013","name":"unity-gameplay-ability-system","desc":"A unified framework for implementing ability systems in Unity","summaryCn":"Unity游戏角色技能系统的通用框架，写技能逻辑从此有章法","reasonCn":"它的核心设计理念是组件化和数据驱动。整个系统由几个关键模块组成：游戏能力管理器负责统筹所有能力的生命周期；能力定义类用数据驱动的方式描述每个技能的属性，比如冷却时间、消耗、触发条件；能力组件挂在角色身上，处理具体的执行逻辑；还有事件系统负责能力之间的通信和状态","stars":1084,"starGrowth":0,"createdAt":"2019-05-19T10:27:49Z","silentRounds":24,"language":"C#","topics":[],"aiDims":["游戏","可视化"],"zone":"工具","funScore":0,"domainTags":["Unity","游戏开发","技能系统","C#"],"domainKey":"Unity","pushedAt":"2022-08-12T19:28:27Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.38,"source":"bigbro","url":"https://github.com/sjai013/unity-gameplay-ability-system","ts":"2022-08-12T19:28:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"edwardkim/rhwp","owner":"edwardkim","name":"rhwp","desc":"아래한글 hwp viewer and editor by rust and wasm","summaryCn":"韩国专有的HWP文档，网页端直接查看编辑的神器","reasonCn":"韩国有自己专有的文字处理格式HWP，类似中国的WPS格式。这个项目用Rust重写解析器，配合WASM让浏览器直接渲染和编辑HWP文件，无需安装韩国本地软件。跨平台支持Mac/Linux/Windows，填补了非韩文环境处理朝鲜文档的工具空白。","stars":3847,"starGrowth":16,"language":"Rust","topics":["hwp","hwpx","rust"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["文档处理","韩文格式","WASM","跨平台"],"domainKey":"文档处理","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"hwp","source":"github","weight":0.7},{"name":"hwpx","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/edwardkim/rhwp","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.06047725500389914,"copyOk":true},{"repo":"openai/codex","owner":"openai","name":"codex","desc":"Lightweight coding agent that runs in your terminal","summaryCn":"终端里的AI编程小秘书，帮你写代码改bug超省心","reasonCn":"OpenAI推出的轻量级编程代理，直接运行在你的终端里。它支持代码生成、调试、重构等多种任务，能理解上下文并给出智能建议。基于Rust构建，启动快资源占用低，与现有工作流无缝衔接，是开发者提升编码效率的好帮手。","stars":125963,"starGrowth":79,"createdAt":"2025-04-13T05:37:54Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["代码助手","AI Agent","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI编程助手","终端","代码生成"],"domainKey":"AI编程助手","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/openai/codex","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.2388815048863555,"copyOk":false},{"repo":"rogerbarton/UnityBuildPipeline","owner":"rogerbarton","name":"UnityBuildPipeline","desc":"Automate Multiple Builds - A Pipeline for Unity using Odin Inspector","summaryCn":"Unity批量构建自动化工具，一键打出多个平台的安装包，告别重复劳动","reasonCn":"核心功能是用Odin Inspector做的配置界面，你可以把各个平台的构建参数存成预设，比如iOS用什么证书、Android用什么签名、WebGL用什么内存配置。配置好了之后点一下按钮就能批量构建，还能根据Git分支自动选择构建目标。","stars":52,"starGrowth":0,"createdAt":"2020-02-11T12:48:26Z","silentRounds":24,"language":"C#","topics":["csharp","odin-inspector","unity","unity-editor"],"aiDims":["游戏","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Unity","游戏开发","自动化构建","C#"],"domainKey":"Unity","pushedAt":"2020-02-12T20:28:05Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"odin-inspector","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity-editor","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.38,"source":"bigbro","url":"https://github.com/rogerbarton/UnityBuildPipeline","ts":"2020-02-12T20:28:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ionic-team/ionic-framework","owner":"ionic-team","name":"ionic-framework","desc":"A powerful cross-platform UI toolkit for building native-quality iOS, Android, and Progressive Web Apps with HTML, CSS, and JavaScript.","summaryCn":"一套Web技术写全平台App的框架，一个项目同时跑iOS和安卓","reasonCn":"基于Web技术栈的跨平台移动开发框架，支持用HTML、CSS和JavaScript构建iOS、Android和PWA应用。内置Material Design和iOS风格组件库，结合Ionic CLI和Capacitor实现原生能力调用，社区活跃且star数超52000，适合快速开发高质量的跨平台应用。","stars":52676,"starGrowth":2,"createdAt":"2013-08-20T23:06:02Z","silentRounds":0,"language":"TypeScript","topics":["mobile","frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["跨平台","移动端开发","UI 框架"],"domainKey":"跨平台","pushedAt":"2026-09-23T20:20:50Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/ionic-team/ionic-framework","ts":"2026-09-23T20:20:50Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mozilla/pontoon","owner":"mozilla","name":"pontoon","desc":"Mozilla's Localization Platform","summaryCn":"Mozilla出品的强大开源翻译平台，团队协作本地化的神器","reasonCn":"这是Mozilla开发的开源本地化平台，已有1600多星。它提供了一个美观的Web界面，让翻译者无需接触代码即可对软件界面进行翻译，并实时预览效果。虽然Mozilla已停止官方维护，但它依然是许多大型开源项目本地化的经典解决方案，界面设计和协作流程非常值得借鉴。","stars":1667,"starGrowth":0,"createdAt":"2011-02-19T11:25:51Z","silentRounds":24,"language":"Python","topics":["design"],"aiDims":["非AI-好玩","游戏","创意工具"],"zone":"工具","funScore":0,"domainTags":["本地化","翻译协作","国际化","Web 应用"],"domainKey":"本地化","pushedAt":"2026-09-23T21:49:12Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/mozilla/pontoon","ts":"2026-09-23T21:49:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Gentleman-Programming/gentle-ai","owner":"Gentleman-Programming","name":"gentle-ai","desc":"Gentle-AI configures the AI coding agents you already use: Claude Code, Cursor, OpenCode, Codex, Pi, and more. Choose persistent memory, Spec-Driven Development, curated skills, MCP servers, personas, and optional bounded review. Open source, no agent lock-in.","summaryCn":"Go语言写的轻量AI工具箱，让开发者轻松接入AI能力不踩坑","reasonCn":"这是一个Go生态下的AI基础设施项目，封装了常用AI调用能力，省去重复造轮子的时间。项目虽无详细文档，但凭借Go语言的简洁优势，适合后端开发者快速集成AI功能，比如大模型调用、向量检索等。5000+的star数说明社区认可度不错，虽然没有炫酷的UI，但作为底层工具非常务实。","stars":7159,"starGrowth":30,"createdAt":"2026-02-27T23:28:01Z","silentRounds":0,"language":"Go","topics":["ai-coding","claude-code","code-review","copilot","cursor","developer-tools","golang","llm","mcp","opencode","persistent-memory","spec-driven-development","agent-skills","ai-agents","codex","coding-agents","pi"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","Claude Code","Cursor","MCP","Go"],"domainKey":"AI编程助手","pushedAt":"2026-09-23T15:48:36Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai-coding","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"code-review","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"persistent-memory","source":"github","weight":0.7},{"name":"spec-driven-development","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"coding-agents","source":"github","weight":0.7},{"name":"pi","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/Gentleman-Programming/gentle-ai","ts":"2026-09-23T15:48:36Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.12308590585277272,"copyOk":true},{"repo":"parse-community/parse-server","owner":"parse-community","name":"parse-server","desc":"Parse Server for Node.js / Express","summaryCn":"后端开发者的瑞士军刀，免写接口直接跑服务","reasonCn":"Parse Server是著名的Parse平台开源版，为移动和Web应用提供即插即用的后端服务。它自动处理用户认证、数据库存储、云代码执行和推送通知，无需从零搭建API。支持灵活的数据模型设计，能大幅缩短开发周期，让开发者专注前端逻辑。","stars":21409,"starGrowth":0,"createdAt":"2016-01-28T18:29:14Z","silentRounds":9,"language":"JavaScript","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["BaaS","Node.js","后端服务","数据库","用户认证"],"domainKey":"BaaS","pushedAt":"2026-09-23T20:25:30Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/parse-community/parse-server","ts":"2026-09-23T20:25:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"scikit-learn/scikit-learn","owner":"scikit-learn","name":"scikit-learn","desc":"scikit-learn: machine learning in Python","summaryCn":"Python机器学习界的瑞士军刀，算法大全一站搞定","reasonCn":"技术层面，它实现了经典的监督学习和无监督学习算法，包括SVM、随机森林、梯度提升、KMeans、PCA等数十种。代码用Python和NumPy优化，运行效率不错。独特之处在于API设计极其一致——fit、predict、transform三个方法走天下，学会了用一种算法就会用所有算法。","stars":67344,"starGrowth":13,"createdAt":"2010-08-17T09:43:38Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["模型与训练","开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["机器学习","Python库","数据科学"],"domainKey":"机器学习","pushedAt":"2026-09-23T17:44:25Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/scikit-learn/scikit-learn","ts":"2026-09-23T17:44:25Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03957911617961511,"copyOk":true},{"repo":"dart-frog-dev/dart_frog","owner":"dart-frog-dev","name":"dart_frog","desc":"A fast, minimalistic backend framework for Dart 🎯","summaryCn":"Dart后端框架，写接口简单，性能强，不卷AI卷开发体验。","reasonCn":"Dart官推的现代后端框架，专为构建高并发REST API设计。零依赖架构让启动极快，内存占用极低，特别适合云原生场景。内置依赖注入、多环境配置和日志中间件，开发者只需关注业务逻辑，无需配置繁琐的框架细节，大幅提升编码效率与运行性能。","stars":2270,"starGrowth":0,"createdAt":"2022-05-02T02:55:21Z","silentRounds":3,"language":"Dart","topics":["backend"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Dart","后端框架","REST API","云原生"],"domainKey":"Dart","pushedAt":"2026-09-01T18:53:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/dart-frog-dev/dart_frog","ts":"2026-09-01T18:53:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"garychowcmu/daizhigev20","owner":"garychowcmu","name":"daizhigev20","desc":"","summaryCn":"garychowcmu/daizhigev20 并非一款软件，而是一个庞大的纯文本资源集合，核","reasonCn":"内容上，该仓库涵盖了诸子百家、经史子集等多类古代典籍，均为 Unicode 编码的 TXT 格式，兼容所有阅读软件。项目说明中特别提醒，部分文件夹（如医藏）过大，在移动端浏览器中打开可能需要较长时间加载，建议电脑端浏览。","stars":3390,"starGrowth":2,"createdAt":"2018-11-26T20:32:45Z","silentRounds":0,"language":"","topics":[],"aiDims":["学习资源","非AI-实用"],"zone":"资源","funScore":0.2,"domainTags":["古代文献","Txt文本","数字图书馆","古籍整理"],"domainKey":"古代文献","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1}],"aiScore":0.2,"source":"search","url":"https://github.com/garychowcmu/daizhigev20","ts":"2026-09-19T18:31:48.187Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PhyAgentOS/PhyAgentOS-core","owner":"PhyAgentOS","name":"PhyAgentOS-core","desc":"","summaryCn":"这是一个专为具身智能任务设计的递归自改进智能体系统基础设施与核心框架。","reasonCn":"PhyAgentOS 解决了具身智能领域中机器人动作执行边界模糊、缺乏有效执行证据链的问题。它通过统一的网关契约隔离底层硬件与智能体，并严格收集动作前后的多模态证据，确保智能体对物理世界的干预安全可控且可被事后验证。","stars":2238,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.38,"domainTags":["具身智能","代理框架","机器人控制"],"domainKey":"具身智能","pushedAt":"2026-09-23T13:30:23.233Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/PhyAgentOS/PhyAgentOS-core","ts":"2026-09-23T13:30:23.233Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jplenio/ComfyUI-MiniMax-Music-Production-Toolkit","owner":"jplenio","name":"ComfyUI-MiniMax-Music-Production-Toolkit","desc":"Complete Production Environment for professional MiniMax Music 3 audio generation, including Prompt Generation, Metadata generation including Cover Art, Sound Enhancement to improve Minimax weaknesses, mastering, and saving in various file formats","summaryCn":"给音乐人准备的AI全流程生产套件，一键搞定编曲封面","reasonCn":"这是一个基于ComfyUI的MiniMax Music 3专业音频生成工作流，集成了提示词生成、元数据创建及封面图制作等功能。它针对MiniMax模型的音频特性，专门开发了声音增强模块以弥补其弱点。开发者利用Python编写，通过工作流节点连接，实现了从文本到完整音乐作品的生产闭环。","stars":62,"starGrowth":1,"createdAt":"2026-08-31T21:01:10Z","silentRounds":1,"language":"Python","topics":["music-production"],"aiDims":["创意工具","非AI-实用"],"zone":"AI","funScore":1,"domainTags":["AI音乐生成","ComfyUI","MiniMax","音频工作流","自动元数据"],"domainKey":"AI音乐生成","pushedAt":"2026-09-20T15:05:54Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music-production","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/jplenio/ComfyUI-MiniMax-Music-Production-Toolkit","ts":"2026-09-20T15:05:54Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alm-tools/alm","owner":"alm-tools","name":"alm","desc":"","summaryCn":"ALM 当初的设计愿景是成为 TypeScript 开发者心目中理想的 IDE，通过解决开","reasonCn":"虽然该项目目前已经宣布停止更新，但这并不妨碍其作为开发工具研究的价值。作者通过构建这一 IDE，深入探讨了如何优化开发者与代码交互的 UI 体验，这些经验目前被整合到了作者的新产品开发中。如果你对 Web 开发工具的历史感兴趣，或者想了解如何通过浏览器构建桌面级代码编写环境","stars":3013,"starGrowth":0,"createdAt":"2015-08-20T04:03:53Z","silentRounds":4,"language":"TypeScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["TypeScript","IDE","编辑器"],"domainKey":"TypeScript","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/alm-tools/alm","ts":"2026-09-19T22:19:04.205Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gokcehan/lf","owner":"gokcehan","name":"lf","desc":"","summaryCn":"lf 诞生于对终端文件管理效率的极致追求，其最显著的技术特征是采用了异步 ","reasonCn":"在部署方面，lf 展现了 Go 语言的典型优势：它是一个单文件二进制程序，不包含任何运行时依赖。这意味着你不需要安装复杂的包管理器或配置环境变量，只需将可执行文件放置在 $PATH 路径下的任意位置即可立即使用。这种“零配置”的特性使其成为配置极简主义环境（如 Docker 容器或裸","stars":9520,"starGrowth":1,"createdAt":"2016-08-13T12:45:44Z","silentRounds":3,"language":"Go","topics":["terminal","file-manager"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["终端文件管理器","Vim键位","命令行工具"],"domainKey":"终端文件管理器","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"terminal","source":"github","weight":0.7},{"name":"file-manager","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/gokcehan/lf","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/shardingsphere","owner":"apache","name":"shardingsphere","desc":"Empowering Data Intelligence with Distributed SQL for Sharding, Scalability, and Security Across All Databases.","summaryCn":"把数据库切分成多个分片，让海量数据读写也能快如闪电","reasonCn":"ShardingSphere是Apache顶级项目，专注分布式数据库中间件。提供数据分片、读写分离、数据加密等核心能力，支持MySQL、PostgreSQL等多种数据库协议。采用插件化架构，灵活配置，帮助开发者轻松应对海量数据和高并发场景。","stars":20802,"starGrowth":1,"createdAt":"2016-01-18T12:49:26Z","silentRounds":5,"language":"Java","topics":["database"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","数据分片","读写分离","数据库中间件","SQL"],"domainKey":"分布式数据库","pushedAt":"2026-09-23T16:25:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/apache/shardingsphere","ts":"2026-09-23T16:25:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"superradcompany/microsandbox","owner":"superradcompany","name":"microsandbox","desc":"🧱 fast branchable microVM for any workload","summaryCn":"给每一个代码片段都穿上一层防弹衣，安全又轻量的沙箱运行器","reasonCn":"Microsandbox 是一个基于 WebAssembly 的高性能沙箱运行时，旨在为本地代码提供安全的执行环境。它通过隔离执行上下文，防止恶意代码对主机系统造成破坏，同时保持了极低的内存开销和启动速度，非常适合 CI/CD 流程、在线代码评测及敏感数据处理的场景。","stars":8399,"starGrowth":15,"createdAt":"2024-10-03T15:57:38Z","language":"Rust","topics":["agents"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["WebAssembly","沙箱","Rust","代码执行","安全隔离"],"domainKey":"WebAssembly","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/superradcompany/microsandbox","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.05035778068817716,"copyOk":true},{"repo":"podman-container-tools/podman","owner":"podman-container-tools","name":"podman","desc":"Podman: A tool for managing OCI containers and pods.","summaryCn":"不用Docker也能玩容器，红帽开源的神级替代品","reasonCn":"这是一个开源的容器引擎，类似于Docker，但更注重安全性和无守护进程的设计。它可以在Linux、macOS和Windows上运行，完全兼容Docker命令，让你无需学习新命令就能迁移。它还支持Rootless模式，解决了容器逃逸的安全隐患，是企业级容器部署的理想选择。","stars":32915,"starGrowth":4,"createdAt":"2017-11-01T15:01:27Z","silentRounds":0,"language":"Go","topics":["docker","kubernetes"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["容器引擎","OCI标准","Rootless容器","无守护进程","Docker替代"],"domainKey":"容器引擎","pushedAt":"2026-09-23T21:38:10Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/podman-container-tools/podman","ts":"2026-09-23T21:38:10Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"leandrowd/react-responsive-carousel","owner":"leandrowd","name":"react-responsive-carousel","desc":"React.js Responsive Carousel (with Swipe)","summaryCn":"很多前端项目都需要一个既好看又易用的轮播图，但市面上要么体积大，要么功","reasonCn":"很多前端项目都需要一个既好看又易用的轮播图，但市面上要么体积大，要么功能单一。react-responsive-carousel 直接面向 React 开发者，提供了即插即用的轮播解决方案。它在移动端支持原生滑动手势，桌面端提供左右箭头和键盘控制，使用者只需在 JSX 中包裹 <Carousel> 即可完成配","stars":2675,"starGrowth":0,"createdAt":"2015-02-08T10:59:28Z","silentRounds":10,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["React组件","轮播图","响应式设计","前端库"],"domainKey":"React组件","pushedAt":"2026-09-16T18:50:47Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.24,"source":"search","url":"https://github.com/leandrowd/react-responsive-carousel","ts":"2026-09-16T18:50:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"trekhleb/javascript-algorithms","owner":"trekhleb","name":"javascript-algorithms","desc":"📝 Algorithms and data structures implemented in JavaScript with explanations and links to further readings","summaryCn":"用JavaScript实现的常见算法和数据结构，边看边练","reasonCn":"项目涵盖了排序算法（冒泡、快排、归并等）、搜索算法（二分查找、BFS、DFS）、数据结构（链表、栈、队列、树、图、哈希表）以及动态规划、贪心算法等核心内容。每个算法都有可视化的动画演示，你能直观看到算法执行过程中数据的变化，比看文字描述容易理解得多。","stars":196797,"starGrowth":8,"createdAt":"2018-03-24T07:47:04Z","silentRounds":0,"language":"JavaScript","topics":["javascript","algorithms","algorithm","javascript-algorithms","computer-science","interview","data-structures","interview-preparation"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["算法","数据结构","JavaScript","可视化","学习教程"],"domainKey":"算法","pushedAt":"2026-08-07T04:07:05.660Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"javascript","source":"github","weight":0.7},{"name":"algorithms","source":"github","weight":0.7},{"name":"algorithm","source":"github","weight":0.7},{"name":"javascript-algorithms","source":"github","weight":0.7},{"name":"computer-science","source":"github","weight":0.7},{"name":"interview","source":"github","weight":0.7},{"name":"data-structures","source":"github","weight":0.7},{"name":"interview-preparation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/trekhleb/javascript-algorithms","ts":"2026-08-07T04:07:05.660Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024121953078553027,"copyOk":true},{"repo":"apache/jmeter","owner":"apache","name":"jmeter","desc":"","summaryCn":"开源性能测试工具，模拟多线程负载压测应用。","reasonCn":"Apache JMeter 是一款用于测量性能和进行负载测试的 Java 应用程序。它能模拟对服务器、网络或对象的繁重负载，以测试其承受力或分析不同负载下的整体表现。项目采用纯 Java 开发，实现了完全的可移植性，并利用多线程技术，允许众多线程同时采样，从而精准模拟真实用户并发行为。它支持 HTTP、HTTPS 等多种协议，能够对静态和动态 Web 应用进行全面的性能评估。","stars":9543,"starGrowth":1,"createdAt":"2010-05-26T23:38:07Z","silentRounds":2,"language":"Java","topics":["java","performance","test"],"aiDims":["非AI-实用","非AI-好玩","开发工具"],"zone":"工具","funScore":0,"domainTags":["性能测试","负载测试","压力测试","Java工具"],"domainKey":"性能测试","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"java","source":"github","weight":0.7},{"name":"performance","source":"github","weight":0.7},{"name":"test","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.26,"source":"search","url":"https://github.com/apache/jmeter","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"infiniflow/infinity","owner":"infiniflow","name":"infinity","desc":"The AI-native database built for LLM applications, providing incredibly fast hybrid search of dense vector, sparse vector, tensor (multi-vector), and full-text.","summaryCn":"让AI拥有搜索引擎般的记忆，知识检索快如闪电","reasonCn":"Infinity是专为大模型应用设计的AI原生数据库，支持密集向量、稀疏向量、张量和全文四种索引的混合搜索，单库可应对千亿级数据。它用C++重写底层存储引擎，摒弃了传统向量数据库依赖外部存储架构，性能远超Milvus等同类项目。原生兼容PostgreSQL协议，开发者几行代码就能接入，直接用于RAG系统或智能搜索场景。","stars":4716,"starGrowth":1,"createdAt":"2022-07-18T13:52:38Z","silentRounds":1,"language":"C++","topics":["vector-db","rag"],"aiDims":["向量数据库","AI基础设施","AI搜索"],"zone":"AI","funScore":0.7,"domainTags":["向量数据库","混合搜索","RAG","PostgreSQL","LLM"],"domainKey":"向量数据库","pushedAt":"2026-09-23T13:38:30Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/infiniflow/infinity","ts":"2026-09-23T13:38:30Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"shadcn-ui/ui","owner":"shadcn-ui","name":"ui","desc":"Composable, accessible components with thoughtful defaults. Build your own component library with code you can customize, extend, and make your own.","summaryCn":"以前开发前端界面，要么去 npm 下载现成的组件库，要么自己画。","reasonCn":"它的设计哲学是“复制粘贴，而非安装”。这意味着你拥有组件的完整源码，可以随意修改、扩展，甚至删除不需要的部分。它基于 Radix UI 提供无障碍访问，基于 Tailwind CSS 提供原子化样式，两者结合让组件既强大又美观。","stars":124423,"starGrowth":53,"createdAt":"2023-01-04T12:43:27Z","silentRounds":0,"language":"TypeScript","topics":["components","nextjs","radix-ui","react","tailwindcss","ui","shadcn","base-ui","laravel","tanstack","vite","react-aria","react-aria-components"],"aiDims":["开发者工具","效率工具","UI组件库"],"zone":"工具","funScore":0,"domainTags":["组件库","前端开发","Tailwind CSS","Radix UI","TypeScript"],"domainKey":"组件库","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"UI组件库","source":"llm","weight":1},{"name":"components","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"radix-ui","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"tailwindcss","source":"github","weight":0.7},{"name":"ui","source":"github","weight":0.7},{"name":"shadcn","source":"github","weight":0.7},{"name":"base-ui","source":"github","weight":0.7},{"name":"laravel","source":"github","weight":0.7},{"name":"tanstack","source":"github","weight":0.7},{"name":"vite","source":"github","weight":0.7},{"name":"react-aria","source":"github","weight":0.7},{"name":"react-aria-components","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/shadcn-ui/ui","ts":"2026-09-17T04:13:51.053Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.16027789878077203,"copyOk":false},{"repo":"serenity-is/Serenity","owner":"serenity-is","name":"Serenity","desc":"Business Apps Made Simple with Asp.Net Core MVC / TypeScript","summaryCn":"ASP.NET快速开发框架，不用重复造轮子直接生成CRUD代码","reasonCn":"Serenity是基于ASP.NET Core MVC的开源业务应用框架，核心卖点是代码生成器。只需要定义实体类和页面配置，就能自动生成完整的增删改查界面和后端代码。内置用户权限管理、数据验证、导出等功能，让开发者把精力集中在业务逻辑而非重复代码上，大幅提升企业级应用开发效率。","stars":2691,"starGrowth":0,"createdAt":"2014-01-31T09:35:59Z","silentRounds":24,"language":"C#","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["ASP.NET","代码生成","业务框架","CRUD","Web开发"],"domainKey":"ASP.NET","pushedAt":"2026-09-23T15:44:56Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/serenity-is/Serenity","ts":"2026-09-23T15:44:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"apache/causeway","owner":"apache","name":"causeway","desc":"Use Apache Causeway™ to rapidly develop domain-driven apps or modular monoliths in Java, on top of the Spring Boot platform. Write your business logic in entities, domain services or view models, and the framework dynamically generates a representation of that domain model as a webapp, GraphQL or RESTful API. For prototyping or production.","summaryCn":"代码写一次业务逻辑，网页自动生成不用动手","reasonCn":"Apache Causeway 是 Java 生态里基于领域驱动设计的应用框架。你用领域实体、服务和视图模型描述业务规则，它能帮你自动生成带 CRUD 功能的响应式 Web 界面和 API。底层对接 Spring Boot，适合想快速搭企业级内嵌业务后台，不想写重复样板代码的 Java 开发者。","stars":906,"starGrowth":0,"createdAt":"2012-11-30T08:00:24Z","silentRounds":24,"language":"Java","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["领域驱动设计","Java","Spring Boot","后台框架","自动生成UI"],"domainKey":"领域驱动设计","pushedAt":"2026-09-23T14:47:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/apache/causeway","ts":"2026-09-23T14:47:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"structr/structr","owner":"structr","name":"structr","desc":"ATTENTION: This repository is a clone of https://gitlab.structr.com/structr/structr. All development/issue tracking has moved there. | Structr is an integrated open-source low-code development and runtime environment that uses a graph database (Neo4j).","summaryCn":"Java低代码平台，可视化构建企业级应用","reasonCn":"Structr是一个开源的低代码开发平台，基于Java技术栈，提供可视化界面构建Web应用。核心功能包括数据建模、工作流引擎、权限管理、REST API自动生成等。适合需要快速开发企业级应用的团队，可以减少大量重复代码的编写工作。","stars":829,"starGrowth":0,"createdAt":"2011-02-01T20:55:18Z","silentRounds":24,"language":"Java","topics":["backend"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["低代码平台","Java开发","企业级应用","工作流引擎","Web构建"],"domainKey":"低代码平台","pushedAt":"2026-09-10T08:57:14Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/structr/structr","ts":"2026-09-10T08:57:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVIDIA/Megatron-LM","owner":"NVIDIA","name":"Megatron-LM","desc":"Ongoing research training transformer models at scale","summaryCn":"NVIDIA官方开源的大规模Transformer语言模型训练框架，专门为在GPU集群上训","reasonCn":"这个项目最早由NVIDIA研究院开发，被广泛应用于各种超大规模语言模型的训练过程中。它解决了分布式训练中面临的核心难题：如何将一个巨大的模型拆分到成百上千张GPU上协同工作。框架支持多种并行策略——数据并行让不同GPU处理不同数据，张量并行将模型层内的矩阵运算拆分到多张卡，","stars":17981,"starGrowth":6,"createdAt":"2019-03-21T16:15:52Z","silentRounds":0,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["大模型","分布式训练","Transformer","GPU"],"domainKey":"大模型","pushedAt":"2026-09-23T21:00:49Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/NVIDIA/Megatron-LM","ts":"2026-09-23T21:00:49Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.019001056670930425,"copyOk":false},{"repo":"databendlabs/databend","owner":"databendlabs","name":"databend","desc":"Data Agent Ready Warehouse : One for  Analytics, Search, AI, Python Sandbox.  — rebuilt from scratch. Unified architecture on your S3.","summaryCn":"专为AI设计的云原生数据仓库，分析搜索一体化","reasonCn":"技术上，Databend采用存算分离的架构，计算引擎和数据存储完全解耦，支持水平扩展。它原生对接S3等对象存储，你可以把数据放在任何云存储上，不需要维护复杂的集群状态。查询引擎兼容MySQL协议，SQL开发者可以无缝迁移。","stars":9451,"starGrowth":0,"createdAt":"2020-10-10T06:04:42Z","silentRounds":1,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["云原生","数据仓库","向量搜索"],"domainKey":"云原生","pushedAt":"2026-09-23T20:24:33Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/databendlabs/databend","ts":"2026-09-23T20:24:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zellij-org/zellij","owner":"zellij-org","name":"zellij","desc":"","summaryCn":"支持多路复用的终端工作区，支持插件与协作。","reasonCn":"Zellij 的核心亮点在于它的可扩展性和用户体验。它内置了 WebAssembly 插件系统，这意味着你可以用任何支持 WASM 的语言编写插件，将外部工具直接嵌入到终端环境中，甚至编写自定义的 UI 组件。此外，Zellij 支持真正的多人协作，团队成员可以实时共享终端会话，共同调试代码或监控","stars":35510,"starGrowth":5,"createdAt":"2020-09-01T14:04:28Z","silentRounds":0,"language":"Rust","topics":["workspace","terminal","multiplexer"],"aiDims":["非AI-实用","非AI-好玩","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["终端复用","协作","WASM插件"],"domainKey":"终端复用","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"workspace","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"multiplexer","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/zellij-org/zellij","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015422416220782878,"copyOk":true},{"repo":"react/react","owner":"react","name":"react","desc":"The library for web and native user interfaces.","summaryCn":"全球开发者都在用的前端框架，搭网页像搭积木一样简单","reasonCn":"它的核心技术亮点有三：虚拟DOM——数据变化时先在内存里生成新树并与旧树对比，找出最小差异再批量更新真实DOM，性能远超直接操作；组件化——把界面拆成独立的可复用模块，每个组件自己管理状态和样式；Hooks——在函数组件里直接使用state、effect等能力，逻辑复用变得超级简洁。","stars":250654,"starGrowth":9,"createdAt":"2013-05-24T16:15:54Z","silentRounds":0,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["前端框架","虚拟DOM","组件化","JavaScript","UI开发"],"domainKey":"前端框架","pushedAt":"2026-09-22T17:36:10Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/react/react","ts":"2026-09-22T17:36:10Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.027107718209164824,"copyOk":true},{"repo":"wechat-article/wechat-article-exporter","owner":"wechat-article","name":"wechat-article-exporter","desc":"一款在线的 微信公众号文章批量下载 工具，支持导出阅读量与评论数据，无需搭建任何环境，可通过 在线网站 使用，支持 docker 私有化部署和 Cloudflare 部署。  支持下载各种文件格式，其中 HTML 格式可100%还原文章排版与样式。","summaryCn":"公众号文章一键批量下载神器，HTML排版100%完美还原原汁原味","reasonCn":"专门针对微信公众号的批量下载工具，支持一键导出文章正文、阅读量和评论数据。HTML格式能100%还原文章原始排版与样式，支持在线网站直接使用，也提供Docker和Cloudflare私有化部署方案，完全免费无需额外环境配置。","stars":12943,"starGrowth":6,"createdAt":"2024-07-17T14:41:12Z","silentRounds":0,"language":"TypeScript","topics":["download","wechat","wechat-article","wechat-download"],"aiDims":["效率工具","实用工具"],"zone":"工具","funScore":0,"domainTags":["微信公众号","批量下载","爬虫"],"domainKey":"微信公众号","pushedAt":"2026-08-07T05:48:10Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"实用工具","source":"llm","weight":1},{"name":"download","source":"github","weight":0.7},{"name":"wechat","source":"github","weight":0.7},{"name":"wechat-article","source":"github","weight":0.7},{"name":"wechat-download","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/wechat-article/wechat-article-exporter","ts":"2026-08-07T05:48:10Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01939071312678668,"copyOk":false},{"repo":"AtsushiSakai/PythonRobotics","owner":"AtsushiSakai","name":"PythonRobotics","desc":"Python sample codes and textbook for robotics algorithms.","summaryCn":"机器人开发听起来很高端，涉及各种复杂的数学公式和物理控制，但这个项目把","reasonCn":"它的厉害之处在于“全面性”和“教学性”。它没有选择那种极其复杂的工业框架，而是用最基础的Python语法，配合Matplotlib进行实时动态演示，把路径规划（比如A*、Dijkstra）、定位（卡尔曼滤波）等核心技术拆解得明明白白。","stars":30566,"starGrowth":5,"createdAt":"2016-03-21T09:34:43Z","silentRounds":0,"language":"Python","topics":["robotics"],"aiDims":["非AI-实用","学习资源","可视化"],"zone":"资源","funScore":0.5,"domainTags":["Robotics","Path Planning","Algorithms"],"domainKey":"Robotics","pushedAt":"2026-09-22T02:57:25Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/AtsushiSakai/PythonRobotics","ts":"2026-09-22T02:57:25Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015490741346594254,"copyOk":false},{"repo":"CapSoftware/Cap","owner":"CapSoftware","name":"Cap","desc":"Open source Loom alternative. Beautiful, shareable screen recordings.","summaryCn":"开源版Loom，把你的屏幕操作录成像Vlog一样好看的教程视频。","reasonCn":"Cap 是一款开源的屏幕录制与演示工具，旨在替代商业产品 Loom。它使用 Rust 语言开发，支持录制屏幕、摄像头、音频，并具备自动生成脚本、视频剪辑和实时分享等功能。项目目前 Star 数超过 2 万，界面设计现代，操作流畅，适合开发者、教育工作者和内容创作者使用。","stars":22637,"starGrowth":16,"createdAt":"2023-11-17T20:31:46Z","silentRounds":0,"language":"Rust","topics":["app","cap","coss","mac","open-source","oss","react","record","screen-capture","screen-recorder","screenshot","tauri","tauri-app","typescript","vite","solidjs","loom","nextjs","rust","nextjs15"],"aiDims":["非AI-实用","创意工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["屏幕录制","视频剪辑","教程制作","开源工具"],"domainKey":"屏幕录制","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"app","source":"github","weight":0.7},{"name":"cap","source":"github","weight":0.7},{"name":"coss","source":"github","weight":0.7},{"name":"mac","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"oss","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"record","source":"github","weight":0.7},{"name":"screen-capture","source":"github","weight":0.7},{"name":"screen-recorder","source":"github","weight":0.7},{"name":"screenshot","source":"github","weight":0.7},{"name":"tauri","source":"github","weight":0.7},{"name":"tauri-app","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"vite","source":"github","weight":0.7},{"name":"solidjs","source":"github","weight":0.7},{"name":"loom","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"nextjs15","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/CapSoftware/Cap","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.050120422317444895,"copyOk":false},{"repo":"0xFA11/MultiplayerNetworkingResources","owner":"0xFA11","name":"MultiplayerNetworkingResources","desc":"A curated list of Multiplayer Game Network Programming Resources","summaryCn":"多人游戏联网编程的入坑指南，从网络同步到框架选型全覆盖","reasonCn":"这是一个精心整理的多人游戏网络编程资源集合，涵盖网络同步、架构设计、帧同步、状态同步等核心技术，还有Unity、Unreal等引擎的实现方案。star超过8600，对想做联网游戏的独立开发者或学生来说是一本实用的技术百科，从基础原理到开源框架都有涉及。","stars":8701,"starGrowth":0,"createdAt":"2018-01-24T09:05:00Z","silentRounds":3,"language":"C","topics":["game"],"aiDims":["游戏","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["多人游戏","网络编程","游戏开发","技术指南","同步"],"domainKey":"多人游戏","pushedAt":"2026-08-27T10:58:50Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/0xFA11/MultiplayerNetworkingResources","ts":"2026-08-27T10:58:50Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"coreybutler/nvm-windows","owner":"coreybutler","name":"nvm-windows","desc":"A node.js version management utility for Windows. Ironically written in Go.","summaryCn":"Windows下管理 Node.js 版本的必备神器，开源项目里的常青树","reasonCn":"如果你在 Windows 上开发过前端或 Node.js 应用，肯定遇到过“这个项目要 Node 16，那个要 Node 18”的版本冲突问题。手动安装卸载不仅麻烦，还容易留下残留文件。nvm-windows 就是为了解决这个问题而生的。它利用 Go 语言的高效编译能力，在 Windows 上实现了类似 Linux 下 nvm 的所","stars":47759,"starGrowth":4,"createdAt":"2014-09-20T16:37:28Z","silentRounds":0,"language":"Go","topics":["node","version-manager","windows","nodejs","nvm","install","node-js"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Node.js","版本管理","Windows"],"domainKey":"Node.js","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"node","source":"github","weight":0.7},{"name":"version-manager","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"nvm","source":"github","weight":0.7},{"name":"install","source":"github","weight":0.7},{"name":"node-js","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/coreybutler/nvm-windows","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opendsm/eeweather","owner":"opendsm","name":"eeweather","desc":"Fetch NCDC ISD, TMY3, or CZ2010 weather data that corresponds to ZIP Code Tabulation Areas or Latitude/Longitude.","summaryCn":"一键抓取美国气象站数据，按邮编或经纬度输出，省事又准。","reasonCn":"eeweather 是用 Python 编写的气象数据抓取库，能够从美国国家气象数据中心（NCDC）的 ISD、TMY3、CZ2010 三类数据源，按邮编对应的 ZCTA 或任意经纬度批量下载历史或典型气候年数据，直接返回 pandas DataFrame，字段完整，便于后续分析、机器学习或可视化。","stars":51,"starGrowth":0,"createdAt":"2018-01-29T21:14:08Z","silentRounds":11,"language":"Python","topics":["weather"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["气象数据","数据分析","Python库","地理信息"],"domainKey":"气象数据","pushedAt":"2026-09-09T20:42:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"weather","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/opendsm/eeweather","ts":"2026-09-09T20:42:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"librosa/librosa","owner":"librosa","name":"librosa","desc":"Python library for audio and music analysis","summaryCn":"你是否好奇过，Spotify是怎么分析你的音乐偏好的？","reasonCn":"核心亮点在于其数学处理的严谨性与功能的全面性。它能够轻松执行快速傅里叶变换（FFT）、计算恒定Q变换（CQT），并能处理复杂的音频加载、重采样和节拍跟踪。与底层C语言库相比，Librosa的接口极其符合Python开发者的直觉，让原本深奥的数字信号处理变得如同拼接积木一样简单明了。","stars":8622,"starGrowth":3,"createdAt":"2012-10-20T14:21:01Z","silentRounds":0,"language":"Python","topics":["music"],"aiDims":["创意工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["音频分析","音乐信息检索","数字信号处理"],"domainKey":"音频分析","pushedAt":"2026-09-21T17:47:07Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/librosa/librosa","ts":"2026-09-21T17:47:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cjpais/Handy","owner":"cjpais","name":"Handy","desc":"A free, open source, and extensible speech-to-text application that works completely offline.","summaryCn":"完全离线好用的语音转文字神器，隐私安全没烦恼","reasonCn":"Handy 是一款基于 Rust 开发的开源语音转文字应用，主打零依赖和完全离线运行，无需联网即可将语音实时转换为文字。它不仅支持多种主流语音识别引擎，还具备高度可扩展的架构，允许用户通过插件自定义功能。项目采用 Rust 编写，保证了极高的运行效率和内存安全性，非常适合对数据隐私有高要求或网络环境受限的用户使用。","stars":32042,"starGrowth":29,"createdAt":"2025-02-13T02:42:29Z","silentRounds":0,"language":"Rust","topics":["speech-to-text","tauri-v2","accessibility","cross-platform"],"aiDims":["AI应用","工具调用","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["语音识别","离线","Whisper"],"domainKey":"语音识别","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"speech-to-text","source":"github","weight":0.7},{"name":"tauri-v2","source":"github","weight":0.7},{"name":"accessibility","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/cjpais/Handy","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08971518631795768,"copyOk":true},{"repo":"Triple-T/gradle-play-publisher","owner":"Triple-T","name":"gradle-play-publisher","desc":"GPP is Android's unofficial release automation Gradle Plugin. It can do anything from building, uploading, and then promoting your App Bundle or APK to publishing app listings and other metadata.","summaryCn":"让安卓应用自动上传Google Play，从此告别手动发布的苦日子","reasonCn":"这是一个Gradle插件，专门帮安卓开发者自动化发布流程。只需配置好，就能一键完成编译、打包、上传到Google Play、更新应用描述和截图等全部操作。支持多版本管理、分阶段发布、灰度发布等功能，还可以集成到CI/CD流水线中，真正实现全自动发布。","stars":4320,"starGrowth":0,"createdAt":"2014-08-07T11:28:01Z","silentRounds":25,"language":"Kotlin","topics":["mobile"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Gradle","Android","自动化","发布工具","CI/CD"],"domainKey":"Gradle","pushedAt":"2026-08-26T21:50:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/Triple-T/gradle-play-publisher","ts":"2026-08-26T21:50:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"seriousm4x/UpSnap","owner":"seriousm4x","name":"UpSnap","desc":"A simple wake on lan web app written with SvelteKit, Go and PocketBase.","summaryCn":"躺着刷手机顺手远程开机，一键唤醒家里所有设备","reasonCn":"UpSnap是一个基于Web的远程唤醒工具，支持通过浏览器或手机轻松唤醒局域网内的电脑、NAS、服务器等设备。采用Go加PocketBase架构，界面简洁直观，部署简单。只需填写设备的MAC地址和IP范围，就能随时远程开机，彻底告别物理触碰。","stars":6299,"starGrowth":4,"createdAt":"2020-11-16T09:27:18Z","silentRounds":0,"language":"Go","topics":["wakeonlan","golang","pocketbase","sveltekit","wol"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Wake on LAN","远程控制","Web应用","Go","PocketBase"],"domainKey":"Wake on LAN","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"wakeonlan","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"pocketbase","source":"github","weight":0.7},{"name":"sveltekit","source":"github","weight":0.7},{"name":"wol","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/seriousm4x/UpSnap","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sebastienros/jint","owner":"sebastienros","name":"jint","desc":"Javascript Interpreter for .NET","summaryCn":"让.NET程序直接跑JS脚本，不用装浏览器也能执行前端代码","reasonCn":"Jint是一个纯C#实现的JavaScript解释器，无需依赖V8引擎或Node.js，就能在.NET应用中执行JavaScript代码。它支持ECMAScript 5.1标准，轻量级、零外部依赖，特别适合需要在服务端动态运行JS脚本、与前端技术栈交互、或执行用户输入代码的场景。","stars":4724,"starGrowth":2,"createdAt":"2014-01-10T01:51:04Z","silentRounds":0,"language":"C#","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript解释器",".NET库","脚本执行"],"domainKey":"JavaScript解释器","pushedAt":"2026-08-18T16:22:44Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/sebastienros/jint","ts":"2026-08-18T16:22:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"immich-app/immich","owner":"immich-app","name":"immich","desc":"High performance self-hosted photo and video management solution.","summaryCn":"自建专属云相册，AI自动识别人脸和物体，照片管理从未如此省心","reasonCn":"开源自托管照片视频管理平台，完美替代Google Photos。支持手机自动备份、AI人脸识别、物体检测、智能相册分类和全文搜索。采用TypeScript和PostgreSQL构建，性能强劲支持多用户多设备协作。隐私优先用户的首选，自托管数据完全掌控，告别云端泄露风险，值得拥有。","stars":114839,"starGrowth":47,"createdAt":"2022-02-03T15:56:27Z","silentRounds":0,"language":"TypeScript","topics":["backup-tool","mobile-app","photo-gallery","photos","self-hosted","videos","flutter","nestjs","nodejs","google-photos-alternative","svelte","sveltekit","google-photos","photos-management","typescript","javascript"],"aiDims":["非AI-实用","AI应用"],"zone":"AI","funScore":0,"domainTags":["自托管","云相册","AI人脸识别","照片管理","Immich"],"domainKey":"自托管","pushedAt":"2026-09-23T21:51:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"backup-tool","source":"github","weight":0.7},{"name":"mobile-app","source":"github","weight":0.7},{"name":"photo-gallery","source":"github","weight":0.7},{"name":"photos","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"videos","source":"github","weight":0.7},{"name":"flutter","source":"github","weight":0.7},{"name":"nestjs","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"google-photos-alternative","source":"github","weight":0.7},{"name":"svelte","source":"github","weight":0.7},{"name":"sveltekit","source":"github","weight":0.7},{"name":"google-photos","source":"github","weight":0.7},{"name":"photos-management","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/immich-app/immich","ts":"2026-09-23T21:51:38Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1422278058847604,"copyOk":false},{"repo":"shiyu-coder/Kronos","owner":"shiyu-coder","name":"Kronos","desc":"Kronos: A Foundation Model for the Language of Financial Markets","summaryCn":"首个开源金融K线基础模型，覆盖45家交易所。","reasonCn":"Kronos是首个专为金融K线数据设计的开源基础模型。它像处理自然语言一样处理金融市场数据，解决了传统时间序列模型难以处理高噪声金融数据的问题。用户可以利用它进行市场趋势预测、量化策略回测，或将其作为金融大模型微调的基座，为量化交易提供强大的数据理解和预测能力。","stars":39346,"starGrowth":10,"createdAt":"2025-07-01T02:54:32Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练","金融行业"],"zone":"AI","funScore":0,"domainTags":["金融大模型","财经NLP","量化分析","语言模型"],"domainKey":"金融大模型","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"金融行业","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/shiyu-coder/Kronos","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030762466324404006,"copyOk":true},{"repo":"mosra/magnum","owner":"mosra","name":"magnum","desc":"Lightweight and modular C++11 graphics middleware for games and data visualization","summaryCn":"轻量级C++图形库，用最少代码打造游戏和可视化效果","reasonCn":"这是一个专注于图形渲染的C++11中间件库，专为游戏开发、科学可视化和实时渲染应用设计。提供2D/3D图形管道、精灵系统、着色器管理和窗口绑定等核心功能。相比OpenGL原始API，Magnum大幅简化了图形编程，代码更简洁，性能更高效，是C++图形开发者的得力工具。","stars":5209,"starGrowth":1,"createdAt":"2010-12-19T22:19:59Z","silentRounds":10,"language":"C++","topics":["game"],"aiDims":["游戏","可视化"],"zone":"工具","funScore":0,"domainTags":["图形渲染","游戏开发","C++库","OpenGL"],"domainKey":"图形渲染","pushedAt":"2026-08-23T10:24:48Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/mosra/magnum","ts":"2026-08-23T10:24:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"abdullahbutt/leben-in-deutschland-test","owner":"abdullahbutt","name":"leben-in-deutschland-test","desc":"German Citizenship Test (Einbürgerungstest) — All 300+160 questions with translations in English & Urdu, correct answers, and explanations","summaryCn":"德国入籍考试的完整题库，包含翻译和答案解析","reasonCn":"这个项目整理了德国入籍考试的全部300多道题目，每道题都配有英文和乌尔都语翻译，并提供正确答案和详细解释。内容覆盖德国法律、社会制度、历史文化等考试要点，按主题分类便于针对性复习。想考德国护照的朋友可以直接用这个准备笔试，不用到处找资料。","stars":24,"starGrowth":0,"createdAt":"2026-04-07T09:09:41Z","silentRounds":13,"language":"","topics":["english"],"aiDims":["学习资源"],"zone":"资源","funScore":0,"domainTags":["德国移民","考试","语言学习"],"domainKey":"德国移民","pushedAt":"2026-09-23T08:28:54Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"english","source":"github","weight":0.7}],"aiScore":0.15,"source":"search","url":"https://github.com/abdullahbutt/leben-in-deutschland-test","ts":"2026-09-23T08:28:54Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bannedbook/fanqiang","owner":"bannedbook","name":"fanqiang","desc":"翻墙-科学上网","summaryCn":"一个汇总各种翻墙方法的开源项目，帮你轻松突破网络限制","reasonCn":"这个仓库收集整理了多种VPN、Shadowsocks、V2Ray等代理工具的配置教程和使用方法，由社区维护更新。项目涵盖Android、iOS、Windows等多平台支持，包含详细步骤说明和常见问题解答。虽然涉及灰色地带，但作为技术爱好者了解网络知识有一定价值。","stars":53659,"starGrowth":11,"createdAt":"2015-01-14T00:34:25Z","silentRounds":0,"language":"Kotlin","topics":["ssr","v2ray","shadowsocks","fanqiang","proxy","goflyway","daze","brook","lightsocks","trojan","psiphon"],"aiDims":["非AI-实用"],"zone":"资源","funScore":0,"domainTags":["网络代理","教程","VPN","配置指南"],"domainKey":"网络代理","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"ssr","source":"github","weight":0.7},{"name":"v2ray","source":"github","weight":0.7},{"name":"shadowsocks","source":"github","weight":0.7},{"name":"fanqiang","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"goflyway","source":"github","weight":0.7},{"name":"daze","source":"github","weight":0.7},{"name":"brook","source":"github","weight":0.7},{"name":"lightsocks","source":"github","weight":0.7},{"name":"trojan","source":"github","weight":0.7},{"name":"psiphon","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/bannedbook/fanqiang","ts":"2026-09-05T02:38:58.868Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03361499468868224,"copyOk":false},{"repo":"anywhere-labs/deepseek-harness-desktop","owner":"anywhere-labs","name":"deepseek-harness-desktop","desc":"为 DeepSeek Harness (DSH) 生态打造的现代化桌面端体验","summaryCn":"把DeepSeek桌面化，本地就能爽玩各种大模型的神器","reasonCn":"这是一个专为DeepSeek Harness生态打造的现代化桌面客户端，利用TypeScript构建，提供了极佳的本地化AI应用体验。它将复杂的AI模型部署与工具调用过程封装在图形界面中，让用户无需繁琐配置即可在本地运行各类大模型，不仅支持模型管理，还能高效调用AI工具，完美平衡了功能深度与易用性。","stars":28510,"starGrowth":75,"createdAt":"2026-08-13T16:30:50Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["AI应用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["桌面客户端","DeepSeek","AI应用"],"domainKey":"桌面客户端","pushedAt":"2026-08-18T23:23:43Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/anywhere-labs/deepseek-harness-desktop","ts":"2026-08-18T23:23:43Z","category":"tool","momentum":["rising","daily"],"fromOfficial":false,"heatScore":0.34933795159593123,"copyOk":false},{"repo":"mlflow/mlflow","owner":"mlflow","name":"mlflow","desc":"The open source AI engineering platform for agents, LLMs, and ML models. MLflow enables teams of all sizes to debug, evaluate, monitor, and optimize production-quality AI applications while controlling costs and managing access to models and data.","summaryCn":"让AI模型从写代码到上线监控全流程丝滑，数据科学家必备神器","reasonCn":"Mlflow是端到端的AI工程化平台，覆盖实验追踪、模型版本管理、部署与监控。支持Python、R等多语言，兼容各大云厂商。无论是跑LLM还是传统机器学习，它都能帮你告别“在我机器上是好的”这种玄学尴尬，是企业级AI落地标配。","stars":28105,"starGrowth":9,"createdAt":"2018-06-05T16:05:58Z","silentRounds":0,"language":"Python","topics":["ml","agents"],"aiDims":["AI基础设施","模型部署","AI应用"],"zone":"AI","funScore":0,"domainTags":["机器学习","模型部署","实验追踪","AI平台","LLM"],"domainKey":"模型部署","pushedAt":"2026-09-23T21:21:36Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/mlflow/mlflow","ts":"2026-09-23T21:21:36Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.027960683152463976,"copyOk":false},{"repo":"darkreader/darkreader","owner":"darkreader","name":"darkreader","desc":"","summaryCn":"Dark Reader 是一款专为浏览器设计的开源扩展程序，致力于解决长时间浏览网","reasonCn":"这个项目的核心优势在于其强大的功能集和极高的自定义能力。用户不仅仅是可以一键开启或关闭深色模式，还可以通过 UI 界面精细地控制对比度、亮度、灰度等参数。这种灵活性使得 Dark Reader 能够适应各种风格的网站，无论是复杂的后台管理系统，还是设计感强烈的个人博客，它都能通","stars":22387,"starGrowth":1,"createdAt":"2014-11-15T14:22:52Z","silentRounds":2,"language":"TypeScript","topics":["dark-theme","chrome-extension","firefox-addon","dark-mode","night-mode","darkmode","nightmode","accessibility","eye-care"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["浏览器扩展","深色模式","护眼"],"domainKey":"浏览器扩展","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"dark-theme","source":"github","weight":0.7},{"name":"chrome-extension","source":"github","weight":0.7},{"name":"firefox-addon","source":"github","weight":0.7},{"name":"dark-mode","source":"github","weight":0.7},{"name":"night-mode","source":"github","weight":0.7},{"name":"darkmode","source":"github","weight":0.7},{"name":"nightmode","source":"github","weight":0.7},{"name":"accessibility","source":"github","weight":0.7},{"name":"eye-care","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.26,"source":"tier","url":"https://github.com/darkreader/darkreader","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"colinlet/PHP-Interview-QA","owner":"colinlet","name":"PHP-Interview-QA","desc":"系统性的 PHP 技术面试问答大全（面经）。涵盖 PHP8、MySQL、Redis、Nginx、系统设计、算法、Laravel 源码及高并发架构原理，助你斩获后端研发 Offer。PHP Interview Questions & Answers.","summaryCn":"PHP后端面试通关宝典，从语法到高并发架构全覆盖。","reasonCn":"这是一个系统性整理PHP技术面试问答的仓库，涵盖PHP8新特性、MySQL优化、Redis缓存策略、Nginx配置、高并发架构原理、Laravel源码解析及算法题。内容非常全面，适合准备后端开发面试的开发者，尤其是PHP技术栈的求职者，能帮你高效梳理知识点，提升面试通过率。","stars":2961,"starGrowth":0,"createdAt":"2018-09-06T10:10:47Z","silentRounds":10,"language":"TypeScript","topics":["backend"],"aiDims":["学习资源","开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["PHP","面试","后端","技术栈","面经"],"domainKey":"PHP","pushedAt":"2026-08-21T16:11:27Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/colinlet/PHP-Interview-QA","ts":"2026-08-21T16:11:27Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"oxnr/awesome-bigdata","owner":"oxnr","name":"awesome-bigdata","desc":"A curated list of awesome big data frameworks, ressources and other awesomeness.","summaryCn":"大数据开发者的宝藏清单，精选各类框架和工具","reasonCn":"这不是一个工具，而是一份精心整理的大数据资源清单，涵盖Hadoop、Spark、Flink等主流框架，以及数据可视化、流处理、存储等方向的开源项目。1.4万star说明很多人把它当参考资料，适合刚入行想快速了解大数据生态的人，或需要搭建平台时查漏补缺的工程师。","stars":14648,"starGrowth":2,"createdAt":"2014-07-04T15:01:03Z","silentRounds":0,"language":"","topics":["data-visualization"],"aiDims":["非AI-实用"],"zone":"资源","funScore":0,"domainTags":["大数据","框架清单","数据可视化","工具集","教程索引"],"domainKey":"大数据","pushedAt":"2026-07-31T12:51:12Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7}],"aiScore":0.35,"source":"search","url":"https://github.com/oxnr/awesome-bigdata","ts":"2026-07-31T12:51:12Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Lifelong-Robot-Learning/LIBERO","owner":"Lifelong-Robot-Learning","name":"LIBERO","desc":"","summaryCn":"面向终身机器人学习的知识迁移基准平台，支持无限任务生成和复杂操作任务。","reasonCn":"项目最大的亮点是它的程序化任务生成管道。理论上，这个管道可以生成无限数量的操纵任务，这意味着研究人员可以在一个动态变化的环境中持续评估算法的泛化能力，而不会遇到训练数据和测试任务重复的问题。这对于研究知识迁移和终身学习来说至关重要，因为现实世界本身就是不断变化","stars":2346,"starGrowth":37,"createdAt":"2023-04-08T18:58:55Z","silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["AI Agent","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["机器人学习","知识迁移","任务生成","强化学习","多任务学习"],"domainKey":"机器人学习","pushedAt":"2026-09-22T20:36:00.388Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/Lifelong-Robot-Learning/LIBERO","ts":"2026-09-22T20:36:00.388Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.1583145780051151,"copyOk":true},{"repo":"apify/crawlee","owner":"apify","name":"crawlee","desc":"Crawlee—A web scraping and browser automation library for Node.js to build reliable crawlers. In JavaScript and TypeScript. Extract data for AI, LLMs, RAG, or GPTs. Download HTML, PDF, JPG, PNG, and other files from websites. Works with Puppeteer, Playwright, Cheerio, JSDOM, and raw HTTP. Both headful and headless mode. With proxy rotation.","summaryCn":"自动爬网抓数据，喂给AI当养料，爬虫界的瑞士军刀","reasonCn":"基于Node.js的爬虫框架，支持JS/TS，自动处理反爬策略和JavaScript渲染页面，内置代理轮换与请求重试机制。专为AI数据收集设计，能提取结构化数据供RAG、LLM或GPTs使用，支持批量下载HTML、PDF、图片等文件，功能全面。","stars":25867,"starGrowth":3,"createdAt":"2016-08-26T18:35:03Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["工具调用","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Web爬虫","Node.js","数据采集"],"domainKey":"Web爬虫","pushedAt":"2026-09-23T16:18:35Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/apify/crawlee","ts":"2026-09-23T16:18:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Dogfalo/materialize","owner":"Dogfalo","name":"materialize","desc":"Materialize, a CSS Framework based on Material Design","summaryCn":"一个自带颜值的CSS框架，轻松搞定酷炫界面无需折腾设计","reasonCn":"这是基于Material Design规范的开源CSS框架，提供现成的精美组件，如卡片、按钮和导航栏。用它开发网页能省不少前端样式时间，适合想快速搭建美观界面的开发者，且完全免费开源，社区支持不错，是前端开发的好帮手。","stars":38803,"starGrowth":0,"createdAt":"2014-09-12T19:35:38Z","silentRounds":3,"language":"JavaScript","topics":["design"],"aiDims":["非AI-好玩","非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["CSS","前端","Material Design","组件库","响应式"],"domainKey":"CSS","pushedAt":"2026-08-20T13:49:11Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/Dogfalo/materialize","ts":"2026-08-20T13:49:11Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lerna/lerna","owner":"lerna","name":"lerna","desc":"Lerna is a fast, modern build system for managing and publishing multiple JavaScript/TypeScript packages from the same repository.","summaryCn":"一个仓库管N个包，代码共享不再头疼，团队协作利器","reasonCn":"基于npm/yarn/pnpm管理多个子包，自动版本同步、发布、依赖更新。支持单仓多包架构，一键构建全部模块，避免重复配置。提供changelog自动生成、git版本控制集成、并行执行脚本等核心能力，特别适合大型前端项目和组件库开发，能显著提升团队开发效率。","stars":36055,"starGrowth":1,"createdAt":"2015-12-04T09:36:55Z","silentRounds":3,"language":"TypeScript","topics":["lerna","monorepo","npm","package","publishing"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["monorepo","npm","TypeScript"],"domainKey":"monorepo","pushedAt":"2026-08-17T09:58:06Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"lerna","source":"github","weight":0.7},{"name":"monorepo","source":"github","weight":0.7},{"name":"npm","source":"github","weight":0.7},{"name":"package","source":"github","weight":0.7},{"name":"publishing","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/lerna/lerna","ts":"2026-08-17T09:58:06Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mojolicious/mojo","owner":"mojolicious","name":"mojo","desc":":sparkles: Mojolicious - Perl real-time web framework","summaryCn":"老牌Perl Web框架，实时通信和自包含架构让开发超省心","reasonCn":"技术上，它最厉害的地方是自包含架构。传统Web开发需要搭Apache或Nginx再加应用服务器，配置麻烦还容易出问题。但Mojolicious内置了HTTP服务器和客户端，一个命令就能启动完整应用，相当于把Web服务器和应用代码打包在一起了。","stars":2748,"starGrowth":1,"createdAt":"2008-10-21T07:50:26Z","silentRounds":25,"language":"Perl","topics":["web-framework"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Perl","Web框架","WebSocket","实时通信","后端开发"],"domainKey":"Perl","pushedAt":"2026-08-09T23:59:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Perl","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/mojolicious/mojo","ts":"2026-08-09T23:59:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"makeplane/plane","owner":"makeplane","name":"plane","desc":"🔥🔥🔥 Open-source Jira, Linear, Monday, and ClickUp alternative. Plane is a modern project management platform to manage tasks, sprints, docs, and triage.","summaryCn":"免费开源的项目管理神器，堪比Jira Linear的平替替代方案","reasonCn":"基于React和Go构建的现代化项目管理平台，支持看板、列表、日历等多种视图。提供Sprint管理、问题分类、文档协作等完整功能。可自托管部署，数据完全自主掌控，适合中小型团队敏捷开发使用，开箱即用且持续迭代活跃。","stars":59757,"starGrowth":21,"createdAt":"2022-11-19T12:55:01Z","silentRounds":0,"language":"TypeScript","topics":["docker"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["项目管理","Jira替代","看板工具"],"domainKey":"项目管理","pushedAt":"2026-09-23T20:46:24Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/makeplane/plane","ts":"2026-09-23T20:46:24Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06405426979266027,"copyOk":true},{"repo":"cube-js/cube","owner":"cube-js","name":"cube","desc":"📊 Cube Core is open-source semantic layer for AI, BI and embedded analytics","summaryCn":"统一语义层让AI和BI系统共享数据口径，告别重复造轮子","reasonCn":"Cube.js 是一个开源语义层，为 AI、BI 和嵌入式分析提供统一的数据抽象层，通过预聚合 SQL 视图实现数据一致性。支持 PostgreSQL、Snowflake 等后端，可对接 Superset、Metabase 等可视化工具，也可提供 REST/GraphQL API 给前端直接调用。开发者可用 Node.js 或 Python 快速构建分析应用，支持 Docker 部署，配置简单。","stars":20891,"starGrowth":4,"createdAt":"2018-09-16T18:58:46Z","silentRounds":0,"language":"Rust","topics":["agents"],"aiDims":["AI基础设施","AI应用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["语义层","数据分析","PostgreSQL"],"domainKey":"语义层","pushedAt":"2026-09-23T21:51:16Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/cube-js/cube","ts":"2026-09-23T21:51:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sinatra/sinatra","owner":"sinatra","name":"sinatra","desc":"Classy web-development dressed in a DSL (official / canonical repo)","summaryCn":"用最少代码写Web应用，Ruby程序员的快乐老家","reasonCn":"Sinatra是Ruby语言中最经典的轻量级Web框架，以极简主义著称。它适合快速构建API、个人项目或小工具。虽然它本身与AI无直接关联，但它是许多早期AI web界面（如Gradio的前身理念）的灵感来源。对于想体验Ruby美学或快速搭建简单工具的开发者来说，它依然优雅。","stars":12454,"starGrowth":1,"createdAt":"2009-01-14T01:27:30Z","silentRounds":7,"language":"Ruby","topics":["web-framework"],"aiDims":["非AI-好玩","开发者工具","教程"],"zone":"工具","funScore":0,"domainTags":["Ruby","微框架","DSL","轻量级","Web开发"],"domainKey":"Ruby","pushedAt":"2026-07-20T09:59:21Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/sinatra/sinatra","ts":"2026-07-20T09:59:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ComicSparks/pikapika","owner":"ComicSparks","name":"pikapika","desc":"A comic browser，support Android / iOS / MacOS / Windows / Linux.","summaryCn":"全平台漫画阅读器，安卓苹果电脑都能用，追番党必备神器","reasonCn":"这是一个支持多平台的漫画浏览器应用，使用Dart语言开发，完全兼容Android、iOS、MacOS、Windows和Linux五大操作系统。界面设计简洁流畅，支持多种漫画源聚合，提供离线缓存下载功能，可以批量保存漫画章节到本地，适合喜欢追漫画的读者使用。项目star数超过8000，社区维护活跃，持续更新中。","stars":8335,"starGrowth":1,"createdAt":"2021-07-27T01:25:07Z","silentRounds":1,"language":"Dart","topics":["game"],"aiDims":["游戏","非AI-好玩","可视化"],"zone":"创意","funScore":1,"domainTags":["漫画阅读器","Flutter","跨平台应用","多媒体","离线阅读"],"domainKey":"漫画阅读器","pushedAt":"2026-07-20T07:14:22Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ComicSparks/pikapika","ts":"2026-07-20T07:14:22Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cocos/cocos-engine-external","owner":"cocos","name":"cocos-engine-external","desc":"external 3rd party modules for cocos-engine","summaryCn":"Cocos游戏引擎的第三方模块合集，扩展更多功能，做游戏更省事","reasonCn":"Cocos引擎的第三方模块仓库，汇集了大量非官方的扩展功能组件。包括游戏物理、渲染、网络通信等各种实用模块，帮助开发者快速集成常用功能，减少重复造轮子。虽然是C++编写，但接口设计友好，适合想要扩展Cocos引擎能力的游戏开发者。","stars":71,"starGrowth":0,"createdAt":"2016-06-29T03:14:21Z","silentRounds":24,"language":"C++","topics":[],"aiDims":["非AI-好玩","游戏","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Cocos","游戏引擎","C++"],"domainKey":"Cocos","pushedAt":"2026-08-14T10:04:13Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/cocos/cocos-engine-external","ts":"2026-08-14T10:04:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hashicorp/terraform","owner":"hashicorp","name":"terraform","desc":"Terraform enables you to safely and predictably create, change, and improve infrastructure. It is a source-available tool that codifies APIs into declarative configuration files that can be shared amongst team members, treated as code, edited, reviewed, and versioned.","summaryCn":"用代码描述并自动搭建任何云基础设施的声明式工具","reasonCn":"Terraform通过声明式配置文件管理云资源，支持AWS、Azure、Google Cloud等上百个云平台。使用HCL语言定义网络、虚拟机、数据库等架构，一键创建、更新或删除，实现基础设施即代码。适合团队协同、版本控制和自动化运维，大幅减少手动操作错误。","stars":49713,"starGrowth":13,"language":"Go","topics":["graph","infrastructure-as-code","terraform","cloud","cloud-management"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["基础设施即代码","云服务","DevOps","配置管理"],"domainKey":"基础设施即代码","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"graph","source":"github","weight":0.7},{"name":"infrastructure-as-code","source":"github","weight":0.7},{"name":"terraform","source":"github","weight":0.7},{"name":"cloud","source":"github","weight":0.7},{"name":"cloud-management","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/hashicorp/terraform","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.039784503047492606,"copyOk":true},{"repo":"camel-ai/camel","owner":"camel-ai","name":"camel","desc":"🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org","summaryCn":"让多个AI角色组团打怪升级，开创多智能体协作新纪元。","reasonCn":"技术上，它基于Python深度构建，设计了优雅的“角色扮演交谈”机制。通过精心设计的提示词和交互协议，让AI在对话中实现自主任务分解和动态修正。它不仅支持单轮问答，还能处理复杂的长周期合作任务，并且能够很好地对接各种主流的大语言模型以及多模态模型。","stars":17757,"starGrowth":4,"createdAt":"2023-03-17T21:41:54Z","silentRounds":0,"language":"Python","topics":["deep-learning"],"aiDims":["AI Agent","多模态Agent","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["多智能体","AI框架","大模型"],"domainKey":"多智能体","pushedAt":"2026-09-20T04:56:23Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/camel-ai/camel","ts":"2026-09-20T04:56:23Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"go-shiori/shiori","owner":"go-shiori","name":"shiori","desc":"Simple bookmark manager built with Go","summaryCn":"命令行下的书签神器，把网页收藏整理变得简单清爽","reasonCn":"这款基于Go语言开发的书签管理器，专为终端用户打造。它支持命令行操作，导入导出功能强大，能帮你把浏览器里散乱的链接统一收纳。虽然不带AI智能标签功能，但其极简架构和资源占用优势，非常适合追求高效管理的极客群体日常使用。","stars":11646,"starGrowth":1,"createdAt":"2018-01-23T13:57:11Z","silentRounds":6,"language":"Go","topics":["command-line"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["书签管理","命令行工具","Web UI","SQLite","极简"],"domainKey":"书签管理","pushedAt":"2026-07-10T06:36:40Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/go-shiori/shiori","ts":"2026-07-10T06:36:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"reagent-project/reagent","owner":"reagent-project","name":"reagent","desc":"A minimalistic ClojureScript interface to React.js","summaryCn":"用Clojure方式写React，简洁优雅的前端开发新体验","reasonCn":"Reagent是用函数式思维封装的React框架，采用类Hiccup的语法替代繁琐的JSX。它将React组件抽象为纯函数，配合Reagent的响应式状态管理，让前端代码像写Clojure一样清爽。相比官方React，它大幅减少样板代码，更适合追求简洁表达的前端开发者。","stars":4888,"starGrowth":0,"createdAt":"2013-12-18T18:12:03Z","silentRounds":24,"language":"Clojure","topics":["frontend"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["ClojureScript","React前端","函数式编程"],"domainKey":"ClojureScript","pushedAt":"2026-08-19T07:18:08Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Clojure","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/reagent-project/reagent","ts":"2026-08-19T07:18:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vueComponent/ant-design-vue","owner":"vueComponent","name":"ant-design-vue","desc":"🌈  An enterprise-class UI components based on Ant Design and Vue. 🐜","summaryCn":"一个超全的Vue企业级UI组件库，开箱即用贼方便","reasonCn":"基于Ant Design设计规范打造，专为Vue3优化，提供80多个高质量组件，覆盖表单、表格、弹窗等高频场景。组件丰富度高、文档详尽，是构建中后台管理系统的首选方案，社区活跃、持续迭代，适合企业级项目开发，能大幅提升前端开发效率。","stars":21662,"starGrowth":1,"createdAt":"2017-09-20T06:04:01Z","silentRounds":2,"language":"Vue","topics":["frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Vue","UI组件库","前端框架","Ant Design"],"domainKey":"Vue","pushedAt":"2026-09-19T01:38:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/vueComponent/ant-design-vue","ts":"2026-09-19T01:38:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"strapi/strapi","owner":"strapi","name":"strapi","desc":"🚀 Strapi is the leading open-source headless CMS. It’s 100% JavaScript/TypeScript, fully customizable, and developer-first.","summaryCn":"傻瓜式内容管理后台，帮你快速搭建网站和API。","reasonCn":"Strapi是一个开源的Headless CMS（无头内容管理系统），提供可视化的后台管理界面和RESTful API，帮助开发者快速构建前端应用。它基于Node.js，高度可定制，支持多种数据类型和插件扩展，是构建AI应用数据层的热门选择。","stars":73215,"starGrowth":4,"createdAt":"2015-09-30T15:34:48Z","silentRounds":0,"language":"TypeScript","topics":["api"],"aiDims":["AI应用","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["CMS","Headless","API","Node.js"],"domainKey":"CMS","pushedAt":"2026-09-23T21:11:20Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/strapi/strapi","ts":"2026-09-23T21:11:20Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"keras-team/keras","owner":"keras-team","name":"keras","desc":"Deep Learning for humans","summaryCn":"人类友好的深度学习库，让AI建模像搭积木一样简单","reasonCn":"Keras是Google维护的高层深度学习API，以简洁易用著称。它支持TensorFlow、JAX、PyTorch等多个后端，让开发者无需关心底层实现即可快速构建神经网络。特别适合初学者入门深度学习，也受经验丰富的研究者青睐。提供大量预训练模型和实用工具，大大缩短了从想法到原型的开发周期。","stars":64328,"starGrowth":0,"createdAt":"2015-03-28T00:35:42Z","silentRounds":1,"language":"Python","topics":["ml"],"aiDims":["模型与训练","大语言模型"],"zone":"AI","funScore":0,"domainTags":["深度学习","神经网络","AI框架"],"domainKey":"深度学习","pushedAt":"2026-09-23T20:15:22Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/keras-team/keras","ts":"2026-09-23T20:15:22Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Universal-Commerce-Protocol/ucp","owner":"Universal-Commerce-Protocol","name":"ucp","desc":"Specification and documentation for the Universal Commerce Protocol (UCP)","summaryCn":"让全球电商互联互通的商业协议，一套标准打通买卖全流程","reasonCn":"这是一个通用商业协议规范项目，致力于解决电商行业中支付、物流、商品标准等跨平台互操作难题。提供完整的API规范、数据模型和集成文档，帮助开发者快速接入分布式商业网络。虽然不涉及AI技术，但作为Web3电商基础设施，对区块链、去中心化商业方向有兴趣的开发者值得了解。","stars":3388,"starGrowth":0,"createdAt":"2025-12-31T02:17:21Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["电商协议","互操作性","Web3","API规范","数据模型"],"domainKey":"电商协议","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/Universal-Commerce-Protocol/ucp","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dagrejs/dagre","owner":"dagrejs","name":"dagre","desc":"Directed graph layout for JavaScript","summaryCn":"JavaScript版流程图神器，再乱的图也能自动排得整整齐齐","reasonCn":"dagre是一个基于Sugiyama算法的JavaScript有向图布局库，支持SVG和Canvas两种输出格式。它能自动计算节点位置和边的路径，支持从左到右、从上到下等多种布局方向。相比Graphviz，它在浏览器端运行，可与D3.js无缝集成，适合需要交互式流程图、依赖关系图、组织架构图的Web开发场景。代码简洁，API友好，npm安装即可使用，支持自定义节点样式和边样式，是前端可视化项目的实用工具。","stars":5799,"starGrowth":0,"createdAt":"2012-09-27T04:39:56Z","silentRounds":2,"language":"TypeScript","topics":[],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["图布局","JavaScript","流程图","D3.js","可视化"],"domainKey":"图布局","pushedAt":"2026-08-08T15:09:12Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/dagrejs/dagre","ts":"2026-08-08T15:09:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"element-plus/element-plus","owner":"element-plus","name":"element-plus","desc":"🎉 A Vue.js 3 UI Library made by Element team","summaryCn":"Vue 3的UI组件库，颜值高、文档全，做前端页面不再头秃","reasonCn":"Element Plus 是 Element UI 的 Vue 3 升级版，提供 80+ 组件，覆盖表单、表格、弹窗等常见场景。采用 TypeScript 编写，完整类型提示，支持按需加载减小打包体积。配套文档中文友好，支持主题定制和暗色模式，是 Vue 3 生态中最成熟的企业级 UI 框架之一。","stars":27780,"starGrowth":1,"createdAt":"2020-07-21T06:51:19Z","silentRounds":2,"language":"TypeScript","topics":["vuejs","component-library","element-ui","vue","vue-components","element-plus"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Vue","UI组件库","前端框架","TypeScript","后台管理"],"domainKey":"Vue","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"vuejs","source":"github","weight":0.7},{"name":"component-library","source":"github","weight":0.7},{"name":"element-ui","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"vue-components","source":"github","weight":0.7},{"name":"element-plus","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/element-plus/element-plus","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Lightning-AI/pytorch-lightning","owner":"Lightning-AI","name":"pytorch-lightning","desc":"Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.","summaryCn":"用最少代码实现PyTorch训练，分布式、混合精度全搞定","reasonCn":"PyTorch Lightning让训练代码从几百行缩减到几十行，自动处理分布式训练、混合精度、日志监控等繁琐工作。只需定义模型结构，框架接管训练循环，支持从单机到上万GPU无缝扩展。与Hugging Face生态深度集成，集成BERT、GPT等预训练模型非常顺滑，是研究者和团队的主流选择。","stars":31357,"starGrowth":2,"createdAt":"2019-03-31T00:45:57Z","silentRounds":0,"language":"Python","topics":["ml","ai"],"aiDims":["模型与训练","微调","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["PyTorch","分布式训练","AI框架","深度学习"],"domainKey":"PyTorch","pushedAt":"2026-09-21T11:44:30Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Lightning-AI/pytorch-lightning","ts":"2026-09-21T11:44:30Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/unilm","owner":"microsoft","name":"unilm","desc":"Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities","summaryCn":"很多时候我们会被各种不同的模型架构搞晕，比如有的专门处理文字，有的专门","reasonCn":"其核心亮点在于创新的Mask机制，通过灵活配置输入序列，模型可以在BERT（理解）、GPT（生成）和Seq2Seq（翻译/对话）之间无缝切换。这种设计打破了模型应用场景的界限，极大降低了为不同任务重复训练大模型的成本。微软在其中集成了最新的跨模态融合技术，让模型在看到图片的同时，","stars":22227,"starGrowth":2,"createdAt":"2019-07-23T04:15:28Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["大语言模型","模型与训练","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["深度学习","预训练模型","多模态","NLP","计算机视觉"],"domainKey":"多模态","pushedAt":"2026-09-21T03:27:00Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/microsoft/unilm","ts":"2026-09-21T03:27:00Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"axmolengine/axmol","owner":"axmolengine","name":"axmol","desc":"Axmol Engine – A Multi-platform Engine for Desktop, XBOX (UWP), WebAssembly and Mobile games.","summaryCn":"跨平台游戏引擎新选择，一次开发同时支持PC主机和网页运行","reasonCn":"技术上，它是cocos2d-x项目的现代化继任者，采用了C++20标准，架构更加清晰。最大的亮点是它支持的运行平台非常广泛——从传统的PC和移动端，到Xbox游戏主机，再到WebAssembly让游戏直接在浏览器中运行。对于想要覆盖多平台的团队来说，这大大降低了开发成本。","stars":1459,"starGrowth":0,"createdAt":"2019-11-23T14:08:10Z","silentRounds":9,"language":"C++","topics":["axmol","axmolengine","cocos2d-x","cpp20","cppwinrt","d3d","d3d11","d3d12","game-engine","metal","opengl","uwp","vulkan","xbox"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","C++","DirectX"],"domainKey":"游戏引擎","pushedAt":"2026-08-17T11:51:46Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"axmol","source":"github","weight":0.7},{"name":"axmolengine","source":"github","weight":0.7},{"name":"cocos2d-x","source":"github","weight":0.7},{"name":"cpp20","source":"github","weight":0.7},{"name":"cppwinrt","source":"github","weight":0.7},{"name":"d3d","source":"github","weight":0.7},{"name":"d3d11","source":"github","weight":0.7},{"name":"d3d12","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"metal","source":"github","weight":0.7},{"name":"opengl","source":"github","weight":0.7},{"name":"uwp","source":"github","weight":0.7},{"name":"vulkan","source":"github","weight":0.7},{"name":"xbox","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/axmolengine/axmol","ts":"2026-08-17T11:51:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mikestefanello/pagoda","owner":"mikestefanello","name":"pagoda","desc":"Rapid, easy full-stack web development starter kit and admin panel in Go","summaryCn":"Go语言写的全栈脚手架，内置管理后台，半小时搭建网站原型","reasonCn":"用Go语言构建的快速全栈Web开发启动模板，包含前端React、后端Go API、PostgreSQL数据库和管理后台。提供用户认证、文件上传、邮件通知等开箱即用功能，无需复杂配置即可部署生产级应用，适合需要快速验证想法的独立开发者和小型团队。","stars":2953,"starGrowth":0,"createdAt":"2021-12-03T11:04:30Z","silentRounds":24,"language":"Go","topics":["web-framework"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["全栈开发","Go","PostgreSQL"],"domainKey":"全栈开发","pushedAt":"2026-08-14T00:36:35Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/mikestefanello/pagoda","ts":"2026-08-14T00:36:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Tresjs/tres","owner":"Tresjs","name":"tres","desc":"Declarative ThreeJS using Vue Components","summaryCn":"用Vue组件写3D场景，告别繁琐的ThreeJS代码","reasonCn":"基于Vue的3D渲染框架，将ThreeJS封装成Vue组件形式。通过简单的标签组合就能搭建3D场景，支持材质、灯光、动画等完整功能。配合Vite实现热更新，开发体验流畅。相比原生ThreeJS写法，3D开发门槛大幅降低，对Vue生态更友好。","stars":3736,"starGrowth":0,"createdAt":"2022-11-27T20:36:37Z","silentRounds":2,"language":"Vue","topics":["composable","custom-renderer","declarative","threejs","vite","vue","webgl"],"aiDims":["可视化","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Vue","ThreeJS","3D渲染"],"domainKey":"Vue","pushedAt":"2026-08-17T22:09:44Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"composable","source":"github","weight":0.7},{"name":"custom-renderer","source":"github","weight":0.7},{"name":"declarative","source":"github","weight":0.7},{"name":"threejs","source":"github","weight":0.7},{"name":"vite","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/Tresjs/tres","ts":"2026-08-17T22:09:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dj-bolt/django-bolt","owner":"dj-bolt","name":"django-bolt","desc":"The fastest Python web framework, built on Django. Rust (Actix Web) HTTP server + async typed handlers + msgspec, with the full Django ORM, Admin, middleware and auth. 300k+ req/s, no gunicorn/uvicorn.","summaryCn":"让Django跑得比Rust还快，Python开发者的性能狂喜神器。","reasonCn":"这是一个基于Rust构建的高性能HTTP服务器，直接替换了传统的Gunicorn或Uvicorn，专为Django框架打造。它利用Actix Web框架，支持异步处理和类型化处理器，并集成了msgspec库。项目保留了Django原生的ORM、Admin后台、中间件和认证系统，同时将性能提升至每秒30万次请求，实现了Python开发体验与高性能Web服务的完美结合。","stars":1693,"starGrowth":2,"createdAt":"2025-09-20T19:04:44Z","silentRounds":0,"language":"Python","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Python","Django","高性能","HTTP服务器","Rust"],"domainKey":"Python","pushedAt":"2026-09-22T23:38:56Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/dj-bolt/django-bolt","ts":"2026-09-22T23:38:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"calesthio/OpenMontage","owner":"calesthio","name":"OpenMontage","desc":"World's first open-source, agentic video production system. 12 production pipelines, 100+ tools, 700+ agent skill and production-knowledge files. Turn your AI coding assistant into a full video production studio.","summaryCn":"把你的代码助手变成电影导演，一键生成大片级视频","reasonCn":"这是一个全球首个开源的代理式视频制作系统，集成了12条生产管线和700+种技能文件，旨在通过AI让视频创作变得像写代码一样简单。它将AI编程助手的功能扩展到了视频领域，提供从分镜、渲染到后期特效的完整自动化解决方案。","stars":60912,"starGrowth":103,"createdAt":"2026-03-29T15:23:22Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["AI应用","多模态Agent","创意工具"],"zone":"AI","funScore":0,"domainTags":["视频生成","AI Agent","内容创作自动化","多模态流水线"],"domainKey":"视频生成","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/calesthio/OpenMontage","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.36608747044917256,"copyOk":true},{"repo":"layabox/LayaAir3.0UnityPlugin","owner":"layabox","name":"LayaAir3.0UnityPlugin","desc":"The tool is used to export Unity assets to Laya","summaryCn":"Unity到LayaAir的资产导出工具，游戏项目一键转换超省心","reasonCn":"这是LayaAir引擎官方推出的Unity资产导出插件，支持将Unity场景、模型、贴图、动画等资源一键转换为LayaAir项目格式。美术和程序可以在Unity中完成内容创作，然后通过插件导出为LayaAir可识别的项目，直接发布到Web、iOS、Android等多端平台，大幅减少跨引擎开发的重复劳动。","stars":39,"starGrowth":0,"createdAt":"2023-03-28T06:29:32Z","silentRounds":25,"language":"C#","topics":[],"aiDims":["游戏","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Unity","LayaAir","游戏引擎","跨平台"],"domainKey":"Unity","pushedAt":"2026-07-27T08:44:41Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/layabox/LayaAir3.0UnityPlugin","ts":"2026-07-27T08:44:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"101arrowz/fflate","owner":"101arrowz","name":"fflate","desc":"High performance (de)compression in an 8kB package","summaryCn":"不到8KB的压缩神器，比pako还小的工业级压缩库","reasonCn":"这是一个高性能的压缩解压缩库，整个包只有8KB，支持DEFLATE、GZIP、ZLIB、ZIP、PNG等多种格式。性能远超同类工具，压缩速度接近zlib原生，解压速度甚至更快。TypeScript编写，零依赖，在浏览器和Node.js环境中都能流畅运行，非常适合对包体积敏感的前端项目和需要快速处理压缩数据的场景。","stars":3020,"starGrowth":1,"createdAt":"2020-08-29T05:58:07Z","silentRounds":3,"language":"TypeScript","topics":["compression","decompression","high-performance","pako","tiny","zlib"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["压缩","解压","TypeScript"],"domainKey":"压缩","pushedAt":"2026-05-16T03:02:11Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"compression","source":"github","weight":0.7},{"name":"decompression","source":"github","weight":0.7},{"name":"high-performance","source":"github","weight":0.7},{"name":"pako","source":"github","weight":0.7},{"name":"tiny","source":"github","weight":0.7},{"name":"zlib","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/101arrowz/fflate","ts":"2026-05-16T03:02:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"framework7io/framework7","owner":"framework7io","name":"framework7","desc":"Full featured HTML framework for building iOS & Android apps","summaryCn":"纯JS写的跨平台移动应用框架，快速构建原生风格App","reasonCn":"Framework7的核心卖点是：UI组件高度还原iOS和Android原生风格，路由系统管理页面跳转，状态管理支持Pinia/Vuex等主流方案，支持PWA离线模式，也可以打包成原生App发布到应用商店。它提供完整的脚手架工具，一行命令就能生成项目模板。","stars":18759,"starGrowth":0,"createdAt":"2014-02-23T12:15:53Z","silentRounds":4,"language":"JavaScript","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["移动应用开发","UI框架","跨平台"],"domainKey":"移动应用开发","pushedAt":"2026-09-08T16:13:19Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/framework7io/framework7","ts":"2026-09-08T16:13:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jeecgboot/JeecgBoot","owner":"jeecgboot","name":"JeecgBoot","desc":"【低代码迈入v2.0时代，一句话即可生成整个系统】企业级AI低代码平台，一键生成前后端代码甚至整个系统。 AI Skills 一句话画流程、设计表单、生成报表、大屏。内置 AI应用平台涵盖：AI聊天、知识库、流程编排、MCP插件等，兼容主流大模型。引领AI低代码「Skills 生成 → 在线配置 → 代码生成 → 手工合并->AI修改」开发模式，解决 Java 项目 90% 重复工作，提高效率又不失灵活。","summaryCn":"用一句话描述需求就能自动生成整套系统的AI低代码平台","reasonCn":"这是国内Star数最高的企业级低代码平台之一，最新v2.0版本深度集成了AI能力。通过自然语言描述即可生成前后端代码、业务流程、报表甚至大屏。内置AI聊天、知识库、流程编排等应用，兼容主流大模型，能覆盖Java项目90%的重复开发工作，大幅降低企业数字化门槛。","stars":47944,"starGrowth":0,"createdAt":"2018-11-26T10:40:00Z","silentRounds":1,"language":"Java","topics":["rag","llm","ai"],"aiDims":["AI应用","代码助手","AI Agent"],"zone":"AI","funScore":0.7,"domainTags":["AI低代码","代码生成","Java开发","大模型应用"],"domainKey":"AI低代码","pushedAt":"2026-09-22T14:26:15Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/jeecgboot/JeecgBoot","ts":"2026-09-22T14:26:15Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dingodb/dingo","owner":"dingodb","name":"dingo","desc":"A multi-modal vector database that supports upserts and vector queries using unified SQL (MySQL-Compatible) on structured and unstructured data, while meeting the requirements of high concurrency and ultra-low latency.","summaryCn":"用SQL就能查向量，多模态数据库让RAG应用开发更简单","reasonCn":"DingoDB是一款面向AI的多模态向量数据库，创新性地用统一SQL语句同时操作向量数据和结构化数据。基于MySQL协议兼容，开发者无需学习新语法，用熟悉的SQL就能完成向量检索。支持高并发场景，适合构建RAG系统和多模态AI应用，降低了向量数据库的使用门槛。","stars":1704,"starGrowth":0,"createdAt":"2021-10-14T02:13:51Z","silentRounds":28,"language":"Java","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["向量数据库","SQL","RAG","多模态检索","MySQL兼容"],"domainKey":"向量数据库","pushedAt":"2026-09-23T15:20:27Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/dingodb/dingo","ts":"2026-09-23T15:20:27Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"runatlantis/atlantis","owner":"runatlantis","name":"atlantis","desc":"Terraform Pull Request Automation","summaryCn":"让代码审查变审批流的DevOps自动化工具","reasonCn":"这是一个把Terraform和Git Pull Request结合的工具。它能自动分析代码变更，生成基建计划，并在PR合并后自动执行。对于搞云原生和K8s的团队来说，这是把“口头说改一下”变成“代码驱动变更”的神器，让基础设施像代码一样可追溯。","stars":9296,"starGrowth":0,"createdAt":"2018-02-06T20:28:06Z","silentRounds":3,"language":"Go","topics":["devops","automation"],"aiDims":["开发者工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Terraform","DevOps","基础设施即代码","Pull Request","CI/CD"],"domainKey":"Terraform","pushedAt":"2026-09-23T03:38:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/runatlantis/atlantis","ts":"2026-09-23T03:38:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tj/n","owner":"tj","name":"n","desc":"Node version management","summaryCn":"一行命令就能切换Node.js版本，前端开发者的时间机器，告别版本困扰","reasonCn":"tj/n是一个轻量级Node.js版本管理工具，用Shell脚本写成，无需额外依赖。通过简单的n命令，开发者可以一键安装、切换不同版本的Node.js，支持LTS、最新稳定版以及自定义路径。相比nvm更简洁，没有复杂的配置，适合追求高效的开发者。","stars":19511,"starGrowth":0,"createdAt":"2011-01-05T14:53:19Z","silentRounds":25,"language":"Shell","topics":[],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Node.js","版本管理","前端开发"],"domainKey":"Node.js","pushedAt":"2026-04-12T09:32:18Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/tj/n","ts":"2026-04-12T09:32:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CommunityToolkit/WindowsCommunityToolkit","owner":"CommunityToolkit","name":"WindowsCommunityToolkit","desc":"ARCHIVE - New Repo: https://aka.ms/toolkit/windows","summaryCn":"微软官方Windows开发工具箱，UI控件封装得明明白白","reasonCn":"这是微软官方维护的Windows应用开发工具包，提供了大量开箱即用的UI控件和扩展功能，包括动画、交互效果、数据绑定等实用组件。基于WinUI3和WinUI2构建，帮助开发者快速搭建现代化Windows界面，告别从零写控件的麻烦。虽然项目已归档，但其设计理念仍影响深远，适合想快速开发Windows桌面应用的团队和个人开发者使用。","stars":6071,"starGrowth":0,"createdAt":"2016-06-17T21:29:46Z","silentRounds":25,"language":"C#","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Windows开发","UI控件","WinUI","C#","桌面应用"],"domainKey":"Windows开发","pushedAt":"2024-10-31T15:52:52Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/CommunityToolkit/WindowsCommunityToolkit","ts":"2024-10-31T15:52:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ludwig-ai/ludwig","owner":"ludwig-ai","name":"ludwig","desc":"Low-code framework for building custom LLMs, neural networks, and other AI models","summaryCn":"想象一下，你手里有一堆数据，想训练一个能预测股价或者识别文档类型的AI，","reasonCn":"技术上，Ludwig 采用声明式设计（Declarative Design）。你编写一个YAML格式的配置文件，指定数据类型（文本、图片、分类等）和模型参数。它会在底层自动构建复杂的计算图，支持分布式训练，并且能够根据任务自动选择最佳的模型架构。","stars":11765,"starGrowth":2,"createdAt":"2018-12-27T23:58:12Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["微调","模型与训练","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AutoML","深度学习","模型训练","低代码框架"],"domainKey":"模型训练","pushedAt":"2026-09-21T21:24:48Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ludwig-ai/ludwig","ts":"2026-09-21T21:24:48Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"onetimesecret/onetimesecret","owner":"onetimesecret","name":"onetimesecret","desc":"Keep passwords and other sensitive information out of your chat logs and inboxes.","summaryCn":"在日常协作中，常常需要把 API 密钥、服务器密码等敏感信息通过聊天工具或邮","reasonCn":"在日常协作中，常常需要把 API 密钥、服务器密码等敏感信息通过聊天工具或邮件临时共享，但一旦泄露或被搜索记录保存，就会带来安全隐患。OneTimeSecret 正是为了解决这一痛点而设计的，它提供了一个简洁的网页界面，用户可以输入任意文本或密码，系统会生成一个唯一的 URL 并对内","stars":2941,"starGrowth":0,"createdAt":"2013-01-30T11:21:31Z","silentRounds":3,"language":"Ruby","topics":["privacy"],"aiDims":["安全工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["密码分享","信息安全","临时密钥"],"domainKey":"密码分享","pushedAt":"2026-09-23T18:47:01Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/onetimesecret/onetimesecret","ts":"2026-09-23T18:47:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Eclmist/Matcha-Editor","owner":"Eclmist","name":"Matcha-Editor","desc":"A game engine editor built in WPF","summaryCn":"用C# WPF做的游戏编辑器，可视化拖拽就能搭游戏关卡","reasonCn":"基于WPF技术栈打造的轻量级游戏引擎编辑器，提供所见即所得的场景编辑能力，开发者无需手写坐标和配置就能拖拽摆放精灵、动画和碰撞体。C#生态对新手友好，适合想做独立游戏或小游戏项目的个人开发者快速搭建原型，目前社区规模较小但上手门槛低。","stars":14,"starGrowth":0,"createdAt":"2021-10-18T18:06:49Z","silentRounds":25,"language":"C#","topics":[],"aiDims":["游戏","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["游戏编辑器","WPF","C#","关卡编辑","独立游戏"],"domainKey":"游戏编辑器","pushedAt":"2023-08-14T10:21:08Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/Eclmist/Matcha-Editor","ts":"2023-08-14T10:21:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"BrandonPotter/SimpleTCP","owner":"BrandonPotter","name":"SimpleTCP","desc":"Straightforward .NET library to handle the repetitive tasks of spinning up and working with TCP sockets (client and server).","summaryCn":"让.NET开发者告别繁琐的TCPSocket编程，几行代码搞定网络连接和通信","reasonCn":"SimpleTCP就是来帮你解决这些麻烦的。它是一个轻量的.NET库，把TCP通信中最常见的操作都封装好了。你不需要关心底层的Socket细节，只需要调用几个简单的方法就能建立连接、发送数据、接收消息。它内置了连接池管理、自动重连、心跳检测等功能，这些都是网络编程里最容易出bug的地方","stars":378,"starGrowth":0,"createdAt":"2015-10-20T01:20:11Z","silentRounds":25,"language":"C#","topics":[],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["TCP","网络编程",".NET","Socket"],"domainKey":"TCP","pushedAt":"2022-11-09T12:06:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/BrandonPotter/SimpleTCP","ts":"2022-11-09T12:06:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vaquarkhan/vaquarkhan","owner":"vaquarkhan","name":"vaquarkhan","desc":"","summaryCn":"一个设计师的个人主页，展示了自己的设计作品和创意项目。","reasonCn":"这是一个个人GitHub仓库，主要内容是设计作品集和个人项目展示，topics标记为design。仓库没有描述信息，也没有提供具体的开源工具或应用。1432的star数表明有一定的关注度，但这更像是个人品牌展示而非可使用的开源项目。项目不包含任何AI功能、实用工具或游戏内容，主要是一个设计师的在线名片和作品集合。","stars":1434,"starGrowth":0,"createdAt":"2016-12-07T12:29:20Z","silentRounds":25,"language":"","topics":["design"],"aiDims":["非AI-好玩"],"zone":"创意","funScore":0,"domainTags":["个人主页","设计作品集","设计师"],"domainKey":"个人主页","pushedAt":"2026-09-23T03:29:28Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7}],"aiScore":0.15,"source":"search","url":"https://github.com/vaquarkhan/vaquarkhan","ts":"2026-09-23T03:29:28Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Leihb/learn-agent-the-hard-way","owner":"Leihb","name":"learn-agent-the-hard-way","desc":"笨办法学 Agent · 亲手打造一个 harness——不用框架，32 个练习亲手写出一个 agent harness 的每一层。系列第一本，全书已完结，三本各自独立不分先后。","summaryCn":"用Go语言从零手写32个练习亲手打造AI大模型代理核心架构","reasonCn":"这是一本极其硬核且干货满满的开源电子书兼实战教程，主打不用任何第三方Agent框架。通过三十二个精心设计的代码练习，作者带着读者用Go语言一步步啃下智能体调度器、状态机、工具绑定和记忆回溯的每一层底层细节。它撕掉了框架的黑盒伪装，是每一个想要真正搞懂AI Agent底层原理的工程师不可多得的案头宝典。","stars":47,"starGrowth":0,"createdAt":"2026-07-31T04:30:54Z","silentRounds":9,"language":"Go","topics":["ebook"],"aiDims":["学习资源","AI Agent","教程"],"zone":"资源","funScore":0.5,"domainTags":["Agent教程","Go语言","AI原理","电子书"],"domainKey":"Agent教程","pushedAt":"2026-09-20T04:12:55Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Leihb/learn-agent-the-hard-way","ts":"2026-09-20T04:12:55Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"practical-tutorials/project-based-learning","owner":"practical-tutorials","name":"project-based-learning","desc":"Curated list of project-based tutorials","summaryCn":"手把手教你做项目的教程合集，从入门到实战全覆盖","reasonCn":"这是GitHub上最受欢迎的编程教程合集之一，按语言和项目类型分类，涵盖Python、JavaScript、Go等多个方向。每个教程都是真实可运行的项目实战，从Web应用到数据分析应有尽有。star超过二十七万，是程序员的宝藏资料库。","stars":284275,"starGrowth":52,"createdAt":"2017-04-12T05:07:46Z","silentRounds":0,"language":"Python","topics":["tutorial"],"aiDims":["学习资源","教程","AI应用"],"zone":"资源","funScore":0.6,"domainTags":["编程教程","项目实战","学习资源"],"domainKey":"编程教程","pushedAt":"2026-09-21T07:11:08Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/practical-tutorials/project-based-learning","ts":"2026-09-21T07:11:08Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.15654876440066837,"copyOk":true},{"repo":"sanic-org/sanic","owner":"sanic-org","name":"sanic","desc":" Accelerate your web app development  | Build fast. Run fast.","summaryCn":"Python异步Web框架，速度快得离谱，写接口就是爽","reasonCn":"Sanic是基于asyncio的高性能Python Web框架，原生支持异步编程，处理高并发请求的能力远超传统框架。它保留了类似Flask的简洁路由风格，上手几乎没有门槛，但底层通过异步I/O大幅提升了吞吐量。支持WebSocket、GraphQL、服务器发送事件等现代开发需求，部署简单，社区活跃，是构建高并发API服务的利器。","stars":18637,"starGrowth":0,"createdAt":"2016-05-26T04:38:22Z","silentRounds":19,"language":"Python","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Python Web框架","异步编程","高并发API服务","WebSocket","RESTful API"],"domainKey":"Python Web框架","pushedAt":"2026-07-29T03:09:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/sanic-org/sanic","ts":"2026-07-29T03:09:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"status-im/status-legacy","owner":"status-im","name":"status-legacy","desc":"a free (libre) open source, mobile OS for Ethereum","summaryCn":"专为以太坊打造的移动操作系统，让区块链应用如手机应用般易用。","reasonCn":"Status Legacy 是一个基于以太坊的移动操作系统，专注于去中心化身份、消息和支付。提供私聊、群聊、转账、DApp 浏览等功能，底层集成 libp2p 网络协议和 IPFS 存储。适合隐私保护倡导者、Web3 开发者和希望摆脱传统社交平台的用户。","stars":4022,"starGrowth":0,"createdAt":"2016-02-22T21:48:48Z","silentRounds":25,"language":"Clojure","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","安全工具"],"zone":"工具","funScore":0,"domainTags":["以太坊","移动操作系统","去中心化应用","IPFS","libp2p"],"domainKey":"以太坊","pushedAt":"2026-07-22T12:43:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Clojure","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/status-im/status-legacy","ts":"2026-07-22T12:43:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Manisso/fsociety","owner":"Manisso","name":"fsociety","desc":"fsociety Hacking Tools Pack – A Penetration Testing Framework","summaryCn":"一键打包上百款黑客工具，安全测试入门必备神器","reasonCn":"这是一个基于Python的渗透测试框架，集成了Metasploit、Nmap、SQLmap、Hydra等数十款经典安全工具。自动部署环境依赖，适合网络安全学习、CTF比赛和授权渗透测试，一键运行即可进入完整的测试环境，省去手动配置的麻烦。","stars":12317,"starGrowth":0,"createdAt":"2016-09-01T03:41:47Z","silentRounds":3,"language":"Python","topics":["desktop"],"aiDims":["安全工具","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["渗透测试","网络安全","工具集","Kali Linux","CTF"],"domainKey":"渗透测试","pushedAt":"2026-07-21T05:22:53Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/Manisso/fsociety","ts":"2026-07-21T05:22:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"comet-ml/opik","owner":"comet-ml","name":"opik","desc":"Debug, evaluate, and monitor your LLM applications, RAG systems, and agentic workflows with comprehensive tracing, automated evaluations, and production-ready dashboards.","summaryCn":"开源LLM全链路观测与评估平台，覆盖追踪、评测到生产监控。","reasonCn":"Opik是专为大语言模型应用打造的观测与评估平台，能追踪AI Agent的每一次调用、评估输出质量、管理提示词并监控生产环境。它由Comet公司开发，采用Apache-2.0开源协议，可免费自托管部署。对任何正在开发或运营LLM应用的团队来说，都是一站式的调试与监控工具。","stars":22199,"starGrowth":7,"createdAt":"2023-05-10T12:57:13Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI基础设施","AI应用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["LLMOps","大模型监控","RAG系统","提示词工程","应用评估"],"domainKey":"LLMOps","pushedAt":"2026-09-19T09:37:41Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/comet-ml/opik","ts":"2026-09-19T09:37:41Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.021945988558043156,"copyOk":true},{"repo":"LudwigKienle/ai-video-production-editor","owner":"LudwigKienle","name":"ai-video-production-editor","desc":"Open-source AI video production editor for filmmakers and AI artists. React + Electron app with node pipelines, image/video generation, storyboarding, previs, editing and creative production workflows.","summaryCn":"做视频最累的不是剪辑，而是前期素材的生成和找素材的过程。","reasonCn":"核心架构采用了 React 加 Electron 的组合，这使得它拥有了类似专业后期软件的交互体验，同时又具备了极强的本地调用能力。通过内置的节点流（Node Pipelines），你可以自定义生成逻辑，比如设定一个“剧本转分镜”的自动流，将文字描述自动推送到指定的模型 API，再将生成后的视频帧","stars":57,"starGrowth":0,"createdAt":"2026-05-03T15:27:38Z","silentRounds":2,"language":"TypeScript","topics":["creative-coding"],"aiDims":["创意工具","多模态Agent","AI应用"],"zone":"创意","funScore":0.8,"domainTags":["视频编辑","AI创作","节点流水线"],"domainKey":"视频编辑","pushedAt":"2026-09-22T11:37:05Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/LudwigKienle/ai-video-production-editor","ts":"2026-09-22T11:37:05Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kataras/iris","owner":"kataras","name":"iris","desc":"The fastest HTTP/2 Go Web Framework. New, modern and easy to learn. Fast development with Code you control. Unbeatable cost-performance ratio :rocket:","summaryCn":"写API像搭积木，Go框架里的速度机器，效率翻倍不是梦","reasonCn":"基于Go语言的HTTP/2 Web框架，以极致性能著称，支持路由分组、中间件链、模板渲染等现代开发需求。内置依赖注入容器，代码组织清晰。比同类框架性能高出数倍，同时保持简洁易用的API设计，适合快速构建生产级Web应用和微服务。","stars":25567,"starGrowth":0,"createdAt":"2016-01-30T04:36:48Z","silentRounds":3,"language":"Go","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Go语言","HTTP/2","高性能框架","微服务","中间件"],"domainKey":"Go语言","pushedAt":"2026-07-27T09:34:12Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/kataras/iris","ts":"2026-07-27T09:34:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ankorstore/yokai","owner":"ankorstore","name":"yokai","desc":"Simple, modular, and observable Go framework for backend applications.","summaryCn":"Go后端脚手架，内置可观测性，微服务开发不再头大","reasonCn":"Yokai是用Go语言编写的轻量级后端框架，主打简洁和模块化设计。它内置了metrics、tracing、健康检查、优雅关闭等生产环境常用能力，开发者无需东拼西凑多个库就能快速搭建可观测的后端服务。相比echo、gin等Web框架，yokai更侧重整体架构的规范性，适合需要长期维护的企业级项目。","stars":843,"starGrowth":0,"createdAt":"2024-01-10T10:39:53Z","silentRounds":25,"language":"Go","topics":["backend"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Go框架","后端脚手架","可观测性","微服务","中间件"],"domainKey":"Go框架","pushedAt":"2026-07-16T08:46:42Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/ankorstore/yokai","ts":"2026-07-16T08:46:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"uber/RIBs","owner":"uber","name":"RIBs","desc":"Uber's cross-platform mobile architecture framework - Android Repository","summaryCn":"Uber大厂内部同款的移动端架构方案，让代码不再一团糟","reasonCn":"RIBs是Uber开源的移动端架构框架，采用Router-Interactor-Builder-View四层模式，帮助开发者构建可测试的大型应用。它通过严格限制组件依赖，强制将业务逻辑、数据获取、UI展示分层分离。与MVC和MVP相比，RIBs提供更细粒度的路由控制和依赖注入机制，适合复杂业务场景的大型团队。","stars":7940,"starGrowth":0,"createdAt":"2017-10-03T18:18:50Z","silentRounds":10,"language":"Kotlin","topics":["mobile"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["移动端架构","Kotlin","分层架构","路由框架","工程化"],"domainKey":"移动端架构","pushedAt":"2026-07-15T22:14:43Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/uber/RIBs","ts":"2026-07-15T22:14:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EvoScientist/EvoScientist","owner":"EvoScientist","name":"EvoScientist","desc":"🔬 Harness Vibe Research with Self-evolving AI Scientists","summaryCn":"利用Vibe Research进行自我进化的AI科学家","reasonCn":"EvoScientist项目专注于利用Vibe Research来训练和发展AI科学家模型。通过自我学习与进化，这些AI模型可以在不断的迭代中提高其性能和准确性。","stars":4974,"starGrowth":7,"createdAt":"2026-01-26T11:19:26Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI科学家","自进化","科研自动化","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-23T15:00:18Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/EvoScientist/EvoScientist","ts":"2026-09-23T15:00:18Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.030266344993968637,"copyOk":false},{"repo":"kubernetes-sigs/kubespray","owner":"kubernetes-sigs","name":"kubespray","desc":"Deploy a Production Ready Kubernetes Cluster","summaryCn":"用Ansible自动化一键部署生产级别K8s集群的硬核运维神器","reasonCn":"技术上，它核心依赖Ansible这一强大的自动化运维工具。通过精心设计的Jinja模板和Playbook，它不仅能自动安装容器运行时、配置网络插件（如Calico、Flannel），还能处理复杂的证书签发、集群拓扑规划和滚动升级。它最大的优势在于通用性和可定制性，无论你的服务器是在AWS云上还是","stars":18745,"starGrowth":1,"createdAt":"2015-10-03T20:18:11Z","silentRounds":2,"language":"Jinja","topics":["kubernetes"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","Ansible","集群部署","容器编排","自动化运维"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T19:44:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Jinja","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/kubernetes-sigs/kubespray","ts":"2026-09-23T19:44:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"servo/servo","owner":"servo","name":"servo","desc":"Servo aims to empower developers with a lightweight, high-performance alternative for embedding web technologies in applications.","summaryCn":"Rust写的实验性浏览器引擎，让Web技术在应用里跑得更快更稳","reasonCn":"Mozilla推出的实验性浏览器渲染引擎，用Rust重写了核心组件。采用并发架构，把布局、样式、脚本执行拆到不同线程并行处理，GPU加速渲染，内存安全无崩溃风险。可嵌入到桌面或移动应用中作为Web视图组件，性能远超传统嵌入式浏览器方案。","stars":38024,"starGrowth":0,"createdAt":"2012-02-08T19:07:25Z","silentRounds":2,"language":"Rust","topics":["servo","rust","browser","webengine","web","webbrowser","webplatform"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["浏览器引擎","Rust","嵌入式Web视图","并发架构"],"domainKey":"浏览器引擎","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"servo","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"browser","source":"github","weight":0.7},{"name":"webengine","source":"github","weight":0.7},{"name":"web","source":"github","weight":0.7},{"name":"webbrowser","source":"github","weight":0.7},{"name":"webplatform","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/servo/servo","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bepu/bepuphysics1","owner":"bepu","name":"bepuphysics1","desc":"Pure C# 3D real time physics simulation library. Repo contains only the 1.X.X versions.","summaryCn":"纯C#写的物理引擎，让游戏里物体碰撞、弹跳、掉落都真实可感","reasonCn":"BEPUPhysics是一款纯C#编写的3D刚体物理模拟库，专注于实时物理计算。它支持刚体动力学、碰撞检测、关节约束求解、软体模拟等核心功能，专为游戏和交互式应用设计。作者Bepu团队在物理引擎领域深耕多年，代码性能优化到位，在同等规模场景下帧率表现优于部分传统引擎。作为纯托管代码库，它不依赖原生DLL，在.NET环境下集成非常友好，配合Unity或MonoGame等框架即可快速上手，非常适合独立开发者制作需要真实物理反馈的游戏Demo或可视化原型。","stars":434,"starGrowth":0,"createdAt":"2016-12-16T22:49:32Z","silentRounds":25,"language":"C#","topics":["physics-simulation-library"],"aiDims":["游戏","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["物理引擎","刚体动力学","碰撞检测","游戏开发","C#"],"domainKey":"物理引擎","pushedAt":"2023-06-06T18:08:15Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"physics-simulation-library","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/bepu/bepuphysics1","ts":"2023-06-06T18:08:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"swc-project/swc","owner":"swc-project","name":"swc","desc":"Rust-based platform for the Web","summaryCn":"用Rust重写比Babel快100倍的编译器，让前端构建快如闪电","reasonCn":"SWC是用Rust编写的超快速JavaScript/TypeScript编译平台，专为前端工程优化。它比传统Babel快约100倍，采用并行处理和JIT编译技术。支持代码压缩、类型转换、插件扩展，被Vite、Next.js、Turbopack等主流框架底层采用。","stars":34206,"starGrowth":0,"createdAt":"2017-12-22T11:40:14Z","silentRounds":1,"language":"Rust","topics":["ecmascript","compiler","rust","javascript","parser","babel","swc","ecmascript-parser","typescript","typescript-compiler","typescript-parser"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript编译器","TypeScript编译","前端构建","Rust","JIT编译"],"domainKey":"JavaScript编译器","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ecmascript","source":"github","weight":0.7},{"name":"compiler","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"parser","source":"github","weight":0.7},{"name":"babel","source":"github","weight":0.7},{"name":"swc","source":"github","weight":0.7},{"name":"ecmascript-parser","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"typescript-compiler","source":"github","weight":0.7},{"name":"typescript-parser","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/swc-project/swc","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MartinDelophy/ai-video-editor","owner":"MartinDelophy","name":"ai-video-editor","desc":"Open-source, local-first video editor where creators and AI agents edit the same real timeline.","summaryCn":"能和AI智能助手在同一条时间轴上并肩剪视频的神奇软件","reasonCn":"技术架构上，它完全采用现代网页技术栈用JavaScript开发，主打一个“本地优先”，你的视频素材不需要上传到云端服务器，直接在你的电脑浏览器或本地客户端里处理，完美保护隐私。它最厉害的地方在于把AI Agent深度嵌入到了时间轴的底层逻辑里，AI不仅能听懂你的自然语言指令去剪辑，","stars":852,"starGrowth":0,"createdAt":"2026-07-09T04:43:11Z","silentRounds":1,"language":"JavaScript","topics":["video-editor"],"aiDims":["创意工具","AI应用"],"zone":"AI","funScore":1,"domainTags":["视频编辑","AI Agent","本地处理"],"domainKey":"AI Agent","pushedAt":"2026-09-21T05:39:14Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/MartinDelophy/ai-video-editor","ts":"2026-09-21T05:39:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"apache/airflow","owner":"apache","name":"airflow","desc":"Apache Airflow - A platform to programmatically author, schedule, and monitor workflows","summaryCn":"让AI模型训练像坐班车一样，到点自动发车，准点到达还不掉线","reasonCn":"Apache Airflow是一个用Python编写的工作流编排平台，通过DAG（有向无环图）将多个任务串联起来。支持定时调度、自动重试、失败告警，能编排AI训练、数据处理、模型评估等复杂流水线。内置多种Executor，可部署在单机或Kubernetes集群上，Web UI直观展示任务执行情况。","stars":46949,"starGrowth":13,"createdAt":"2015-04-13T18:04:58Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["工作流编排","DAG","调度系统","Python","ETL"],"domainKey":"工作流编排","pushedAt":"2026-09-23T21:47:33Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/apache/airflow","ts":"2026-09-23T21:47:33Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03983068861956591,"copyOk":false},{"repo":"chengkehan/GPUSkinning","owner":"chengkehan","name":"GPUSkinning","desc":"","summaryCn":"用GPU加速骨骼蒙皮渲染，大量角色同屏也不卡顿的秘诀","reasonCn":"这个项目的核心技术是用Unity的Compute Shader实现GPU蒙皮。基本原理是：你的3D模型每个顶点都有关联的骨骼权重，动画播放时，顶点位置要根据所有影响它的骨骼的变换来计算。传统做法是在CPU上算好每个顶点的新位置，然后每帧上传到GPU。","stars":1064,"starGrowth":0,"createdAt":"2017-02-15T12:33:38Z","silentRounds":25,"language":"C#","topics":["gpu-skinning","gpuskinning"],"aiDims":["游戏","可视化"],"zone":"工具","funScore":0,"domainTags":["GPU蒙皮","Unity","Compute Shader","图形优化","骨骼动画"],"domainKey":"GPU蒙皮","pushedAt":"2019-07-21T14:47:31Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"gpu-skinning","source":"github","weight":0.7},{"name":"gpuskinning","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/chengkehan/GPUSkinning","ts":"2019-07-21T14:47:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lauridskern/open-runde","owner":"lauridskern","name":"open-runde","desc":"A soft, rounded variant of Inter","summaryCn":"给 Inter 字体穿上软萌圆角外衣，让文字瞬间变温柔的设计神器","reasonCn":"Open Runde 是 Inter 字体的软圆角变体，保留了 Inter 的清晰结构和现代感，同时给每个字符加了柔和的圆角处理，让文字视觉更友好温暖。支持多字重和 OpenType 特性，可直接嵌入网页和 UI 设计，适合需要亲和感界面风格的场景。","stars":1550,"starGrowth":0,"createdAt":"2023-07-26T14:53:46Z","silentRounds":3,"language":"","topics":["design"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["字体设计","UI设计","Inter字体","OpenType","网页设计"],"domainKey":"字体设计","pushedAt":"2026-07-25T15:28:04Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7}],"aiScore":0.35,"source":"search","url":"https://github.com/lauridskern/open-runde","ts":"2026-07-25T15:28:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"opentk/opentk","owner":"opentk","name":"opentk","desc":"The Open Toolkit library is a fast, low-level C# wrapper for OpenGL, OpenAL & OpenCL. It also includes windowing, mouse, keyboard and joystick input and a robust and fast math library, giving you everything you need to write your own renderer or game engine. OpenTK can be used standalone or inside a GUI on Windows, Linux, Mac.","summaryCn":"C#开发3D游戏和图形应用的瑞士军刀，一站式搞定渲染、音效和物理计算","reasonCn":"OpenTK 是 C# 生态中一个老牌且实用的图形开发库，封装了 OpenGL、OpenAL 和 OpenCL 三大底层技术。它让开发者不用面对复杂难懂的原生 C API，就能在 .NET 环境下快速构建跨平台 3D 应用、游戏或可视化工具。除了图形渲染，还内置了窗口管理、键盘鼠标输入和强大的数学计算库，相当于把一套完整的图形开发基础能力打包给你，非常适合想做独立游戏、仿真模拟或专业图形软件的技术开发者。","stars":3538,"starGrowth":0,"createdAt":"2013-11-12T01:34:24Z","silentRounds":25,"language":"C#","topics":["c-sharp","game-development","game-engine","graphics","graphics-library","math-library","openal","opencl","opengl","opengl-es","scientific-visualization"],"aiDims":["游戏","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["OpenGL","OpenAL","OpenCL","C#","3D开发"],"domainKey":"OpenGL","pushedAt":"2026-08-05T12:29:20Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"c-sharp","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"graphics","source":"github","weight":0.7},{"name":"graphics-library","source":"github","weight":0.7},{"name":"math-library","source":"github","weight":0.7},{"name":"openal","source":"github","weight":0.7},{"name":"opencl","source":"github","weight":0.7},{"name":"opengl","source":"github","weight":0.7},{"name":"opengl-es","source":"github","weight":0.7},{"name":"scientific-visualization","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/opentk/opentk","ts":"2026-08-05T12:29:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vespa-engine/vespa","owner":"vespa-engine","name":"vespa","desc":"The AI search platform","summaryCn":"雅虎开源的AI搜索平台，向量检索和机器学习一站式搞定","reasonCn":"Vespa的核心架构分三层：存储层负责数据的持久化和索引，计算层负责向量相似度计算和机器学习推理，查询层负责接收查询请求并返回结果。它原生支持稠密向量和稀疏向量混合搜索，能同时利用关键词匹配和语义理解。机器学习方面，支持在线推理，可以将模型直接部署到搜索链路中，实现","stars":7106,"starGrowth":1,"createdAt":"2016-06-03T20:54:20Z","silentRounds":1,"language":"Java","topics":["vector-db","rag"],"aiDims":["AI搜索","向量数据库","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["向量检索","机器学习","搜索引擎"],"domainKey":"向量检索","pushedAt":"2026-09-23T16:22:09Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/vespa-engine/vespa","ts":"2026-09-23T16:22:09Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ioBroker/ioBroker.admin","owner":"ioBroker","name":"ioBroker.admin","desc":"user interface for configuration and administration","summaryCn":"简单来说，这就是一个帮你管理家里成百上千个智能硬件的“中控大屏”。","reasonCn":"技术架构上，它基于TypeScript和现代Web前端框架构建，采用了前后端分离的设计理念。其核心优势在于极高的扩展性，通过插件化架构支持数千种不同的设备驱动和适配器。界面响应迅速，不仅支持实时数据刷新，还提供了美观的图表展示和详细的系统日志追踪功能。","stars":302,"starGrowth":0,"createdAt":"2014-08-20T15:23:02Z","silentRounds":17,"language":"TypeScript","topics":["home-automation"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["智能家居","物联网","中控面板","设备管理"],"domainKey":"智能家居","pushedAt":"2026-09-23T20:23:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/ioBroker/ioBroker.admin","ts":"2026-09-23T20:23:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ticlo/rc-dock","owner":"ticlo","name":"rc-dock","desc":"Dock Layout for React Component","summaryCn":"给React应用装上拖拽面板的神器，告别复杂布局","reasonCn":"它采用自定义hook的架构设计，核心是布局协议——你只需要描述面板之间的层级关系和约束，库负责处理所有的拖拽、缩放、吸附等交互细节。相比其他布局库，它的独特之处在于支持多层嵌套布局和动态面板增删，同时提供完整的键盘快捷键支持，对无障碍友好。","stars":813,"starGrowth":0,"createdAt":"2019-03-07T07:48:48Z","silentRounds":25,"language":"TypeScript","topics":["dock","drag","drag-and-drop","drop","layout","panel","react","react-dom","reactjs","tabs"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["React","拖拽布局","UI组件库"],"domainKey":"React","pushedAt":"2026-06-29T21:31:31Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"dock","source":"github","weight":0.7},{"name":"drag","source":"github","weight":0.7},{"name":"drag-and-drop","source":"github","weight":0.7},{"name":"drop","source":"github","weight":0.7},{"name":"layout","source":"github","weight":0.7},{"name":"panel","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"react-dom","source":"github","weight":0.7},{"name":"reactjs","source":"github","weight":0.7},{"name":"tabs","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/ticlo/rc-dock","ts":"2026-06-29T21:31:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wy-luke/All-Chinese-Character-Set","owner":"wy-luke","name":"All-Chinese-Character-Set","desc":"致力于维护一个包含全部中文字符（汉字+符号）的字符集，包罗万象，含3500 / 7000 常用汉字集、常用符号集。A Charater Set contains all Chinese charaters.","summaryCn":"收集全部中文字符的字典库，写程序再也不怕生僻字渲染不出或乱码了","reasonCn":"这个项目整理了从基本汉字到Unicode扩展区的生僻字，涵盖3500常用字、7000通用字和各类符号。提供JSON、CSV等多种格式导出，对中文NLP、字体开发、输入法构建等项目非常实用。再也不用到处找零散的字符表，一个仓库就能搞定所有中文字符需求，省去大量数据整理工作。","stars":1947,"starGrowth":0,"createdAt":"2021-11-03T11:45:15Z","silentRounds":25,"language":"JavaScript","topics":["character-set","chinese","chinese-character","chinese-characters","chinese-simplified","font","textmeshpro","unity","unity-3d","unity3d"],"aiDims":["非AI-实用","开发者工具"],"zone":"资源","funScore":0,"domainTags":["中文","字符集","字体开发"],"domainKey":"中文","pushedAt":"2026-04-21T15:23:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"character-set","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"chinese-character","source":"github","weight":0.7},{"name":"chinese-characters","source":"github","weight":0.7},{"name":"chinese-simplified","source":"github","weight":0.7},{"name":"font","source":"github","weight":0.7},{"name":"textmeshpro","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity-3d","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/wy-luke/All-Chinese-Character-Set","ts":"2026-04-21T15:23:21Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Nition/UnityOctree","owner":"Nition","name":"UnityOctree","desc":"A dynamic, loose octree implementation for Unity written in C#","summaryCn":"游戏开发必备：超快空间划分算法，让渲染效率翻倍","reasonCn":"动态松八叉树算法，专为Unity游戏开发设计，用于在3D空间中快速定位和管理游戏对象。支持碰撞检测、视锥剔除、对象查找等核心功能。与传统八叉树相比，松八叉树允许节点间存在重叠区域，大幅减少动态场景中的节点重构开销。纯C#编写，零外部依赖，直接复制到项目就能跑，非常适合中小规模游戏项目。","stars":1009,"starGrowth":0,"createdAt":"2014-05-08T05:07:09Z","silentRounds":25,"language":"C#","topics":["c-sharp","dynamic","loose-octrees","octree","unity-3d","unity-scripts","unity3d"],"aiDims":["游戏","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Unity","八叉树","空间划分"],"domainKey":"Unity","pushedAt":"2021-06-07T22:22:16Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"c-sharp","source":"github","weight":0.7},{"name":"dynamic","source":"github","weight":0.7},{"name":"loose-octrees","source":"github","weight":0.7},{"name":"octree","source":"github","weight":0.7},{"name":"unity-3d","source":"github","weight":0.7},{"name":"unity-scripts","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/Nition/UnityOctree","ts":"2021-06-07T22:22:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zumerlab/snapdom","owner":"zumerlab","name":"snapdom","desc":"High-performance engine for capturing, modifying, and converting DOM elements into any format.","summaryCn":"在前端开发里，我们经常需要把某个网页组件导出成图片或者是生成报告，传统","reasonCn":"它的核心技术亮点在于其高效的渲染处理管道。它不是简单的调用浏览器内核截图，而是对DOM节点进行深度克隆与计算，通过精细化的CSS处理逻辑，保证了在转换成其他格式时不会出现布局崩塌。在高并发场景下，它的性能表现极其优秀，能快速完成复杂结构的解析，这对于需要批量生成页面","stars":8152,"starGrowth":3,"createdAt":"2025-04-27T11:01:19Z","silentRounds":0,"language":"JavaScript","topics":["frontend"],"aiDims":["可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["DOM操作","网页渲染","前端工具","截图导出"],"domainKey":"DOM操作","pushedAt":"2026-09-23T14:34:27Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/zumerlab/snapdom","ts":"2026-09-23T14:34:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"netdata/netdata","owner":"netdata","name":"netdata","desc":"The fastest path to AI-powered full stack observability, even for lean teams.","summaryCn":"给服务器装个智能体检仪，AI自动发现异常告警，运维再也不用熬夜盯屏幕","reasonCn":"技术上，Netdata采用Go和C混合编写，性能极高，对系统资源占用极低，agent模式部署在每台机器上，数据实时推送。它的架构特点是自动发现机制，安装后会自动扫描系统检测出数百到数千项指标，包括应用程序、数据库、服务等各种维度。","stars":80626,"starGrowth":7,"createdAt":"2013-06-17T18:39:10Z","silentRounds":0,"language":"Go","topics":["ml","ai"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["服务器监控","可观测性","时序数据","告警系统"],"domainKey":"服务器监控","pushedAt":"2026-09-23T20:23:11Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.68,"source":"trending","url":"https://github.com/netdata/netdata","ts":"2026-09-23T20:23:11Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.021260461885744056,"copyOk":false},{"repo":"dyn4j/dyn4j","owner":"dyn4j","name":"dyn4j","desc":"Java Collision Detection and Physics Engine","summaryCn":"Java物理引擎，让游戏里的物体碰撞反弹真实又流畅","reasonCn":"dyn4j是Java语言编写的2D物理引擎，专注于碰撞检测和物理模拟。它支持连续碰撞检测CCD来防止高速物体穿透，内置凸包分解算法处理任意形状多边形。相比Box2D等引擎，dyn4j设计更轻量，API简洁易用，在Java游戏开发社区中颇受欢迎。","stars":538,"starGrowth":0,"createdAt":"2015-03-14T03:49:36Z","silentRounds":25,"language":"Java","topics":["collision-detection","continuous-collision-detection","convex-decomposition","convex-hull-algorithms","dyn4j","java","physics-engine","simulation"],"aiDims":["非AI-好玩","游戏","可视化"],"zone":"工具","funScore":0,"domainTags":["物理引擎","碰撞检测","2D物理","Java","游戏开发"],"domainKey":"物理引擎","pushedAt":"2026-07-18T04:07:12Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"collision-detection","source":"github","weight":0.7},{"name":"continuous-collision-detection","source":"github","weight":0.7},{"name":"convex-decomposition","source":"github","weight":0.7},{"name":"convex-hull-algorithms","source":"github","weight":0.7},{"name":"dyn4j","source":"github","weight":0.7},{"name":"java","source":"github","weight":0.7},{"name":"physics-engine","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.35,"source":"bigbro","url":"https://github.com/dyn4j/dyn4j","ts":"2026-07-18T04:07:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"c3js/c3","owner":"c3js","name":"c3","desc":":bar_chart: A D3-based reusable chart library","summaryCn":"基于D3的图表库，用JavaScript轻松画出各种图表。","reasonCn":"c3是一个基于D3.js的可复用图表库，提供简单易用的API，支持折线图、柱状图、饼图、表格等多种图表类型。它的设计哲学是让你能用最少的代码创建漂亮的图表，同时保留了D3的强大定制能力，适合快速构建数据可视化界面。","stars":9351,"starGrowth":0,"createdAt":"2013-07-18T06:51:40Z","silentRounds":24,"language":"JavaScript","topics":["data-visualization"],"aiDims":["非AI-好玩","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据可视化","JavaScript","图表库","D3.js","前端组件"],"domainKey":"数据可视化","pushedAt":"2026-09-14T06:23:04Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/c3js/c3","ts":"2026-09-14T06:23:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hackersandslackers/flask-sqlalchemy-tutorial","owner":"hackersandslackers","name":"flask-sqlalchemy-tutorial","desc":":snake: :information_source: Create and manage data in your Flask app via a SQL database.","summaryCn":"手把手教你在Flask里轻松搞定数据库操作的保姆级教程","reasonCn":"这是一个针对Python初学者的实战教程，专注于如何在使用Flask框架的应用中集成SQLAlchemy ORM。它通过具体的代码示例，演示了如何定义数据模型、创建数据库表结构，以及执行常见的增删改查操作。适合想要快速掌握Web后端数据库连接机制的开发者参考学习。","stars":399,"starGrowth":0,"createdAt":"2019-02-06T05:40:52Z","silentRounds":13,"language":"Python","topics":["tutorial"],"aiDims":["教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["Flask","SQLAlchemy","ORM","教程","Python Web"],"domainKey":"Flask","pushedAt":"2026-09-07T17:33:57Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/hackersandslackers/flask-sqlalchemy-tutorial","ts":"2026-09-07T17:33:57Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lets-all-be-stupid-forever/circuit-artist","owner":"lets-all-be-stupid-forever","name":"circuit-artist","desc":"Circuit Artist is a digital circuit drawing and simulation game.","summaryCn":"用画画的方式随手搭电路并实时模拟运行的数字电路小游戏","reasonCn":"技术层面上，它采用纯C语言编写，底层渲染依托于大名鼎鼎、轻量且极速的Raylib图形库。整个程序极其小巧精悍，几乎没有任何臃肿的依赖包，启动速度堪称秒开。它内置了高效的数字逻辑实时仿真引擎，能够在鼠标拖拽和连线的瞬间，计算出各个逻辑门之间的电平跳变和状态更新，画面丝滑","stars":478,"starGrowth":0,"createdAt":"2024-09-10T22:29:24Z","silentRounds":24,"language":"C","topics":["raylib"],"aiDims":["非AI-好玩","游戏","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["数字电路","逻辑门仿真","raylib","C语言游戏","极客玩具"],"domainKey":"数字电路","pushedAt":"2026-09-20T21:01:14Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"raylib","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/lets-all-be-stupid-forever/circuit-artist","ts":"2026-09-20T21:01:14Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"travsteward/openwriter","owner":"travsteward","name":"openwriter","desc":"The open-source writing surface for AI agents. Markdown-native editor with pending change review — your agent writes, you accept or reject.","summaryCn":"AI写手的写作助手，支持编辑和审核在云端","reasonCn":"OpenWriter 是一个面向 AI 代理的 Markdown 编辑器，提供实时写作与变更审阅功能。用户可在编辑时随时接受或驳回代理生成的内容，确保文本质量与创作方向一致。它利用 TypeScript 构建，支持插件化，方便开发者将自定义 AI 接入，打造写作助手。","stars":30,"starGrowth":0,"createdAt":"2026-02-17T16:50:36Z","silentRounds":15,"language":"TypeScript","topics":["writing"],"aiDims":["AI Agent","AI写作","工具调用"],"zone":"工具","funScore":0,"domainTags":["Markdown编辑器","AI写作","变更审阅","插件化"],"domainKey":"Markdown编辑器","pushedAt":"2026-09-23T19:07:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.73,"source":"search","url":"https://github.com/travsteward/openwriter","ts":"2026-09-23T19:07:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lissy93/personal-security-checklist","owner":"lissy93","name":"personal-security-checklist","desc":"🔒 A compiled checklist of 300+ tips for protecting digital security and privacy in 2026","summaryCn":"一份超全的数字安全生存指南，保护你的隐私","reasonCn":"这是一个为2026年量身定制的网络安全隐私防护清单，包含300多条实用技巧。涵盖了从强密码管理到网络钓鱼识别，再到设备加密等全方位保护措施。项目由TypeScript编写，代码结构清晰，提供了可视化的进度追踪功能，帮助用户系统性地提升数字安全水平。","stars":22371,"starGrowth":1,"createdAt":"2018-03-02T21:12:37Z","silentRounds":1,"language":"TypeScript","topics":["security","awesome","privacy"],"aiDims":["安全工具"],"zone":"资源","funScore":0,"domainTags":["网络安全","隐私保护","清单","TypeScript"],"domainKey":"网络安全","pushedAt":"2026-09-07T11:54:46Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/lissy93/personal-security-checklist","ts":"2026-09-07T11:54:46Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ossu/computer-science","owner":"ossu","name":"computer-science","desc":"🎓 Path to a free self-taught education in Computer Science!","summaryCn":"免费CS教育路径图，从入门到精通全套课程指南","reasonCn":"这是一个开源的计算机科学自学路线图，整合了全球顶尖大学（如MIT、斯坦福）的免费课程资源。它不是代码库，而是一份精心策划的学习计划，涵盖数学、编程、系统、AI等所有核心领域。对于想系统转型AI或夯实基础的人来说，它是最好的导航地图。","stars":209377,"starGrowth":21,"createdAt":"2014-05-04T00:18:39Z","silentRounds":0,"language":"HTML","topics":["computer-science","awesome-list","courses","curriculum"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["计算机科学","自学","课程指南","MIT","学习路径"],"domainKey":"计算机科学","pushedAt":"2026-08-07T04:07:05.660Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"computer-science","source":"github","weight":0.7},{"name":"awesome-list","source":"github","weight":0.7},{"name":"courses","source":"github","weight":0.7},{"name":"curriculum","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/ossu/computer-science","ts":"2026-08-07T04:07:05.660Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06330089264818962,"copyOk":false},{"repo":"FortAwesome/vue-fontawesome","owner":"FortAwesome","name":"vue-fontawesome","desc":"Font Awesome Vue component","summaryCn":"Vue项目里一键引入Font Awesome图标的懒人福音","reasonCn":"这是Font Awesome官方出品的Vue 3组件库，让你在Vue项目中用最少的代码引入和使用图标。通过npm安装后，直接在模板中用组件渲染图标，支持动态切换图标、颜色和大小，还能应用CSS动画效果。集成简单，API清晰，适合需要管理大量图标的Vue前端项目，让图标使用变得优雅而高效。","stars":2406,"starGrowth":0,"createdAt":"2017-08-11T19:12:13Z","silentRounds":23,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Vue","图标库","前端组件","Font Awesome"],"domainKey":"Vue","pushedAt":"2026-09-13T15:30:59Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/FortAwesome/vue-fontawesome","ts":"2026-09-13T15:30:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"emilianavt/OpenSeeFace","owner":"emilianavt","name":"OpenSeeFace","desc":"","summaryCn":"这是一个专注于面部特征点检测与动作捕捉的计算机视觉开源项目。","reasonCn":"需要明确的是，该项目本身是一个纯粹的底层追踪库，而不是一个开箱即用的独立虚拟形象驱动程序。尽管如此，它强大的追踪能力在整个虚拟主播生态中扮演了基石般的角色。许多知名的面捕软件，例如用于驱动 Live2D 模型的 VTube Studio 以及用于动画化 VRM 和 VSFAvatar 3D 模型的 VS","stars":2056,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["多模态Agent","AI应用"],"zone":"AI","funScore":0.62,"domainTags":["面部追踪","计算机视觉","虚拟主播","ONNX运行时"],"domainKey":"面部追踪","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/emilianavt/OpenSeeFace","ts":"2026-09-23T18:23:13.861Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MaartenGr/BERTopic","owner":"MaartenGr","name":"BERTopic","desc":"Leveraging BERT and c-TF-IDF to create easily interpretable topics. ","summaryCn":"能把海量文本自动分类出主题的AI分析利器","reasonCn":"在技术实现上，它并没有采用传统的 LDA 模型，而是采用了现代的深度学习方案。首先通过 BERT 等 Transformer 模型将文本转化为高维向量，然后利用 UMAP 进行降维以保留局部结构，最后使用 HDBSCAN 进行聚类。最精妙之处在于引入了 c-TF-IDF，它能为每个聚类提取出最具代表性的关键","stars":7849,"starGrowth":0,"createdAt":"2020-09-22T14:19:29Z","silentRounds":1,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","可视化","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["NLP","主题建模","BERT","聚类分析"],"domainKey":"NLP","pushedAt":"2026-09-22T08:11:03Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/MaartenGr/BERTopic","ts":"2026-09-22T08:11:03Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jj-vcs/jj","owner":"jj-vcs","name":"jj","desc":"A Git-compatible VCS that is both simple and powerful","summaryCn":"新一代Git替代品，让版本控制更简洁高效","reasonCn":"Rust编写的Git兼容版本控制系统，设计目标是解决Git复杂操作的学习曲线问题。支持撤销提交、重新排序历史、无缝合并等高级功能，命令比Git更直观。由Google前工程师创建，采用现代Rust开发，性能优秀。适合追求高效工作流的开发者，尤其是团队协作场景下简化代码审查流程。","stars":31700,"starGrowth":2,"createdAt":"2020-12-18T04:05:27Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["版本控制","Git","VCS","Rust"],"domainKey":"版本控制","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/jj-vcs/jj","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Tencent/Hippy","owner":"Tencent","name":"Hippy","desc":"Hippy is designed to easily build cross-platform dynamic apps. 👏","summaryCn":"一套代码同时跑iOS、Android和Web的跨平台开发框架","reasonCn":"Tencent推出的跨平台框架，用JavaScript开发可以同时编译到iOS、Android和Web端，适合追求一次开发多端运行的团队，技术栈是TypeScript/C++。\n\n技术层面，Hippy采用自研的桥接方案而非WebView渲染，性能接近原生。它提供统一的组件API，底层自动适配各平台差异。核心优势是动态化能力——无需发版即可更新UI逻辑，同时保持较好的渲染性能。相比React Native，它在动态更新和包体积上有优化。\n\n使用方法：安装Tencent Hippy CLI工具，创建新项目后编写JavaScript代码，通过命令行编译到目标平台。支持VS Code插件，开发体验接近原生。\n\n适合需要同时覆盖iOS、Android和Web的产品团队，尤其是需要频繁更新UI但不想依赖各端发版的场景。需要一定的JavaScript和移动开发基础。\n\n项目star8500+，由腾讯维护，在国内有较多落地案例。在跨平台框架中属于性能较好的选择，适合对包体积和动态化有要求的团队。竞品包括React Native、Flutter，Hippy的差异化在于更强的动态能力和腾讯内部验证的稳定性。","stars":8563,"starGrowth":1,"createdAt":"2019-11-15T01:55:31Z","silentRounds":2,"language":"C++","topics":["mobile","frontend"],"aiDims":["非AI-实用","工具调用"],"zone":"工具","funScore":0,"domainTags":["跨平台开发","JavaScript","动态化","原生渲染","前端框架"],"domainKey":"跨平台开发","pushedAt":"2026-09-09T09:21:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.28,"source":"search","url":"https://github.com/Tencent/Hippy","ts":"2026-09-09T09:21:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"apache/pouchdb","owner":"apache","name":"pouchdb","desc":":kangaroo: - PouchDB is a pocket-sized database.","summaryCn":"能装进口袋的JavaScript数据库，浏览器和本地都能跑","reasonCn":"PouchDB是JavaScript实现的轻量级数据库，专为浏览器和移动端设计。它采用CouchDB兼容的同步协议，能在客户端存储数据，也能与后端CouchDB同步。支持索引查询、实时监听数据变化。对于前端开发构建离线应用、移动应用数据管理非常实用，是轻量级数据存储的优选方案。","stars":17612,"starGrowth":0,"createdAt":"2010-06-10T18:34:24Z","silentRounds":2,"language":"JavaScript","topics":["database"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["IndexedDB","离线存储","CouchDB","实时同步","前端开发"],"domainKey":"IndexedDB","pushedAt":"2026-09-13T07:11:34Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/apache/pouchdb","ts":"2026-09-13T07:11:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Sollimann/bonsai","owner":"Sollimann","name":"bonsai","desc":"Rust implementation of behavior trees for deterministic AI (now with Python bindings)","summaryCn":"在游戏开发中，NPC的行为往往让人头疼。如果完全随机，敌人可能会像傻子一样","reasonCn":"Bonsai是纯Rust编写的，性能极高，非常适合对性能要求苛刻的游戏引擎。它最大的特点就是“确定性”，同样的输入和状态，AI永远会做出一样的反应，这对于游戏平衡性至关重要。它还提供了Python绑定，这意味着即使你不懂Rust，也能在Python游戏引擎（如Unity、Unreal的Python接口或Pyg","stars":1082,"starGrowth":0,"createdAt":"2022-05-23T20:15:23Z","silentRounds":9,"language":"Rust","topics":["robotics"],"aiDims":["AI应用","游戏","创意工具"],"zone":"AI","funScore":0.6,"domainTags":["游戏 AI","行为树","决策系统","Rust","Python 绑定"],"domainKey":"游戏 AI","pushedAt":"2026-09-22T08:00:51Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/Sollimann/bonsai","ts":"2026-09-22T08:00:51Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"komari-monitor/komari","owner":"komari-monitor","name":"komari","desc":"A simple server monitor tool.","summaryCn":"颜值超高的服务器监控面板，一眼看清主机健康状态","reasonCn":"基于Go和Vue构建的轻量级服务器监控仪表盘，可视化效果出众。支持多节点聚合，通过简单的二进制部署即可运行，无需复杂配置。对运维开发者而言，它是替代昂贵商业软件的极佳免费方案，让服务器状态一目了然，告别命令行焦虑。","stars":6208,"starGrowth":2,"createdAt":"2025-04-19T16:34:16Z","silentRounds":0,"language":"Go","topics":["monitoring"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["服务器监控","可视化仪表盘","运维面板"],"domainKey":"服务器监控","pushedAt":"2026-09-23T06:05:27Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/komari-monitor/komari","ts":"2026-09-23T06:05:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"syncthing/syncthing-android","owner":"syncthing","name":"syncthing-android","desc":"","summaryCn":"该项目在过去的时间里，成功连接了数以万计的个人设备，实现了无需依赖第三","reasonCn":"对于仍在使用该应用的用户，目前虽然核心功能依然能够运行，但缺少了后续的安全性补丁与系统适配更新，这意味着它可能无法兼容未来的 Android 版本。这也是开源项目在面临商业环境挑战时的一个典型缩影，即便是如此普及的工具，在缺乏长期维护动力后，也需要用户及时评估迁移策略。","stars":4282,"starGrowth":0,"createdAt":"2014-02-22T17:15:19Z","silentRounds":3,"language":"Java","topics":["android","java","synchronization","peer-to-peer","p2p"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["文件同步","Android应用","私有云"],"domainKey":"文件同步","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"android","source":"github","weight":0.7},{"name":"java","source":"github","weight":0.7},{"name":"synchronization","source":"github","weight":0.7},{"name":"peer-to-peer","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.2,"source":"tier","url":"https://github.com/syncthing/syncthing-android","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"segmentio/myth","owner":"segmentio","name":"myth","desc":"","summaryCn":"在实际前端开发中，尽管CSS规范的许多高级特性（如自定义属性、更高效的数学","reasonCn":"Myth的独特之处在于其“无侵入性”的理念。它宣称让用户能够以“最初想象的方式”编写CSS，这意味着开发者不需要记忆新的预处理器特有语法，只需使用标准的CSS变量语法和原生数学函数。对于已经熟悉CSS语言结构的开发者来说，学习曲线极平，甚至可以直接在现有的CSS风格上无缝迁移，减","stars":4285,"starGrowth":0,"createdAt":"2013-12-16T09:22:19Z","silentRounds":4,"language":"JavaScript","topics":[],"aiDims":["非AI-实用","开发者工具","前端工程"],"zone":"工具","funScore":0.2,"domainTags":["CSS预处理","样式变量","前端构建","浏览器兼容"],"domainKey":"CSS预处理","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"前端工程","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"tier","url":"https://github.com/segmentio/myth","ts":"2026-09-18T00:01:25.926Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"iina/iina","owner":"iina","name":"iina","desc":"The modern video player for macOS.","summaryCn":"颜值爆表的macOS视频播放器，界面优雅功能强大","reasonCn":"基于mpv内核打造的全新一代macOS视频播放器，界面采用原生Swift开发，与macOS风格完美融合。支持HDR视频、杜比全景声、画中画播放等高级功能，同时支持字幕渲染、手势控制和播放列表管理。通过GPU加速确保流畅播放，命令行配置灵活，各种视频格式都能完美兼容。","stars":46454,"starGrowth":10,"createdAt":"2016-12-19T07:18:45Z","silentRounds":0,"language":"Swift","topics":["hacktoberfest","macos","mpv","swift","video","video-player"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["视频播放","macOS","mpv","Swift","播放器"],"domainKey":"视频播放","pushedAt":"2026-08-18T02:17:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"mpv","source":"github","weight":0.7},{"name":"swift","source":"github","weight":0.7},{"name":"video","source":"github","weight":0.7},{"name":"video-player","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.3,"source":"bigbro","url":"https://github.com/iina/iina","ts":"2026-08-18T02:17:06Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030645800146381366,"copyOk":false},{"repo":"agents-flex/agents-flex","owner":"agents-flex","name":"agents-flex","desc":"Agents-flex is A lightweight Java AI agent development framework (positioned as a counterpart to Spring AI). It supports features such as RAG, MCP, Skills, Text2SQL , LLM Wiki, Sub-agents, Web Search, TTS (synchronous and streaming), and STT.","summaryCn":"Java开发者的AI Agent瑞士军刀，RAG和工具调用全搞定","reasonCn":"从技术角度看，agents-flex的能力栈相当完整。它不只支持基础的LLM对话，还集成了RAG检索增强生成、MCP（Model Context Protocol）协议、Skills技能系统、Text2SQL文本转SQL查询、LLM Wiki知识问答、子智能体协作以及网页搜索等多种高级能力。","stars":1068,"starGrowth":0,"createdAt":"2024-01-12T08:23:40Z","silentRounds":3,"language":"Java","topics":["chatbot"],"aiDims":["AI Agent","工具调用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["Java","AI Agent","RAG"],"domainKey":"AI Agent","pushedAt":"2026-09-23T03:03:31Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/agents-flex/agents-flex","ts":"2026-09-23T03:03:31Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kjtsune/embyToLocalPlayer","owner":"kjtsune","name":"embyToLocalPlayer","desc":"etlp - Emby/Jellyfin 调用外部本地播放器，并回传播放记录。适配 Plex。","summaryCn":"让你用MPV或PotPlayer直接播放Emby视频，并同步观影进度。","reasonCn":"该工具通过脚本拦截Emby、Jellyfin及Plex的播放指令，将视频流无缝调用至MPV或PotPlayer等本地强大播放器。它解决了网页端解码性能差的问题，并利用API与WebHook实现了播放进度的毫秒级实时回传，既享有顶级本地硬解码与渲染增强，又能无缝同步云端媒体库观影记录。","stars":1437,"starGrowth":0,"createdAt":"2022-07-29T13:57:48Z","silentRounds":3,"language":"Python","topics":["jellyfin"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["私有媒体库","视频播放器桥接","Emby","Jellyfin","观影进度同步"],"domainKey":"私有媒体库","pushedAt":"2026-09-23T09:54:24Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/kjtsune/embyToLocalPlayer","ts":"2026-09-23T09:54:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"konstaui/konsta","owner":"konstaui","name":"konsta","desc":"Mobile UI components made with Tailwind CSS","summaryCn":"用Tailwind CSS打造的移动UI组件库，一套代码同时支持iOS和Android风格","reasonCn":"技术上，它最大的特色是完全基于Tailwind CSS，这意味着你可以灵活定制样式，不用担心被组件库绑死。它支持React、Vue 3和Svelte三个主流框架，每个框架都有独立的版本。组件包括导航栏、标签栏、卡片、按钮、输入框等移动端常用元素，全部设计精致，交互流畅。","stars":4256,"starGrowth":0,"createdAt":"2021-01-18T19:52:48Z","silentRounds":3,"language":"JavaScript","topics":["mobile"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["移动端UI","Tailwind CSS","组件库","PWA"],"domainKey":"移动端UI","pushedAt":"2026-09-02T16:43:24Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/konstaui/konsta","ts":"2026-09-02T16:43:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tuono-labs/tuono","owner":"tuono-labs","name":"tuono","desc":"⚡ Modern fullstack web framework based on Rust and React","summaryCn":"Rust写的Web框架，性能超强，React后端全栈开发利器","reasonCn":"Tuono是基于Rust和React的全栈Web框架，将两者优势结合，提供极致的构建速度和运行时性能。它利用Rust的编译期优化和React的组件化能力，自动生成服务器端渲染代码，无需手动配置路由。开发者可以用熟悉的React语法编写前端，同时享受Rust后端带来的接近原生的执行效率，特别适合追求性能和开发体验并重的全栈项目。","stars":1140,"starGrowth":1,"createdAt":"2024-04-28T08:52:07Z","silentRounds":26,"language":"Rust","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Rust 框架","全栈开发","React","服务端渲染"],"domainKey":"Rust 框架","pushedAt":"2026-09-21T22:24:37Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/tuono-labs/tuono","ts":"2026-09-21T22:24:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Rabadash8820/UnityAssemblies","owner":"Rabadash8820","name":"UnityAssemblies","desc":"A NuGet package for simple, forward-compatible references to any Unity DLL on any platform.","summaryCn":"Unity开发者神器，一行代码搞定跨平台DLL引用","reasonCn":"Unity开发中最头疼的问题之一就是引用Unity DLL，尤其在跨平台场景下路径各不相同。这个NuGet包解决了这个痛点，通过简单的包引用就能访问任意Unity平台的DLL，兼容性好且维护省心，彻底告别手动复制DLL的繁琐操作，大幅提升Unity项目的构建和依赖管理效率。","stars":124,"starGrowth":0,"createdAt":"2019-06-17T03:30:12Z","silentRounds":25,"language":"ShaderLab","topics":["csharp","msbuild","nuget","unity","unity-apis","unity-assemblies","unity-binaries","unity-hub","unity2d","unity3d","unityeditor","unityengine","unityextensions","visual-studio"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["Unity","DLL 引用","NuGet","跨平台","游戏开发"],"domainKey":"Unity","pushedAt":"2026-07-06T20:07:25Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"msbuild","source":"github","weight":0.7},{"name":"nuget","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity-apis","source":"github","weight":0.7},{"name":"unity-assemblies","source":"github","weight":0.7},{"name":"unity-binaries","source":"github","weight":0.7},{"name":"unity-hub","source":"github","weight":0.7},{"name":"unity2d","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"unityeditor","source":"github","weight":0.7},{"name":"unityengine","source":"github","weight":0.7},{"name":"unityextensions","source":"github","weight":0.7},{"name":"visual-studio","source":"github","weight":0.7},{"name":"ShaderLab","source":"language","weight":0.5}],"aiScore":0.32,"source":"bigbro","url":"https://github.com/Rabadash8820/UnityAssemblies","ts":"2026-07-06T20:07:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"valeriansaliou/sonic","owner":"valeriansaliou","name":"sonic","desc":"🦔 Fast, lightweight & schema-less search backend. An alternative to Elasticsearch that runs on a few MBs of RAM.","summaryCn":"Elasticsearch的轻量替代品，几兆内存就能跑全文搜索引擎","reasonCn":"一款用Rust编写的极速轻量级搜索后端，内存占用仅需几MB，却能提供亚毫秒级的全文搜索能力。作为Elasticsearch的替代品，它支持拼音搜索、自动补全、模糊匹配等特性，无需复杂的集群管理，单机部署即可满足大多数中小型应用的搜索需求。","stars":21345,"starGrowth":0,"createdAt":"2019-02-26T13:16:58Z","silentRounds":4,"language":"Rust","topics":["database","backend"],"aiDims":["AI搜索","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["全文搜索","搜索引擎","Rust","倒排索引","数据库"],"domainKey":"全文搜索","pushedAt":"2026-09-23T21:21:04Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"backend","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/valeriansaliou/sonic","ts":"2026-09-23T21:21:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"supranim/supranim","owner":"supranim","name":"supranim","desc":"⚡️ Just a bloody web framework written in Nim 👑","summaryCn":"一个用 Nim 语言写的超快 Web 框架，自带炫酷特效。","reasonCn":"Supranim 是一个用高性能编程语言 Nim 编写的现代 Web 框架。它主打极致的渲染性能和丰富的交互特效，能够帮助开发者快速构建出视觉效果出众的单页应用。框架设计理念先进，语法简洁，特别适合那些既看重开发效率又追求极致用户体验的项目。","stars":42,"starGrowth":0,"createdAt":"2021-07-31T06:39:58Z","silentRounds":12,"language":"Nim","topics":["web-framework"],"aiDims":["创意工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Nim","Web框架","后端开发"],"domainKey":"Nim","pushedAt":"2026-09-23T17:43:53Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Nim","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/supranim/supranim","ts":"2026-09-23T17:43:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"observablehq/plot","owner":"observablehq","name":"plot","desc":"A concise API for exploratory data visualization implementing a layered grammar of graphics","summaryCn":"简洁优雅的 JavaScript 数据可视化库","reasonCn":"由 Observable 团队开发，Plot 是一个基于语法化图形思想的 JavaScript 可视化库。它提供了简洁的 API，支持快速生成线图、柱状图、散点图等常见图表。与 D3.js 相比，Plot 更易上手，代码更简洁，适合快速原型开发和数据探索。","stars":5386,"starGrowth":0,"createdAt":"2020-10-29T22:30:31Z","silentRounds":3,"language":"HTML","topics":["data-visualization"],"aiDims":["非AI-实用","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据可视化","JavaScript","语法化图形","交互式图表"],"domainKey":"数据可视化","pushedAt":"2026-09-01T17:19:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/observablehq/plot","ts":"2026-09-01T17:19:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"amaurote/NodeGraphControl","owner":"amaurote","name":"NodeGraphControl","desc":"Node-Graph user-interface control for .NET WinForms.","summaryCn":"在Windows桌面应用里画节点图的控件，做可视化编程和流程图都好用","reasonCn":"技术实现上用GDI+做渲染，支持节点拖拽、选择高亮、连线动态更新这些基本交互。节点可以嵌套子节点，连线支持贝塞尔曲线。扩展性不错，你可以通过继承Node基类来创建自定义节点，然后组合成复杂的工作流图。对于想做可视化编程工具或者数据流编辑器的开发者来说，这个库省了不少事","stars":48,"starGrowth":0,"createdAt":"2021-02-20T14:17:13Z","silentRounds":25,"language":"C#","topics":["csharp","dotnet-framework","gdi-plus","graph","node","windows-forms"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["节点图","WinForms","流程图","可视化编程","GDI+"],"domainKey":"节点图","pushedAt":"2021-03-10T15:22:03Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"csharp","source":"github","weight":0.7},{"name":"dotnet-framework","source":"github","weight":0.7},{"name":"gdi-plus","source":"github","weight":0.7},{"name":"graph","source":"github","weight":0.7},{"name":"node","source":"github","weight":0.7},{"name":"windows-forms","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.32,"source":"bigbro","url":"https://github.com/amaurote/NodeGraphControl","ts":"2021-03-10T15:22:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"javalin/javalin","owner":"javalin","name":"javalin","desc":"A simple and modern Java and Kotlin web framework","summaryCn":"超轻量Java Web框架，代码少功能强","reasonCn":"专为 Kotlin 和 Java 设计的现代 Web 框架，以极简 API 著称。它没有繁重的注解扫描，启动速度极快，非常适合构建微服务或快速原型。相比 Spring Boot，它剔除了大量不必要的依赖，让开发者能专注于业务逻辑而非配置，是追求轻量级解决方案的绝佳选择。","stars":8349,"starGrowth":0,"createdAt":"2017-04-02T20:01:28Z","silentRounds":12,"language":"Kotlin","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Java","Kotlin","Web框架"],"domainKey":"Java","pushedAt":"2026-09-21T18:35:39Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/javalin/javalin","ts":"2026-09-21T18:35:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"metabase/metabase","owner":"metabase","name":"metabase","desc":"The easy-to-use open source Business Intelligence and Embedded Analytics tool that lets everyone work with data :bar_chart:","summaryCn":"让公司里谁都能轻松看懂数据、做分析的开源神器","reasonCn":"Metabase是一款开源的BI分析工具，主打零门槛上手，无需写SQL就能创建图表和看板。支持PostgreSQL、MySQL等主流数据库直连，界面简洁直观，非技术人员也能五分钟拖拽出数据看板。提供嵌入式分析能力，可集成到自有产品中。","stars":49376,"starGrowth":9,"createdAt":"2015-02-02T19:25:47Z","silentRounds":0,"language":"Clojure","topics":["database","data-visualization"],"aiDims":["非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["BI工具","数据可视化","嵌入式分析"],"domainKey":"BI工具","pushedAt":"2026-09-23T21:50:58Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"Clojure","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/metabase/metabase","ts":"2026-09-23T21:50:58Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.027546824368114065,"copyOk":false},{"repo":"CodebuffAI/freebuff","owner":"CodebuffAI","name":"freebuff","desc":"The free coding agent","summaryCn":"免费且永不限速的 coding agent，让编程助手不再吃土。","reasonCn":"CodebuffAI 推出的 Coding Agent，特点是真正免费且不限速，解决了商业 AI 编程工具高昂订阅费的痛点。它集成了 VS Code、JetBrains 等主流 IDE，支持 MCP 协议扩展工具能力，让开发者能零成本获得高效的 AI 辅助编码体验。","stars":12577,"starGrowth":31,"createdAt":"2024-07-09T21:21:56Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","开发者工具","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程","免费","VS Code","MCP","Coding Agent"],"domainKey":"AI编程","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/CodebuffAI/freebuff","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.100394450186849,"copyOk":false},{"repo":"sparckles/Robyn","owner":"sparckles","name":"Robyn","desc":"Robyn is a Super Fast Async Python Web Framework with a Rust runtime.","summaryCn":"用Rust跑起来的Python框架，速度比Flask快十倍不是梦","reasonCn":"Robyn是一个结合Python易用性和Rust高性能的异步Web框架。它通过Rust运行时处理底层网络IO，同时保留Python的装饰器语法，让开发者用熟悉的方式写出高性能API。对于追求极致速度的Python后端项目，这是一个极具吸引力的新选择。","stars":7406,"starGrowth":0,"createdAt":"2021-06-18T09:32:00Z","silentRounds":4,"language":"Python","topics":["backend"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Python","Web框架","异步"],"domainKey":"Python","pushedAt":"2026-09-21T18:18:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/sparckles/Robyn","ts":"2026-09-21T18:18:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"litu54/DevOps-Interview-Guide","owner":"litu54","name":"DevOps-Interview-Guide","desc":"DevOps Interview Guide","summaryCn":"DevOps面试全攻略，从基础到进阶知识一网打尽","reasonCn":"这是一份全面的DevOps面试准备指南，涵盖了Linux基础、网络原理、CI/CD流水线、容器化技术、Kubernetes编排、云平台使用、监控告警系统等核心知识点。内容包括常见面试题、解题思路和扩展阅读，是准备DevOps岗位面试的实用参考资料。","stars":1470,"starGrowth":3,"createdAt":"2025-04-06T16:47:44Z","silentRounds":0,"language":"","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["DevOps","面试准备","运维"],"domainKey":"DevOps","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1}],"aiScore":0.3,"source":"trending","url":"https://github.com/litu54/DevOps-Interview-Guide","ts":"2026-08-09T00:58:42.474Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"malisper/pgrust","owner":"malisper","name":"pgrust","desc":"Postgres rewritten in Rust, now faster than Postgres and Clickhouse","summaryCn":"用Rust重写的Postgres数据库，速度比原版和Clickhouse都快","reasonCn":"核心技术亮点在于并行查询执行引擎。传统Postgres的查询执行是串行的，而pgrust将查询计划分解为可并行执行的子任务，充分利用多核CPU。同时使用SIMD指令加速向量计算，这对分析型查询尤其有效。它保留了PostgreSQL协议兼容性，意味着现有工具如psql、pgAdmin可以直接连接使用，迁","stars":5114,"starGrowth":3,"createdAt":"2026-04-20T03:35:59Z","silentRounds":0,"language":"Rust","topics":["ai-assisted-development","database","postgres","postgresql","rust"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["Rust","PostgreSQL","数据库","性能优化","查询引擎"],"domainKey":"Rust","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai-assisted-development","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"postgres","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/malisper/pgrust","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"VerifyTests/Verify","owner":"VerifyTests","name":"Verify","desc":"Verify is a snapshot testing tool that simplifies the assertion of complex data models and documents.","summaryCn":"一个简化复杂数据模型和文档的测试工具。","reasonCn":"它提供了一个易于使用的界面，让用户能够轻松地编写和执行测试用例，确保软件的功能和数据准确性。","stars":3467,"starGrowth":0,"createdAt":"2019-11-21T23:37:12Z","silentRounds":1,"language":"C#","topics":["testing"],"aiDims":["测试","自动化测试"],"zone":"工具","funScore":0,"domainTags":["快照测试","断言工具","数据模型验证"],"domainKey":"快照测试","pushedAt":"2026-09-23T21:12:45Z","tags":[{"name":"测试","source":"llm","weight":1},{"name":"自动化测试","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/VerifyTests/Verify","ts":"2026-09-23T21:12:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zinja-coder/jadx-ai-mcp","owner":"zinja-coder","name":"jadx-ai-mcp","desc":"Plugin for JADX to integrate MCP server","summaryCn":"给安卓逆向工具装上AI大脑，自动分析APK代码","reasonCn":"jadx-ai-mcp是一款为JADX反编译工具打造的MCP协议插件，让AI大模型能够通过自然语言直接操控APK逆向流程。它将JADX暴露为MCP服务器，使AI助手可直接调用反编译API，读取Java源码、分析代码结构、定位关键逻辑，实现智能代码审查与漏洞挖掘。相比传统手动逆向，开发者用一句话就能完成APK功能解析，大幅降低移动端安全研究门槛。","stars":2817,"starGrowth":3,"createdAt":"2025-04-07T17:28:54Z","silentRounds":0,"language":"Java","topics":["mobile"],"aiDims":["工具调用","AI应用","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["APK逆向","MCP协议","Android安全","代码分析"],"domainKey":"APK逆向","pushedAt":"2026-09-23T07:51:12Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/zinja-coder/jadx-ai-mcp","ts":"2026-09-23T07:51:12Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"veops/oneterm","owner":"veops","name":"oneterm","desc":"Provide secure access and control over all infrastructure","summaryCn":"运维人员的安全访问入口，集中管理所有服务器权限告别账号混乱","reasonCn":"Go开发的基础设施安全访问控制平台，支持SSH、VNC、RDP等多种协议统一接入。提供细粒度权限管理、操作审计日志和审批流程，避免运维人员使用个人账号直连生产环境。Web端界面简洁，适合云原生团队集中管控运维入口。","stars":1926,"starGrowth":4,"createdAt":"2024-01-30T01:58:43Z","silentRounds":0,"language":"Go","topics":["golang","ssh","terminal","vue","oneterm","rdp","vnc","bastion"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["运维","服务器管理","SSH管理","堡垒机","权限控制"],"domainKey":"运维","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"golang","source":"github","weight":0.7},{"name":"ssh","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"oneterm","source":"github","weight":0.7},{"name":"rdp","source":"github","weight":0.7},{"name":"vnc","source":"github","weight":0.7},{"name":"bastion","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/veops/oneterm","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ctf-wiki/ctf-wiki","owner":"ctf-wiki","name":"ctf-wiki","desc":"Come and join us, we need you!","summaryCn":"中文CTF竞赛全品类知识库，从入门到实战一站式学习平台","reasonCn":"这是中文世界最全面的CTF（Capture The Flag）竞赛知识库，涵盖Web安全、密码学、逆向工程、二进制漏洞利用、Misc等多个方向。每个知识点都配有原理讲解、实战例题和详细解题思路，由社区持续维护更新。9500+星，是安全爱好者和CTF选手的必备资源。","stars":9651,"starGrowth":1,"createdAt":"2017-06-04T15:29:18Z","silentRounds":3,"language":"Python","topics":["mobile"],"aiDims":["学习资源","教程","安全工具"],"zone":"资源","funScore":0,"domainTags":["CTF","网络安全","知识库","Web安全","实战演练"],"domainKey":"CTF","pushedAt":"2026-08-23T03:50:38Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/ctf-wiki/ctf-wiki","ts":"2026-08-23T03:50:38Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"labstack/echo","owner":"labstack","name":"echo","desc":"High performance, minimalist Go web framework","summaryCn":"Go语言里最精简快活的Web框架，代码少到让你怀疑人生","reasonCn":"Echo是Go生态中最著名的微型Web框架之一，以高性能和极简主义著称。它提供了中间件、路由绑定和优雅的错误处理机制，适合构建REST API和微服务。相比Gin，Echo的上下文设计更加灵活，且在基准测试中常年保持顶级速度。","stars":32724,"starGrowth":2,"createdAt":"2015-03-01T17:43:01Z","silentRounds":0,"language":"Go","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Go 框架","Web 开发","REST API","微服务"],"domainKey":"Go 框架","pushedAt":"2026-09-21T04:33:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/labstack/echo","ts":"2026-09-21T04:33:17Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dmlc/xgboost","owner":"dmlc","name":"xgboost","desc":"Scalable, Portable and Distributed Gradient Boosting (GBDT, GBRT or GBM) Library,  for Python, R, Java, Scala, C and more. Runs on single machine, Hadoop, Spark, Dask, Flink and DataFlow","summaryCn":"机器学习竞赛夺冠利器，梯度提升算法的工业级实现","reasonCn":"技术上，XGBoost实现了高度优化的梯度提升框架。核心创新包括正则化项防止过拟合、二阶泰勒展开精确优化、稀疏数据自动学习、并行树学习加速等。它支持多种目标函数和评估指标，可以处理回归、分类、排序等多种任务。","stars":28786,"starGrowth":1,"createdAt":"2014-02-06T17:28:03Z","silentRounds":2,"language":"C++","topics":["ml"],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["梯度提升","机器学习","表格数据"],"domainKey":"梯度提升","pushedAt":"2026-09-23T18:31:50Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/dmlc/xgboost","ts":"2026-09-23T18:31:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"plbrault/youre-the-os","owner":"plbrault","name":"youre-the-os","desc":"A game where you are a computer's OS and you have to manage processes, memory and I/O events.","summaryCn":"把自己变成电脑管家，手忙脚乱排任务的小游戏","reasonCn":"这是一款用纯 Python 写成的小型系统模拟游戏。玩家接管一个迷你操作系统控制台，要手动处理进程创建、优先级调度、内存分配、外设输入输出和异常事件。它把抽象的调度算法、上下文切换、缓冲区和中断概念变成可操作的紧张时刻，适合边玩边理解计算机基础和系统思维。","stars":2683,"starGrowth":0,"createdAt":"2022-11-17T01:33:59Z","silentRounds":6,"language":"Python","topics":["game"],"aiDims":["游戏","非AI-好玩","学习资源"],"zone":"创意","funScore":0.7,"domainTags":["操作系统","教学游戏","进程调度","内存管理","Python"],"domainKey":"操作系统","pushedAt":"2026-09-23T00:08:14Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/plbrault/youre-the-os","ts":"2026-09-23T00:08:14Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"chinedufn/percy","owner":"chinedufn","name":"percy","desc":"Build frontend browser apps with Rust + WebAssembly. Supports server side rendering.","summaryCn":"用Rust写Web前端，性能拉满还能服务端渲染黑科技","reasonCn":"用Rust和WebAssembly构建前端浏览器应用的新范式，支持服务端渲染和静态站点生成，兼顾极致性能和传统前端开发体验。通过Wasm将Rust代码编译为浏览器可执行格式，实现接近原生速度的Web应用，适合追求性能极致的前端开发者。","stars":2315,"starGrowth":0,"createdAt":"2018-07-16T11:14:12Z","silentRounds":25,"language":"Rust","topics":["frontend"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WebAssembly","Rust","前端开发"],"domainKey":"WebAssembly","pushedAt":"2026-08-25T00:27:33Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/chinedufn/percy","ts":"2026-08-25T00:27:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"betapop/Beta-MAS-lbe","owner":"betapop","name":"Beta-MAS-lbe","desc":"a submod that adds new topics, spritepacks, interactions and more!","summaryCn":"视觉小说模组，给老游戏加新剧情和互动彩蛋","reasonCn":"在技术上，它完全运行在Ren'Py这个强大的视觉小说引擎之上。Ren'Py使用Python脚本控制游戏流程，这就意味着开发者可以灵活地插入新的分支、定义新的角色立绘（Spritepacks）以及设计复杂的互动反馈。这个模块的价值在于它整合了多个创作者的素材，包括新绘制的立绘、新的背景图以及","stars":14,"starGrowth":0,"createdAt":"2024-02-02T22:35:02Z","silentRounds":9,"language":"Ren'Py","topics":["english"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["Ren'Py","视觉小说","游戏模组","互动叙事"],"domainKey":"Ren'Py","pushedAt":"2026-09-16T15:50:37Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"english","source":"github","weight":0.7},{"name":"Ren'Py","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/betapop/Beta-MAS-lbe","ts":"2026-09-16T15:50:37Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"remnawave/panel","owner":"remnawave","name":"panel","desc":"A powerful proxy management tool, built on top of Xray-core, with a focus on simplicity and ease of use.","summaryCn":"Xray代理的一键管理面板，网页操作就能搞定复杂配置","reasonCn":"技术上，它基于Xray-core这个强大的代理内核构建，Xray是目前功能最全面的开源代理工具之一，支持VMess、VLESS、Trojan、Shadowsocks等多种协议。panel在此基础上提供了一个现代化的Web管理界面，让你通过浏览器就能完成所有配置操作，完全不需要碰命令行。","stars":5100,"starGrowth":3,"createdAt":"2025-01-10T15:50:17Z","silentRounds":0,"language":"TypeScript","topics":["censorship-resistant","nestjs","proxy-management","remnawave","trojan-gfw","typescript","vless","xray","xtls-core"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["代理管理","Xray","网络工具"],"domainKey":"代理管理","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"censorship-resistant","source":"github","weight":0.7},{"name":"nestjs","source":"github","weight":0.7},{"name":"proxy-management","source":"github","weight":0.7},{"name":"remnawave","source":"github","weight":0.7},{"name":"trojan-gfw","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"vless","source":"github","weight":0.7},{"name":"xray","source":"github","weight":0.7},{"name":"xtls-core","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.3,"source":"trending","url":"https://github.com/remnawave/panel","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wasm3/wasm3","owner":"wasm3","name":"wasm3","desc":"🚀 A fast WebAssembly interpreter and the most universal WASM runtime","summaryCn":"超快 WebAssembly 解释器，在浏览器和设备上跑代码的神器。","reasonCn":"它的核心优势在于“通用”和“快”。它不需要依赖任何外部库，编译后体积很小，可以直接嵌入到你的项目中。无论是浏览器环境，还是嵌入式设备，它都能提供稳定的运行时支持。 安装和使用非常方便，你可以直接从 GitHub 下载源码编译，或者通过包管理器安装。","stars":8031,"starGrowth":0,"createdAt":"2019-10-01T17:06:03Z","silentRounds":1,"language":"C","topics":["devops","iot"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["WebAssembly","解释器","跨平台","性能优化","嵌入式开发"],"domainKey":"WebAssembly","pushedAt":"2026-09-23T16:54:27Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"iot","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/wasm3/wasm3","ts":"2026-09-23T16:54:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gitify-app/gitify","owner":"gitify-app","name":"gitify","desc":"Git notifications on your menu bar. Available on macOS, Windows & Linux.","summaryCn":"想象一下，你作为开发者，每天要盯着好几个项目进度，网页版的GitHub通知系","reasonCn":"技术上，它采用TypeScript开发，基于Electron框架构建，确保了跨平台的原生体验。它巧妙地利用了GitHub提供的GraphQL和REST API，通过轮询机制实时抓取通知状态。它的架构轻巧，内存占用极低，不会在后台抢占宝贵的开发资源。","stars":5353,"starGrowth":2,"createdAt":"2015-05-21T18:15:28Z","silentRounds":0,"language":"TypeScript","topics":["git"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["GitHub","Notification","Desktop App"],"domainKey":"GitHub","pushedAt":"2026-09-23T12:10:23Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/gitify-app/gitify","ts":"2026-09-23T12:10:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kubernetes-sigs/kind","owner":"kubernetes-sigs","name":"kind","desc":"Kubernetes IN Docker - local clusters for testing Kubernetes","summaryCn":"Docker里跑K8s集群，本地测试部署的神器","reasonCn":"kind全称Kubernetes IN Docker，是Kubernetes官方SIG项目。它的核心思想是把每个K8s节点运行在一个Docker容器里，这样你本地只要有Docker就能跑K8s集群。相比minikube只跑单节点，kind支持多节点集群模拟；相比kubeadm需要虚拟机，kind更轻量启动更快。","stars":15502,"starGrowth":0,"createdAt":"2018-09-12T21:45:05Z","silentRounds":1,"language":"Go","topics":["docker","kubernetes"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","Docker","测试环境","DevOps"],"domainKey":"Kubernetes","pushedAt":"2026-09-04T13:08:26Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.28,"source":"trending","url":"https://github.com/kubernetes-sigs/kind","ts":"2026-09-04T13:08:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"filebrowser/filebrowser","owner":"filebrowser","name":"filebrowser","desc":"File Browser provides a file managing interface within a specified directory and it can be used to upload, delete, preview and edit your files.","summaryCn":"自托管私有网盘的颜值担当，手机电脑随时传文件","reasonCn":"File Browser是一个自带Web界面的轻量级文件管理器，部署在一台服务器上，就能在手机或电脑的浏览器里管理文件——上传、下载、预览、编辑，甚至共享文件夹给他人。它基于Go语言开发，单二进制文件，零依赖，支持MySQL/PostgreSQL/SQLite等多种数据库，自带用户权限管理，是自建私有云盘或团队文件中心的绝佳选择。","stars":35939,"starGrowth":2,"createdAt":"2015-09-11T21:21:58Z","silentRounds":0,"language":"Go","topics":["self-hosted"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文件管理","Web界面","私有云盘","自托管","Go语言"],"domainKey":"文件管理","pushedAt":"2026-07-31T10:06:58Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/filebrowser/filebrowser","ts":"2026-07-31T10:06:58Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cocos/cocos4","owner":"cocos","name":"cocos4","desc":"COCOS 4","summaryCn":"Cocos游戏引擎第四代版本，2D3D游戏开发利器","reasonCn":"这是Cocos游戏引擎的第四代版本，用于开发2D和3D游戏。Cocos是知名的国产游戏引擎，支持跨平台发布，适合开发移动端和网页游戏。star数1400+，说明有一定用户基础。虽然不是AI项目，但作为游戏开发工具，对游戏开发者来说很实用。","stars":1471,"starGrowth":0,"createdAt":"2025-12-18T15:01:19Z","silentRounds":3,"language":"C++","topics":[],"aiDims":["游戏","创意工具"],"zone":"工具","funScore":0,"domainTags":["游戏引擎","C++","2D/3D 开发","跨平台","移动端"],"domainKey":"游戏引擎","pushedAt":"2026-08-17T10:01:13Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.3,"source":"bigbro","url":"https://github.com/cocos/cocos4","ts":"2026-08-17T10:01:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/langextract","owner":"google","name":"langextract","desc":"A Python library for extracting structured information from unstructured text using LLMs with precise source grounding and interactive visualization.","summaryCn":"大模型终于学会‘翻原文’了，不再是张口就来的瞎编侠","reasonCn":"谷歌开源的 Python 库，利用 LLM 从杂乱无章的非结构化文本里提取关键信息。它最牛的地方是‘源头锚定’——每提取一条信息，都能精准标注原文位置。自带可视化界面，让你直观看到模型是怎么从一堆废话里捞出干货的，特别适合需要高可信度的知识处理场景。","stars":38644,"starGrowth":2,"createdAt":"2025-07-08T20:46:06Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["文档理解","AI应用","提示工程"],"zone":"AI","funScore":0,"domainTags":["LLM","信息提取","源头锚定"],"domainKey":"LLM","pushedAt":"2026-09-21T20:54:13Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/google/langextract","ts":"2026-09-21T20:54:13Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"advplyr/audiobookshelf","owner":"advplyr","name":"audiobookshelf","desc":"Self-hosted audiobook and podcast server","summaryCn":"打造私人有声书库，把你的播客和有声书全部私有化","reasonCn":"它的架构是典型的自托管（Self-hosted）模式。后端负责文件的扫描、索引和进度跟踪，前端提供精美的 Web 界面。它最强的地方在于对‘书籍’这一特殊格式的优化：支持多章节管理、记录每本书读到具体哪一秒、支持多种音频格式转码，并且能自动从网上下载书籍封面和作者信息，让你的私","stars":14414,"starGrowth":9,"createdAt":"2021-08-17T22:00:40Z","silentRounds":0,"language":"JavaScript","topics":["audiobookshelf","audiobooks","self-hosted","audiobook-manager","podcasts"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["有声书","自托管","播客","媒体服务器","元数据管理"],"domainKey":"有声书","pushedAt":"2026-09-22T23:00:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"audiobookshelf","source":"github","weight":0.7},{"name":"audiobooks","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"audiobook-manager","source":"github","weight":0.7},{"name":"podcasts","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/advplyr/audiobookshelf","ts":"2026-09-22T23:00:08Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02887317885389205,"copyOk":true},{"repo":"tuyoogame/YooAsset","owner":"tuyoogame","name":"YooAsset","desc":"unity3d resources management  system","summaryCn":"Unity游戏开发者的资源管理神器，告别打包混乱和内存爆炸","reasonCn":"YooAsset的核心设计包括资源包管理、依赖分析、异步加载和热更新。它把资源按业务模块打包，自动分析资源依赖关系，支持增量更新和回滚。技术上采用Addressable的改进方案，但提供了更友好的API和可视化编辑器工具。","stars":3370,"starGrowth":0,"createdAt":"2022-03-01T02:16:07Z","silentRounds":1,"language":"C#","topics":[],"aiDims":["游戏","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Unity","资源管理","游戏开发"],"domainKey":"Unity","pushedAt":"2026-08-11T08:05:44Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"bigbro","url":"https://github.com/tuyoogame/YooAsset","ts":"2026-08-11T08:05:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"everfu/hexo-theme-solitude","owner":"everfu","name":"hexo-theme-solitude","desc":"一款设计师风格的 Hexo 主题，支持懒加载、PWA、Latex以及多种评论系统。","summaryCn":"一键打造专业级技术博客，支持PWA和论文公式","reasonCn":"这款Hexo主题是专为设计师和开发者打造的个人博客模板，内置懒加载提升页面速度，支持PWA离线访问，还能渲染LaTeX数学公式，适配多种评论系统如Disqus和Gitalk。主题采用Stylus预处理器，提供深色模式和响应式布局，让你零代码基础也能快速搭建高颜值的学术或技术博客。","stars":1189,"starGrowth":0,"createdAt":"2023-10-25T01:10:13Z","silentRounds":25,"language":"Stylus","topics":["design"],"aiDims":["非AI-实用","创意工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Hexo博客","PWA","LaTeX公式"],"domainKey":"Hexo博客","pushedAt":"2026-09-03T00:33:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Stylus","source":"language","weight":0.5}],"aiScore":0.28,"source":"search","url":"https://github.com/everfu/hexo-theme-solitude","ts":"2026-09-03T00:33:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kujian/frontendDaily","owner":"kujian","name":"frontendDaily","desc":"前端开发博客，分享互联网最精彩的前端技术，欢迎关注我微信公众号：前端开发博客，回复 1024，领取前端进阶资料包，回复 加群，与大神一起交流学习。","summaryCn":"前端学习博客，收录互联网优质文章并送资料包","reasonCn":"这是一个前端技术博客项目，汇聚了互联网上精彩的HTML、CSS和JavaScript教程，覆盖Vue、React等主流框架。作者运营公众号分享优质内容，并整理了进阶资料包供读者免费领取。虽然完全非AI项目，但对开发者的技能提升有帮助。","stars":2385,"starGrowth":0,"createdAt":"2016-05-27T06:07:42Z","silentRounds":25,"language":"HTML","topics":["frontend"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["前端开发","技术博客","学习资源"],"domainKey":"前端开发","pushedAt":"2026-08-11T13:28:03Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.3,"source":"search","url":"https://github.com/kujian/frontendDaily","ts":"2026-08-11T13:28:03Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"daylen/stockfish-mac","owner":"daylen","name":"stockfish-mac","desc":"Beautiful, powerful chess app for the Mac","summaryCn":"功能强大的 macOS 国际象棋引擎，配合 UI 即可对战。","reasonCn":"这是著名的开源国际象棋引擎Stockfish在macOS平台上的移植版本。Stockfish被公认为目前全球最强的开源象棋AI，拥有极高的棋力。该版本针对Apple Silicon进行了深度优化，能够充分利用M系列芯片的性能，提供流畅的棋局分析体验。","stars":541,"starGrowth":0,"createdAt":"2014-01-11T08:15:17Z","silentRounds":15,"language":"C++","topics":["chess"],"aiDims":["非AI-实用","游戏"],"zone":"创意","funScore":1,"domainTags":["国际象棋","AI 对弈","引擎优化","Mac 平台","算法博弈"],"domainKey":"国际象棋","pushedAt":"2026-09-22T19:34:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/daylen/stockfish-mac","ts":"2026-09-22T19:34:22Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fishercoder1534/Leetcode","owner":"fishercoder1534","name":"Leetcode","desc":"Solutions to LeetCode problems; updated daily. Subscribe to my YouTube channel for more.","summaryCn":"每日更新 LeetCode 刷题解法，Java 面试刷题的神器。","reasonCn":"项目最大的价值在于“时效性”。随着 LeetCode 题库的更新，这里的代码也会同步更新。无论是经典的二叉树、链表算法，还是最新的动态规划、贪心算法，你都能在这里找到经过验证的解决方案。每个题目的解析通常都包含“题目解析”、“代码实现”和“复杂度分析”三个部分，让你不仅知道“怎么","stars":4001,"starGrowth":0,"createdAt":"2016-07-31T05:28:15Z","silentRounds":8,"language":"Java","topics":["interview"],"aiDims":["代码助手","学习资源"],"zone":"资源","funScore":0,"domainTags":["算法","面试准备","LeetCode","Java","题解"],"domainKey":"算法","pushedAt":"2026-09-20T14:14:21Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/fishercoder1534/Leetcode","ts":"2026-09-20T14:14:21Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yoosuf/Messenger","owner":"yoosuf","name":"Messenger","desc":"Relational messaging database schema for one-to-one and group chat apps with attachments and MySQL Workbench model.","summaryCn":"聊天应用的数据库设计模板，直接抄就能建消息系统","reasonCn":"技术上，它采用经典的关系型数据库设计模式。核心表包括用户表、消息表、群组表、群成员关系表等，支持一对一私聊和多人群聊两种场景。消息表通常采用自关联设计，方便查询会话历史。附件通过外键关联到消息记录，支持图片、文件等多种类型。","stars":611,"starGrowth":0,"createdAt":"2015-05-28T07:44:20Z","silentRounds":25,"language":"","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"资源","funScore":0,"domainTags":["数据库设计","MySQL","聊天应用","Schema","即时通讯"],"domainKey":"数据库设计","pushedAt":"2026-07-29T02:36:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7}],"aiScore":0.3,"source":"search","url":"https://github.com/yoosuf/Messenger","ts":"2026-07-29T02:36:08Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"typescript-cheatsheets/react","owner":"typescript-cheatsheets","name":"react","desc":"Cheatsheets for experienced React developers getting started with TypeScript","summaryCn":"帮React老手快速上手TypeScript的全套速查表，省时省力","reasonCn":"很多前端团队在转向 TypeScript 时，往往要在已有的 React 代码库里逐步加入类型标注，过程中会遇到很多细节问题。这个速查表项目正是为了解决这类痛点，收集了从组件属性定义、事件处理到 Hook 使用的常见场景，并配以简明示例，让开发者不必翻阅官方文档就能快速定位答案。","stars":47099,"starGrowth":0,"createdAt":"2018-06-02T04:08:16Z","silentRounds":1,"language":"TypeScript","topics":["cheatsheet"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["TypeScript","React","速查表"],"domainKey":"TypeScript","pushedAt":"2026-09-09T17:26:52Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/typescript-cheatsheets/react","ts":"2026-09-09T17:26:52Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ConduitPlatform/Conduit","owner":"ConduitPlatform","name":"Conduit","desc":"Batteries-included backend that works with any stack.","summaryCn":"开箱即用的全栈后端，自带用户认证实时通信不用从零搭","reasonCn":"Conduit是一个开箱即用的后端框架，集成了用户认证、实时通信、文件存储等常用功能。采用TypeScript编写，支持多种部署方式。通过插件机制灵活扩展，开发者可以直接使用现成功能或按需定制，适合快速搭建API服务和复杂后端系统，节省重复开发时间。","stars":925,"starGrowth":0,"createdAt":"2022-01-30T15:53:02Z","silentRounds":24,"language":"TypeScript","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["后端框架","用户认证","实时通信"],"domainKey":"后端框架","pushedAt":"2026-09-21T00:59:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/ConduitPlatform/Conduit","ts":"2026-09-21T00:59:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"chenyme/grok2api","owner":"chenyme","name":"grok2api","desc":"Multi-account API gateway for Grok Build, Grok Web, and Grok Console","summaryCn":"把Grok智能聊天接入口袋，多账号聚合API超省心","reasonCn":"这是一个基于Go语言开发的网关服务，专门用来聚合Grok的多账号API接口。它支持Grok Build、Grok Web和Grok Console三种入口，开发者可以通过统一的API端点调用不同账号的额度，实现负载均衡和高并发访问，非常适合需要批量处理任务或突破单账号频率限制的应用场景。","stars":7730,"starGrowth":11,"language":"Go","topics":["grok","grok-imagine","grok-build","grok-video","grok-console","grok-web"],"aiDims":["AI应用","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["API聚合","Grok","网关服务","模型调用"],"domainKey":"API聚合","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"grok","source":"github","weight":0.7},{"name":"grok-imagine","source":"github","weight":0.7},{"name":"grok-build","source":"github","weight":0.7},{"name":"grok-video","source":"github","weight":0.7},{"name":"grok-console","source":"github","weight":0.7},{"name":"grok-web","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/chenyme/grok2api","ts":"2026-09-23T21:53:43.262Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03726908150064683,"copyOk":false},{"repo":"digitalsamba/claude-code-video-toolkit","owner":"digitalsamba","name":"claude-code-video-toolkit","desc":"","summaryCn":"AI原生视频生产工作区，用Copilot自然语言指令生成完整视频。","reasonCn":"这是一个专为 Claude Code 设计的 AI 视频制作工具包。用户只需自然语言描述需求，AI 代理即自动完成脚本撰写、配音、配乐及画面生成，最终渲染为 MP4。它利用低成本开源模型，单次视频生成成本极低，实现了从创意到成片的全流程自动化，大幅降低了视频制作门槛。","stars":2078,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","多模态Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI视频生成","语音合成","视频渲染","Claude集成","开源模型"],"domainKey":"AI视频生成","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/digitalsamba/claude-code-video-toolkit","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"opentk/GLWpfControl","owner":"opentk","name":"GLWpfControl","desc":"A fast native control for OpenTK 4.x + 3.x on WPF.","summaryCn":"在WPF窗口里流畅跑OpenGL的C#控件，3D图形渲染利器","reasonCn":"这是一个专为WPF设计的OpenGL渲染控件，基于OpenTK库实现硬件加速的3D图形渲染。它让开发者能在Windows桌面应用中轻松嵌入OpenGL画布，支持OpenTK 3.x和4.x两个版本。通过NuGet包管理，一行代码即可添加到项目。相比传统软件渲染方案，这个控件提供原生GPU加速，适合实时渲染场景。对于需要可视化、游戏开发或3D建模工具的人来说，这是非常实用的基础设施。","stars":230,"starGrowth":0,"createdAt":"2019-06-16T13:50:45Z","silentRounds":25,"language":"C#","topics":["3d","glcontrol","graphics","nuget","opengl","opentk","wpf"],"aiDims":["非AI-好玩","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["3D渲染","OpenGL","WPF","图形库","控件"],"domainKey":"3D渲染","pushedAt":"2026-04-14T23:51:39Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"3d","source":"github","weight":0.7},{"name":"glcontrol","source":"github","weight":0.7},{"name":"graphics","source":"github","weight":0.7},{"name":"nuget","source":"github","weight":0.7},{"name":"opengl","source":"github","weight":0.7},{"name":"opentk","source":"github","weight":0.7},{"name":"wpf","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.3,"source":"bigbro","url":"https://github.com/opentk/GLWpfControl","ts":"2026-04-14T23:51:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xMiden/miden-vm","owner":"0xMiden","name":"miden-vm","desc":"STARK-based virtual machine","summaryCn":"零知识虚拟机，验证代码完整性，提升安全性的可信","reasonCn":"Miden VM 基于 STARK 零知识证明实现安全、可扩展的虚拟机，支持 Rust 代码执行与验证，兼容 WebAssembly 运行。\n核心采用自定义指令集与状态管理，能够与以太坊等区块链兼容，提供高性能、低成本的验证路径，简化链下计算。\n适合区块链合约、隐私保护去中心化应用的开发者，快速构建可信执行环境，减少 Gas 费用和安全风险。","stars":774,"starGrowth":0,"createdAt":"2021-09-10T05:08:56Z","silentRounds":11,"language":"Rust","topics":["cryptography"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["零知识证明","虚拟机","区块链","密码学"],"domainKey":"零知识证明","pushedAt":"2026-09-23T19:58:22Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cryptography","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.12,"source":"search","url":"https://github.com/0xMiden/miden-vm","ts":"2026-09-23T19:58:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"raspberrypi/firmware","owner":"raspberrypi","name":"firmware","desc":"","summaryCn":"这是树莓派（Raspberry Pi）官方的固件与预编译二进制文件仓库，主要服务于","reasonCn":"仓库内容涵盖多个关键部分：boot 目录下包含 GPU 固件文件（start*.elf、fixup*.dat 等）、引导加载器（bootcode.bin）以及 Linux 内核镜像（kernel.img）。这些组件构成了树莓派启动和运行的基础。dtbs 和 overlay 文件则用于设备树配置，让系统能够适配不同型号的树莓派硬件。","stars":5575,"starGrowth":0,"createdAt":"2012-02-27T23:06:53Z","silentRounds":6,"language":"","topics":[],"aiDims":["其他"],"zone":"资源","funScore":0,"domainTags":["树莓派","固件","嵌入式"],"domainKey":"树莓派","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"其他","source":"llm","weight":1}],"aiScore":0.2,"source":"tier","url":"https://github.com/raspberrypi/firmware","ts":"2026-09-15T22:54:20.912Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Doriandarko/deepseek-engineer","owner":"Doriandarko","name":"deepseek-engineer","desc":"","summaryCn":"终端里的AI编程搭档，对话加函数调用直接操作文件写代码","reasonCn":"这是一个将DeepSeek大模型接入终端环境的AI编程助手，核心亮点是函数调用架构：AI能直接调用工具执行文件操作和代码修改，而不只是给出文字建议。你可以在交互式终端里用自然语言提需求，AI会实际动手改代码、分析项目。需要配置DeepSeek API密钥后使用。适合希望用AI真正干活而不是陪聊的开发者，尤其适合命令行习惯者。","stars":2235,"starGrowth":0,"createdAt":"2024-12-26T20:22:01Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.38,"domainTags":["编程助手","命令行工具","DeepSeek","函数调用"],"domainKey":"编程助手","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/Doriandarko/deepseek-engineer","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"raspberrypi/rpi-imager","owner":"raspberrypi","name":"rpi-imager","desc":"The home of Raspberry Pi Imager, a user-friendly tool for creating bootable media for Raspberry Pi devices.","summaryCn":"给树莓派装系统的一键傻瓜工具，比烧录卡更省心","reasonCn":"在技术架构上，它采用C++编写，保证了极高的运行效率和对底层硬件的精准控制。不同于传统的图形化工具，它内置了强大的存储管理引擎，能够实时监控写入进度、计算校验和，甚至在写入完成后自动清除敏感数据。它支持从本地、云端（通过HTTP/HTTPS）或直接从树莓派设备读取镜像，这意","stars":2612,"starGrowth":0,"createdAt":"2019-09-30T12:49:05Z","silentRounds":3,"language":"C++","topics":["raspberry-pi"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Raspberry Pi","系统部署","镜像烧录"],"domainKey":"Raspberry Pi","pushedAt":"2026-09-23T15:05:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/raspberrypi/rpi-imager","ts":"2026-09-23T15:05:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"romannurik/AndroidAssetStudio","owner":"romannurik","name":"AndroidAssetStudio","desc":"","summaryCn":"工具的设计理念是将枯燥的UI素材适配工作自动化，通过上传一张基础图片，它","reasonCn":"若想在本地运行该工具，开发者需要确保环境中已配置好Node.js及npm。通过简单的 npm install 和 npm start 命令即可启动本地服务，在浏览器界面中直接使用各项资产生成功能。虽然工具已不再更新，但其处理资源缩放与目录分类的逻辑仍非常清晰，适合需要自定义资源生成流程的开发者","stars":6593,"starGrowth":0,"createdAt":"2014-05-19T14:02:00Z","silentRounds":6,"language":"JavaScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["安卓开发","图标生成","UI资源"],"domainKey":"安卓开发","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"tier","url":"https://github.com/romannurik/AndroidAssetStudio","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AR-js-org/AR.js","owner":"AR-js-org","name":"AR.js","desc":"","summaryCn":"无需安装即可在浏览器里运行增强现实的轻量级JavaScript库。","reasonCn":"AR.js是一个轻量级Web AR库，无需安装任何应用或插件，直接在浏览器里就能跑增强现实效果。支持三种核心模式：标记跟踪（对准特定图案触发AR）、图像跟踪（识别真实世界中的图片）和基于位置的AR（根据GPS定位在场景中叠加虚拟物体）。它基于Three.js等成熟图形库构建，开发者可以用几行代码就把3D模型、视频等内容叠加到真实画面上，适合做互动展览、营销玩法和教学演示。","stars":5991,"starGrowth":0,"createdAt":"2020-02-28T09:03:36Z","silentRounds":2,"language":"JavaScript","topics":["augmented-reality","webar","ar","aframe","threejs","hacktoberfest","hacktoberfest2020"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"工具","funScore":0.44,"domainTags":["Web AR","增强现实","JavaScript库","图像跟踪","标记跟踪","浏览器端"],"domainKey":"Web AR","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"augmented-reality","source":"github","weight":0.7},{"name":"webar","source":"github","weight":0.7},{"name":"ar","source":"github","weight":0.7},{"name":"aframe","source":"github","weight":0.7},{"name":"threejs","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"hacktoberfest2020","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/AR-js-org/AR.js","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"electronicarts/CnC_Red_Alert","owner":"electronicarts","name":"CnC_Red_Alert","desc":"","summaryCn":"项目不仅是一段代码的搬运，更是一次对经典游戏底层的溯源，它包含了当年开","reasonCn":"安装和使用方面，由于代码依赖了大量如Greenleaf Communications和HMI声音处理系统等过时库，想在现代Windows开发机上复现构建过程需要开发者手动替换或剔除这些老旧依赖。对于大部分玩家而言，这并非直接玩游戏的入口，而是给那些希望深入挖掘RTS游戏底层机制、进行逆向工程或深度","stars":6703,"starGrowth":0,"createdAt":"2024-08-22T01:35:23Z","silentRounds":6,"language":"C++","topics":["command-and-conquer","open-source","real-time-strategy","rts"],"aiDims":["开发者工具"],"zone":"创意","funScore":1,"domainTags":["游戏源码","RTS","红色警戒","逆向工程"],"domainKey":"游戏源码","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"command-and-conquer","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"real-time-strategy","source":"github","weight":0.7},{"name":"rts","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.2,"source":"tier","url":"https://github.com/electronicarts/CnC_Red_Alert","ts":"2026-09-15T22:54:20.912Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"britcruise9/GrowBot","owner":"britcruise9","name":"GrowBot","desc":"What's the simplest possible AI robot that learns everything from scratch?","summaryCn":"从零开始教机器人学习生存，这个最简易的AI伴侣正等着你去亲手创造。","reasonCn":"GrowBot致力于探索AI与物理实体结合的最简单可能。它不仅提供AI学习的基础框架，还深度融合了3D打印技术，让你能够自己动手制造机器人的躯壳。通过这种方式，它让AI不仅存在于虚拟服务器，更能通过传感器与现实世界产生交互。这是一个将AI实验与实体制作完美统一的极客项目，充满想象力。","stars":252,"starGrowth":1,"createdAt":"2026-06-04T14:54:11Z","silentRounds":2,"language":"Python","topics":["3d-printing"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["3D打印","强化学习","DIY机器人"],"domainKey":"3D打印","pushedAt":"2026-09-21T20:42:57Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/britcruise9/GrowBot","ts":"2026-09-21T20:42:57Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alibaba/EasyNLP","owner":"alibaba","name":"EasyNLP","desc":"","summaryCn":"阿里开源的易用 NLP 训练与部署框架。","reasonCn":"EasyNLP 是阿里云推出的高性能 NLP 开发工具包，旨在解决大模型落地难的问题。它集成了知识蒸馏、少样本学习等算法，能将庞大的预训练模型压缩并适配到各种业务场景中。无论是文本分类、信息抽取还是多模态任务，开发者都可以通过这个框架快速完成从训练到部署的全流程，大大降低了使用大模型的门槛。","stars":2186,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","大语言模型"],"zone":"工具","funScore":0.32,"domainTags":["NLP 工具包","预训练模型","知识蒸馏","Few-Shot Learning"],"domainKey":"NLP 工具包","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/alibaba/EasyNLP","ts":"2026-09-22T22:51:50.275Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"andrew--r/frontend-case-studies","owner":"andrew--r","name":"frontend-case-studies","desc":"💼 A curated list of talks and articles about real world frontend development","summaryCn":"前端开发者的宝藏库，精选大厂实战经验和避坑指南","reasonCn":"这个项目收录了来自Netflix、Google、Shopify等科技公司的真实前端案例，包括架构设计、性能优化、错误处理等实战经验。每篇文章都经过精心挑选，作者来自一线工程师，涵盖React、Vue、CSS等技术栈。适合想提升前端工程能力的开发者系统学习。","stars":7430,"starGrowth":0,"createdAt":"2018-02-25T14:15:12Z","silentRounds":24,"language":"Astro","topics":["frontend"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["前端开发","案例分析","最佳实践","技术分享"],"domainKey":"前端开发","pushedAt":"2026-08-19T22:07:45Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"Astro","source":"language","weight":0.5}],"aiScore":0.28,"source":"search","url":"https://github.com/andrew--r/frontend-case-studies","ts":"2026-08-19T22:07:45Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"amberframework/amber","owner":"amberframework","name":"amber","desc":"A Crystal web framework that makes building applications fast, simple, and enjoyable. Get started with quick prototyping, less bugs, and blazing fast performance.","summaryCn":"Crystal语言写的极速Web框架，上手超简单","reasonCn":"Amber是一个基于Crystal语言的Web框架，Crystal是类Ruby语法的系统级语言，编译后性能接近C。Amber提供了完整的MVC架构、ORM、路由、模板引擎等开箱即用的组件。其核心卖点是开发体验好、运行速度快、bug少，适合需要高性能Web应用的场景。","stars":2604,"starGrowth":0,"createdAt":"2017-04-17T18:02:35Z","silentRounds":25,"language":"Crystal","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Crystal 语言","Web 框架","MVC","高性能"],"domainKey":"Crystal 语言","pushedAt":"2026-08-13T17:04:24Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Crystal","source":"language","weight":0.5}],"aiScore":0.28,"source":"search","url":"https://github.com/amberframework/amber","ts":"2026-08-13T17:04:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"radarsimx/radarsimpy","owner":"radarsimx","name":"radarsimpy","desc":"Radar Simulator built with Python and C++","summaryCn":"Python+ C++ 雷达模拟器，逼真波形可视化","reasonCn":"radarsimpy 是基于 Python 与 C++ 的雷达波形模拟器，支持多频段、双通道仿真，能生成逼真波形图。\n其核心模块采用高性能 FFT 与可视化渲染，支持用户自定义探测器参数和环境噪声，实现实时交互。\n适合航空、航天研发人员、雷达爱好者，以及教学实验室使用，快速验证设计方案与算法性能，降低实验成本。","stars":577,"starGrowth":0,"createdAt":"2020-07-27T16:48:19Z","silentRounds":11,"language":"Python","topics":["simulation"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["雷达仿真","波形可视化","信号处理","FFT"],"domainKey":"雷达仿真","pushedAt":"2026-09-17T01:07:29Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.18,"source":"search","url":"https://github.com/radarsimx/radarsimpy","ts":"2026-09-17T01:07:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DepthAnything/Video-Depth-Anything","owner":"DepthAnything","name":"Video-Depth-Anything","desc":"","summaryCn":"任意长度视频的深度估计模型，兼具扩散模型的精度与更快速度。","reasonCn":"Video-Depth-Anything是由字节跳动团队开发的视频深度估计工具，基于Depth Anything V2构建。它能处理任意长度的输入视频，同时保持帧间深度的连续性和一致性。相比扩散模型，推理速度更快、参数更少，而且精度并不打折扣。","stars":2165,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","模型与训练"],"zone":"AI","funScore":0.32,"domainTags":["视频深度估计","计算机视觉","人工智能","字节跳动"],"domainKey":"视频深度估计","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/DepthAnything/Video-Depth-Anything","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"CaviraOSS/LongMemory","owner":"CaviraOSS","name":"LongMemory","desc":"Local persistent memory store for LLM applications including claude desktop, github copilot, codex, antigravity, etc.","summaryCn":"给AI装上长期记忆，告别每次对话像失忆。","reasonCn":"专为LLM应用打造的本地持久化向量存储，支持Claude Desktop、GitHub Copilot等多平台。通过向量检索实现长期上下文记忆，解决对话遗忘问题。提供RESTful API，兼容OpenAI兼容接口。支持CRUD操作，适合需要记忆上下文的应用场景。","stars":4512,"starGrowth":1,"createdAt":"2025-10-19T16:12:49Z","silentRounds":2,"language":"TypeScript","topics":["vector-db","rag"],"aiDims":["向量数据库","AI应用","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["向量数据库","LLM记忆","本地持久化","RESTful API","TypeScript"],"domainKey":"向量数据库","pushedAt":"2026-09-20T11:57:46Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/CaviraOSS/LongMemory","ts":"2026-09-20T11:57:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"FoalTS/foal","owner":"FoalTS","name":"foal","desc":"Full-featured Node.js framework 🚀","summaryCn":"用TypeScript写后端API的瑞士军刀，从路由到数据库全包了","reasonCn":"FoalTS是一个基于TypeScript的全栈Node.js框架，提供从路由、依赖注入、验证到数据库访问的完整能力。相比Express需要手动拼凑中间件，它内置了装饰器、ORM和权限管理，开箱即用。项目使用现代TypeScript特性，类型安全，适合追求代码质量和开发效率的团队构建REST API或BFF层。","stars":1935,"starGrowth":0,"createdAt":"2017-09-17T18:20:25Z","silentRounds":25,"language":"TypeScript","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Node.js框架","REST API","TypeScript"],"domainKey":"Node.js框架","pushedAt":"2026-07-29T14:49:29Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.28,"source":"search","url":"https://github.com/FoalTS/foal","ts":"2026-07-29T14:49:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"FlameskyDexive/FastBugly","owner":"FlameskyDexive","name":"FastBugly","desc":"快速接入新版Bugly到unity，支持unity2021","summaryCn":"Unity游戏接入Bugly崩溃上报，几分钟就搞定","reasonCn":"Unity游戏上线后崩溃排查是最头疼的事之一，这个工具帮你快速接入腾讯Bugly崩溃上报服务。支持Unity 2021版本，集成简单，能自动捕获异常、上传堆栈日志，让游戏运营者第一时间知道出了什么问题，不需要手动收集用户反馈，大幅提升问题定位效率。","stars":123,"starGrowth":0,"createdAt":"2022-09-16T07:33:33Z","silentRounds":25,"language":"C#","topics":["bug","bugly","crash","crash-reporting","exception","firebase","log","reports","unity"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["Unity","崩溃上报","Bugly"],"domainKey":"Unity","pushedAt":"2022-12-06T16:16:33Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"bug","source":"github","weight":0.7},{"name":"bugly","source":"github","weight":0.7},{"name":"crash","source":"github","weight":0.7},{"name":"crash-reporting","source":"github","weight":0.7},{"name":"exception","source":"github","weight":0.7},{"name":"firebase","source":"github","weight":0.7},{"name":"log","source":"github","weight":0.7},{"name":"reports","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.28,"source":"bigbro","url":"https://github.com/FlameskyDexive/FastBugly","ts":"2022-12-06T16:16:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"linkwarden/linkwarden","owner":"linkwarden","name":"linkwarden","desc":"⚡️⚡️⚡️ Self-hosted collaborative bookmark manager to collect, read, annotate, and fully preserve what matters, all in one place.","summaryCn":"自托管的书签神器，把喜欢的网页全部收藏整理统一管理","reasonCn":"它的核心亮点是自托管——你用自己的服务器运行，数据100%属于你。它不只是存链接，还能完整保存网页内容（包括图片和排版），即使原文被删除你也能查看。支持标签分类、收藏夹分组、全文搜索，找到以前保存的内容超方便。","stars":19824,"starGrowth":4,"createdAt":"2022-04-09T07:53:01Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["书签管理","自托管","知识管理"],"domainKey":"书签管理","pushedAt":"2026-09-10T18:53:30Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.22,"source":"search","url":"https://github.com/linkwarden/linkwarden","ts":"2026-09-10T18:53:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"datajuicer/data-juicer","owner":"datajuicer","name":"data-juicer","desc":"Data processing for and with foundation models!  🍎 🍋 🌽 ➡️ ➡️🍸 🍹 🍷","summaryCn":"阿里开源的大模型数据管道，一键清洗混合训练数据","reasonCn":"阿里达摩院开源的 AI 数据处理框架，专为大模型训练设计。支持文本清洗、去重、过滤、混合等全套数据流水线操作，提供 100 多种数据处理器，可通过 YAML 配置文件灵活编排数据处理流程，让大模型团队不再为脏数据头疼。","stars":7083,"starGrowth":1,"createdAt":"2023-08-01T09:16:41Z","silentRounds":1,"language":"Python","topics":["data-visualization","data-science"],"aiDims":["AI基础设施","模型与训练"],"zone":"AI","funScore":0,"domainTags":["数据清洗","大模型","数据处理"],"domainKey":"数据清洗","pushedAt":"2026-09-22T03:10:18Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/datajuicer/data-juicer","ts":"2026-09-22T03:10:18Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"polterguy/magic","owner":"polterguy","name":"magic","desc":"Deterministic AI agents and generated backends — MIT-licensed, self-hosted, MCP-native","summaryCn":"C#开发者的确定性AI代理框架，MCP原生支持让Agent调用更可靠","reasonCn":"架构设计上，Magic采用分层结构：核心引擎负责Agent状态管理和执行流程，工具层提供标准化的工具注册和调用接口，记忆层支持短期上下文和长期持久化，MCP协议层实现与外部系统的标准化通信。与LangChain等框架相比，Magic的独特之处在于它的确定性模型——通过状态机和验证规则，确保","stars":1194,"starGrowth":0,"createdAt":"2019-03-16T04:57:01Z","silentRounds":8,"language":"C#","topics":["backend"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","确定性模型","C#","MCP协议","状态机"],"domainKey":"AI Agent","pushedAt":"2026-09-23T11:56:48Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/polterguy/magic","ts":"2026-09-23T11:56:48Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Jakubantalik/border-beam","owner":"Jakubantalik","name":"border-beam","desc":"Animated border beam effect for React","summaryCn":"给React元素加一圈流光边框，炫酷动画效果拉满","reasonCn":"这是一个纯前端CSS动画组件，专为React设计。通过关键帧动画实现类似极光或赛博朋克的流动边框效果。代码轻量，无外部依赖，只需几行代码即可为按钮、卡片等元素增添视觉冲击力，极大提升界面设计感，适合追求极致UI体验的前端开发者。","stars":3629,"starGrowth":14,"createdAt":"2026-04-12T17:27:34Z","silentRounds":0,"language":"TypeScript","topics":["design"],"aiDims":["可视化","非AI-好玩"],"zone":"创意","funScore":0.5,"domainTags":["React","CSS动画","UI特效"],"domainKey":"React","pushedAt":"2026-07-28T15:23:07Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/Jakubantalik/border-beam","ts":"2026-07-28T15:23:07Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.06428812344998622,"copyOk":false},{"repo":"YMFE/yapi","owner":"YMFE","name":"yapi","desc":"","summaryCn":"可视化接口管理平台，支持Mock调试和自动化测试，可内网部署","reasonCn":"YApi是可视化接口管理平台，帮助开发、产品、测试人员创建和管理API。它基于Json5和Mockjs定义接口返回数据结构，支持类似Postman的接口调试、自动化测试以及Mock期望功能。支持从postman、har、swagger导入数据，免费开源可内网部署保障信息安全。","stars":27723,"starGrowth":0,"createdAt":"2016-05-30T09:09:09Z","silentRounds":2,"language":"JavaScript","topics":["api","api-manager","api-management","api-test"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["API 管理","Mock 调试","接口测试"],"domainKey":"API 管理","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"api-manager","source":"github","weight":0.7},{"name":"api-management","source":"github","weight":0.7},{"name":"api-test","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.2,"source":"tier","url":"https://github.com/YMFE/yapi","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"testing-library/user-event","owner":"testing-library","name":"user-event","desc":"🐕 Simulate user events","summaryCn":"前端测试的神器，模拟真实用户点击输入，专治各种测不通过的组件","reasonCn":"这是 testing-library 的官方插件，专门用来模拟真实用户的交互行为，比如点击、输入、悬停等。它弥补了 Jest、pytest 等底层库在处理 UI 事件时的不足，能让你的前端组件测试更接近真实场景，是 React/Vue 开发者的必备利器。","stars":2328,"starGrowth":0,"createdAt":"2018-09-20T16:23:35Z","silentRounds":25,"language":"TypeScript","topics":["frontend"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["前端测试","TypeScript","UI 交互","单元测试","Jest"],"domainKey":"前端测试","pushedAt":"2026-09-02T01:55:10Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/testing-library/user-event","ts":"2026-09-02T01:55:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"certimate-go/certimate","owner":"certimate-go","name":"certimate","desc":"An open-source and free self-hosted SSL certificates ACME tool, automates the full-cycle of issuance, deployment, renewal, and monitoring visually. 完全开源免费的自托管 SSL 证书 ACME 工具，申请、部署、续期、监控全流程自动化可视化，支持各大主流云厂商。","summaryCn":"网站SSL证书免费申请全自动，再也不用担心到期续期的烦恼","reasonCn":"Certimate是一款完全开源免费的自托管SSL证书管理工具，基于ACME协议实现证书的自动化全生命周期管理。它支持Let's Encrypt等主流CA，通过Web可视化界面完成证书申请、验证、部署和续期，无需手动敲命令。支持Nginx、Apache、Traefik等多种服务器部署，还能对接Cloudflare等DNS服务商实现DNS-01验证，到期自动续期并平滑重载服务，彻底告别手动续证书的噩梦。","stars":9321,"starGrowth":0,"createdAt":"2024-08-21T03:54:39Z","silentRounds":2,"language":"Go","topics":["devops","self-hosted"],"aiDims":["开发者工具","安全工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["SSL证书","ACME","运维自动化"],"domainKey":"SSL证书","pushedAt":"2026-09-23T14:23:48Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/certimate-go/certimate","ts":"2026-09-23T14:23:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"datawhalechina/deepagents-in-action","owner":"datawhalechina","name":"deepagents-in-action","desc":"📚 《Deep Agents 实战》—— LangChain 官方大使出品，基于 LangChain / LangGraph 生态，从零构建生产级 AI Agent 的完整指南","summaryCn":"想象一下，你不再需要手动点鼠标处理繁琐的工作，而是通过构建一个AI Agent","reasonCn":"该项目的核心亮点在于它不仅是“怎么做”，更是“为什么这么做”。它基于LangGraph这一当下最前沿的状态管理框架，完美解决了传统链式调用无法处理循环逻辑的痛点。相比于散落在各处的碎片化教程，它构建了一套完整的生产级范式，包括任务编排、多代理协作以及鲁棒的错误处理机制。","stars":2499,"starGrowth":2,"createdAt":"2026-05-05T08:14:44Z","silentRounds":0,"language":"Astro","topics":["tutorial","course"],"aiDims":["AI Agent","教程"],"zone":"资源","funScore":0,"domainTags":["AI Agent教程","LangGraph实战","生产级开发"],"domainKey":"AI Agent教程","pushedAt":"2026-09-20T10:11:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"course","source":"github","weight":0.7},{"name":"Astro","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/datawhalechina/deepagents-in-action","ts":"2026-09-20T10:11:42Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dbeaver/dbeaver","owner":"dbeaver","name":"dbeaver","desc":"Free universal database tool and SQL client","summaryCn":"万能数据库客户端，连通所有数据库只需这一个软件","reasonCn":"DBeaver是一款功能极其强大的免费通用数据库管理工具，支持MySQL、PostgreSQL、Oracle等几十种主流数据库。它提供直观的数据编辑器、SQL开发环境和ER图生成功能，一键连接各种数据库，是开发者和DBA日常连接、管理数据的必备利器。","stars":51841,"starGrowth":10,"createdAt":"2015-10-21T08:26:28Z","silentRounds":0,"language":"Java","topics":["database","ai"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["SQL","数据库","客户端"],"domainKey":"SQL","pushedAt":"2026-09-23T20:12:35Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.1,"source":"search","url":"https://github.com/dbeaver/dbeaver","ts":"2026-09-23T20:12:35Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03057869254065315,"copyOk":true},{"repo":"realth000/tsdm_client","owner":"realth000","name":"tsdm_client","desc":"天使动漫论坛官方跨平台客户端","summaryCn":"天使动漫论坛官方Flutter客户端，支持多平台登录、阅读、评论","reasonCn":"该项目使用 Flutter 框架实现，支持 iOS、Android、桌面等多平台，提供登录、帖子浏览、评论、收藏、搜索等完整功能。通过官方 REST API，实时同步帖子内容与用户操作，保持与网页版一致。只需配置 API token，即可在本地快速体验全新论坛界面。","stars":349,"starGrowth":0,"createdAt":"2023-03-03T18:14:48Z","silentRounds":11,"language":"Dart","topics":["anime"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["论坛","动漫","跨平台","Flutter客户端"],"domainKey":"论坛","pushedAt":"2026-09-09T11:45:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.22,"source":"search","url":"https://github.com/realth000/tsdm_client","ts":"2026-09-09T11:45:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CodeWithHarry/Sigma-Web-Dev-Course","owner":"CodeWithHarry","name":"Sigma-Web-Dev-Course","desc":"Source Code for Sigma Web Development Course","summaryCn":"免费Web开发宝藏课程，从HTML到数据库都能学到","reasonCn":"这个项目是印度开发者CodeWithHarry推出的免费Web开发课程体系，完整覆盖HTML、CSS、JavaScript、后端编程和数据库等内容。课程以视频形式呈现，适合编程零基础的新手系统学习，配合源码练习能真正动手掌握网页开发技能，在GitHub上积累了上万star，说明学习人数众多、口碑不错。","stars":11934,"starGrowth":3,"createdAt":"2023-09-26T20:34:11Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["Web开发","前端教程","源码"],"domainKey":"Web开发","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"trending","url":"https://github.com/CodeWithHarry/Sigma-Web-Dev-Course","ts":"2026-08-11T09:45:50.026Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"stepfun-ai/gelab-zero","owner":"stepfun-ai","name":"gelab-zero","desc":"","summaryCn":"首个兼具模型与基础设施、支持本地隐私控制的完全开源 GUI 智能体。","reasonCn":"作为一款突破性的开源图形界面智能体解决方案，该项目实现了模型与基础设施的双料全开源。它主打零云端依赖的即插即用工程特性，为用户提供百分之百的隐私控制权。通过集成 MCP 服务器支持多设备管理与任务分发，并发布了轻量化的 4B 模型及 AndroidDaily 基准测试任务，让本地化智能操作触手可及。","stars":2265,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.62,"domainTags":["GUI智能体","完全开源","本地隐私控制","MCP服务器"],"domainKey":"GUI智能体","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/stepfun-ai/gelab-zero","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"apexcharts/apexcharts.js","owner":"apexcharts","name":"apexcharts.js","desc":"📊 Interactive JavaScript Charts built on SVG","summaryCn":"几行代码就能搞定炫酷交互式图表，数据可视化神器","reasonCn":"技术上它基于SVG渲染，这意味着图表在任意分辨率下都清晰锐利，不会出现Canvas那种缩放模糊的问题。支持35种以上的图表类型，从基础折线图、柱状图、饼图到高级的热力图、雷达图、3D散点图等。它的实时数据更新功能特别适合做监控大屏——数据来了自动刷新，动画效果丝滑。","stars":15159,"starGrowth":0,"createdAt":"2018-07-24T15:22:23Z","silentRounds":10,"language":"JavaScript","topics":["data-visualization"],"aiDims":["可视化","效率工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["SVG图表","数据可视化","交互式图表","实时监控","前端组件"],"domainKey":"SVG图表","pushedAt":"2026-09-23T12:51:52Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/apexcharts/apexcharts.js","ts":"2026-09-23T12:51:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jondot/awesome-react-native","owner":"jondot","name":"awesome-react-native","desc":"Awesome React Native components, news, tools, and learning material!","summaryCn":"全网最全React Native资源导航站，组件教程一站式找齐","reasonCn":"这是一份精心整理的React Native开源资源汇总清单，涵盖组件库、开发工具、学习资料等多个维度。收录了UI组件、状态管理、导航、测试框架等优质项目，帮助你快速定位所需技术栈，节省大量搜索时间，堪称移动开发领域的开源地图。","stars":35704,"starGrowth":0,"createdAt":"2015-03-26T19:58:06Z","silentRounds":1,"language":"","topics":["mobile"],"aiDims":["学习资源","开发者工具"],"zone":"资源","funScore":0,"domainTags":["React Native","开源清单","移动开发"],"domainKey":"React Native","pushedAt":"2026-08-26T14:11:18Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7}],"aiScore":0.25,"source":"search","url":"https://github.com/jondot/awesome-react-native","ts":"2026-08-26T14:11:18Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"RandyGaul/cute_headers","owner":"RandyGaul","name":"cute_headers","desc":"Collection of cross-platform one-file C/C++ libraries with no dependencies, primarily used for games","summaryCn":"一堆零依赖的单文件C/C++库，写小游戏原型飞快","reasonCn":"这个仓库把常用的游戏开发功能拆成几十个独立头文件，每个都只有几百到几千行代码，没有第三方依赖。涵盖2D图形渲染、输入处理、物理模拟、碰撞检测、粒子系统、音频播放等。作者Randy Gaul是知名独立游戏开发者，他把多年游戏开发经验浓缩成这套极简方案，特别适合快速原型和独立游戏开发。","stars":5072,"starGrowth":3,"createdAt":"2016-06-05T00:19:59Z","silentRounds":0,"language":"C","topics":["game"],"aiDims":["游戏","非AI-好玩","开发者工具"],"zone":"创意","funScore":1,"domainTags":["C/C++","游戏引擎","原型开发"],"domainKey":"C/C++","pushedAt":"2026-08-05T00:10:44Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/RandyGaul/cute_headers","ts":"2026-08-05T00:10:44Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nanxiaobei/hugo-paper","owner":"nanxiaobei","name":"hugo-paper","desc":"🪺 A simple, clean, customizable Hugo theme","summaryCn":"极简Hugo博客主题，代码清爽好上手，帮你快速搭建个人网站","reasonCn":"这是一个专为Hugo静态网站生成器设计的博客主题，主打极简清爽的设计风格，内置响应式布局和多种代码高亮方案。通过修改配置即可切换配色和布局，无需改代码就能搭出一个精致的个人技术博客或作品集页面，适合追求效率的开发者。","stars":2416,"starGrowth":1,"createdAt":"2018-01-08T16:26:30Z","silentRounds":25,"language":"CSS","topics":["design"],"aiDims":["非AI-实用","创意工具","可视化"],"zone":"资源","funScore":0,"domainTags":["Hugo","静态网站","博客主题"],"domainKey":"Hugo","pushedAt":"2026-08-21T02:22:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/nanxiaobei/hugo-paper","ts":"2026-08-21T02:22:09Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"LancerLab/croqtile","owner":"LancerLab","name":"croqtile","desc":"A Next-Gen AI-native Kernel programming DSL for Maximizing Productivity","summaryCn":"程序员专用的智能布局神器，让窗口自动排队不挡眼","reasonCn":"这是一个基于 C++ 开发的下一代 AI 原生内核编程 DSL 项目。它不只是简单的窗口排版工具，而是引入了领域特定语言（DSL）的概念，让开发者能用代码描述窗口布局规则，实现高度定制化的屏幕管理体验，堪称程序员生产力进阶的神器。","stars":62,"starGrowth":0,"createdAt":"2026-01-27T03:47:08Z","silentRounds":3,"language":"C++","topics":["llm-model"],"aiDims":["非AI-实用","工具调用","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["窗口管理","DSL","平铺窗口","C++"],"domainKey":"窗口管理","pushedAt":"2026-09-22T16:31:20Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/LancerLab/croqtile","ts":"2026-09-22T16:31:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"anthropics/claude-desktop-buddy","owner":"anthropics","name":"claude-desktop-buddy","desc":"","summaryCn":"这款项目利用Nordic UART Service协议，为Claude搭建了一座通往物理世界的桥","reasonCn":"项目核心在于其轻量级的协议设计，它并没有强制要求使用特定代码，而是提供了一套通用的JSON模式与传输协议。这意味着你不仅可以参考官方用ESP32制作的桌面宠物，还能基于此协议自由开发任何能显示提示的设备。无论是屏幕组件还是各类指示灯，都能通过极简的协议实现与Claude的实时","stars":2605,"starGrowth":12,"createdAt":"2026-04-09T18:08:19Z","silentRounds":0,"language":"C++","topics":[],"aiDims":["AI Agent","工具调用","非AI-好玩"],"zone":"创意","funScore":0.68,"domainTags":["桌面宠物","ESP32","硬件集成","BLE协议","Claude API"],"domainKey":"桌面宠物","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/anthropics/claude-desktop-buddy","ts":"2026-09-21T23:17:36.060Z","category":"fun","momentum":["daily"],"fromOfficial":true,"heatScore":0.059783493282149706,"copyOk":false},{"repo":"LuaJIT/LuaJIT","owner":"LuaJIT","name":"LuaJIT","desc":"Mirror of the LuaJIT git repository","summaryCn":"用C语言写的超快Lua虚拟机，游戏脚本的性能加速器","reasonCn":"LuaJIT是业界公认最快的Lua实现，采用JIT即时编译技术，能将Lua代码编译为原生机器码执行，性能比标准Lua虚拟机提升数倍。它被广泛用于游戏开发、嵌入式系统和网络服务中，是许多知名项目如OpenResty、Roblox的核心引擎。虽然与AI无直接关联，但在高性能脚本场景下是不可或缺的利器。","stars":5777,"starGrowth":1,"createdAt":"2015-08-06T13:33:38Z","silentRounds":3,"language":"C","topics":[],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["JIT","脚本引擎","游戏开发"],"domainKey":"JIT","pushedAt":"2026-08-18T13:11:43Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.25,"source":"bigbro","url":"https://github.com/LuaJIT/LuaJIT","ts":"2026-08-18T13:11:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PowerShell/PowerShell","owner":"PowerShell","name":"PowerShell","desc":"PowerShell for every system!","summaryCn":"微软开源的跨平台命令行工具，让所有系统都能高效干活","reasonCn":"PowerShell是微软开源的自动化和配置管理工具，现已支持Windows、Linux和macOS。它不仅能执行命令，还能管理对象，提供强大的脚本能力。对于系统管理员、DevOps工程师以及需要频繁操作终端的用户来说，它是提升工作流效率的必备利器。","stars":55489,"starGrowth":2,"createdAt":"2016-01-13T23:41:35Z","silentRounds":0,"language":"C#","topics":["command-line"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["命令行","自动化","跨平台"],"domainKey":"命令行","pushedAt":"2026-09-23T21:33:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.1,"source":"search","url":"https://github.com/PowerShell/PowerShell","ts":"2026-09-23T21:33:33Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"luckyframework/lucky","owner":"luckyframework","name":"lucky","desc":"A full-featured Crystal web framework that catches bugs for you, runs incredibly fast, and helps you write code that lasts.","summaryCn":"这个框架能提前帮你抓住代码里的bug，写Web应用更省心","reasonCn":"Lucky是基于Crystal语言的全功能Web框架，最大亮点是在编译阶段就能捕获常见错误，大幅减少运行时Bug。它通过代码生成技术自动生成模型、路由和测试代码，内置了类型安全的模板系统、资产管道和迁移工具，性能接近原生C语言。和Ruby on Rails相比，它用Crystal替代了Ruby，速度提升明显，类型安全让大型项目更易维护。安装需要通过Crystal包管理器，运行lucky setup命令初始化项目，然后lucky dev启动开发服务器即可，API设计和路由配置都很直观。适合追求高性能Web应用的后端开发者、喜欢类型安全的工程师，以及想尝试新兴语言替代Ruby的创业团队。项目2018年开始活跃，GitHub star约2700，Crystal社区规模较小，但框架本身维护稳定。相比Rails、Hanami等同类框架，Lucky在编译时错误检测方面独树一帜，适合对代码质量要求高的场景。","stars":2729,"starGrowth":1,"createdAt":"2017-01-06T20:58:57Z","silentRounds":25,"language":"Crystal","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Web框架","Crystal语言","类型安全"],"domainKey":"Web框架","pushedAt":"2026-08-13T14:19:00Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Crystal","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/luckyframework/lucky","ts":"2026-08-13T14:19:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/vscode","owner":"microsoft","name":"vscode","desc":"Visual Studio Code","summaryCn":"微软出品的全球最火代码编辑器，插件丰富功能强大","reasonCn":"它的核心优势在于插件生态。扩展市场有几万个插件，从主题美化到语言支持，从Git管理到容器部署，什么都有。内置功能也很强：智能代码补全、错误检测、重构工具、内置终端、调试器，开箱即用。最近几版深度集成了AI功能，Copilot能帮你写代码、解释代码、生成单元测试，写代码速度","stars":192798,"starGrowth":15,"createdAt":"2015-09-03T20:23:38Z","silentRounds":0,"language":"TypeScript","topics":["editor"],"aiDims":["开发者工具","效率工具","代码助手"],"zone":"工具","funScore":0,"domainTags":["代码编辑器","IDE","插件生态","TypeScript"],"domainKey":"代码编辑器","pushedAt":"2026-09-23T21:51:47Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/microsoft/vscode","ts":"2026-09-23T21:51:47Z","category":"tool","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.045233404910839325,"copyOk":true},{"repo":"lucidrains/audiolm-pytorch","owner":"lucidrains","name":"audiolm-pytorch","desc":"","summaryCn":"该项目是对Google Research所提出的AudioLM模型的开源实现，标志着声音生成","reasonCn":"项目仓库内还包含了基于MIT协议的SoundStream实现，并兼容EnCodec，这为音频数据的编解码提供了底层支撑。这种高度整合的设计，让科研人员和AI开发者能够在一个统一的环境中完成从音频预处理到模型训练的全部工作。不仅如此，随着MusicLM等后续工作的开展，AudioLM的方法论已经被证","stars":2624,"starGrowth":0,"createdAt":"2022-09-09T21:55:02Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.5,"domainTags":["音频生成","深度学习","TTS","音乐生成"],"domainKey":"音频生成","pushedAt":"2026-09-21T19:19:33.440Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/lucidrains/audiolm-pytorch","ts":"2026-09-21T19:19:33.440Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gongxh0901/bit-framework","owner":"gongxh0901","name":"bit-framework","desc":"bit框架","summaryCn":"TypeScript轻量开发框架，让写代码更简洁高效","reasonCn":"这是一个TypeScript开发框架，从项目名称看可能提供了一些常用的开发工具或模板。虽然信息相对有限，但对于TypeScript开发者来说，如果框架设计合理、文档清晰，可以作为日常开发的辅助工具。适合追求简洁代码结构、希望提升开发效率的TypeScript项目。","stars":88,"starGrowth":0,"createdAt":"2025-12-25T02:50:22Z","silentRounds":25,"language":"TypeScript","topics":[],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["TypeScript","Web框架","开发工具"],"domainKey":"TypeScript","pushedAt":"2026-08-18T02:21:06Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"bigbro","url":"https://github.com/gongxh0901/bit-framework","ts":"2026-08-18T02:21:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jaredhanson/passport","owner":"jaredhanson","name":"passport","desc":"Simple, unobtrusive authentication for Node.js.","summaryCn":"Node.js万能身份认证插件，一行代码搞定登录验证","reasonCn":"passport.js是Node.js生态中最流行的身份认证中间件，支持OAuth、OpenID、SAML等多种协议。通过简洁的Strategy模式，开发者可以轻松集成微信登录、GitHub授权或企业SSO，无需重复造轮子。23531star证明其可靠性，是构建用户系统的必备工具。","stars":23531,"starGrowth":1,"createdAt":"2011-10-08T22:38:32Z","silentRounds":12,"language":"JavaScript","topics":["express","nodejs","oauth","oauth2","openid","openid-connect","passport","saml"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["身份认证","OAuth","Node.js","中间件","JWT"],"domainKey":"身份认证","pushedAt":"2024-08-16T21:56:14Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"express","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"oauth","source":"github","weight":0.7},{"name":"oauth2","source":"github","weight":0.7},{"name":"openid","source":"github","weight":0.7},{"name":"openid-connect","source":"github","weight":0.7},{"name":"passport","source":"github","weight":0.7},{"name":"saml","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"bigbro","url":"https://github.com/jaredhanson/passport","ts":"2024-08-16T21:56:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cmderdev/cmder","owner":"cmderdev","name":"cmder","desc":"Lovely console emulator package for Windows","summaryCn":"Windows下的神级命令行神器，颜值实用双在线，终端控必装","reasonCn":"Cmder是Windows平台上备受推崇的便携式命令行模拟器，完美解决了原生CMD和PowerShell界面简陋、功能受限的痛点。它基于ConEmu深度定制，原生支持Tab补全、语法高亮、多标签切换及自定义主题，无需安装即可直接运行。虽然完全脱离AI范畴，但作为开发者效率基石，它能显著降低日常编码交互门槛，让你的终端操作体验直接媲美Linux/macOS，属于实用类工具中不可或缺的必备之选。","stars":27009,"starGrowth":0,"createdAt":"2013-07-09T07:44:22Z","silentRounds":1,"language":"PowerShell","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0.6,"domainTags":["Windows终端","命令行","PowerShell","ConEmu"],"domainKey":"Windows终端","pushedAt":"2026-09-21T15:53:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"PowerShell","source":"language","weight":0.5}],"aiScore":0.12,"source":"search","url":"https://github.com/cmderdev/cmder","ts":"2026-09-21T15:53:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"getlago/lago","owner":"getlago","name":"lago","desc":"Open Source Metering and Usage Based Billing API ⭐️ Consumption tracking, Subscription management, Pricing iterations, Payment orchestration &amp; Revenue analytics","summaryCn":"帮你搞定计费系统的开源神器，订阅收费全搞定","reasonCn":"这是一个完全开源的计费和用量计费 API 系统，支持消费追踪、订阅管理、定价迭代、支付编排和收入分析。它用 Go 语言编写，旨在替代昂贵的 SaaS 计费工具，提供灵活的定价模型和精准的财务报表，帮助开发者快速构建 SaaS 业务。","stars":10595,"starGrowth":1,"createdAt":"2022-02-28T09:22:45Z","silentRounds":1,"language":"Go","topics":["self-hosted"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["计费系统","订阅管理","API","开源软件"],"domainKey":"计费系统","pushedAt":"2026-09-22T13:04:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/getlago/lago","ts":"2026-09-22T13:04:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"awjuliani/DeepRL-Agents","owner":"awjuliani","name":"DeepRL-Agents","desc":"","summaryCn":"强化学习概念抽象，代码实现往往伴随复杂的数学推导，DeepRL-Agents 通过提","reasonCn":"该项目的最大特色在于其教育导向，每个 Notebook 均与作者在 Medium 上连载的教程系列一一对照。这种“代码+深度解析”的模式，解决了网上大多只给代码不给原理，或只讲原理缺乏实操的弊端。例如在 Q-Network 部分，你可以清晰看到损失函数如何在动作选择中发挥作用，以及探索与利用","stars":2279,"starGrowth":0,"createdAt":"2016-06-14T22:25:31Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["AI应用","模型与训练","教程"],"zone":"资源","funScore":0.2,"domainTags":["强化学习","Q-Learning","TensorFlow","算法实现"],"domainKey":"强化学习","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/awjuliani/DeepRL-Agents","ts":"2026-09-22T21:05:02.747Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"projectdiscovery/nuclei-templates","owner":"projectdiscovery","name":"nuclei-templates","desc":"Community curated list of templates for the nuclei engine to find security vulnerabilities.","summaryCn":"网络安全界的社区版漏洞扫描器模板库，帮你发现各种安全弱点","reasonCn":"这是Nuclei漏洞扫描引擎的社区模板库，收录了上万种针对常见组件、框架和应用的漏洞检测规则。支持REST API、数据库、中间件等多种场景，模板用YAML编写，易扩展。结合nuclei工具可自动化批量扫描，发现配置错误、未授权访问、XSS等高危问题，是安全从业者必备资产。","stars":13013,"starGrowth":2,"createdAt":"2020-04-04T16:21:34Z","silentRounds":0,"language":"JavaScript","topics":["security"],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["安全扫描","漏洞检测","YAML模板"],"domainKey":"安全扫描","pushedAt":"2026-09-23T18:29:17Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.1,"source":"trending","url":"https://github.com/projectdiscovery/nuclei-templates","ts":"2026-09-23T18:29:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"remirror/remirror","owner":"remirror","name":"remirror","desc":"ProseMirror toolkit for React 🎉","summaryCn":"React开发者的富文本编辑器神器，搭积木一样做在线文档","reasonCn":"基于ProseMirror构建的React编辑器框架，提供声明式API，通过组合独立组件即可快速搭建富文本编辑器。支持加粗、斜体、图片插入等常用功能，内置撤销重做、快捷键、光标管理等模块。插件系统完善，可扩展性强。虽然与AI无直接关联，但可作为AI写作助手或文档类应用的前端核心。","stars":3036,"starGrowth":0,"createdAt":"2019-01-21T08:59:53Z","silentRounds":25,"language":"TypeScript","topics":["mobile"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["React","富文本编辑器","TypeScript","ProseMirror","前端组件库"],"domainKey":"React","pushedAt":"2026-07-12T19:11:06Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/remirror/remirror","ts":"2026-07-12T19:11:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Pylons/pyramid","owner":"Pylons","name":"pyramid","desc":"Pyramid - A Python web framework","summaryCn":"轻量灵活的Python Web框架，按需组装不求人","reasonCn":"Pyramid的设计哲学是\"渐进式复杂性\"——小项目用很少的配置就能跑起来，大项目可以逐步引入更多组件。它使用装饰器或配置来定义URL路由，视图函数返回响应，支持多种模板引擎和数据持久化方案。与Django不同，Pyramid不强制使用特定ORM或模板系统，你可以自由选择SQLAlchemy或Peewee","stars":4101,"starGrowth":0,"createdAt":"2010-10-24T22:00:35Z","silentRounds":25,"language":"Python","topics":["web-framework"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["Python","Web框架","后端开发","WSGI","渐进式架构"],"domainKey":"Python","pushedAt":"2026-08-04T21:13:50Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/Pylons/pyramid","ts":"2026-08-04T21:13:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"JustinPerea/nebula-nodes","owner":"JustinPerea","name":"nebula-nodes","desc":"Nebula Nodes — local-first, BYOK node-based AI media canvas for image, video, audio, and 3D workflows, with Claude and Codex agents that build graphs in plain language.","summaryCn":"本地化多模态AI画布，用自然语言构建3D媒体","reasonCn":"从技术架构来看，Nebula Nodes 采用了“节点式”的工作流设计。它将AI的能力封装成一个个可拖拽的节点，比如“文本生成图像”、“音频转文字”、“3D模型生成”等。用户通过将这些节点连接起来，就能构建出复杂的逻辑链。底层的AI引擎支持多模态，意味着它能理解文本指令，并生成对应的图像","stars":19,"starGrowth":0,"createdAt":"2026-04-23T10:50:37Z","silentRounds":10,"language":"Python","topics":["creative-coding"],"aiDims":["多模态Agent","AI应用","创意工具"],"zone":"创意","funScore":1,"domainTags":["创意编程","AI媒体","3D生成","节点编辑器","本地AI"],"domainKey":"创意编程","pushedAt":"2026-09-20T14:50:36Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/JustinPerea/nebula-nodes","ts":"2026-09-20T14:50:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alvinreal/awesome-opensource-ai","owner":"alvinreal","name":"awesome-opensource-ai","desc":"Curated list of the best truly open-source AI projects, models, tools, and infrastructure. Daily updated.","summaryCn":"每日自动更新的顶级开源人工智能项目藏宝图，汇聚全网最优质的AI干货。","reasonCn":"一个高质量且保持每日更新的开源人工智能项目与模型工具精选列表，专注于真正开源的AI基础设施和应用。它系统性地梳理了当前大模型生态中的核心资源，为开发者和研究人员提供了一份省时省力的导航指南，是快速了解人工智能前沿动态的必备宝库。","stars":4770,"starGrowth":2,"createdAt":"2026-03-24T15:58:22Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["学习资源","AI基础设施","大语言模型"],"zone":"资源","funScore":0,"domainTags":["AI","开源项目","导航","资源库"],"domainKey":"AI","pushedAt":"2026-09-21T09:33:34Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/alvinreal/awesome-opensource-ai","ts":"2026-09-21T09:33:34Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"davidjerleke/embla-carousel","owner":"davidjerleke","name":"embla-carousel","desc":"A lightweight carousel library with fluid motion and great swipe precision.","summaryCn":"这是一个纯粹的UI组件库，和AI完全没有关系。","reasonCn":"从技术角度看，Embla Carousel的核心优势在于它的架构设计。它不直接操作DOM，而是通过一层抽象来管理滑动逻辑，这意味着它可以在不同框架中保持一致的行为。它使用了现代CSS和JavaScript的特性来实现60fps的流畅动画，同时支持触摸、鼠标和键盘多种输入方式。","stars":8422,"starGrowth":0,"createdAt":"2019-03-18T18:34:50Z","silentRounds":1,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["轮播组件","前端库","TypeScript","动画","Web开发"],"domainKey":"轮播组件","pushedAt":"2026-09-18T02:05:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/davidjerleke/embla-carousel","ts":"2026-09-18T02:05:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"douyu/jupiter","owner":"douyu","name":"jupiter","desc":"Jupiter: Governance-oriented Microservice Framework.","summaryCn":"Go语言界的微服务管家，搞定分布式系统治理，后端开发者的效率神器","reasonCn":"这是一个基于Go语言的微服务治理框架，专为解决分布式系统中的复杂管理难题而生。它集成了服务发现、负载均衡、熔断降级、链路追踪等核心功能，帮助开发者在构建大规模服务架构时，不再从零造轮子，而是直接获得一套成熟、稳定且高性能的基础设施支持，大幅提升开发效率与系统可靠性。","stars":4566,"starGrowth":0,"createdAt":"2020-02-27T12:50:47Z","silentRounds":10,"language":"Go","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Go","微服务","服务治理"],"domainKey":"Go","pushedAt":"2026-09-17T22:39:20Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/douyu/jupiter","ts":"2026-09-17T22:39:20Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Laravel-Backpack/CRUD","owner":"Laravel-Backpack","name":"CRUD","desc":"Build custom admin panels. Fast!","summaryCn":"Laravel框架下的后台面板快速生成器，敲几行代码搞定增删改查界面","reasonCn":"基于Laravel框架的CRUD构建工具，通过极简配置自动生成后台管理界面，支持表格、表单、图表等多种UI组件。内置权限管理、路由配置、搜索过滤等功能，大幅减少重复开发工作量，适合需要快速搭建管理后台的PHP开发者。","stars":3372,"starGrowth":0,"createdAt":"2016-03-10T12:03:46Z","silentRounds":25,"language":"PHP","topics":["backend"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Laravel","CRUD","后台管理","PHP","代码生成"],"domainKey":"Laravel","pushedAt":"2026-09-17T19:35:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/Laravel-Backpack/CRUD","ts":"2026-09-17T19:35:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"1186258278/OpenClawChineseTranslation","owner":"1186258278","name":"OpenClawChineseTranslation","desc":"🦞 OpenClaw (Clawdbot/Moltbot) 汉化版 - 开源个人 AI 助手中文版 | Claude/ChatGPT LLM 接入 | WhatsApp/Telegram/Discord 多平台 | 每小时自动同步 | CLI + Dashboard 全中文 | 全流程搭建教程，以及排错指南！","summaryCn":"把火爆的开源个人人工智能助手完全汉化，让你零门槛拥有全平台私家秘书。","reasonCn":"针对海外热门开源智能助理项目进行的深度全方位汉化版本，完美支持命令行和控制面板全中文交互。不仅打通了各大主流聊天软件，还贴心地提供了详细的搭建教程与排错指南，极大地降低了普通中文用户上手私人人工智能助手的门槛，让你轻松摆脱语言障碍。","stars":3802,"starGrowth":0,"createdAt":"2026-01-30T13:12:57Z","silentRounds":16,"language":"JavaScript","topics":["chatbot","ai-agent"],"aiDims":["AI Agent","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["AI助手","汉化","开源项目"],"domainKey":"AI助手","pushedAt":"2026-09-19T11:59:30Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/1186258278/OpenClawChineseTranslation","ts":"2026-09-19T11:59:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pallets/flask","owner":"pallets","name":"flask","desc":"The Python micro framework for building web applications.","summaryCn":"Python 最经典的轻量级 Web 框架，灵活易用经典不过时。","reasonCn":"Flask 是 Python 生态中最受欢迎、最经典的轻量级 Web 框架。它以微内核架构著称，简洁优雅，扩展性强，几乎被所有 Python 开发者学习和使用。虽然不直接涉及 AI，但它是构建 AI 模型 Web 界面（如 Gradio、Streamlit 底层）的基础设施之一，经久不衰。","stars":74768,"starGrowth":4,"createdAt":"2010-04-06T11:11:59Z","silentRounds":0,"language":"Python","topics":["web-framework"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Python","Web框架","后端开发"],"domainKey":"Python","pushedAt":"2026-09-08T16:40:53Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/pallets/flask","ts":"2026-09-08T16:40:53Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HeyPuter/puter","owner":"HeyPuter","name":"puter","desc":"🌐 The Internet Computer! Free, Open-Source, and Self-Hostable.","summaryCn":"把整个电脑装进浏览器，无需安装就能在线办公的神器","reasonCn":"Puter是一款开源的在线操作系统，完全基于Web技术实现。它允许用户在浏览器中运行应用程序、管理文件，甚至支持本地部署。该项目集成了多种常用工具，如文本编辑、代码编辑器、终端模拟器等，无需安装任何软件即可使用。","stars":43577,"starGrowth":4,"createdAt":"2024-03-03T01:44:32Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["Web 操作系统","浏览器应用","虚拟桌面","文件管理","在线办公"],"domainKey":"Web 操作系统","pushedAt":"2026-09-23T21:48:19Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/HeyPuter/puter","ts":"2026-09-23T21:48:19Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ester987222-max/AniForge-Hub","owner":"ester987222-max","name":"AniForge-Hub","desc":"Best Open Source Anime & Manga Streaming App 2026","summaryCn":"把动漫番剧和漫画整合在一起看，界面清爽又方便管理的开源神器。","reasonCn":"基于HTML构建的动漫漫画聚合流媒体应用，旨在提供最佳的2026年开源视听体验。它聚合了多维度的动漫资源，不仅支持在线流畅播放，还具备漫画阅读功能。通过前端技术优化加载速度，让用户在本地或私有云环境中享受高质量的内容管理，解决了多平台资源分散的痛点，是ACG爱好者构建个人媒体库的理想选择。","stars":118,"starGrowth":0,"createdAt":"2026-06-28T19:22:32Z","silentRounds":14,"language":"HTML","topics":["anime"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["动漫","漫画","流媒体","聚合","ACG"],"domainKey":"动漫","pushedAt":"2026-09-17T07:30:22Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/ester987222-max/AniForge-Hub","ts":"2026-09-17T07:30:22Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"saschagrunert/webapp.rs","owner":"saschagrunert","name":"webapp.rs","desc":"A web application completely written in Rust. 🌍 ","summaryCn":"全栈Rust Web应用样板，前后端都用Rust写出高性能应用","reasonCn":"技术上，它展示了Rust Web开发的完整可能性。前端用Yew框架构建，通过WebAssembly在浏览器中运行；后端用Axum和Diesel处理HTTP请求和数据库操作。这种全栈Rust的架构不仅能跑在服务器上，还能编译成WebAssembly直接嵌入网页，性能接近原生应用。","stars":2247,"starGrowth":0,"createdAt":"2018-06-09T10:14:05Z","silentRounds":25,"language":"Rust","topics":["backend","frontend"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Rust","WebAssembly","全栈开发"],"domainKey":"Rust","pushedAt":"2026-09-17T05:48:28Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/saschagrunert/webapp.rs","ts":"2026-09-17T05:48:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"emilk/egui","owner":"emilk","name":"egui","desc":"egui: an easy-to-use immediate mode GUI in Rust that runs on both web and native","summaryCn":"Rust写的即时模式GUI库，网页和本地都能跑","reasonCn":"即时模式GUI框架，单次帧循环内完成所有绘制调用，无需状态管理。内置丰富控件，支持热重载代码修改界面实时生效。可编译为WASM运行于浏览器，也能打包为原生应用。在egui官方示例里有太空大战、粒子模拟等炫酷demo，代码简洁直观。","stars":30669,"starGrowth":18,"createdAt":"2019-01-13T15:39:15Z","silentRounds":0,"language":"Rust","topics":["egui","game-development","gamedev","gui","imgui","rust","wasm"],"aiDims":["可视化","游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Rust","GUI库","即时模式","WebAssembly","跨平台"],"domainKey":"Rust","pushedAt":"2026-09-23T19:40:46Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"egui","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"imgui","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.6,"source":"bigbro","url":"https://github.com/emilk/egui","ts":"2026-09-23T19:40:46Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05576073559620463,"copyOk":true},{"repo":"songloft-org/songloft","owner":"songloft-org","name":"songloft","desc":"🎵 插件化、无广告的免费音乐服务器","summaryCn":"自己搭个没广告的音乐服务器，想听啥听啥，隐私安心","reasonCn":"基于Go语言开发的轻量级自托管音乐服务器，主打插件化架构与零广告体验。支持通过插件扩展音效增强、歌词同步等功能，提供Web界面管理本地或网络流媒体音乐库。相比传统音乐播放器，它拥有更开放的生态，允许用户深度定制播放逻辑，是音乐发烧友搭建私人流媒体基站的高性价比选择，兼顾性能与扩展性。","stars":1785,"starGrowth":0,"createdAt":"2026-02-06T07:20:42Z","silentRounds":1,"language":"Go","topics":["music"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["音乐流媒体","自托管","插件系统"],"domainKey":"音乐流媒体","pushedAt":"2026-09-23T08:50:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/songloft-org/songloft","ts":"2026-09-23T08:50:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bikini/exploitarium","owner":"bikini","name":"exploitarium","desc":"A single archive of public exploit PoCs and vulnerability research writeups. At the time I post these, none have been reported. Feel free to report them yourself and take credit for the CVE if handed out lulz. Please do not abuse these. I do this so to allure people into the field, and I've always found this is the most efficient way.","summaryCn":"这个项目是网络安全研究者的宝藏——它收集了大量漏洞利用的概念验证（PoC）代","reasonCn":"技术上，项目按漏洞类型组织：Web应用漏洞、缓冲区溢出、提权漏洞、零日漏洞等。每个PoC都附带详细说明：受影响版本、利用前提条件、执行步骤、防御建议。代码以Python、Bash、C为主，力求简洁可复现。相比Metasploit的庞大框架，exploitarium更侧重教育性和研究性，而非自动化攻击","stars":5132,"starGrowth":2,"createdAt":"2026-06-23T05:14:11Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["非AI-实用","安全工具"],"zone":"资源","funScore":0,"domainTags":["漏洞利用","网络安全研究","渗透测试","PoC","CTF"],"domainKey":"漏洞利用","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.2,"source":"trending","url":"https://github.com/bikini/exploitarium","ts":"2026-09-07T22:38:09.240Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"RamonUnch/AltSnap","owner":"RamonUnch","name":"AltSnap","desc":"Maintained continuation of Stefan Sundin's AltDrag","summaryCn":"窗口拖拽自由如橡皮泥，多屏办公党的救星，从此告别束缚人的固定布局","reasonCn":"AltSnap是经典AltDrag项目的精神续作，完全用C语言重写。它支持任意窗口拖拽、固定比例缩放、智能对齐边对边，甚至能在拖拽时实时预览位置。对于同时开着十几个窗口、还要用多屏协作的用户来说，这是提升桌面生产力的必备神器。","stars":5235,"starGrowth":0,"createdAt":"2020-11-23T22:13:36Z","silentRounds":2,"language":"C","topics":["desktop"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["窗口管理","Windows","桌面效率","多屏","快捷键"],"domainKey":"窗口管理","pushedAt":"2026-09-16T19:13:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/RamonUnch/AltSnap","ts":"2026-09-16T19:13:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Nerwyn/universal-remote-card","owner":"Nerwyn","name":"universal-remote-card","desc":"A completely customizable universal remote card for Home Assistant. Supports multiple platforms out of the box.","summaryCn":"Home Assistant 万能遥控卡，支持多平台自定义","reasonCn":"通过 TypeScript 编写，`universal-remote-card` 让 Home Assistant 用户一键生成支持灯光、音响、空调等多种设备的遥控卡。预置多平台遥控模板，只需在 YAML 配置中修改按钮映射即可快速部署，支持自定义颜色、图标和动作，极大提升了智能家居操作的便捷度。","stars":609,"starGrowth":1,"createdAt":"2023-06-16T03:14:31Z","silentRounds":2,"language":"TypeScript","topics":["jellyfin"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Home Assistant","智能家居","遥控卡","TypeScript","自定义组件"],"domainKey":"Home Assistant","pushedAt":"2026-09-20T02:12:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.12,"source":"search","url":"https://github.com/Nerwyn/universal-remote-card","ts":"2026-09-20T02:12:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bigcode-project/starcoder2","owner":"bigcode-project","name":"starcoder2","desc":"","summaryCn":"StarCoder2在训练数据上下了功夫——使用了The Stack v2数据集，覆盖超过600种","reasonCn":"在技术架构上，StarCoder2采用了分组查询注意力（GQA），配合16384 token的上下文窗口和4096的滑动窗口注意力。这意味着它能连贯地处理中等规模代码文件，在生成长代码序列时也更高效。对于需要理解跨文件依赖或者长函数体的场景，这样的设计显然经过深思熟虑。","stars":2094,"starGrowth":2,"createdAt":"2023-12-08T08:46:25Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","代码助手","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["代码生成","大语言模型","开源模型","多语言编程"],"domainKey":"大语言模型","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/bigcode-project/starcoder2","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"milvus-io/pymilvus","owner":"milvus-io","name":"pymilvus","desc":"Python SDK for Milvus Vector Database","summaryCn":"用Python轻松操控海量向量数据库，让AI拥有超强记忆力","reasonCn":"Milvus本身是一个开源的向量数据库，专为AI场景设计，支持多种索引算法（IVF、HNSW、SCANN等）和相似度计算方法（余弦、欧氏、内积等）。pymilvus作为官方Python SDK，把这些复杂操作封装成简洁的API——创建集合、插入向量、构建索引、执行搜索，几行代码就能完成。","stars":1409,"starGrowth":0,"createdAt":"2019-06-13T11:38:34Z","silentRounds":26,"language":"Python","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["向量数据库","向量检索","Python SDK","AI数据库","相似度搜索"],"domainKey":"向量数据库","pushedAt":"2026-09-23T03:21:37Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/milvus-io/pymilvus","ts":"2026-09-23T03:21:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVlabs/parrot","owner":"NVlabs","name":"parrot","desc":"Parrot is an array fusion GPU library built on NVIDIA's CCCL libaries (Thrust/CUB).","summaryCn":"NVlabs 的 Parrot 是一个专为 NVIDIA 设备设计的 GPU 数组融合库，目标是把","reasonCn":"NVlabs 的 Parrot 是一个专为 NVIDIA 设备设计的 GPU 数组融合库，目标是把多维数组操作的性能提升到极致。无论是矩阵乘、卷积还是自定义运算，Parrot 都能将它们映射成高效的 Thrust/CUB 调用，显著减少内存跳转。\\n\\n它采用了 LLVM 的多态编译技术，生成针对不同 GPU 架构的 PTX","stars":288,"starGrowth":0,"createdAt":"2025-10-07T18:43:59Z","silentRounds":16,"language":"Cuda","topics":["algorithms"],"aiDims":["开发者工具","效率工具"],"zone":"AI","funScore":0,"domainTags":["CUDA","GPU","高性能计算"],"domainKey":"CUDA","pushedAt":"2026-09-03T14:42:15Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"algorithms","source":"github","weight":0.7},{"name":"Cuda","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/NVlabs/parrot","ts":"2026-09-03T14:42:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ant-design/ant-design-mobile","owner":"ant-design","name":"ant-design-mobile","desc":"Essential UI blocks for building mobile web apps.","summaryCn":"阿里出品移动端组件库，React构建高质量移动Web应用必备","reasonCn":"架构上采用组件化设计，每个组件都是独立的UI模块，按需引入减少打包体积。核心特色包括：主题定制系统，支持颜色、字体、间距等设计Token的配置，轻松实现品牌化定制；完善的TypeScript支持，类型定义完整，IDE体验优秀；无障碍访问支持，让残障用户也能友好使用。","stars":12052,"starGrowth":1,"createdAt":"2015-11-30T03:33:06Z","silentRounds":21,"language":"TypeScript","topics":["mobile"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["React组件库","移动端UI","TypeScript"],"domainKey":"React组件库","pushedAt":"2026-09-14T09:06:19Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/ant-design/ant-design-mobile","ts":"2026-09-14T09:06:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"moderndive/ModernDive_book","owner":"moderndive","name":"ModernDive_book","desc":"Statistical Inference via Data Science: A ModernDive into R and the Tidyverse","summaryCn":"用 R 语言学统计学的开源教材，配交互式可视化，边学边练效果极佳","reasonCn":"这是一套基于 R 语言和 Tidyverse 生态的统计学现代教材，采用数据科学视角讲解统计推断。书中大量使用交互式可视化和真实数据集，将抽象的统计概念转化为可操作的代码示例。配套提供 R Markdown 文件和完整代码，读者可以直接在 RStudio 中运行和修改。内容覆盖回归分析、假设检验、贝叶斯方法等核心主题。","stars":805,"starGrowth":0,"createdAt":"2016-08-29T06:43:02Z","silentRounds":15,"language":"HTML","topics":["data-visualization"],"aiDims":["学习资源","可视化"],"zone":"资源","funScore":0,"domainTags":["R语言","统计学","数据科学"],"domainKey":"R语言","pushedAt":"2026-08-26T23:56:25Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/moderndive/ModernDive_book","ts":"2026-08-26T23:56:25Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"czy0729/Bangumi","owner":"czy0729","name":"Bangumi","desc":":electron: An unofficial https://bgm.tv ui first app client for Android and iOS, built with React Native. 一个无广告、以爱好为驱动、不以盈利为目的、专门做 ACG 的类似豆瓣的追番记录，bgm.tv 第三方客户端。为移动端重新设计，内置大量加强的网页端难以实现的功能，且提供了相当的自定义选项。 目前已适配 iOS / Android。","summaryCn":"无广告追番神器，Bangumi 第三方客户端让你舒服看番","reasonCn":"基于 React Native 开发的 Bangumi.tv 第三方客户端，无广告、纯爱好驱动。内置番剧记录、追更提醒、评分评论等完整功能，针对移动端做了全新设计，支持丰富的自定义主题和布局选项，ACG 爱好者体验 Bangumi 的最佳移动端选择。","stars":5949,"starGrowth":1,"createdAt":"2019-05-08T15:23:07Z","silentRounds":1,"language":"TypeScript","topics":["design"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":0,"domainTags":["番剧追踪","ACG社区","移动应用"],"domainKey":"番剧追踪","pushedAt":"2026-09-22T08:44:05Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/czy0729/Bangumi","ts":"2026-09-22T08:44:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"gmh5225/awesome-game-security","owner":"gmh5225","name":"awesome-game-security","desc":"awesome game security [Welcome to PR]","summaryCn":"游戏安全领域最全资源合集，3000+星认证的知识宝库","reasonCn":"内容方面，awesome列表通常会按主题分类整理优质资源。这个项目从topic来看，应该涵盖了游戏反作弊技术、游戏客户端加密、数据包分析、内存保护、作弊检测算法、逆向工程工具等多个方向。3000多个star说明内容比较全面且质量有保障，是社区共同认可的知识集合。","stars":3534,"starGrowth":3,"createdAt":"2022-04-04T06:07:57Z","silentRounds":0,"language":"Python","topics":["game"],"aiDims":["安全工具","游戏","学习资源"],"zone":"资源","funScore":0.5,"domainTags":["游戏安全","反作弊","逆向工程"],"domainKey":"游戏安全","pushedAt":"2026-09-23T21:47:46Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/gmh5225/awesome-game-security","ts":"2026-09-23T21:47:46Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vi/websocat","owner":"vi","name":"websocat","desc":"Command-line client for WebSockets, like netcat (or curl) for ws:// with advanced socat-like functions","summaryCn":"像netcat一样的WebSocket命令行工具，调试网络接口的瑞士军刀","reasonCn":"技术上，它的核心架构借鉴了经典的netcat和socat工具，但专门针对WebSocket协议做了优化。它支持ws://和wss://协议，可以自动处理认证、压缩扩展、自定义头部等。它还支持将WebSocket转换为TCP连接、HTTP代理、甚至管道到子进程，功能非常灵活。","stars":8695,"starGrowth":0,"createdAt":"2016-11-23T13:36:22Z","silentRounds":3,"language":"Rust","topics":["command-line"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["WebSocket","命令行","网络调试"],"domainKey":"WebSocket","pushedAt":"2026-08-13T10:17:56Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/vi/websocat","ts":"2026-08-13T10:17:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dariubs/GoBooks","owner":"dariubs","name":"GoBooks","desc":"List of Golang books","summaryCn":"收集了大量Golang编程书籍的学习资源库","reasonCn":"这是一个专门整理Go语言编程书籍的集合仓库，收录了从入门到进阶的各种经典教材和实战指南。涵盖了Go并发编程、系统设计、微服务架构等多个领域，帮助开发者系统学习Go语言。虽然不是AI相关项目，但对提升编程技能很有帮助。","stars":19682,"starGrowth":1,"createdAt":"2015-05-05T10:45:36Z","silentRounds":2,"language":"Go","topics":["devops"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["Go语言","编程学习","书单"],"domainKey":"Go语言","pushedAt":"2026-07-13T16:08:02Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.2,"source":"search","url":"https://github.com/dariubs/GoBooks","ts":"2026-07-13T16:08:02Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"frontendbr/forum","owner":"frontendbr","name":"forum","desc":":beer: Portando discussões feitas em grupos (Facebook, Google Groups, Slack, Disqus) para o GitHub Discussions","summaryCn":"巴西前端社区精华讨论搬运到GitHub方便查阅","reasonCn":"项目本质上是一个社区运营工具，把原本散落在多个平台的讨论集中到GitHub Discussions。好处是讨论内容 searchable、可star、可fork，并且利用GitHub的协作生态。内容覆盖React、Vue、Node.js、CSS等各种前端话题。 使用方法：直接在GitHub页面浏览和参与讨论即可，不需要额外安装","stars":4235,"starGrowth":0,"createdAt":"2016-02-01T12:15:40Z","silentRounds":25,"language":"","topics":["frontend"],"aiDims":["学习资源","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["前端","社区","论坛"],"domainKey":"前端","pushedAt":"2026-07-20T12:25:08Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7}],"aiScore":0.2,"source":"search","url":"https://github.com/frontendbr/forum","ts":"2026-07-20T12:25:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"feigeCode/navop","owner":"feigeCode","name":"navop","desc":"A native, all-in-one workspace for databases, SSH, SFTP, terminals, remote desktop, monitoring, and AI.","summaryCn":"一站式本地终端+数据库+AI助手，省去切换工具的麻烦。","reasonCn":"Navop 采用 Rust 编写，提供原生 UI，整合 SSH、SFTP、终端、数据库管理、远程桌面、系统监控和 AI 辅助功能，支持多协议一键连接，界面流畅且资源占用低。跨平台支持 Windows、macOS、Linux，插件系统可自行扩展，内置 AI 大模型实现自然语言查询。","stars":1469,"starGrowth":0,"createdAt":"2026-07-10T06:01:54Z","silentRounds":1,"language":"Rust","topics":["ai-agents","clickhouse","dameng","duckdb","iotdb","kingbase","markdown","mermaid","mssql","mysql","notes","oracle","port-forwarding","serial","sftp","sqlite","ssh","terminal","openguass"],"aiDims":["AI应用","开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["数据库管理","SSH","开发工具"],"domainKey":"数据库管理","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"clickhouse","source":"github","weight":0.7},{"name":"dameng","source":"github","weight":0.7},{"name":"duckdb","source":"github","weight":0.7},{"name":"iotdb","source":"github","weight":0.7},{"name":"kingbase","source":"github","weight":0.7},{"name":"markdown","source":"github","weight":0.7},{"name":"mermaid","source":"github","weight":0.7},{"name":"mssql","source":"github","weight":0.7},{"name":"mysql","source":"github","weight":0.7},{"name":"notes","source":"github","weight":0.7},{"name":"oracle","source":"github","weight":0.7},{"name":"port-forwarding","source":"github","weight":0.7},{"name":"serial","source":"github","weight":0.7},{"name":"sftp","source":"github","weight":0.7},{"name":"sqlite","source":"github","weight":0.7},{"name":"ssh","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"openguass","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/feigeCode/navop","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PrefectHQ/prefect","owner":"PrefectHQ","name":"prefect","desc":"Prefect is a workflow orchestration framework for building resilient data pipelines in Python.","summaryCn":"让数据管道像搭积木一样简单，失败自动重试超省心","reasonCn":"这个Python工作流编排工具帮开发者轻松管理数据流水线。支持自动重试、依赖追踪、定时调度等，配合可观测性面板能实时查看每个节点状态。特别适合机器学习模型训练、ETL任务等场景，和Airflow相比更现代化，API设计优雅，学习曲线平缓。","stars":23899,"starGrowth":10,"createdAt":"2018-06-29T21:59:26Z","silentRounds":0,"language":"Python","topics":["automation","data-science"],"aiDims":["工具调用","AI基础设施","效率工具"],"zone":"工具","funScore":0,"domainTags":["工作流编排","数据管道","机器学习","ETL"],"domainKey":"工作流编排","pushedAt":"2026-09-23T14:06:20Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/PrefectHQ/prefect","ts":"2026-09-23T14:06:20Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03125528264780953,"copyOk":false},{"repo":"commitizen/cz-cli","owner":"commitizen","name":"cz-cli","desc":"The commitizen command line utility. #BlackLivesMatter","summaryCn":"智能Git提交工具，告别乱写commit message的噩梦","reasonCn":"这是一个命令行工具，用于规范Git提交信息。它通过交互式向导引导你按照约定式提交（Conventional Commits）格式填写commit message。虽然不直接涉及AI，但它能极大提升团队协作效率，减少代码审查中的沟通成本，是每个开发者都值得拥有的实用效率神器。","stars":17500,"starGrowth":0,"createdAt":"2015-04-20T01:11:39Z","silentRounds":5,"language":"JavaScript","topics":["command-line","git"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Git","命令行","约定式提交","提交规范","JavaScript"],"domainKey":"Git","pushedAt":"2026-09-04T00:09:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.18,"source":"search","url":"https://github.com/commitizen/cz-cli","ts":"2026-09-04T00:09:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"arthurspk/guiadobackend","owner":"arthurspk","name":"guiadobackend","desc":"Guia de Back-end: trilhas, cursos, livros, canais, ferramentas e comunidades para você entrar e evoluir na área.","summaryCn":"程序员必备的后端学习宝库，从入门到精通全覆盖","reasonCn":"这个项目把后端开发的各类学习资源一站式整理好了，包括学习路线、免费课程、经典书籍、优质频道和实用工具，全用葡萄牙语编写。它采用目录结构按技术领域分类，每类资源标注难度和推荐等级，方便不同阶段的开发者快速定位内容，是自学后端的省心选择。","stars":2442,"starGrowth":0,"createdAt":"2022-04-02T17:04:44Z","silentRounds":22,"language":"","topics":["backend"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0,"domainTags":["后端开发","学习路线","资源索引"],"domainKey":"后端开发","pushedAt":"2026-09-02T13:46:23Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7}],"aiScore":0.18,"source":"search","url":"https://github.com/arthurspk/guiadobackend","ts":"2026-09-02T13:46:23Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dofy/learn-vim","owner":"dofy","name":"learn-vim","desc":"Learning Vim. A Hands-On Tutorial of Vim.","summaryCn":"Vim学习手把手教程，零基础也能玩转编辑器","reasonCn":"核心亮点：\n· TypeScript编写的教学脚本，易于阅读和维护；\n· 交互式练习，实时反馈错误和建议；\n· 章节涵盖常用命令、插件管理、键位自定义、Vimscript基础，甚至是高级技巧如宏录制和多光标操作；\n· 代码仓库与学习脚本分离，便于快速更新与扩展。","stars":1730,"starGrowth":1,"createdAt":"2016-01-13T08:14:10Z","silentRounds":12,"language":"TypeScript","topics":["editor","tutorial","learning"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["Vim","编辑器教程","交互式学习"],"domainKey":"Vim","pushedAt":"2026-09-10T06:23:46Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"tutorial","source":"github","weight":0.7},{"name":"learning","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/dofy/learn-vim","ts":"2026-09-10T06:23:46Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"trailofbits/claude-code-config","owner":"trailofbits","name":"claude-code-config","desc":"","summaryCn":"安全专家为 AI 编程助手定制的配置集，强化权限控制与审计工作流。","reasonCn":"这是知名安全公司 Trail of Bits 公开的一套 Claude Code 配置模板。它不只包含基础设置，更深层面是规定了沙箱隔离、权限管理和特定工作流。旨在让 AI 编程助手在安全审计、开发和研究场景下，以受控且高效的方式运行，避免潜在的安全风险。","stars":2108,"starGrowth":0,"silentRounds":0,"language":"Shell","topics":[],"aiDims":["AI Agent","安全工具"],"zone":"AI","funScore":0.26,"domainTags":["安全审计","AI 编程助手","沙箱隔离"],"domainKey":"安全审计","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/trailofbits/claude-code-config","ts":"2026-09-23T18:23:13.861Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MewPurPur/GodSVG","owner":"MewPurPur","name":"GodSVG","desc":"A vector graphics application for structured SVG editing, available on all major desktop platforms and on web. Currently in late alpha.","summaryCn":"矢量图形编辑工具，适用于所有主流桌面平台","reasonCn":"GodSVG是一个面向桌面用户的矢量图形编辑工具，它提供了一个直观的用户界面，使用户能够轻松地创建和编辑SVG文件。它支持多种矢量图形格式，如SVG、EPS和CGM，并且可以与其他矢量图形软件（如Inkscape）无缝协作。此外，它还提供了一些高级功能，如图层管理、路径编辑和节点编辑，","stars":2751,"starGrowth":0,"createdAt":"2023-06-17T09:24:11Z","silentRounds":1,"language":"GDScript","topics":["godot"],"aiDims":["向量图形","结构化SVG"],"zone":"工具","funScore":0,"domainTags":["矢量图形","SVG编辑","跨平台","桌面应用","图形软件"],"domainKey":"矢量图形","pushedAt":"2026-09-23T10:26:14Z","tags":[{"name":"向量图形","source":"llm","weight":1},{"name":"结构化SVG","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/MewPurPur/GodSVG","ts":"2026-09-23T10:26:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"polarsquad/krops","owner":"polarsquad","name":"krops","desc":"Kubernetes-Native Resource Operations","summaryCn":"在K8s里直接用Rust写的资源批量管理工具","reasonCn":"Kubernetes 集群里经常要对大量自定义资源进行统一增删改查，传统的 kubectl 脚本写起来繁琐且缺乏类型安全。krops 诞生的初衷就是把这些重复的运维任务交给 Rust 编写的原生工具来完成，让管理员可以像写代码一样声明批量操作，省去手动编辑大量 YAML 的痛苦。","stars":147,"starGrowth":0,"createdAt":"2026-04-09T18:24:55Z","silentRounds":2,"language":"Rust","topics":[],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","Rust","DevOps"],"domainKey":"Kubernetes","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/polarsquad/krops","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fmtlib/fmt","owner":"fmtlib","name":"fmt","desc":"A modern formatting library","summaryCn":"C++格式化标准库，比printf更安全更优雅","reasonCn":"现代C++格式化库，提供类型安全的格式化函数fmt::format，语法类似Python的str.format。支持自定义类型格式化、编译期字符串检查、高性能输出。已被C++20标准采纳为std::format的基础，是C++开发者的必备工具库。适合所有使用C++的项目，尤其适合需要调试输出和日志记录的场景。","stars":25801,"starGrowth":1,"createdAt":"2012-12-07T16:26:46Z","silentRounds":1,"language":"C++","topics":["formatting","printf","output","chrono","performance","multiplatform","cross-platform","cpp","ranges","floating-point","unicode","c-plus-plus"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["C++","字符串格式化","编译期检查"],"domainKey":"C++","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"formatting","source":"github","weight":0.7},{"name":"printf","source":"github","weight":0.7},{"name":"output","source":"github","weight":0.7},{"name":"chrono","source":"github","weight":0.7},{"name":"performance","source":"github","weight":0.7},{"name":"multiplatform","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"ranges","source":"github","weight":0.7},{"name":"floating-point","source":"github","weight":0.7},{"name":"unicode","source":"github","weight":0.7},{"name":"c-plus-plus","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/fmtlib/fmt","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"backend-pt/vagas","owner":"backend-pt","name":"vagas","desc":"🔬 Espaço para divulgação de vagas para desenvolvedores back-end em Portugal","summaryCn":"专门给在葡萄牙找后端开发工作的人打造的职位聚合平台","reasonCn":"这个项目是葡萄牙开发者社区精心打造的职位信息聚合平台，专门为寻找后端开发岗位的小伙伴提供服务。它采用完全开源的协作模式，通过简单的Markdown文件记录招聘信息，任何人都能通过提交PR的方式分享职位机会。虽然界面朴素，但信息收集功能非常实用，为在葡萄牙发展的技术人搭建了一个可靠的求职桥梁。","stars":567,"starGrowth":0,"createdAt":"2019-08-30T22:56:11Z","silentRounds":10,"language":"","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"资源","funScore":0,"domainTags":["后端开发","葡萄牙","求职平台","Markdown","开源社区"],"domainKey":"后端开发","pushedAt":"2026-08-31T20:00:44Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7}],"aiScore":0.15,"source":"search","url":"https://github.com/backend-pt/vagas","ts":"2026-08-31T20:00:44Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"evancohen/smart-mirror","owner":"evancohen","name":"smart-mirror","desc":"The fairest of them all. A DIY voice controlled smart mirror with IoT integration.","summaryCn":"动手做一面会说话的智能镜子，语音控制还能连IoT设备。","reasonCn":"Smart Mirror 项目提供一个开源的 DIY 智能镜子方案，基于 Raspberry Pi 与 Node.js，实现语音识别、天气、新闻等信息的实时展示，并通过 MQTT 或 HTTP 与其他 IoT 设备互联，可自定义插件扩展功能，兼容多种语音服务，让普通开发者也能轻松搭建会说话的家庭助手。","stars":2822,"starGrowth":0,"createdAt":"2015-09-12T02:41:53Z","silentRounds":11,"language":"JavaScript","topics":["raspberry-pi","home-automation"],"aiDims":["AI Agent","非AI-好玩"],"zone":"创意","funScore":0,"domainTags":["智能镜子","Raspberry Pi","语音控制"],"domainKey":"智能镜子","pushedAt":"2026-09-22T17:13:38Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"home-automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/evancohen/smart-mirror","ts":"2026-09-22T17:13:38Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"crate/crate","owner":"crate","name":"crate","desc":"CrateDB is a distributed and scalable SQL database for storing and analyzing massive amounts of data in near real-time, even with complex queries. It is PostgreSQL-compatible, and based on Lucene.","summaryCn":"一个能用SQL查询向量数据的分布式数据库，AI应用的数据库好选择","reasonCn":"从技术角度看，CrateDB基于Apache Lucene和Elasticsearch构建，继承了它们的分布式搜索能力，同时通过CrateDB协议层实现了PostgreSQL兼容性。它的向量搜索功能基于HNSW算法，支持余弦距离、欧氏距离等多种相似度度量。","stars":4436,"starGrowth":0,"createdAt":"2013-04-10T09:17:16Z","silentRounds":24,"language":"Java","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","向量搜索","PostgreSQL兼容","实时分析"],"domainKey":"分布式数据库","pushedAt":"2026-09-23T20:52:02Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/crate/crate","ts":"2026-09-23T20:52:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/typescript-go","owner":"microsoft","name":"typescript-go","desc":"Staging repo for development of native port of TypeScript","summaryCn":"把TypeScript编译成Go语言的神器，开发效率直接起飞","reasonCn":"技术上，它通过将TS类型系统、AST（抽象语法树）和编译流程深度集成到Go的编译器中，实现了一站式的代码转换。它保留了TS的所有特性，包括接口、泛型、装饰器等，并能自动生成符合Go规范的代码结构。相比传统的“TS转JS再转Go”的方案，它直接在TS源头进行转换，避免了中间层的性能损","stars":26175,"starGrowth":0,"createdAt":"2024-09-19T20:25:12Z","silentRounds":25,"language":"Go","topics":[],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["TypeScript","编译器","Go语言","微软","开发工具"],"domainKey":"TypeScript","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/microsoft/typescript-go","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"robertohuertasm/SQLite4Unity3d","owner":"robertohuertasm","name":"SQLite4Unity3d","desc":"SQLite made easy for Unity3d","summaryCn":"让Unity3D游戏轻松读写SQLite数据库的神器","reasonCn":"这个库把SQLite核心编译成Unity3D支持的DLL格式，解决了跨平台兼容性痛点。开发者无需折腾原生编译，在Unity编辑器里直接引用就能读写数据库。无论是存档系统、玩家数据、配置缓存还是本地搜索功能，都能用熟悉的SQL语句实现，比JSON文件更结构化、查询更高效，同时比引入完整数据库引擎轻量很多，特别适合中小规模数据管理场景。","stars":1410,"starGrowth":0,"createdAt":"2014-10-10T08:11:07Z","silentRounds":25,"language":"C#","topics":[],"aiDims":["游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Unity3D","SQLite","游戏开发","跨平台","C#"],"domainKey":"Unity3D","pushedAt":"2019-05-25T09:01:13Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.15,"source":"bigbro","url":"https://github.com/robertohuertasm/SQLite4Unity3d","ts":"2019-05-25T09:01:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"esengine/lawn-mower-demo","owner":"esengine","name":"lawn-mower-demo","desc":"","summaryCn":"一个草坪割草机的仿真演示，看看机器人怎么智能割草","reasonCn":"这个项目用TypeScript开发了一个草坪割草机器人的仿真演示环境。虽然仓库没有任何描述和topic标签，但从名称可以推测它模拟了自动割草机器人的运行逻辑。项目仅有23个star，属于比较冷门的新项目。对于想了解机器人路径规划或仿真控制逻辑的开发者来说，这是一个可以借鉴的小项目，但目前信息较少，实际价值有待进一步观察。","stars":23,"starGrowth":0,"createdAt":"2025-07-02T16:31:54Z","silentRounds":25,"language":"TypeScript","topics":[],"aiDims":["非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["机器人仿真","路径规划","TypeScript","可视化","教育演示"],"domainKey":"机器人仿真","pushedAt":"2025-12-31T13:49:05Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"bigbro","url":"https://github.com/esengine/lawn-mower-demo","ts":"2025-12-31T13:49:05Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"carlvellotti/free-ai-courses","owner":"carlvellotti","name":"free-ai-courses","desc":"","summaryCn":"把AI课程放进Claude里，让AI亲自教你用AI的交互式课堂。","reasonCn":"这是一套完全免费的交互式AI课程，奇特之处在于课程不依赖网站视频，而是直接运行在Claude、Cursor等AI应用内部。你只需把一段初始提示词粘贴进去，AI就会自动安装课程内容并开始一对一教学。课程涵盖AI写作、Cursor使用、产品经理提效等主题，边做边学，学完就能用。","stars":2183,"starGrowth":7,"createdAt":"2025-10-17T21:41:16Z","silentRounds":0,"language":"HTML","topics":[],"aiDims":["教程","学习资源"],"zone":"资源","funScore":0.5,"domainTags":["AI课程","Claude","Cursor","产品经理","教学","交互式学习"],"domainKey":"AI课程","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/carlvellotti/free-ai-courses","ts":"2026-09-22T22:51:50.275Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.03674374713696748,"copyOk":true},{"repo":"espanso/espanso","owner":"espanso","name":"espanso","desc":"A Privacy-first, Cross-platform Text Expander written in Rust","summaryCn":"如果你每天都要反复输入邮箱、地址、模板化的回复或者大段重复代码，那你的","reasonCn":"技术上，它采用了高性能的Rust编写，后台占用几乎可以忽略不计。最强大之处在于它的可编程性：它不仅仅是简单的文字替换，还支持YAML配置文件，甚至可以挂载本地的Python或Shell脚本。这意味着当你输入一个指令时，它可以自动运行脚本抓取实时数据并填入，这种灵活性是其他静态文本","stars":14524,"starGrowth":1,"createdAt":"2019-08-30T07:16:58Z","silentRounds":1,"language":"Rust","topics":["productivity"],"aiDims":["效率工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文本替换","生产力","Rust","快捷键"],"domainKey":"文本替换","pushedAt":"2026-09-20T00:25:10Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/espanso/espanso","ts":"2026-09-20T00:25:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lodash/lodash","owner":"lodash","name":"lodash","desc":"A modern JavaScript utility library delivering modularity, performance, &amp; extras.","summaryCn":"前端代码里最好用的工具箱，告别手写重复代码","reasonCn":"作为JavaScript生态中最受欢迎的工具库之一，它提供了60多个经过高度优化的函数，覆盖数组、对象、函数等常见操作。通过模块化设计，开发者可以按需引入特定功能，避免引入整个库带来的体积负担。项目维护超过10年，拥有超过6万个star，是前端开发者的必备依赖。","stars":61270,"starGrowth":0,"createdAt":"2012-04-07T04:11:46Z","silentRounds":25,"language":"JavaScript","topics":["lodash","utilities","javascript","modules"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["JavaScript","工具库","函数式编程","前端开发","npm包"],"domainKey":"JavaScript","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"lodash","source":"github","weight":0.7},{"name":"utilities","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"modules","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/lodash/lodash","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ydcss/vue-ydui","owner":"ydcss","name":"vue-ydui","desc":"A mobile components Library with Vue2.js. 一只基于Vue2.x的移动端组件库。","summaryCn":"复古又好用的老牌Vue2移动端组件库，手速党必藏","reasonCn":"这是一个专攻Vue2的移动端UI组件库，拥有两千多星的超高人气。内置丰富的移动端交互组件，样式设计复古实用，文档清晰。虽然技术栈停留在Vue2，但在老项目维护或需要快速搭建移动端页面时，它依然是非常可靠且好用的实用工具。","stars":2725,"starGrowth":0,"createdAt":"2017-01-18T07:22:15Z","silentRounds":25,"language":"Vue","topics":["mobile"],"aiDims":["非AI-好玩","游戏","可视化"],"zone":"工具","funScore":0,"domainTags":["Vue2","移动端UI","组件库","前端开发"],"domainKey":"Vue2","pushedAt":"2026-07-27T13:15:15Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/ydcss/vue-ydui","ts":"2026-07-27T13:15:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"baidu/san","owner":"baidu","name":"san","desc":"A fast, portable, flexible JavaScript component framework","summaryCn":"百度出品的轻量前端组件框架，写起来优雅跑起来飞快，前端工程师的心头好","reasonCn":"技术层面，san 采用了自己的差异计算引擎，不是完全照搬虚拟 DOM 方案，而是在 diff 算法上做了优化。它支持组件化开发，组件之间状态管理清晰，支持数据驱动的视图更新。相比 Vue 和 React，san 的体积更小，API 更精简，学习成本更低，适合中小型项目或需要快速上手的场景。","stars":4735,"starGrowth":0,"createdAt":"2016-08-29T09:27:23Z","silentRounds":11,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["前端框架","JavaScript","组件化","轻量级"],"domainKey":"前端框架","pushedAt":"2026-07-15T11:59:18Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/baidu/san","ts":"2026-07-15T11:59:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lucidrains/lion-pytorch","owner":"lucidrains","name":"lion-pytorch","desc":"","summaryCn":"Lion 优化器是 Google Brain 最新研发的 EvoLved Sign Momentum（EvoLved S","reasonCn":"项目文档特别强调了与 AdamW 的超参数差异：Lion 的学习率需要比 AdamW 小 3‑10 倍，而对应的 decoupled weight decay λ 则要大 3‑10 倍，以保持相同的正则化强度。这一调参经验来源于作者在真实训练任务中的实验观察，帮助使用者快速上手而无需反复调参。","stars":2204,"starGrowth":1,"createdAt":"2023-02-15T04:24:19Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0.62,"domainTags":["PyTorch","优化器","机器学习训练"],"domainKey":"PyTorch","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/lucidrains/lion-pytorch","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"darkzOGx/youtube-automation-agent","owner":"darkzOGx","name":"youtube-automation-agent","desc":"🎬 Fully automated YouTube channel management with AI agents. Creates, optimizes &amp; publishes videos 24/7. Works with FREE Gemini API or OpenAI. No coding required!","summaryCn":"全自动托管你的YouTube频道，AI24小时不睡觉自动生产爆款视频","reasonCn":"这是一个利用大模型Agent技术实现YouTube全流程自动化的项目。它通过调用AI模型处理视频创意、脚本撰写、画面生成、音频合成及多平台发布等任务，支持Gemini和OpenAI两种免费API。用户只需配置好账号和API Key，系统就能独立完成从0到1的频道运营，极大降低了内容生产门槛。","stars":3647,"starGrowth":17,"createdAt":"2025-08-14T19:35:56Z","silentRounds":0,"language":"JavaScript","topics":["ai-agents","ai-powered","automation","content-automation","content-creation","content-strategy","free-tool","google-gemini","javascript","nodejs","openai","seo-optimization","social-media-automation","thumbnail-generator","video-generation","youtube-api","youtube-automation","youtube-bot","youtube-channel","youtube-uploader"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["YouTube运营","视频自动化","AI Agent","内容生成"],"domainKey":"AI Agent","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-powered","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"content-automation","source":"github","weight":0.7},{"name":"content-creation","source":"github","weight":0.7},{"name":"content-strategy","source":"github","weight":0.7},{"name":"free-tool","source":"github","weight":0.7},{"name":"google-gemini","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"seo-optimization","source":"github","weight":0.7},{"name":"social-media-automation","source":"github","weight":0.7},{"name":"thumbnail-generator","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"youtube-api","source":"github","weight":0.7},{"name":"youtube-automation","source":"github","weight":0.7},{"name":"youtube-bot","source":"github","weight":0.7},{"name":"youtube-channel","source":"github","weight":0.7},{"name":"youtube-uploader","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/darkzOGx/youtube-automation-agent","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.06498409651768577,"copyOk":true},{"repo":"tachyons-css/tachyons","owner":"tachyons-css","name":"tachyons","desc":"Functional css for humans","summaryCn":"小而美的CSS工具库，帮你快速写出干净又好看的页面样式","reasonCn":"Tachyons是一组极简的原子化CSS类，每个类只负责一个样式属性，比如颜色、间距或字体。它不依赖JavaScript框架，纯粹用HTML类名组合就能构建响应式界面。相比Bootstrap等重型框架，Tachyons更注重灵活性和可维护性，适合追求极致性能的开发者，但学习曲线稍陡峭，需要习惯新的思维模式。","stars":11724,"starGrowth":0,"createdAt":"2013-11-04T10:41:59Z","silentRounds":24,"language":"CSS","topics":["design","frontend"],"aiDims":["非AI-实用","可视化","创意工具"],"zone":"工具","funScore":0,"domainTags":["CSS","原子化CSS","前端样式","实用工具"],"domainKey":"CSS","pushedAt":"2026-07-20T15:11:15Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.15,"source":"search","url":"https://github.com/tachyons-css/tachyons","ts":"2026-07-20T15:11:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"satnaing/shadcn-admin","owner":"satnaing","name":"shadcn-admin","desc":"Admin Dashboard UI built with Shadcn and Vite.","summaryCn":"基于Shadcn的高质量后台管理界面模板，开箱即用超省心","reasonCn":"技术上，这个项目基于Shadcn UI组件库和Vite构建，使用的是TypeScript。Shadcn的特点是组件完全可控，你拿走就能改，不像Ant Design那样有固定风格约束。模板里包含了登录页、Dashboard、用户管理、角色权限、设置中心、表格、图表等各种常用页面，布局是经典的侧边栏+顶栏结构，响","stars":14282,"starGrowth":0,"createdAt":"2024-01-26T02:46:42Z","silentRounds":1,"language":"TypeScript","topics":["admin-dashboard","admin-panel","dashboard","shadcn-ui","shadcn-admin"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["前端模板","后台管理","Shadcn UI","TypeScript","Vite"],"domainKey":"前端模板","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"admin-dashboard","source":"github","weight":0.7},{"name":"admin-panel","source":"github","weight":0.7},{"name":"dashboard","source":"github","weight":0.7},{"name":"shadcn-ui","source":"github","weight":0.7},{"name":"shadcn-admin","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.15,"source":"trending","url":"https://github.com/satnaing/shadcn-admin","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"marionettejs/backbone.marionette","owner":"marionettejs","name":"backbone.marionette","desc":"The Backbone Framework","summaryCn":"给老牌前端框架Backbone打补丁的MVC组件库，当年很火现在已过气","reasonCn":"Backbone.Marionette旨在解决Backbone.js在构建大型应用时的复杂性，通过引入View、Region、Router等MVC模式组件，让前端代码更结构化。它是一个纯粹的工程化工具，完全独立于现代AI生态，目前社区活跃度极低，属于历史遗迹。","stars":7033,"starGrowth":0,"createdAt":"2011-12-12T16:36:09Z","silentRounds":25,"language":"JavaScript","topics":["frontend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript","MVC","Backbone"],"domainKey":"JavaScript","pushedAt":"2026-09-13T12:07:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.1,"source":"search","url":"https://github.com/marionettejs/backbone.marionette","ts":"2026-09-13T12:07:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"questdb/questdb","owner":"questdb","name":"questdb","desc":"QuestDB is a high performance, open-source, time-series database","summaryCn":"专门给物联网和实时数据准备的“超速”数据库，写数据快得离谱","reasonCn":"QuestDB 是一款专为时间序列数据打造的高性能开源关系型数据库，采用行列混合存储引擎，每秒可写入数百万条数据。它原生支持 SQL 查询，无需额外建模即可处理海量传感器日志或金融 tick 数据，是构建实时监控大屏和时序分析平台的基础设施。","stars":17341,"starGrowth":2,"createdAt":"2014-04-28T23:29:15Z","silentRounds":0,"language":"Java","topics":["database"],"aiDims":["非AI-实用","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["时序数据库","IoT","SQL","高吞吐","实时数据"],"domainKey":"时序数据库","pushedAt":"2026-09-23T20:57:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/questdb/questdb","ts":"2026-09-23T20:57:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/SynapseML","owner":"microsoft","name":"SynapseML","desc":"Simple and Distributed Machine Learning Python Library porting ML algorithms for Spark","summaryCn":"处理大数据和做AI模型往往是两拨人的事，但 SynapseML 试图把这二者统一起来","reasonCn":"技术上，它利用了 Apache Spark 的分布式计算能力，将机器学习算法转化为高效的 Spark 算子。它最大的特点是实现了对各类 AI 服务的无缝集成，比如你可以一行代码调用 Azure 的认知服务接口，或者将 Hugging Face 的模型直接部署到 Spark 数据管道中进行推理。","stars":5246,"starGrowth":0,"createdAt":"2017-06-05T08:23:44Z","silentRounds":21,"language":"Scala","topics":["data-science"],"aiDims":["AI基础设施","数据科学"],"zone":"AI","funScore":0,"domainTags":["分布式机器学习","Apache Spark","模型部署","大数据流水线","Azure集成"],"domainKey":"模型部署","pushedAt":"2026-09-22T23:00:24Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"数据科学","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Scala","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/microsoft/SynapseML","ts":"2026-09-22T23:00:24Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"ory/hydra","owner":"ory","name":"hydra","desc":"Internet-scale OpenID Certified™ OpenID Connect and OAuth2.1 provider that integrates with your user management through headless APIs. Solve OIDC/OAuth2 user cases over night. Consume as a service on Ory Network or self-host. Trusted by OpenAI and many others for scale and security. Written in Go.","summaryCn":"一键搞定OAuth2登录，几分钟就能上生产","reasonCn":"技术上，它是Go语言写的，性能很好，资源占用低。支持完整的OAuth2授权码流程、客户端凭证流程，以及OIDC的登录和 userinfo 端点。它支持PKCE防授权码劫持、JWT Profile等现代安全特性。它不是把所有东西都集成在一起的单体应用，而是采用无状态设计，认证数据存储在你自己的数据库","stars":17557,"starGrowth":2,"createdAt":"2015-05-22T12:42:08Z","silentRounds":0,"language":"Go","topics":["security"],"aiDims":["安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["OAuth2","OIDC","Identity Provider","Go","Authentication"],"domainKey":"OAuth2","pushedAt":"2026-07-29T09:30:11Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.12,"source":"search","url":"https://github.com/ory/hydra","ts":"2026-07-29T09:30:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sinhaparth5/awesome-tech-books","owner":"sinhaparth5","name":"awesome-tech-books","desc":"Foundational AI/ML papers, GPU and parallel programming, the mathematics underneath it all, and the systems books that explain how the machine actually works.","summaryCn":"一份精选的AI与系统底层硬核技术书籍导航清单。","reasonCn":"内容极具深度，分为三大板块：AI/ML基础论文、GPU/并行计算、底层数学。比如GPU部分不只讲CUDA API，还深入到内存层次结构和指令流水线；AI部分则追溯Transformer原始论文及关键变体。它的独特之处在于强调“为什么”，比如为什么梯度下降在这里收敛，为什么浮点运算会有误差，这种对","stars":46,"starGrowth":0,"createdAt":"2024-12-24T23:42:06Z","silentRounds":15,"language":"","topics":["mathematics"],"aiDims":["学习资源"],"zone":"资源","funScore":0.6,"domainTags":["AI书籍","系统底层","GPU编程","技术书单"],"domainKey":"AI书籍","pushedAt":"2026-09-06T17:15:31Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7}],"aiScore":0.1,"source":"search","url":"https://github.com/sinhaparth5/awesome-tech-books","ts":"2026-09-06T17:15:31Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"eslint/eslint","owner":"eslint","name":"eslint","desc":"Find and fix problems in your JavaScript code.","summaryCn":"JavaScript代码的自动质检员，帮你提前发现并修复代码隐患","reasonCn":"ESLint是业界最流行的JavaScript代码质量检测工具，能自动扫描代码中的错误、风格问题和潜在bug，支持高度自定义规则配置。它可以在编辑器中实时提示问题，帮助团队统一代码风格，减少代码审查成本，是前端工程化的必备利器。","stars":27519,"starGrowth":3,"createdAt":"2013-06-29T23:59:48Z","silentRounds":0,"language":"JavaScript","topics":["eslint","javascript","ecmascript","static-code-analysis","linter"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["JavaScript","Linting","AST","Code Quality","ESLint"],"domainKey":"JavaScript","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"eslint","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"ecmascript","source":"github","weight":0.7},{"name":"static-code-analysis","source":"github","weight":0.7},{"name":"linter","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.1,"source":"trending","url":"https://github.com/eslint/eslint","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xinsir6/ControlNetPlus","owner":"xinsir6","name":"ControlNetPlus","desc":"","summaryCn":"在 AI 图像生成的浪潮中，ControlNet 技术让人类能够精确控制画面的构图、姿","reasonCn":"该项目最核心的亮点在于其“无损叠加”的能力。传统的多条件生成往往需要串行执行多个 ControlNet 模块，显存开销巨大，而 ControlNetPlus 提出的新模块允许同时输入多种条件而不增加计算卸载（computation offload）。","stars":2139,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["多模态Agent","工具调用","AI应用"],"zone":"AI","funScore":0.32,"domainTags":["图像生成","ControlNet","多条件控制","Stable Diffusion"],"domainKey":"图像生成","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/xinsir6/ControlNetPlus","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"shadowlzt/ironclaw-ai-vision","owner":"shadowlzt","name":"ironclaw-ai-vision","desc":"Ultimate AI Fitness Coach PWA 2026: Iron Mike Workout & Smart Recipe Agent","summaryCn":"一个能盯着你看你练得对不对的 AI 健身私教","reasonCn":"结合计算机视觉与大模型构建的 PWA 应用，能够实时分析人体姿态并提供健身指导。项目整合了 Iron Mike 训练法与智能食谱 Agent，通过前端视觉识别与后端 AI 逻辑，为用户提供从训练动作纠正到营养饮食规划的全流程服务。","stars":117,"starGrowth":0,"createdAt":"2026-06-28T20:03:41Z","silentRounds":20,"language":"HTML","topics":["fitness"],"aiDims":["多模态Agent","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["AI健身","姿态识别","计算机视觉"],"domainKey":"AI健身","pushedAt":"2026-09-23T18:49:54Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/shadowlzt/ironclaw-ai-vision","ts":"2026-09-23T18:49:54Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jorge07/symfony-7-es-cqrs-boilerplate","owner":"jorge07","name":"symfony-7-es-cqrs-boilerplate","desc":"Symfony 7 DDD ES CQRS backend boilerplate.","summaryCn":"PHP后端开发者的架构神器，DDD+CQRS+事件溯源一站配齐","reasonCn":"这是一个基于Symfony 7的PHP后端架构样板项目，完整集成了领域驱动设计（DDD）、事件溯源（Event Sourcing）和CQRS命令查询职责分离模式。采用干净的架构分层，包含实体、值对象、聚合根、领域事件等标准组件，适合中大型后端系统快速起步。","stars":1087,"starGrowth":0,"createdAt":"2018-01-02T22:39:20Z","silentRounds":25,"language":"PHP","topics":["backend"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["PHP","Symfony","DDD","CQRS","Event Sourcing"],"domainKey":"PHP","pushedAt":"2026-08-09T20:30:44Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.1,"source":"search","url":"https://github.com/jorge07/symfony-7-es-cqrs-boilerplate","ts":"2026-08-09T20:30:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"chandansgowda/learn-python-in-kannada","owner":"chandansgowda","name":"learn-python-in-kannada","desc":"Your one-stop resource to learn Python from scratch in namma kannada 💛❤️","summaryCn":"用卡纳达语学 Python，零基础也能快速入门","reasonCn":"这份资源为卡纳达语使用者提供从零开始学习 Python 的完整路线。内容涵盖基本语法、数据结构、函数、文件操作和常用标准库，配有示例代码与练习，帮助读者在实际项目中快速上手。项目以 Markdown 文档形式发布，便于离线阅读和快速查阅，适合想在母语环境下学习编程的新人。","stars":300,"starGrowth":0,"createdAt":"2024-09-08T19:03:18Z","silentRounds":3,"language":"Python","topics":["learning"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["Python","Kannada","Tutorial","Markdown","Education"],"domainKey":"Python","pushedAt":"2026-09-07T07:11:02Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.04,"source":"search","url":"https://github.com/chandansgowda/learn-python-in-kannada","ts":"2026-09-07T07:11:02Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tw93/Waza","owner":"tw93","name":"Waza","desc":"🥷 Engineering habits you already know, turned into skills Claude can run.","summaryCn":"把你的习惯变成AI技能，让Claude自动帮你养成好习惯，好玩又实用","reasonCn":"Waza是一个将工程习惯转化为AI可执行技能的开源项目。它把一些你早已知道的优秀习惯（如每日复盘、代码审查 checklist）封装成Claude能运行的技能包。用户只需导入技能，Claude就能在对话中自动提醒、执行这些习惯。比如设置一个\"代码审查习惯\"，Claude就会在每次你提交代码前主动询问是否检查了边界条件。","stars":7077,"starGrowth":1,"createdAt":"2026-03-12T00:35:56Z","silentRounds":1,"language":"Python","topics":["design"],"aiDims":["AI应用","非AI-好玩","效率工具"],"zone":"AI","funScore":0,"domainTags":["Claude","AI技能包","习惯养成"],"domainKey":"Claude","pushedAt":"2026-09-22T09:24:57Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/tw93/Waza","ts":"2026-09-22T09:24:57Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"dkhamsing/open-source-ios-apps","owner":"dkhamsing","name":"open-source-ios-apps","desc":":iphone: Collaborative List of Open-Source iOS Apps","summaryCn":"苹果生态里的开源宝库，免费好用的iOS应用一网打尽","reasonCn":"这个仓库是由全球开发者社区共同维护的iOS开源应用百科全书，收录了数千款高质量的开源iOS应用。每个项目都附有GitHub仓库链接、功能简介和分类标签，涵盖游戏、效率工具、社交媒体、媒体播放等各个类别。开发者可以快速发现可复用的开源代码，或参考优秀项目的架构设计。","stars":52235,"starGrowth":6,"createdAt":"2015-01-26T23:32:34Z","silentRounds":0,"language":"","topics":["game"],"aiDims":["非AI-好玩","非AI-实用","游戏"],"zone":"资源","funScore":0,"domainTags":["iOS","开源应用","导航"],"domainKey":"iOS","pushedAt":"2026-09-23T12:18:47Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7}],"aiScore":0.65,"source":"search","url":"https://github.com/dkhamsing/open-source-ios-apps","ts":"2026-09-23T12:18:47Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.018344596534890397,"copyOk":false},{"repo":"EveryInc/compound-engineering-plugin","owner":"EveryInc","name":"compound-engineering-plugin","desc":"Official Compound Engineering plugin for Claude Code, Codex, Cursor, and more","summaryCn":"第一段：写代码总会遇到找不到合适 API、重复写模板代码的尴尬时刻，尤其切","reasonCn":"第一段：写代码总会遇到找不到合适 API、重复写模板代码的尴尬时刻，尤其切换项目时要记忆大量库的用法。Compound Engineering 插件把 AI 当成你的即时编程伙伴，直接在编辑器里把思路变成可运行代码，省掉去文档或搜索引擎的时间。","stars":25214,"starGrowth":25,"createdAt":"2025-10-09T19:43:46Z","silentRounds":0,"language":"TypeScript","topics":["compound","engineering"],"aiDims":["代码助手","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI插件","代码补全","Claude","Cursor"],"domainKey":"AI插件","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"compound","source":"github","weight":0.7},{"name":"engineering","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.86,"source":"trending","url":"https://github.com/EveryInc/compound-engineering-plugin","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.09356944554612516,"copyOk":false},{"repo":"dandavison/delta","owner":"dandavison","name":"delta","desc":"A syntax-highlighting pager for git, diff, grep, rg --json, and blame output","summaryCn":"Delta是一个简单的Rust程序，用于在Git仓库中对文件进行语法高亮显示。","reasonCn":"Delta是一个简单的Rust程序，用于在Git仓库中对文件进行语法高亮显示。它可以自动检测文件中的关键词并进行高亮，同时也可以方便地对整个文件进行语法高亮。这对于开发者来说是一个很好的辅助工具，可以帮助他们更快地定位代码问题并提高代码可读性。","stars":32306,"starGrowth":8,"createdAt":"2019-06-24T15:00:13Z","silentRounds":0,"language":"Rust","topics":["git"],"aiDims":["Git","语法高亮页码输出"],"zone":"工具","funScore":0,"domainTags":["git","代码diff","语法高亮","命令行"],"domainKey":"git","pushedAt":"2026-09-19T22:05:38Z","tags":[{"name":"Git","source":"llm","weight":1},{"name":"语法高亮页码输出","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/dandavison/delta","ts":"2026-09-19T22:05:38Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0247428960564601,"copyOk":false},{"repo":"alganzory/HaramBlur","owner":"alganzory","name":"HaramBlur","desc":"","summaryCn":"HaramBlur 是一个深度集成 AI 视觉识别的浏览器插件，其核心在于通过人脸检","reasonCn":"作为一款专注于提升隐私与减少干扰的工具，它提供了灵活的控制面板，允许用户根据自身需求定制过滤的深度与范围。用户无需对复杂的深度学习模型有任何了解，只需安装扩展程序，即可享受自动化过滤带来的“护眼”体验。","stars":2076,"starGrowth":8,"createdAt":"2023-09-22T17:28:25Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["AI应用","非AI-好玩"],"zone":"AI","funScore":0.5,"domainTags":["浏览器扩展","内容过滤","人脸检测","NSFW检测"],"domainKey":"浏览器扩展","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/alganzory/HaramBlur","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.0355606936416185,"copyOk":false},{"repo":"qdrant/qdrant-client","owner":"qdrant","name":"qdrant-client","desc":"Python client for Qdrant vector search engine","summaryCn":"Qdrant向量数据库的Python客户端，让AI搜索和推荐变得超简单","reasonCn":"这是Qdrant向量搜索引擎的官方Python客户端，支持高维向量相似度搜索、带过滤条件的检索、批量操作等核心功能。和LangChain、LlamaIndex等框架无缝集成，是构建RAG系统和语义搜索的必备组件，1300多star是核心库中稳健的选择。","stars":1361,"starGrowth":0,"createdAt":"2021-02-09T01:33:20Z","silentRounds":3,"language":"Python","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["向量数据库","向量搜索","RAG","语义检索","Python SDK"],"domainKey":"向量数据库","pushedAt":"2026-09-22T09:46:59Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/qdrant/qdrant-client","ts":"2026-09-22T09:46:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rancher/rancher","owner":"rancher","name":"rancher","desc":"Complete container management platform","summaryCn":"给K8s集群装个大脑，一键管理上百个容器应用。","reasonCn":"Rancher是业界领先的Kubernetes容器管理平台，提供统一的多集群视图和完整的生命周期管理功能。它内置了认证授权、监控告警、网络策略等企业级功能，极大地简化了Kubernetes的使用复杂度。通过Rancher，运维人员可以在一个界面中管理来自不同云厂商的K8s集群，支持从应用部署到故障排查的全流程操作，有效降低运维门槛。","stars":25930,"starGrowth":1,"createdAt":"2014-11-07T20:49:31Z","silentRounds":1,"language":"Go","topics":["docker","kubernetes"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","容器管理","DevOps","多集群管理"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T21:38:18Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/rancher/rancher","ts":"2026-09-23T21:38:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Bouwles/DriftSync","owner":"Bouwles","name":"DriftSync","desc":"A real-time machine learning system that predicts when a user is likely to make a mistake before it occurs.","summaryCn":"实时预测玩家失误的 AI 系统，游戏中提前预警让你少踩坑。","reasonCn":"在竞技游戏或教学类游戏中，玩家往往因为瞬间的操作失误导致整体表现受挫。DriftSync 通过在游戏运行时持续捕获玩家的键盘、鼠标或手柄输入序列，将其送入训练好的时序模型（如 LSTM 或 Transformer），预测接下来几百毫秒内的失误概率，并在界面上以颜色或图标方式提前提示。","stars":187,"starGrowth":0,"createdAt":"2026-03-03T08:59:21Z","silentRounds":12,"language":"Python","topics":["pygame"],"aiDims":["AI应用","工具调用","效率工具"],"zone":"AI","funScore":0,"domainTags":["实时预测","时序模型","pygame","游戏开发","输入分析"],"domainKey":"实时预测","pushedAt":"2026-09-21T17:43:06Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"pygame","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.79,"source":"search","url":"https://github.com/Bouwles/DriftSync","ts":"2026-09-21T17:43:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JuliaLang/julia","owner":"JuliaLang","name":"julia","desc":"The Julia Programming Language","summaryCn":"专为科学计算打造的编程语言，AI和数据分析效率超高","reasonCn":"Julia是一门为高性能数值计算和科学计算而生的编程语言，语法像Python一样简洁，速度却接近C。在机器学习、数值模拟、统计学等领域表现出色，拥有Pluto.jl交互式笔记本、Flux.jl深度学习框架等丰富生态。48000+星，是学术界和工业界都在用的成熟项目。","stars":49139,"starGrowth":3,"createdAt":"2011-04-21T07:01:50Z","silentRounds":0,"language":"Julia","topics":["ml"],"aiDims":["模型与训练","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Julia","科学计算","编程语言"],"domainKey":"Julia","pushedAt":"2026-09-23T21:21:49Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Julia","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/JuliaLang/julia","ts":"2026-09-23T21:21:49Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MemoriLabs/Memori","owner":"MemoriLabs","name":"Memori","desc":"Memori is agent-native memory infrastructure. A LLM-agnostic layer that turns agent execution and conversation into structured, persistent state for production systems. Built for enterprise, Memori works with the data infrastructure you already run, no rip-and-replace, and deploys across managed cloud, single-tenant cloud, VPC, and on-premises.","summaryCn":"给AI Agent装上持久记忆，跨会话也能认出你是谁","reasonCn":"专为AI Agent设计的记忆基础设施层，支持多种大语言模型，能将对话历史和用户偏好结构化存储。提供语义搜索和持久化状态管理，让Agent跨会话保持连续性，解决当前Agent缺乏长期记忆的痛点，适合生产级企业应用部署。","stars":16905,"starGrowth":12,"createdAt":"2025-07-24T07:07:51Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI Agent","AI基础设施","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["RAG","Memory Infrastructure","Agent"],"domainKey":"RAG","pushedAt":"2026-09-18T00:03:15Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/MemoriLabs/Memori","ts":"2026-09-18T00:03:15Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03812954747116237,"copyOk":false},{"repo":"lklynet/aurral","owner":"lklynet","name":"aurral","desc":"The best in open-source music discovery.","summaryCn":"想象一下，你不想听算法猜给你的歌，而是想真正地“探索”音乐。","reasonCn":"从技术上，它利用了Spotify的公开API作为数据源，结合了Elasticsearch进行快速检索。最大的亮点在于它的交互设计，不是传统的列表式播放器，而是基于时间轴和风格的网格化探索界面。它支持多种数据源的混合，比如同时显示Last.fm的标签和Spotify的元数据，打破了信息孤岛。","stars":1684,"starGrowth":2,"createdAt":"2026-01-15T20:27:08Z","silentRounds":0,"language":"JavaScript","topics":["music"],"aiDims":["AI搜索","非AI-好玩","创意工具"],"zone":"工具","funScore":0.7,"domainTags":["音乐发现","交互式浏览","流媒体","音乐探索"],"domainKey":"音乐发现","pushedAt":"2026-09-23T21:08:08Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/lklynet/aurral","ts":"2026-09-23T21:08:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"grafana/pyroscope","owner":"grafana","name":"pyroscope","desc":"Continuous Profiling Platform. Debug performance issues down to a single line of code","summaryCn":"性能调优神器，精准定位代码卡顿的罪魁祸首，比X光还清楚","reasonCn":"这是一个持续性能剖析平台，能收集应用运行时的CPU和内存使用数据，生成火焰图。开发者可以直观看到哪行代码消耗最多资源，从而快速修复性能瓶颈。相比传统APM，它更轻量且与Prometheus生态无缝集成，适合云原生环境下的长期性能监控和趋势分析。","stars":11668,"starGrowth":0,"createdAt":"2020-12-31T05:47:19Z","silentRounds":3,"language":"Go","topics":["devops","developer-tools"],"aiDims":["非AI-实用","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["性能剖析","eBPF","Prometheus","云原生监控","火焰图"],"domainKey":"性能剖析","pushedAt":"2026-09-23T21:14:21Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/grafana/pyroscope","ts":"2026-09-23T21:14:21Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pmbstyle/Alice","owner":"pmbstyle","name":"Alice","desc":"Alice is a voice-first desktop AI assistant application built with Vue.js, Vite, and Electron. Advanced memory system, function calling, MCP support, optional fully local use, and more.","summaryCn":"把各种大模型和本地电脑无缝融合的多功能桌面语音助手。","reasonCn":"技术架构上，它前端基于Vue.js和Vite开发，保证了极高的运行流畅度和丝滑的交互动画，底层通过Electron实现跨平台桌面端适配。它最厉害的地方在于集成了先进的长期记忆系统，能够记住你聊过的话题和习惯；同时原生支持MCP协议和函数调用，这意味着它可以像触手一样外挂各种工具，随","stars":318,"starGrowth":0,"createdAt":"2024-07-31T22:45:54Z","silentRounds":12,"language":"TypeScript","topics":["vector-db"],"aiDims":["AI应用","AI Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["桌面AI助手","语音交互","MCP","本地部署","Electron"],"domainKey":"桌面AI助手","pushedAt":"2026-09-18T21:57:16Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/pmbstyle/Alice","ts":"2026-09-18T21:57:16Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"documenso/documenso","owner":"documenso","name":"documenso","desc":"The Open Source DocuSign Alternative.","summaryCn":"开源的电子签名工具，像 DocuSign 那样好用。","reasonCn":"这是一款基于 Web 的开源电子签名平台，允许用户上传文档、添加签名或印章、设置签署流程，并生成具有法律效力的 PDF 文件。支持多级审批、邮件通知和自定义品牌颜色。用户只需部署一次，即可作为 SaaS 服务供团队或客户使用，完全掌控数据隐私。","stars":15141,"starGrowth":3,"createdAt":"2023-03-12T16:28:56Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted"],"aiDims":["AI应用","工具调用"],"zone":"工具","funScore":0,"domainTags":["电子签名","开源SaaS","PDF处理"],"domainKey":"电子签名","pushedAt":"2026-09-23T20:08:57Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/documenso/documenso","ts":"2026-09-23T20:08:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/skills","owner":"google","name":"skills","desc":"Agent Skills for Google products and technologies","summaryCn":"谷歌官方出的AI技能库，一键激活Docs等产品的强大能力","reasonCn":"这是谷歌官方推出的开源技能集合，专为AI智能体设计。通过标准化的接口，让Agent能直接调用Google Docs、Sheets、Maps等数十款产品的核心能力。底层基于Python实现，支持自定义技能开发，是构建多模态、多工具AI助手的关键基础设施，极大降低了开发门槛。","stars":20290,"starGrowth":21,"createdAt":"2026-03-31T19:02:20Z","silentRounds":0,"language":"Python","topics":["google","googlecloud","skills"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","技能包","Google API","插件开发","多工具集成"],"domainKey":"AI Agent","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"google","source":"github","weight":0.7},{"name":"googlecloud","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/google/skills","ts":"2026-09-17T04:13:51.053Z","category":"ai","momentum":["daily"],"fromOfficial":true,"heatScore":0.07932597338590439,"copyOk":false},{"repo":"Fincept-Corporation/FinceptTerminal","owner":"Fincept-Corporation","name":"FinceptTerminal","desc":"FinceptTerminal is a modern finance application offering advanced market analytics, investment research, and economic data tools, designed for interactive exploration and data-driven decision-making in a user-friendly environment.","summaryCn":"AI驱动的金融分析终端，实时市场数据和智能投资研究工具","reasonCn":"FinceptTerminal是一款现代金融应用程序，结合AI和机器学习技术提供高级市场分析、投资研究和经济数据工具。它支持交互式数据探索、实时行情展示和自动化投资建议，帮助投资者和数据分析师做出更明智的决策。界面美观且功能强大，是金融领域的AI实用工具。","stars":31924,"starGrowth":16,"createdAt":"2024-08-29T14:38:08Z","silentRounds":0,"language":"C++","topics":["ml"],"aiDims":["AI应用","工具调用","可视化"],"zone":"AI","funScore":0.7,"domainTags":["金融分析","机器学习","投资决策","市场数据","C++应用"],"domainKey":"金融分析","pushedAt":"2026-09-19T12:41:43Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/Fincept-Corporation/FinceptTerminal","ts":"2026-09-19T12:41:43Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04950357098108007,"copyOk":true},{"repo":"opencardev/crankshaft","owner":"opencardev","name":"crankshaft","desc":"Crankshaft: A turnkey GNU/Linux solution that transforms a Raspberry Pi to an Android Auto head unit.","summaryCn":"把树莓派秒变安卓车机，支持 CarPlay 和语音助手。","reasonCn":"Crankshaft 的技术优势在于它的“开箱即用”性。它集成了 Android Auto 的核心服务，并针对树莓派的硬件资源进行了优化。它支持 HDMI 输出，连接车载屏幕就能获得完整的交互体验。此外，它还内置了语音助手和触控键盘，解决了在车载环境下操作手机的不便。","stars":2678,"starGrowth":0,"createdAt":"2018-02-23T17:45:14Z","silentRounds":3,"language":"Python","topics":["raspberry-pi"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["树莓派","Android Auto","车机系统","DIY","Linux"],"domainKey":"树莓派","pushedAt":"2026-09-20T07:54:02Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/opencardev/crankshaft","ts":"2026-09-20T07:54:02Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GitHubDaily/ChatGPT-Prompt-Engineering-for-Developers-in-Chinese","owner":"GitHubDaily","name":"ChatGPT-Prompt-Engineering-for-Developers-in-Chinese","desc":"","summaryCn":"自 ChatGPT 推出以来，如何高效编写提示词已成为人机交互的分水岭，熟练掌握","reasonCn":"项目目录结构清晰，核心字幕存放在专属文件夹中，同时分离出了纯中文字幕与英文字幕，方便不同学习需求的读者进行对照阅读。此外，仓库还同步收录了从课程中复制至本地的 Notebook 代码文件，并结合了 Datawhale 提供的中文版配套笔记，形成了一套完整的学习闭环。","stars":2125,"starGrowth":0,"silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程","提示工程"],"zone":"资源","funScore":0.2,"domainTags":["提示词工程","ChatGPT","双语字幕","人工智能教程"],"domainKey":"提示词工程","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/GitHubDaily/ChatGPT-Prompt-Engineering-for-Developers-in-Chinese","ts":"2026-09-22T22:51:50.275Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"directus/directus","owner":"directus","name":"directus","desc":"The flexible backend for all your projects 🐰 Turn your DB into a headless CMS, admin panels, or apps with a custom UI, instant APIs, auth & more.","summaryCn":"把数据库秒变后台管理系统，零代码搭建各类应用","reasonCn":"Directus是一个开源的Headless CMS平台，底层基于任何SQL数据库运行。它会自动扫描数据库表结构，生成REST和GraphQL API接口，同时提供可视化数据管理界面和完整的用户权限系统。支持实时订阅、文件存储管理、OAuth认证等能力，兼容PostgreSQL、MySQL、SQLite等多种数据库引擎。","stars":37943,"starGrowth":2,"createdAt":"2012-12-12T01:35:36Z","silentRounds":0,"language":"TypeScript","topics":["database","data-visualization"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["无头CMS","数据库","API生成"],"domainKey":"无头CMS","pushedAt":"2026-09-23T16:44:55Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/directus/directus","ts":"2026-09-23T16:44:55Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"coleam00/mcp-crawl4ai-rag","owner":"coleam00","name":"mcp-crawl4ai-rag","desc":"","summaryCn":"在构建 AI Agent 时，如何让智能体拥有“获取最新信息”的能力一直是个难题。","reasonCn":"该项目的核心优势在于其“MCP 原生”的设计。随着 Anthropic 推出 Model Context Protocol，AI 助手（如 Claude）调用外部工具的标准正在形成。这个服务器正是顺应这一趋势，让 AI 编程助手能够像调用内置函数一样自然地调用网页爬取和知识库查询服务。","stars":2255,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["RAG与知识","AI Agent","工具调用"],"zone":"AI","funScore":0.2,"domainTags":["网页爬取","RAG检索","MCP协议"],"domainKey":"网页爬取","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/coleam00/mcp-crawl4ai-rag","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"simonlin1212/vibe-astock","owner":"simonlin1212","name":"vibe-astock","desc":"A 股短线复盘看板：涨停池·连板梯队·龙虎榜·板块资金一屏看完，赚钱效应/晋级率/梯队断层/情绪周期等派生指标纯计算直出（不经过 AI），AI 只把数据串成能读的盘面研判。全本地运行，可用 Claude/Codex 订阅免 API key。| A-share short-term daily-review dashboard: derived sentiment metrics computed locally, AI only writes the narrative. No API key needed.","summaryCn":"把A股封板龙虎榜数据变成直观情绪看盘的本地化智能分析板。","reasonCn":"技术架构非常务实。数据层采用纯Python计算，确保涨停梯队、晋级率等核心指标绝对准确，不依赖AI概率，杜绝了幻觉风险。表现层调用本地AI（支持Claude/Codex），仅负责将结构化数据转化为流畅的中文复盘文案。这种“数据硬、文案软”的混合模式，既保证了金融数据的严谨性，又提供了","stars":408,"starGrowth":1,"createdAt":"2026-07-29T23:40:47Z","silentRounds":1,"language":"Python","topics":["trading"],"aiDims":["AI应用","可视化"],"zone":"工具","funScore":0.5,"domainTags":["A股交易","复盘看板","本地AI","股市情绪","短线交易"],"domainKey":"A股交易","pushedAt":"2026-09-20T12:07:55Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/simonlin1212/vibe-astock","ts":"2026-09-20T12:07:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"chartbrew/chartbrew","owner":"chartbrew","name":"chartbrew","desc":"Open-source reporting platform to build and share live dashboards from APIs, SQL and NoSQL databases, with powerful AI assistant, scheduling, and embeddable charts 📈📊","summaryCn":"一键连数据库和API生成交互仪表盘，还能让AI自动写图表。","reasonCn":"Chartbrew 基于Node.js与React构建的开源报表平台，支持从RESTAPI、SQL、NoSQL多源拉取数据，内置AI助手可依据自然语言自动生成图表配置，具备定时刷新、共享链接和可嵌入的可视化组件，帮助业务快速搭建实时监控看板，无需写前端代码。","stars":4062,"starGrowth":1,"createdAt":"2019-08-29T13:14:17Z","silentRounds":3,"language":"JavaScript","topics":["data-visualization"],"aiDims":["AI应用","可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["数据可视化","仪表盘","报表平台"],"domainKey":"数据可视化","pushedAt":"2026-09-21T03:28:47Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.81,"source":"search","url":"https://github.com/chartbrew/chartbrew","ts":"2026-09-21T03:28:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AIGCDesignGroup/ReplaceAnything","owner":"AIGCDesignGroup","name":"ReplaceAnything","desc":"","summaryCn":"ReplaceAnything是阿里巴巴智能计算研究院推出的一项突破性图像编辑技术。","reasonCn":"这个项目非常适合需要进行快速素材处理的设计师或艺术创作者。在安装使用方面，项目已经集成了HuggingFace和魔搭社区的体验链接，用户无需在本地配置繁杂的深度学习环境，即可直接上手操作。通过简单的遮罩选区，即可启动智能重绘引擎，系统会自动识别图片意图并输出高质量的替换结","stars":2424,"starGrowth":0,"createdAt":"2024-01-05T05:45:11Z","silentRounds":1,"language":"","topics":[],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.2,"domainTags":["图像修复","内容替换","虚拟模型","AIGC"],"domainKey":"图像修复","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1}],"aiScore":0.8,"source":"search","url":"https://github.com/AIGCDesignGroup/ReplaceAnything","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"datawhalechina/deep-learning-notes","owner":"datawhalechina","name":"deep-learning-notes","desc":"Personal deep learning study notes and tutorial-style notebooks","summaryCn":"国内顶尖开源社区打造的保姆级深度学习自学笔记与教程。 (28字)","reasonCn":"技术架构上，它完全基于Python语言，深度结合了PyTorch和Jupyter Notebook。内容涵盖了从最基础的线性回归、多层感知机，到大火的卷积神经网络、循环神经网络，甚至包含现代大模型的基础理论。每一章不仅有硬核的公式推导解释，更配有可以直接运行的互动代码，真正做到理论与实践无","stars":690,"starGrowth":0,"createdAt":"2026-04-21T04:06:43Z","silentRounds":3,"language":"Python","topics":["tutorial"],"aiDims":["学习资源","教程","模型与训练"],"zone":"资源","funScore":0,"domainTags":["深度学习","教程","PyTorch","笔记"],"domainKey":"深度学习","pushedAt":"2026-09-21T17:57:22Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/datawhalechina/deep-learning-notes","ts":"2026-09-21T17:57:22Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"camenduru/text-generation-webui-colab","owner":"camenduru","name":"text-generation-webui-colab","desc":"","summaryCn":"在Colab上轻松运行大语言模型网页界面，一键部署文本生成工具。","reasonCn":"这是一个让你在谷歌Colab上免配置运行文本生成WebUI的项目。它把原本复杂的大模型部署过程打包成一键式笔记本，你只需打开就能用上LLaMA等模型。项目适配了免费和付费的Colab账号，还配套了视频教程和社区支持，对想要快速体验大模型的新手非常友好。","stars":2093,"starGrowth":3,"createdAt":"2023-04-08T22:04:37Z","silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["大语言模型","模型部署","AI应用"],"zone":"AI","funScore":0.38,"domainTags":["大语言模型","文本生成","Colab","深度学习"],"domainKey":"大语言模型","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/camenduru/text-generation-webui-colab","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hunglc007/tensorflow-yolov4-tflite","owner":"hunglc007","name":"tensorflow-yolov4-tflite","desc":"","summaryCn":"这个深度学习工程完整复现了经典的 YOLOv4 以及 YOLOv4-tiny 目标检测网络，","reasonCn":"它的核心亮点在于提供了极其便捷的权重转换能力，能够把官方的 YOLOv4、YOLOv3 及其 tiny 版本的权重文件无缝转换为 .pb、.tflite 甚至是适用于英伟达显卡的 TensorRT（trt）格式。这使得原本沉重庞大的神经网络模型，得以顺利压缩并迁移到手机移动端、嵌入式边缘设备或高性能服务","stars":2254,"starGrowth":0,"createdAt":"2020-04-24T10:22:52Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型部署","推理引擎"],"zone":"AI","funScore":0.2,"domainTags":["目标检测","YOLOv4","TensorFlow Lite","模型转换"],"domainKey":"目标检测","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/hunglc007/tensorflow-yolov4-tflite","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"approximatelabs/sketch","owner":"approximatelabs","name":"sketch","desc":"","summaryCn":"Sketch 的核心定位是消除数据分析师与代码之间的摩擦。","reasonCn":"上手体验方面，Sketch 最大的卖点是极低的部署成本。官方承诺无需修改现有的 IDE 配置或安装繁琐的插件，用户只需按照指引简单配置，几秒钟内即可开始交互。这种轻量级设计使得它非常适合那些不想被开发环境束缚，只想快速获得代码帮助的数据从业者。","stars":2282,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","代码助手"],"zone":"AI","funScore":0.2,"domainTags":["Pandas","数据分析","自然语言编程"],"domainKey":"Pandas","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/approximatelabs/sketch","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/neural-tangents","owner":"google","name":"neural-tangents","desc":"","summaryCn":"在无限宽度神经网络的研究中，Neural Tangents 解决的核心痛点是计算复杂度","reasonCn":"该库的核心亮点在于其统一的 API 设计。无论是有限的还是无限宽度的网络，用户都使用同一套接口进行定义、训练和评估。根据项目文档，这一库已被超过 100 篇学术论文使用，这证明了其在学术界分析神经网络理论特性时的广泛接受度。","stars":2383,"starGrowth":0,"createdAt":"2019-04-08T16:48:48Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0.2,"domainTags":["神经网络","高斯过程","无限宽度","理论分析","JAX"],"domainKey":"神经网络","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.752,"source":"search","url":"https://github.com/google/neural-tangents","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"SearchSavior/OpenArc","owner":"SearchSavior","name":"OpenArc","desc":"Inference engine for Intel devices. Serve LLMs, VLMs, Whisper, Kokoro-TTS, Embedding and Rerank models over OpenAI endpoints.","summaryCn":"Inference engine for Intel devices. It serves LLMs, VLMs, Whisper, Kokoro-TTS, Embedding and Rerank models over OpenAI endpoints.","reasonCn":"这个项目是一个基于Transformers的推理引擎，它可以在Intel设备上运行，服务于各种LLM、VLM、Whisper、Kokoro-TTS、Embedding和Rerank模型，并支持OpenAI端点。","stars":530,"starGrowth":0,"createdAt":"2025-02-11T05:46:58Z","silentRounds":2,"language":"Python","topics":["transformers"],"aiDims":["Transformers"],"zone":"AI","funScore":0.5,"domainTags":["推理引擎","LLM","Intel设备","OpenAI兼容","模型部署"],"domainKey":"推理引擎","pushedAt":"2026-09-23T01:31:14Z","tags":[{"name":"Transformers","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/SearchSavior/OpenArc","ts":"2026-09-23T01:31:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Armur-Ai/Pentest-Swarm-AI","owner":"Armur-Ai","name":"Pentest-Swarm-AI","desc":"","summaryCn":"开源AI渗透测试集群，通过多智能体并发寻找漏洞，被称为XBOW替代品。","reasonCn":"该软件明确面向专业技术人员，标签显示其构建目标是服务于渗透测试人员（pentesters）、Bug赏金猎人（bug bounty hunters）、红队成员（red teamers）以及安全研究员。文档中特别提到了“Swarm vs. Multi-Agent”（集群与多智能体）的文档章节，暗示其在架构设计上对“集群协作”有特定","stars":2564,"starGrowth":2,"createdAt":"2024-03-26T06:40:36Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI Agent","多模态Agent","安全工具"],"zone":"AI","funScore":0.5,"domainTags":["渗透测试","漏洞扫描","红队模拟","Bug赏金"],"domainKey":"渗透测试","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/Armur-Ai/Pentest-Swarm-AI","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"666ghj/MiroFish","owner":"666ghj","name":"MiroFish","desc":"A Simple and Universal Swarm Intelligence Engine, Predicting Anything. 简洁通用的群体智能引擎，预测万物","summaryCn":"一群小鱼游来游去，就能预测房价股价和天气","reasonCn":"核心技术是群体智能（Swarm Intelligence），主要算法包括鱼群算法（FSA）、粒子群优化（PSO）等变种。它把问题空间想象成一片水域，每条\"鱼\"代表一个可能的解，鱼群通过信息素、食物搜索、躲避天敌等模拟行为，在解空间中不断进化搜索，最终收敛到最优或近似最优解。","stars":74303,"starGrowth":35,"createdAt":"2025-11-26T04:23:02Z","silentRounds":0,"language":"Python","topics":["agent-memory","financial-forecasting","future-prediction","knowledge-graph","llms","multi-agent-simulation","public-opinion-analysis","python3","social-prediction","swarm-intelligence"],"aiDims":["AI应用","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["群体智能优化","鱼群算法","粒子群优化","黑盒优化求解器"],"domainKey":"群体智能优化","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agent-memory","source":"github","weight":0.7},{"name":"financial-forecasting","source":"github","weight":0.7},{"name":"future-prediction","source":"github","weight":0.7},{"name":"knowledge-graph","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"multi-agent-simulation","source":"github","weight":0.7},{"name":"public-opinion-analysis","source":"github","weight":0.7},{"name":"python3","source":"github","weight":0.7},{"name":"social-prediction","source":"github","weight":0.7},{"name":"swarm-intelligence","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/666ghj/MiroFish","ts":"2026-09-17T04:13:51.053Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.12769575925601925,"copyOk":true},{"repo":"vikashplus/robohive","owner":"vikashplus","name":"robohive","desc":"A unified framework for robot learning","summaryCn":"机器人学习的统一仿真框架，强化学习训练机械臂的开源利器","reasonCn":"核心技术上，它基于 Python 开发，集成了 MuJoCo 和 Isaac Gym 等主流物理仿真后端，支持连续动作空间和离散动作空间的任务。框架提供了统一的环境接口，同一个算法可以直接在不同仿真环境里跑，方便做跨环境对比实验。","stars":632,"starGrowth":0,"createdAt":"2018-07-06T16:55:41Z","silentRounds":17,"language":"Python","topics":["simulation"],"aiDims":["AI基础设施","模型与训练","AI Agent"],"zone":"AI","funScore":0,"domainTags":["机器人强化学习","MuJoCo仿真","Isaac Gym","Sim2Real"],"domainKey":"机器人强化学习","pushedAt":"2026-09-22T14:39:19Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/vikashplus/robohive","ts":"2026-09-22T14:39:19Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xonsh/xonsh","owner":"xonsh","name":"xonsh","desc":"🐚 Python-powered shell. Full-featured, cross-platform and AI-friendly.","summaryCn":"用Python语法重写命令行，开发者的终极效率武器库","reasonCn":"Xonsh是一个革命性的shell环境，用Python语法完全替代传统bash/zsh。它保留了所有shell命令的同时，让你能在命令行里直接写Python代码、调用Python库。跨平台支持Windows/Mac/Linux，还能与AI工具无缝集成，堪称开发者的瑞士军刀。","stars":9651,"starGrowth":1,"createdAt":"2015-01-21T22:05:27Z","silentRounds":1,"language":"Python","topics":["devops","raspberry-pi"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Shell","Python","命令行"],"domainKey":"Shell","pushedAt":"2026-09-22T09:21:00Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/xonsh/xonsh","ts":"2026-09-22T09:21:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"calcom/sans","owner":"calcom","name":"sans","desc":"The home for our Cal Sans font.","summaryCn":"想象一下，你正在做一个精美的网页或者PPT，却发现默认字体太单调，看着没档","reasonCn":"从技术角度看，Cal Sans 是一款高度优化的屏幕字体。它的设计目标是“易读性”和“美感”的平衡。它不像传统印刷字体那样过于花哨，而是针对高分辨率屏幕做了特别优化。项目提供了完整的字重家族，从轻量到粗体一应俱全，还支持斜体和连字，能够满足排版布局的各种需求。","stars":1618,"starGrowth":0,"createdAt":"2021-11-01T20:02:51Z","silentRounds":1,"language":"Python","topics":["design"],"aiDims":["创意工具","非AI-实用"],"zone":"创意","funScore":0.6,"domainTags":["开源字体","屏幕字体","排版","设计","易读性"],"domainKey":"开源字体","pushedAt":"2026-09-23T16:59:12Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/calcom/sans","ts":"2026-09-23T16:59:12Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"duckframework/duck","owner":"duckframework","name":"duck","desc":"Build high-performance, scalable, server-side reactive web applications — without a separate frontend framework or a complex JavaScript stack.","summaryCn":"不用写JS也能写后端，让前端开发者的效率起飞。","reasonCn":"它的核心思想是“统一”。你用Python写前端组件，用Python写后端API，用Python写数据库操作。所有的代码都在同一个项目里，甚至同一个文件里，通过路由和装饰器就能区分。它内置了类似React/Vue的响应式组件系统，但语法是Python的，不需要学JS框架。","stars":46,"starGrowth":0,"createdAt":"2022-12-04T11:54:27Z","silentRounds":10,"language":"Python","topics":["web-framework"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Web框架","Python","全栈开发","响应式组件"],"domainKey":"Web框架","pushedAt":"2026-09-22T19:07:45Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/duckframework/duck","ts":"2026-09-22T19:07:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kantord/SeaGOAT","owner":"kantord","name":"SeaGOAT","desc":"local-first semantic code search engine","summaryCn":"让代码搜索像问AI一样智能，本地运行语义匹配再也不用记函数名","reasonCn":"核心技术上，它用Embedding模型把代码转换成向量，存入本地向量数据库，实现语义级别的相似度搜索。不同于传统搜索引擎按关键字匹配，它能理解代码的意图和上下文关系。它还支持Git历史追踪，能看到代码变更脉络，跨语言搜索也能工作，Python代码里能搜到JavaScript的相似实现。","stars":1310,"starGrowth":1,"createdAt":"2023-06-20T14:38:16Z","silentRounds":10,"language":"Python","topics":["vector-db"],"aiDims":["AI搜索","向量数据库","代码助手"],"zone":"工具","funScore":0.6,"domainTags":["代码搜索","语义分析","向量数据库","本地检索","Git追踪"],"domainKey":"代码搜索","pushedAt":"2026-09-22T11:46:44Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/kantord/SeaGOAT","ts":"2026-09-22T11:46:44Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/differential-privacy","owner":"google","name":"differential-privacy","desc":"Google's differential privacy libraries.","summaryCn":"在如今这个数据驱动的时代，我们既想利用大数据训练出聪明的AI，又害怕隐私","reasonCn":"核心亮点在于它严谨的数学模型。差分隐私的核心是添加噪声（Noise），但添加多少是门艺术。谷歌的这个库封装了复杂的计算过程，提供了包括均值、计数、直方图等常用的差分隐私函数。它不仅遵循严格的隐私预算（Privacy Budget）控制，还针对工业生产进行了高性能优化，比手写实现要","stars":3362,"starGrowth":0,"createdAt":"2019-09-04T13:04:15Z","silentRounds":9,"language":"Go","topics":["privacy"],"aiDims":["安全工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["差分隐私","密码学","统计","数据保护"],"domainKey":"差分隐私","pushedAt":"2026-09-23T21:53:13Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/google/differential-privacy","ts":"2026-09-23T21:53:13Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"chaitin/SafeLine","owner":"chaitin","name":"SafeLine","desc":"SafeLine is a self-hosted WAF(Web Application Firewall) / reverse proxy to protect your web apps from attacks and exploits.","summaryCn":"给你的网站穿上一件智能防弹衣，黑客来了也打不穿","reasonCn":"长亭科技出品的开源Web应用防火墙，内置智能攻击识别引擎，能自动拦截SQL注入、XSS跨站脚本等常见Web漏洞利用。支持反向代理模式部署，提供可视化管控面板，规则自定义灵活，Docker一键部署，配置简单上手快，社区活跃且持续迭代。","stars":22663,"starGrowth":1,"createdAt":"2023-04-12T11:30:14Z","silentRounds":1,"language":"Go","topics":["security","self-hosted"],"aiDims":["安全工具"],"zone":"工具","funScore":0,"domainTags":["Web应用防火墙","网络安全","逆向代理","攻击拦截"],"domainKey":"Web应用防火墙","pushedAt":"2026-09-22T09:04:10Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/chaitin/SafeLine","ts":"2026-09-22T09:04:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lmnr-ai/index","owner":"lmnr-ai","name":"index","desc":"","summaryCn":"lmnr-ai/index 是一个旨在让 AI 助手具备“手”的浏览器 Agent 框架。","reasonCn":"其核心亮点在于对多种顶级推理模型的支持，包括 Gemini 2.5 Pro、Claude 3.7 Sonnet 和 OpenAI o4-mini。用户可以根据任务复杂度和成本预算灵活切换模型：对于简单任务使用高速低成本的 Flash 模型，对于复杂逻辑推理使用 Pro 或 Sonnet 以确保证据的可靠性。","stars":2444,"starGrowth":1,"createdAt":"2024-11-30T19:34:02Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["多模态Agent","AI Agent","工具调用"],"zone":"AI","funScore":0.38,"domainTags":["浏览器 Agent","自动化任务","LLM 集成"],"domainKey":"浏览器 Agent","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/lmnr-ai/index","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"keplergl/kepler.gl","owner":"keplergl","name":"kepler.gl","desc":"Kepler.gl is a powerful open source geospatial analysis tool for large-scale data sets.","summaryCn":"导入数据就能在地图上画出炫酷的动态数据可视化","reasonCn":"技术架构上，Kepler.gl基于Deck.gl构建，这是一个专门针对大规模地理数据优化的 WebGL可视化框架。它支持多种可视化类型：热力图（heatmap）展示密度分布、轨迹动画（trips）还原移动路径、3D建筑（3D buildings）展示城市轮廓、点图层（point layer）标记位置数据。","stars":12024,"starGrowth":0,"createdAt":"2018-02-28T21:36:16Z","silentRounds":1,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["地理空间","数据可视化","WebGL"],"domainKey":"地理空间","pushedAt":"2026-09-23T15:56:25Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/keplergl/kepler.gl","ts":"2026-09-23T15:56:25Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"makeyourownneuralnetwork/makeyourownneuralnetwork","owner":"makeyourownneuralnetwork","name":"makeyourownneuralnetwork","desc":"","summaryCn":"该项目从最基础的数学概念入手，带你理解神经网络如何通过权重调整来“学习”","reasonCn":"项目内容覆盖了从数学模型推导到Python代码实现的完整流程。你不需要高深的数学背景，只要具备基础的程序编写能力，就能通过阅读其文档和运行代码来完成一个识别手写数字的小型神经网络。作者不仅提供了清晰的逻辑引导，还通过详细的注释让代码极具可读性，确保即便是非专业出身的","stars":2496,"starGrowth":1,"createdAt":"2016-02-22T15:58:18Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["模型与训练","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["神经网络","深度学习","Python教程"],"domainKey":"神经网络","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/makeyourownneuralnetwork/makeyourownneuralnetwork","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"scylladb/scylladb","owner":"scylladb","name":"scylladb","desc":"NoSQL data store using the Seastar framework, compatible with Apache Cassandra and Amazon DynamoDB","summaryCn":"比Cassandra快10倍的开源数据库，适合AI训练数据高并发读写","reasonCn":"ScyllaDB是一款兼容Cassandra和DynamoDB协议的高性能NoSQL数据库，基于C++和Seastar框架构建，采用无共享架构实现微秒级延迟。相比传统Java数据库，它内存管理更高效，吞吐量大幅提升，适合AI场景下向量索引、特征存储等高频读写需求，Star数超1.5万，社区活跃。","stars":15773,"starGrowth":0,"createdAt":"2014-12-24T13:16:33Z","silentRounds":1,"language":"C++","topics":["database"],"aiDims":["非AI-实用","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["NoSQL数据库","分布式存储","高并发读写","Cassandra兼容","向量索引"],"domainKey":"NoSQL数据库","pushedAt":"2026-09-23T20:30:42Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/scylladb/scylladb","ts":"2026-09-23T20:30:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SnailSploit/Claude-Red","owner":"SnailSploit","name":"Claude-Red","desc":"claude-red is a curated library of offensive security skills designed for the Claude skills system. Each skill is a structured SKILL.md file that primes Claude with expert-level methodology for a specific attack surface — from SQLi to shellcode, EDR evasion to exploit development.","summaryCn":"把渗透测试技巧写进Claude，让AI直接输出专家级攻击方案","reasonCn":"Claude-Red提供数百条结构化SKILL.md文件，每条覆盖渗透测试的背景、步骤、工具和指纹，供Claude在对话中直接调用，实现从信息收集到漏洞利用的完整链路，兼容Claude 2并支持自定义提示，既可本地也可云端部署，适合红队与渗透培训使用。","stars":6765,"starGrowth":34,"createdAt":"2026-03-04T05:52:35Z","silentRounds":0,"language":"Python","topics":["claude-ai","claude-skills","redteam","redteam-tools","skills","claude-pt"],"aiDims":["AI Agent","安全工具","工具调用"],"zone":"AI","funScore":0,"domainTags":["渗透测试","红队技能","Claude技能"],"domainKey":"渗透测试","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"claude-ai","source":"github","weight":0.7},{"name":"claude-skills","source":"github","weight":0.7},{"name":"redteam","source":"github","weight":0.7},{"name":"redteam-tools","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"claude-pt","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/SnailSploit/Claude-Red","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.14049312638580932,"copyOk":false},{"repo":"frostnova721/animestream","owner":"frostnova721","name":"animestream","desc":"An app to stream and download anime","summaryCn":"专为二次元动漫迷打造的聚合流媒体播放与下载神器。","reasonCn":"技术上，它完全采用Flutter跨平台框架构建，这意味着开发者只用写一套代码，就能完美适配Android、iOS甚至桌面端。它内置了高效的视频流解析引擎，能够智能抓取各大动漫源的视频地址，并提供流畅的倍速播放、历史记录同步以及离线缓存下载功能，架构设计轻量且优雅。","stars":436,"starGrowth":1,"createdAt":"2024-02-15T20:10:18Z","silentRounds":3,"language":"Dart","topics":["anime"],"aiDims":["非AI-好玩","创意工具","非AI-实用"],"zone":"创意","funScore":0,"domainTags":["动漫播放","视频流","Flutter","跨平台应用"],"domainKey":"动漫播放","pushedAt":"2026-09-21T18:15:57Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/frostnova721/animestream","ts":"2026-09-21T18:15:57Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"huginn/huginn","owner":"huginn","name":"huginn","desc":"Create agents that monitor and act on your behalf.  Your agents are standing by!","summaryCn":"老牌自动化监控神器，像雇了个数字员工帮你盯网页","reasonCn":"被誉为开源版的IFTTT/Zapier，允许用户创建“代理”来监控网站、API、发送通知。Ruby开发，功能极其强大且灵活，可实现复杂的多条件自动化工作流。虽然不直接使用大模型，但它是自动化领域的经典之作，配合AI可实现智能监控场景。","stars":49983,"starGrowth":1,"createdAt":"2013-03-10T06:00:31Z","silentRounds":3,"language":"Ruby","topics":["automation"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["自动化监控","工作流","数据抓取"],"domainKey":"自动化监控","pushedAt":"2026-09-23T18:25:51Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/huginn/huginn","ts":"2026-09-23T18:25:51Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Erudika/para","owner":"Erudika","name":"para","desc":"Multitenant backend server for building web and mobile apps rapidly. The backend for busy developers. (self-hosted or hosted)","summaryCn":"给开发者准备的快速构建应用的后端服务器，自建省心。","reasonCn":"Para 是一个多租户后端服务器，专为快速开发Web和移动应用而设计，支持自托管或云端部署。它提供了一套完整的后端解决方案，包括数据库、认证、会话管理和RESTful API，帮助开发者从繁琐的基础设施配置中解脱出来，专注于业务逻辑的实现。Java语言编写，Star数超过500，在开发者工具领域具有较高的实用性和关注度。","stars":574,"starGrowth":0,"createdAt":"2014-03-09T22:31:26Z","silentRounds":27,"language":"Java","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["后端框架","多租户","Java","REST API"],"domainKey":"后端框架","pushedAt":"2026-09-22T17:30:08Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/Erudika/para","ts":"2026-09-22T17:30:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"modelscope/mcore-bridge","owner":"modelscope","name":"mcore-bridge","desc":"MCore-Bridge: Providing Megatron-Core model definitions for state-of-the-art large models and making Megatron training as simple as Transformers — with support for 300+ large language models (Qwen3-Next, GLM-5.2, Deepseek-V4, MiniMax-2.7, ...) and 200+ multimodal large models (Qwen3.5, Qwen3-Omni, Gemma4, ...).","summaryCn":"大模型训练往往像是一场烧钱的噩梦，不仅环境配置复杂，并行计算的代码写起","reasonCn":"技术上，它通过封装Megatron-Core的核心组件，解决了PyTorch分布式处理中的各种适配问题。它最厉害的地方在于兼容了Hugging Face Transformers的定义风格，这意味着你不需要重写复杂的模型结构代码，直接调用项目中的桥接层，就能利用Megatron强大的张量并行、流水线并行能力，极大","stars":101,"starGrowth":0,"createdAt":"2026-03-16T02:19:38Z","silentRounds":11,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","大语言模型"],"zone":"AI","funScore":0,"domainTags":["大模型训练","Megatron","分布式训练","PyTorch"],"domainKey":"大模型训练","pushedAt":"2026-09-21T03:53:43Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/modelscope/mcore-bridge","ts":"2026-09-21T03:53:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenWebGAL/WebGAL","owner":"OpenWebGAL","name":"WebGAL","desc":"A brand new web Visual Novel engine | 全新的网页端视觉小说引擎","summaryCn":"用网页就能制作和游玩视觉小说，零门槛打造属于你的互动故事。","reasonCn":"基于 TypeScript 开发的 Web 视觉小说引擎，无需安装客户端，直接在浏览器中运行和发布作品。支持剧情分镜、角色立绘、对话系统、音乐背景等完整功能，提供可视化编辑器降低创作难度。适合独立游戏开发者、互动叙事爱好者和快速原型制作。","stars":3979,"starGrowth":0,"createdAt":"2021-09-16T23:51:26Z","silentRounds":3,"language":"TypeScript","topics":["game"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":0.9,"domainTags":["视觉小说","网页游戏","互动叙事","TypeScript","可视化编辑器"],"domainKey":"视觉小说","pushedAt":"2026-09-23T15:02:39Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/OpenWebGAL/WebGAL","ts":"2026-09-23T15:02:39Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vitejs/vite","owner":"vitejs","name":"vite","desc":"Next generation frontend tooling. It's fast!","summaryCn":"前端神器，秒级启动告别漫长等待，体验真香","reasonCn":"Vite是基于原生ES模块的下一代前端构建工具，利用浏览器原生支持import语法，无需打包即可开发，热更新毫秒级响应。底层采用esbuild预构建依赖，比Webpack快10-100倍，支持Vue、React、Svelte等多种框架，插件体系丰富。","stars":82950,"starGrowth":4,"createdAt":"2020-04-21T05:03:57Z","silentRounds":0,"language":"TypeScript","topics":["frontend"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["前端构建","开发服务器","热更新","esbuild"],"domainKey":"前端构建","pushedAt":"2026-09-23T06:52:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/vitejs/vite","ts":"2026-09-23T06:52:57Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"giancarloerra/SocratiCode","owner":"giancarloerra","name":"SocratiCode","desc":"Enterprise-grade (40m+ LOC) codebase intelligence, zero-setup, local & private Plugin/Skill/Extension or MCP: hybrid semantic search, polyglot dependency graphs, symbol-level impact analysis & call-flow, interactive HTML viewer, cross-project & branch-aware search, DB/API/infra knowledge. 61% less tokens, 84% fewer calls, 37x faster. Cloud in beta.","summaryCn":"读懂整个代码库的智能分析工具，零配置即可用","reasonCn":"SocratiCode是一个企业级代码库智能分析工具，支持零配置本地运行。利用向量数据库进行混合语义搜索，构建多语言依赖关系图，可分析符号级别的代码变更影响。适用于大型代码库，能快速定位代码、理解架构、评估修改风险。4000万行代码的企业级项目也能流畅运行，是开发者的代码智能助手。","stars":3318,"starGrowth":0,"createdAt":"2026-02-26T16:12:26Z","silentRounds":1,"language":"TypeScript","topics":["vector-db"],"aiDims":["向量数据库","AI搜索","代码助手"],"zone":"工具","funScore":0,"domainTags":["代码分析","向量搜索","依赖图","静态分析","代码库"],"domainKey":"代码分析","pushedAt":"2026-09-23T17:51:50Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/giancarloerra/SocratiCode","ts":"2026-09-23T17:51:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"whyour/qinglong","owner":"whyour","name":"qinglong","desc":"支持 Python3、JavaScript、Shell、Typescript 的定时任务管理平台（Timed task management platform supporting Python3, JavaScript, Shell, Typescript）","summaryCn":"很多朋友在网上找了一些很有用的自动化脚本，但每次都要手动在电脑上跑，或","reasonCn":"核心技术上，它构建了一个基于 Node.js 的调度系统，能够灵活地管理不同语言的运行环境。它不仅支持简单的定时触发，还内置了完善的日志记录系统，你可以实时看到脚本运行到了哪一步，哪里报错了。同时，它支持环境变量配置，这意味着你可以把 API 密钥等敏感信息统一管理，而不需","stars":19892,"starGrowth":1,"createdAt":"2021-03-13T15:21:04Z","silentRounds":3,"language":"TypeScript","topics":["docker"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["定时任务","脚本管理","Docker","自动化"],"domainKey":"定时任务","pushedAt":"2026-09-22T14:23:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/whyour/qinglong","ts":"2026-09-22T14:23:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AI-Hypercomputer/maxtext","owner":"AI-Hypercomputer","name":"maxtext","desc":"","summaryCn":"基于纯 Python JAX 的高性能大模型训练库。","reasonCn":"MaxText 是一个专为高性能和可扩展性设计的开源大语言模型库，它使用纯 Python 和 JAX 语言编写，并针对 Google Cloud 的 TPU 和 GPU 硬件进行了优化。该项目提供了一个高性能模型库供用户选择，其中包括 Gemma、Llama、DeepSeek、Qwen 和 Mistral 等知名模型。通过将底层硬件优化与主流的开源模型相结合，MaxText 致力于为开发者提供一个既强大又易于使用的 AI 训练解决方案。","stars":2427,"starGrowth":0,"createdAt":"2023-02-28T19:47:29Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","大语言模型","推理引擎"],"zone":"AI","funScore":0.26,"domainTags":["大语言模型","Python","JAX","云原生"],"domainKey":"大语言模型","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/AI-Hypercomputer/maxtext","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"matiaskorhonen/paper-age","owner":"matiaskorhonen","name":"paper-age","desc":"Easy and secure paper backups of secrets","summaryCn":"在数字时代，把密码写在纸上听起来很复古，但Paper Age却把它做成了最安全的","reasonCn":"技术上，它利用了Rust的高性能和内存安全特性。它采用了高强度的加密算法，将你的加密密钥转换成可视化的“纸面密码”。这个转换过程是不可逆的，你无法从生成的图片反推出原始密码。这意味着即使你的电脑中了木马，黑客也无法解密你的数据。","stars":614,"starGrowth":0,"createdAt":"2023-01-26T14:15:13Z","silentRounds":18,"language":"Rust","topics":["cryptography"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["密码学","Rust","离线安全"],"domainKey":"密码学","pushedAt":"2026-09-22T16:09:27Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"cryptography","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/matiaskorhonen/paper-age","ts":"2026-09-22T16:09:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jihe520/MathModelAgent","owner":"jihe520","name":"MathModelAgent","desc":"🤖📐专为数学建模设计的 Agent &amp; skills ,自动完成数学建模，生成一份完整的可以直接提交的论文。 An Agent Designed for Mathematical Modeling ,Automatically complete mathmodel and generate a complete paper ready for submission.","summaryCn":"数学建模神器！Agent自动搭模型写论文，比赛作业一条龙全包了","reasonCn":"专为数学建模打造的AI Agent，内置多种建模skill，能从题目理解、数据清洗、模型选择到结果可视化全流程自动化。直接输入题目就能生成符合学术规范的完整论文，支持多种数学建模语言。对于数学竞赛、课程论文、科研攻关非常实用，能把几天工作量压缩到几分钟。","stars":6010,"starGrowth":57,"createdAt":"2025-01-30T14:02:20Z","silentRounds":0,"language":"Python","topics":["agent","llm","mathmodel","skills"],"aiDims":["AI Agent","AI写作","AI应用"],"zone":"AI","funScore":1,"domainTags":["数学建模","AI Agent","论文生成","LangChain"],"domainKey":"AI Agent","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mathmodel","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/jihe520/MathModelAgent","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.19945257903494173,"copyOk":true},{"repo":"facebookresearch/chameleon","owner":"facebookresearch","name":"chameleon","desc":"","summaryCn":"Chameleon是Meta FAIR团队发布的一款多模态基础模型，该仓库提供了运行该模","reasonCn":"在运行环境方面，项目明确指向高性能计算需求。文档指出，运行推理组件和输入输出查看器需要配备CUDA兼容的GPU。这意味着用户需要准备 NVIDIA 显卡及相应的驱动环境才能体验其原生推理代码的极速性能。虽然限制了硬件门槛，但这保证了推理过程的高效流畅，符合生产级多模态模型对实","stars":2102,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","多模态Agent","推理引擎"],"zone":"AI","funScore":0.2,"domainTags":["多模态模型","GPU推理","模型评估","Meta FAIR"],"domainKey":"多模态模型","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/facebookresearch/chameleon","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Keralots/BambuHelper","owner":"Keralots","name":"BambuHelper","desc":"ESP32 desk monitor for Bambu Lab printers. Real-time MQTT dashboard on a color TFT - temps, progress, fans, AMS, HMS errors. LAN or Cloud, dual printers, OTA, web config portal. Flash from the browser.","summaryCn":"给Bambu打印机装个屏幕，实时看温度进度超爽","reasonCn":"这是一个专为Bambu Lab 3D打印机设计的ESP32桌面显示器。它通过MQTT协议连接打印机，在TFT彩色屏幕上实时显示打印进度、温度、风扇转速、AMS耗材状态及HMS错误码。支持局域网和云端连接、双机管理、OTA升级和Web配置界面，完全浏览器化操作。","stars":459,"starGrowth":0,"createdAt":"2026-03-07T16:27:12Z","silentRounds":2,"language":"C","topics":["arduino"],"aiDims":["开发工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["ESP32","3D打印","MQTT"],"domainKey":"ESP32","pushedAt":"2026-09-19T08:50:40Z","tags":[{"name":"开发工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/Keralots/BambuHelper","ts":"2026-09-19T08:50:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"whiteguo233/OpenBiliClaw","owner":"whiteguo233","name":"OpenBiliClaw","desc":"本地私有、开源的自进化跨平台 AI 内容发现 Agent：先理解你，再主动从 B站、小红书、抖音、YouTube、X、知乎、Reddit、微博等平台与开放 Web 寻找内容。（支持 deepseek harness 插件） | Local-first open-source cross-platform AI content discovery agent: understands you, then proactively finds content across Bilibili, Xiaohongshu, Douyin, YouTube, X, Zhihu, Reddit, Weibo and the open web.（support deepseek harness plugin）","summaryCn":"像个私人助理，帮你全网找好内容并自动整理","reasonCn":"这是一个基于大模型的自进化内容发现Agent，专为解决信息过载设计。它首先通过深度理解你的兴趣偏好，然后主动从B站、小红书、抖音、YouTube、X、知乎、Reddit、微博等主流平台及开放Web抓取数据，利用RAG技术生成个性化内容摘要。项目集成了DeepSeek Harness插件，支持本地私有化部署，确保数据安全且完全免费。","stars":3344,"starGrowth":1,"createdAt":"2026-03-07T13:46:10Z","silentRounds":2,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","AI搜索","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","内容聚合","RAG","社交媒体","本地部署"],"domainKey":"AI Agent","pushedAt":"2026-09-18T17:35:16Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/whiteguo233/OpenBiliClaw","ts":"2026-09-18T17:35:16Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kawaiiDango/pano-scrobbler","owner":"kawaiiDango","name":"pano-scrobbler","desc":"Feature rich scrobbler for Windows, Linux & Android. Supports Last.fm, ListenBrainz, Libre.fm & Pleroma. With regex edits, charts & Discord Rich Presence on PC.","summaryCn":"想象一下，你每天在手机、电脑和各种播放器间来回切换，听了无数首好歌，却","reasonCn":"从技术架构来看，它采用了轻量级的后台驻留机制，能够实时捕获不同系统的媒体播放通知，通过高效的正则匹配引擎对元数据进行二次清洗和归一化处理。这种处理方式比传统的插件更稳定，因为它直接对接操作系统的多媒体框架，从而确保无论你是听Spotify、YouTube Music还是本地音乐，","stars":2295,"starGrowth":0,"createdAt":"2017-07-04T20:21:51Z","silentRounds":2,"language":"Kotlin","topics":["music"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["音乐库管理","Last.fm 同步","跨平台统计","媒体元数据"],"domainKey":"音乐库管理","pushedAt":"2026-09-23T13:20:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/kawaiiDango/pano-scrobbler","ts":"2026-09-23T13:20:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nanochess/bootOS","owner":"nanochess","name":"bootOS","desc":"","summaryCn":"512字节极限OS，兼容8088经典平台，展示汇编极致艺术","reasonCn":"bootOS 是一个将完整操作系统内核压缩进 512 字节启动扇区的极限编程项目。它专为兼容原始 IBM PC 平台的 8088 处理器设计，能在极小内存中加载、执行和保存程序，并维护简单的文件系统。这不仅是一个技术挑战，更是对汇编语言与底层硬件控制的极致展示，适合对计算机底层原理感兴趣的极客体验复古计算的奇妙世界。","stars":2141,"starGrowth":0,"silentRounds":0,"language":"Assembly","topics":[],"aiDims":["非AI-好玩"],"zone":"创意","funScore":0.5599999999999999,"domainTags":["操作系统","汇编语言","极限编程","复古计算"],"domainKey":"操作系统","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Assembly","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/nanochess/bootOS","ts":"2026-09-22T22:51:50.275Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"break-into-data/ai-engineer-toolkit","owner":"break-into-data","name":"ai-engineer-toolkit","desc":"","summaryCn":"面向工程师的AI应用构建指南，涵盖从基础到生产环境的框架与资源清单。","reasonCn":"这是一个帮助软件工程师和非技术人员从零开始构建生产级AI应用的工具包。它整理了当前流行的AI框架、工具和资源，旨在让开发者能够一步步交付AI项目。无论是熟悉Agent工程基础，还是掌握全栈Agent开发，这里都有对应的学习路径和示例。特别适合想要系统性提升AI工程能力，或者想将AI应用落地到生产环境的开发者和团队。","stars":2189,"starGrowth":0,"silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["AI工程","Agent开发","软件架构"],"domainKey":"AI工程","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/break-into-data/ai-engineer-toolkit","ts":"2026-09-22T22:51:50.275Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zed-industries/zed","owner":"zed-industries","name":"zed","desc":"Code at the speed of thought – Zed is a high-performance, multiplayer code editor from the creators of Atom and Tree-sitter.","summaryCn":"Atom原班人马打造的高性能多人编辑器，打字如飞","reasonCn":"由Atom和Tree-sitter的原作者们开发的新一代代码编辑器，主打极速性能和多人实时协作。基于Rust编写，原生支持LSP，解决了传统编辑器卡顿的痛点。虽然核心非AI，但其内置的AI辅助功能和流畅体验使其成为开发者提升效率的强力工具，适合作为VS Code的现代化替代品。","stars":90728,"starGrowth":17,"createdAt":"2021-02-20T03:01:06Z","silentRounds":0,"language":"Rust","topics":["text-editor","zed","gpui","rust-lang"],"aiDims":["开发者工具","效率工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["代码编辑器","多人协作","Rust","跨平台"],"domainKey":"代码编辑器","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"text-editor","source":"github","weight":0.7},{"name":"zed","source":"github","weight":0.7},{"name":"gpui","source":"github","weight":0.7},{"name":"rust-lang","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/zed-industries/zed","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05156211974252712,"copyOk":false},{"repo":"TecharoHQ/anubis","owner":"TecharoHQ","name":"anubis","desc":"Weighs the soul of incoming HTTP requests to stop AI crawlers","summaryCn":"用灵魂称重来拦AI爬虫，给网站穿上隐形铠甲","reasonCn":"开源AI爬虫防护工具，通过JavaScript挑战在HTTP请求阶段识别并拦截自动化AI爬虫。核心机制是让AI爬虫完成需要人类级推理才能解开的智能挑战，而非简单验证码。支持Docker一键部署，兼容Nginx和Traefik等反向代理，几行配置即可为网站披上防护外衣，有效抵御AI工具爬取数据。","stars":22609,"starGrowth":10,"createdAt":"2025-03-17T17:35:28Z","silentRounds":0,"language":"Go","topics":["security"],"aiDims":["安全工具","开发者工具","AI基础设施"],"zone":"工具","funScore":0.8,"domainTags":["AI爬虫防护","HTTP中间件","JavaScript挑战","安全防护"],"domainKey":"AI爬虫防护","pushedAt":"2026-09-21T00:38:31Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/TecharoHQ/anubis","ts":"2026-09-21T00:38:31Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.031326905214737494,"copyOk":false},{"repo":"microsoft/AI-For-Beginners","owner":"microsoft","name":"AI-For-Beginners","desc":"12 Weeks, 24 Lessons, AI for All!","summaryCn":"微软出品的 AI 入门教程，12周24课从零学机器学习不花一分钱","reasonCn":"简单来说，这是微软专门为完全没接触过 AI 的人设计的一套完整课程，相当于免费的 AI 大学预科班。不需要你有任何数学或编程基础，从最基础的概念讲起，一步一步带你走完机器学习、深度学习、自然语言处理、计算机视觉和强化学习五大模块的入门之旅。","stars":68879,"starGrowth":28,"createdAt":"2021-03-03T16:27:36Z","silentRounds":0,"language":"Jupyter Notebook","topics":["ml","ai"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["Machine Learning","Deep Learning","Course"],"domainKey":"Machine Learning","pushedAt":"2026-09-16T20:17:56Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/microsoft/AI-For-Beginners","ts":"2026-09-16T20:17:56Z","category":"learning","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.08521952990025988,"copyOk":false},{"repo":"velero-io/velero","owner":"velero-io","name":"velero","desc":"Backup and migrate Kubernetes applications and their persistent volumes","summaryCn":"给K8s集群装个自动备份系统，数据安全不用愁","reasonCn":"这是一个Kubernetes灾难恢复解决方案，核心功能是备份和恢复集群应用及持久卷。它将K8s集群状态、应用配置和重要数据打包成快照，并存储到兼容S3的对象存储中。通过插件机制支持AWS、Azure等多种云平台，提供快照、复制、恢复等完整数据生命周期管理功能，是运维团队的必备工具。","stars":10309,"starGrowth":1,"createdAt":"2017-08-02T17:22:11Z","silentRounds":1,"language":"Go","topics":["kubernetes"],"aiDims":["非AI-实用","开发工具","安全工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","数据备份","灾难恢复"],"domainKey":"Kubernetes","pushedAt":"2026-09-23T13:59:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/velero-io/velero","ts":"2026-09-23T13:59:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cytostack/openwolf","owner":"cytostack","name":"openwolf","desc":"","summaryCn":"为AI编程助手提供持久化项目记忆，避免重复上下文消耗。","reasonCn":"这个项目的核心亮点在于它的记忆和索引功能完全在本地运行，不需要依赖额外的模型调用。同时它还会统计不同AI代理和模型在会话中消耗了多少令牌，帮助开发者了解token花费情况。对于经常使用AI编程助手的开发者来说，这能减少很多重复的上下文提示工作。","stars":2282,"starGrowth":0,"silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手","AI应用"],"zone":"AI","funScore":0.32,"domainTags":["AI编程助手","上下文管理","任务记忆","本地存储"],"domainKey":"AI编程助手","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/cytostack/openwolf","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"simonlin1212/a-stock-data","owner":"simonlin1212","name":"a-stock-data","desc":"A股全栈数据工具包 · 10层架构 · 43端点(含3官方备胎) · 15数据源 · 行情/研报/资金面/筹码/公告/打板/ETF期权/舆情互动全覆盖+备用源降级 | China A-Share full-stack data toolkit (43 endpoints)","summaryCn":"AI投资助手必备！一键接入A股全量数据的神级工具包","reasonCn":"为AI Agent构建A股市场数据获取能力而生的全栈工具包，采用10层分层架构，提供43个标准化API端点，覆盖行情、研报、资金流向、筹码分布、龙虎榜、ETF期权和舆情互动等全场景数据。内置15个数据源和官方备用源自动降级机制，保障数据获取稳定性，是搭建智能投顾和股票研究Agent的核心基础设施。","stars":10090,"starGrowth":26,"createdAt":"2026-05-11T01:08:15Z","silentRounds":0,"language":"","topics":["ai-agent"],"aiDims":["AI Agent","工具调用"],"zone":"工具","funScore":0,"domainTags":["A股数据","金融数据","股票分析","API接口"],"domainKey":"A股数据","pushedAt":"2026-09-22T16:11:05Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7}],"aiScore":0.6,"source":"search","url":"https://github.com/simonlin1212/a-stock-data","ts":"2026-09-22T16:11:05Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.1028765113974232,"copyOk":true},{"repo":"kaixxx/noScribe","owner":"kaixxx","name":"noScribe","desc":"Cutting edge AI technology for automated audio transcription. A nice GUI for OpenAIs Whisper and pyannote (speaker identification)","summaryCn":"把录音秒变精美文字稿的AI转录神器，自带说话人识别","reasonCn":"技术核心是组合了目前最强的开源音频方案：使用 OpenAI 的 Whisper 模型负责语音到文本的转换，确保极高的识别准确率；同时引入 pyannote.audio 进行说话人日志（Speaker Diarization）分析。通过这种组合，它实现了从原始音频到结构化文本的完整闭环，且通过 GUI 界面屏蔽了复杂的","stars":2171,"starGrowth":2,"createdAt":"2023-05-12T06:25:03Z","silentRounds":0,"language":"Python","topics":["interview"],"aiDims":["AI应用","文档理解","效率工具"],"zone":"AI","funScore":0,"domainTags":["语音转录","音频处理","Whisper"],"domainKey":"语音转录","pushedAt":"2026-09-19T16:43:19Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/kaixxx/noScribe","ts":"2026-09-19T16:43:19Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"huggingface/optimum-habana","owner":"huggingface","name":"optimum-habana","desc":"Easy and lightning fast training of 🤗 Transformers on Habana Gaudi processor (HPU)","summaryCn":"想象一下，你正在用昂贵的显卡跑模型，突然发现有个专门的芯片能让你的训练","reasonCn":"它的核心亮点在于深度整合了Habana PyTorch集成层，通过高度优化的算子库将Tensor运算映射到HPU上。与传统的CUDA环境不同，它通过图形编译技术将模型计算图提前优化，减少了CPU与加速器之间的通信开销。这种架构设计保证了在处理大规模语言模型时，显存管理和计算密度都能达到行业","stars":213,"starGrowth":0,"createdAt":"2022-03-01T13:43:02Z","silentRounds":20,"language":"Python","topics":["transformers"],"aiDims":["AI基础设施","模型与训练"],"zone":"AI","funScore":0,"domainTags":["模型训练","Habana Gaudi","硬件加速","PyTorch"],"domainKey":"模型训练","pushedAt":"2026-09-23T14:17:28Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/huggingface/optimum-habana","ts":"2026-09-23T14:17:28Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"andreas128/RePaint","owner":"andreas128","name":"RePaint","desc":"","summaryCn":"RePaint 是一个开创性的图像修复工具，其核心机制是应用去噪扩散概率模型填","reasonCn":"为了保证修复的可信度，它预训练了多个针对不同数据集的模型，包括复杂的自然风景、标准的城市街景以及高精度的人脸特征。这意味着无论是恢复人物面部的光影细节，还是重建建筑结构的几何连贯性，模型都展现出了极高的保真度。","stars":2269,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","AI应用"],"zone":"AI","funScore":0.38,"domainTags":["图像修复","生成式AI","扩散模型","计算机视觉"],"domainKey":"图像修复","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/andreas128/RePaint","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"samchon/nestia","owner":"samchon","name":"nestia","desc":"NestJS Helper + AI Chatbot Development","summaryCn":"NestJS开发者的AI辅助神器，一键生成类型安全的前后端","reasonCn":"Nestia专为NestJS框架打造，能根据后端DTO自动生成前端TypeScript类型和HTTP客户端代码。对于做AI应用开发来说，它能大幅减少前后端联调时的类型错误，特别适合需要频繁调用LLM API的场景，让AI助手类应用开发更丝滑高效。","stars":2180,"starGrowth":0,"createdAt":"2021-01-20T09:12:03Z","silentRounds":4,"language":"TypeScript","topics":["llm-model"],"aiDims":["AI应用","工具调用","代码助手"],"zone":"工具","funScore":0,"domainTags":["代码生成","TypeScript","NestJS","API 开发","类型安全"],"domainKey":"代码生成","pushedAt":"2026-09-23T21:23:32Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/samchon/nestia","ts":"2026-09-23T21:23:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yegor256/takes","owner":"yegor256","name":"takes","desc":"True Object-Oriented Java Web Framework without NULLs, Static Methods, Annotations, and Mutable Objects","summaryCn":"给Java代码来个彻底的大扫除，告别垃圾代码","reasonCn":"Java 开发经常被繁琐的代码折磨，而 Takes 框架主打“零冗余”。它彻底抛弃了传统的面向过程和可变对象设计，要求所有方法返回新对象而非修改原对象，从而彻底消除空指针异常。同时，它也不依赖注解和静态方法，让依赖注入和单元测试变得极其简单。","stars":878,"starGrowth":0,"createdAt":"2015-01-07T16:57:47Z","silentRounds":25,"language":"Java","topics":["web-framework"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Java","Web框架","面向对象编程","不可变对象","依赖注入"],"domainKey":"Java","pushedAt":"2026-09-23T09:52:11Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/yegor256/takes","ts":"2026-09-23T09:52:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EvoMap/evolver","owner":"EvoMap","name":"evolver","desc":"The GEP-powered self-evolving engine for AI agents. Auditable evolution with Genes, Capsules, and Events. | evomap.ai","summaryCn":"让AI智能体像生物一样自我进化，基因胶囊事件可审计","reasonCn":"这是一个基于GEP（基因表达编程）的AI Agent自进化引擎。智能体不再是一次性训练的静态模型，而是像生物一样，通过基因、胶囊和事件记录来不断进化。每次进化都有据可查，特别适合需要持续学习、自我优化的复杂AI系统。","stars":9112,"starGrowth":1,"createdAt":"2026-02-01T05:59:24Z","silentRounds":1,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","AI基础设施","推理引擎"],"zone":"AI","funScore":0,"domainTags":["AI Agent","基因表达编程","自进化系统","模型优化"],"domainKey":"AI Agent","pushedAt":"2026-09-18T17:22:26Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/EvoMap/evolver","ts":"2026-09-18T17:22:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fmhy/edit","owner":"fmhy","name":"edit","desc":"Make changes to FMHY","summaryCn":"汇聚全网优质免费资源和实用干货的超大百宝箱项目","reasonCn":"它的核心特点是内容极度全面且更新迅速。通过精细的目录结构，把教育、开发、多媒体、隐私保护、阅读等几十个领域的优质资源整合在一起。它本身不是一个软件，而是一个巨大的资源清单源文件，采用Markdown格式编写，任何人都可以提交修改和补充。","stars":12010,"starGrowth":15,"createdAt":"2021-08-09T00:49:24Z","silentRounds":0,"language":"JavaScript","topics":["freemediaheckyeah","fmhy"],"aiDims":["非AI-实用","学习资源"],"zone":"资源","funScore":0,"domainTags":["免费资源","软件导航","互联网宝藏","资源清单"],"domainKey":"免费资源","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"freemediaheckyeah","source":"github","weight":0.7},{"name":"fmhy","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/fmhy/edit","ts":"2026-09-22T04:07:30.369Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.048746877601998334,"copyOk":false},{"repo":"tonhowtf/omniget","owner":"tonhowtf","name":"omniget","desc":"Download Udemy and Hotmart courses, YouTube videos, music and books — 1,800+ sites, no terminal. Free open-source desktop app for Windows, macOS and Linux, with a built-in course player, PDF/EPUB reader and music library. Powered by yt-dlp. Your files stay on your computer.","summaryCn":"万能下载神器，把网课、音乐、电子书全搞定","reasonCn":"这是一款跨平台的桌面软件，号称支持1800多个网站的内容下载，包括Udemy网课、Hotmart课程、YouTube视频、音乐和电子书。它完全基于Rust开发，无需打开终端，内置了阅读器，支持PDF和EPUB格式，适合不想折腾命令行的用户。","stars":14192,"starGrowth":34,"createdAt":"2026-02-11T22:44:56Z","silentRounds":0,"language":"Rust","topics":["download-manager","instagram-downloader","media-downloader","reddit-downloader","tiktok-downloader","twitter-downloader","video-downloader","youtube-downloader","hotmart-downloader","udemy-downloader","yt-dlp","bilibili-downloader","course-downloader","downloader","yt-dlp-gui","epub-reader","spaced-repetition","telegram-downloader","twitch-downloader","subtitle-downloader"],"aiDims":["非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["课程下载","视频下载","电子书","桌面应用"],"domainKey":"课程下载","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"download-manager","source":"github","weight":0.7},{"name":"instagram-downloader","source":"github","weight":0.7},{"name":"media-downloader","source":"github","weight":0.7},{"name":"reddit-downloader","source":"github","weight":0.7},{"name":"tiktok-downloader","source":"github","weight":0.7},{"name":"twitter-downloader","source":"github","weight":0.7},{"name":"video-downloader","source":"github","weight":0.7},{"name":"youtube-downloader","source":"github","weight":0.7},{"name":"hotmart-downloader","source":"github","weight":0.7},{"name":"udemy-downloader","source":"github","weight":0.7},{"name":"yt-dlp","source":"github","weight":0.7},{"name":"bilibili-downloader","source":"github","weight":0.7},{"name":"course-downloader","source":"github","weight":0.7},{"name":"downloader","source":"github","weight":0.7},{"name":"yt-dlp-gui","source":"github","weight":0.7},{"name":"epub-reader","source":"github","weight":0.7},{"name":"spaced-repetition","source":"github","weight":0.7},{"name":"telegram-downloader","source":"github","weight":0.7},{"name":"twitch-downloader","source":"github","weight":0.7},{"name":"subtitle-downloader","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/tonhowtf/omniget","ts":"2026-09-17T04:13:51.053Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.13102457722660654,"copyOk":false},{"repo":"KylinMountain/TradingAgents-AShare","owner":"KylinMountain","name":"TradingAgents-AShare","desc":"A 股多智能体智能投研系统 — 基于 TradingAgents 架构，15 名 AI Agent 模拟机构协作与实时辩论对抗，全流程可视化，支持 OpenClaw / Claude Code 集成，Docker 一键部署。","summaryCn":"如果你对股市的决策过程充满好奇，又觉得普通的炒股软件太枯燥，那这个项目","reasonCn":"技术架构上，该项目采用了高度可扩展的多智能体协作框架，每个Agent都具备独立的决策逻辑和知识库，能够进行跨层级的联动交流。它巧妙地集成了可视化看板，让你能实时看到每一个Agent在思考什么、为什么对某只股票产生争议。","stars":839,"starGrowth":0,"createdAt":"2026-03-02T03:33:41Z","silentRounds":2,"language":"Python","topics":["finance"],"aiDims":["AI Agent","AI应用","可视化"],"zone":"AI","funScore":0.8,"domainTags":["量化交易","多智能体系统","A股投研","金融AI Agent"],"domainKey":"量化交易","pushedAt":"2026-09-18T05:06:33Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/KylinMountain/TradingAgents-AShare","ts":"2026-09-18T05:06:33Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"arianrhodsandlot/nostalgist","owner":"arianrhodsandlot","name":"nostalgist","desc":"A JavaScript library used for running emulators of retro consoles inside browsers.","summaryCn":"想象一下，你正在浏览一个普通的网页，突然能直接玩起小时候卡带机里的经典","reasonCn":"技术层面上，它巧妙地结合了WebAssembly和各种主流开源模拟器核心，把原本需要独立客户端才能跑起来的模拟器塞进了浏览器沙盒中。它支持自动加载游戏ROM、处理复杂的音频输出、渲染高帧率画面，并且对键盘和各种外设手柄提供了极佳的兼容性，确保复古游戏在现代浏览器里也能流畅运","stars":986,"starGrowth":0,"createdAt":"2023-10-03T13:09:21Z","silentRounds":17,"language":"TypeScript","topics":["emulator"],"aiDims":["非AI-好玩","开发者工具"],"zone":"创意","funScore":1,"domainTags":["复古模拟器","WebAssembly","浏览器游戏"],"domainKey":"复古模拟器","pushedAt":"2026-09-20T07:25:24Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/arianrhodsandlot/nostalgist","ts":"2026-09-20T07:25:24Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"farion1231/cc-switch","owner":"farion1231","name":"cc-switch","desc":"A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build &amp; Hermes Agent. Only official website: ccswitch.io","summaryCn":"一个把所有AI编程助手塞进一个桌面的神仙工具，再也不用切换窗口了","reasonCn":"用Rust构建的跨平台桌面聚合工具，支持Claude Code、Codex、OpenCode、OpenClaw、Grok Build等多个AI编程助手统一管理。提供可视化界面控制各AI工具的运行状态、对话历史和配置参数，支持一键切换不同AI助手，极大提升多模型协作开发效率。","stars":134194,"starGrowth":51,"createdAt":"2025-08-04T14:16:16Z","silentRounds":0,"language":"Rust","topics":["ai-tools","claude-code","desktop-app","open-source","rust","tauri","codex","mcp","provider-management","wsl-support","opencode","skills","skills-management","openclaw","openclaw-ui","hermes","hermes-agent","grok","grokbuild","pi"],"aiDims":["代码助手","AI Agent","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["AI编程助手","开发效率","多模型管理","Rust","桌面应用"],"domainKey":"AI编程助手","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai-tools","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"desktop-app","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"tauri","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"provider-management","source":"github","weight":0.7},{"name":"wsl-support","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"skills-management","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"openclaw-ui","source":"github","weight":0.7},{"name":"hermes","source":"github","weight":0.7},{"name":"hermes-agent","source":"github","weight":0.7},{"name":"grok","source":"github","weight":0.7},{"name":"grokbuild","source":"github","weight":0.7},{"name":"pi","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/farion1231/cc-switch","ts":"2026-09-11T22:17:29.879Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.15414014039375828,"copyOk":false},{"repo":"xuanyustudio/LocalMiniDrama","owner":"xuanyustudio","name":"LocalMiniDrama","desc":"🎬 seedance2接入 开源本地 AI 短剧 &amp; 漫剧生成工具 —— 从故事到成片一站式完成，数据不出本机，短剧工作流管理平台，高灵活度，AI真人剧，AI漫剧本地搞定。 Open-source local AI short drama maker: story → storyboard → video, fully offline, your data stays yours. 纳米流水线","summaryCn":"本地生成AI短剧的神器，故事剧本一键变视频。","reasonCn":"LocalMiniDrama是一款基于Stable Diffusion的本地化短剧生成平台，用户无需联网即可完成从文本故事到成片的制作。项目集成了Story-to-Storyboard和Story-to-Video工作流，支持高灵活度的角色与场景定制，确保本地数据安全。","stars":1815,"starGrowth":7,"createdAt":"2026-02-10T07:44:18Z","silentRounds":0,"language":"JavaScript","topics":["ai-mini-drama","offline-short-drama","ai-agent","short-video","short-video-maker","mini-drama","ai-video","storyboard","video-generation","ai","electron","javascript","nodejs","offline","open-source","vue","vue3"],"aiDims":["AI应用","多模态Agent","创意工具"],"zone":"创意","funScore":1,"domainTags":["AI视频生成","短剧制作","本地部署"],"domainKey":"AI视频生成","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai-mini-drama","source":"github","weight":0.7},{"name":"offline-short-drama","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"short-video","source":"github","weight":0.7},{"name":"short-video-maker","source":"github","weight":0.7},{"name":"mini-drama","source":"github","weight":0.7},{"name":"ai-video","source":"github","weight":0.7},{"name":"storyboard","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"offline","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"vue","source":"github","weight":0.7},{"name":"vue3","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/xuanyustudio/LocalMiniDrama","ts":"2026-09-18T10:23:42.801Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.03908429752066116,"copyOk":false},{"repo":"OpenCSGs/csghub","owner":"OpenCSGs","name":"csghub","desc":"CSGHub is a brand-new open-source platform for managing LLMs, developed by the OpenCSG team. It offers both open-source and on-premise/SaaS solutions, with features comparable to Hugging Face. Gain full control over the lifecycle of LLMs, datasets, and agents, with Python SDK compatibility with Hugging Face. Join us! ⭐️","summaryCn":"像 Hugging Face 一样管理模型，支持私有部署。","reasonCn":"从架构上看，CSGHub 采用了现代化的前端架构（Vue.js）和后端服务，支持用户友好的 Web 界面。它不仅是一个模型仓库，还集成了在线推理能力，用户可以直接在网页上测试模型效果。其核心功能包括模型版本管理、数据集管理和推理 API 服务，这些功能都是 Hugging Face 的核心功能的平","stars":4112,"starGrowth":0,"createdAt":"2024-01-12T09:44:48Z","silentRounds":5,"language":"Vue","topics":["git"],"aiDims":["AI基础设施","模型部署","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["模型管理","私有部署","HuggingFace替代"],"domainKey":"模型管理","pushedAt":"2026-09-19T03:32:15Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/OpenCSGs/csghub","ts":"2026-09-19T03:32:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fastai/fastai","owner":"fastai","name":"fastai","desc":"The fastai deep learning library","summaryCn":"用几行代码就能训练AI的神器，深度学习从此不痛苦","reasonCn":"基于PyTorch构建的深度学习库，抽象了大量训练细节，用高层API封装了学习率调度、混合精度等复杂技术。提供开箱即用的数据管道、模型构建和训练循环，配合完善的课程资源，让零基础用户也能快速上手训练图像分类、文本处理等任务。","stars":28188,"starGrowth":1,"createdAt":"2017-09-09T17:43:36Z","silentRounds":1,"language":"Jupyter Notebook","topics":["ml"],"aiDims":["模型与训练","微调","教程"],"zone":"AI","funScore":0,"domainTags":["深度学习","PyTorch","机器学习"],"domainKey":"深度学习","pushedAt":"2026-09-21T10:41:46Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/fastai/fastai","ts":"2026-09-21T10:41:46Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"knightcrawler25/GLSL-PathTracer","owner":"knightcrawler25","name":"GLSL-PathTracer","desc":"","summaryCn":"这是一个将复杂的物理路径追踪算法塞进显卡片元着色器运行的轻量级渲染项目","reasonCn":"在功能特性上，该项目不仅支持常见的全套纹理贴图映射，还引入了多重重要性采样和随机 alpha 测试，甚至对 MagicaVoxel 风格的同质体积提供了支持。这些丰富的特性让它不仅仅是一个简单的代码玩具，而是一个具备实际参考价值的底层图形学实现。","stars":2058,"starGrowth":0,"silentRounds":0,"language":"C++","topics":[],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":0.62,"domainTags":["路径追踪","图形学","GLSL着色器","物理渲染"],"domainKey":"路径追踪","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/knightcrawler25/GLSL-PathTracer","ts":"2026-09-23T18:23:13.861Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"syxanash/awesome-web-desktops","owner":"syxanash","name":"awesome-web-desktops","desc":"Websites, web apps, portfolios that look like desktop operating systems","summaryCn":"把网页变成桌面系统，刷新浏览器就有新玩趣","reasonCn":"这个仓库收集了大量把网站、作品集或应用界面做成操作系统桌面的开源项目。它不是AI工具，但设计感极强，互动体验新颖。用纯前端技术模拟窗口、任务栏，让人在浏览器里体验到“第二空间”的乐趣，非常适合喜欢折腾网页设计和追求新鲜感的开发者与设计师收藏。","stars":2090,"starGrowth":2,"createdAt":"2019-10-08T20:49:41Z","silentRounds":0,"language":"Ruby","topics":["design","desktop"],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"资源","funScore":0,"domainTags":["Web桌面","前端设计","拟物化","创意编程"],"domainKey":"Web桌面","pushedAt":"2026-09-23T18:22:27Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/syxanash/awesome-web-desktops","ts":"2026-09-23T18:22:27Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ycm-core/ycmd","owner":"ycm-core","name":"ycmd","desc":"A code-completion & code-comprehension server","summaryCn":"经久不衰、全能且极速的底层代码自动补全与理解服务器。","reasonCn":"在底层技术实现上，它采用高性能的C++编写服务器核心，利用Clang等成熟的编译器前端来解析抽象语法树，从而提供精准到符号级别的语义理解。它采用客户端-服务器架构，编辑器作为前端只负责展示，所有的重度计算和索引都在后台这个纯C++引擎中飞速运转，内存占用低且响应极其迅速。","stars":1741,"starGrowth":0,"createdAt":"2014-05-15T19:30:17Z","silentRounds":18,"language":"C++","topics":["editor"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["代码补全","LSP","Vim插件","C++引擎"],"domainKey":"代码补全","pushedAt":"2026-09-23T05:27:47Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/ycm-core/ycmd","ts":"2026-09-23T05:27:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ComposioHQ/awesome-claude-skills","owner":"ComposioHQ","name":"awesome-claude-skills","desc":"A curated list of awesome Claude Skills, resources, and tools for customizing Claude AI workflows","summaryCn":"Claude AI的瑞士军刀库，内置技能让大模型干啥都行","reasonCn":"这是一个精心整理的 Claude Skills 精选列表，Skills 是 Claude 的扩展机制，相当于给 AI 装配各种专业工具。仓库收集了大量现成的 Skill 配置和教程，涵盖代码编写、数据分析、创意写作等场景。通过加载这些 Skill，Claude 能在特定领域表现得像专家一样精准，大幅扩展了 AI 的工作能力边界。","stars":75494,"starGrowth":23,"createdAt":"2025-10-17T07:15:01Z","silentRounds":0,"language":"Python","topics":["claude","claude-code","agent-skills","ai-agents","antigravity","automation","codex","composio","cursor","gemini-cli","mcp","rube","saas","skill","workflow-automation","developer-tools","openai-codex"],"aiDims":["AI Agent","工具调用","提示工程"],"zone":"资源","funScore":0,"domainTags":["Claude Skills","AI工作流","提示工程","资源导航"],"domainKey":"Claude Skills","pushedAt":"2026-09-18T07:16:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"composio","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"gemini-cli","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"rube","source":"github","weight":0.7},{"name":"saas","source":"github","weight":0.7},{"name":"skill","source":"github","weight":0.7},{"name":"workflow-automation","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"openai-codex","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/ComposioHQ/awesome-claude-skills","ts":"2026-09-18T07:16:11Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08389677590272075,"copyOk":false},{"repo":"hepengwei/visualization-collection","owner":"hepengwei","name":"visualization-collection","desc":"🌈 一个专注于前端视觉效果的集合应用，包含CSS动效、Canvas动画、Three.js3D、人工智能应用等上百个案例（持续更新）","summaryCn":"前端视觉效果百宝箱，上百个AI动效案例直接抄","reasonCn":"这个仓库收集了上百个前端视觉效果案例，涵盖CSS动画、Canvas绘图、Three.js 3D场景，还有AI相关的视觉效果应用。每个案例都有源码和在线演示，按类别整理得清清楚楚。虽然是纯前端项目，但其中包含的AI可视化案例很有意思，比如用Three.js做的3D模型展示、神经网络可视化等，对想了解AI如何呈现给普通用户的人来说是很好的学习材料。","stars":4593,"starGrowth":0,"createdAt":"2022-10-20T06:44:36Z","silentRounds":5,"language":"TypeScript","topics":["frontend"],"aiDims":["AI应用","可视化","创意工具"],"zone":"资源","funScore":1,"domainTags":["CSS动画","Canvas","Three.js","前端可视化"],"domainKey":"CSS动画","pushedAt":"2026-09-23T10:12:45Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/hepengwei/visualization-collection","ts":"2026-09-23T10:12:45Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JustArchiNET/ArchiSteamFarm","owner":"JustArchiNET","name":"ArchiSteamFarm","desc":"C# application with primary purpose of farming Steam cards from multiple accounts simultaneously.","summaryCn":"多账号一键挂机刷Steam卡牌，解放双手赚钱包的神器","reasonCn":"它用C#写成自动化脚本，支持同时登多个Steam账号，自动处理登录、验证和抓取卡牌。虽然不涉及人工智能，但逻辑极强，堪称数字生活管家。对于收藏党、低价区玩家而言，它是必备的效率杠杆，省去大量重复点击的时间成本。","stars":13697,"starGrowth":0,"createdAt":"2015-10-25T05:12:33Z","silentRounds":1,"language":"C#","topics":["automation"],"aiDims":["非AI-好玩","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Steam","自动化","刷卡","多账号管理"],"domainKey":"Steam","pushedAt":"2026-09-23T04:42:03Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/JustArchiNET/ArchiSteamFarm","ts":"2026-09-23T04:42:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"insin/control-panel-for-twitter","owner":"insin","name":"control-panel-for-twitter","desc":"Browser extension which gives you more control over your Twitter timeline and adds missing features and UI improvements - for desktop and mobile","summaryCn":"还你一个清爽的推特！强力清理垃圾信息，定制属于你的高效沉浸式阅读界面。","reasonCn":"这是一款功能极为强大的浏览器插件，能够精准去除推特时间线中的广告推荐、关注建议以及令人焦虑的各种干扰项。它赋予用户高度的定制自由，让你可以隐藏特定的UI元素、调整布局，甚至找回那些被强制改变的原始功能，是重度信息筛选者的必备利器，用过就再也回不去原版了。","stars":2627,"starGrowth":0,"createdAt":"2019-07-23T05:22:58Z","silentRounds":9,"language":"JavaScript","topics":["productivity"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Twitter定制","浏览器扩展","DOM注入","信息过滤"],"domainKey":"Twitter定制","pushedAt":"2026-09-22T06:38:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/insin/control-panel-for-twitter","ts":"2026-09-22T06:38:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yantrikos/yantrikdb-server","owner":"yantrikos","name":"yantrikdb-server","desc":"Cognitive memory database for AI agents — consolidates duplicates, detects contradictions, fades stale memories via temporal decay. Rust, Apache-2.0, ships as library / MCP server / HTTP cluster.","summaryCn":"一个先进的向量数据库，用于AI应用中的数据存储和管理。","reasonCn":"这个系统采用Rust编程语言编写，能够有效地处理大量数据，并支持高效的数据索引和查询。","stars":174,"starGrowth":0,"createdAt":"2026-04-05T18:08:33Z","silentRounds":3,"language":"Rust","topics":["vector-db"],"aiDims":["vector-db"],"zone":"AI","funScore":0.5,"domainTags":["向量数据库","Rust","AI Agent","MCP Server"],"domainKey":"向量数据库","pushedAt":"2026-09-19T20:58:39Z","tags":[{"name":"vector-db","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/yantrikos/yantrikdb-server","ts":"2026-09-19T20:58:39Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Vinzent03/obsidian-git","owner":"Vinzent03","name":"obsidian-git","desc":"Integrate Git version control with automatic commit-and-sync and other advanced features in Obsidian.md","summaryCn":"让你的笔记软件拥有像代码一样强大的版本管理","reasonCn":"该插件将 Git 的版本控制能力深度集成到 Obsidian 中，支持自动提交与同步，可通过配置定时任务实现笔记的无感备份。它解决了知识库在多端同步时的冲突问题，让用户能随时回溯任何一个历史修改版本，确保笔记绝对安全。","stars":12025,"starGrowth":3,"createdAt":"2020-10-27T16:08:16Z","silentRounds":0,"language":"TypeScript","topics":["git"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Git同步","笔记备份","Obsidian插件","版本控制","自动化"],"domainKey":"Git同步","pushedAt":"2026-09-23T21:47:54Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/Vinzent03/obsidian-git","ts":"2026-09-23T21:47:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kepano/obsidian-skills","owner":"kepano","name":"obsidian-skills","desc":"Agent skills for Obsidian. Teach your agent to use Obsidian CLI and open formats including Markdown, Bases, JSON Canvas.","summaryCn":"让你的 Obsidian 变身超智能 AI 助手，一键调用命令和文件。","reasonCn":"这是一个基于 AI Agent 框架的插件，专为 Obsidian 打造。它能让大模型不仅能帮你写，还能直接操作 Obsidian 的 CLI 命令，读取 JSON Canvas 和 Markdown 文件内容。通过 Skill 的定义，Agent 可以学习并执行一系列复杂操作，比如整理笔记、生成图表，实现从“写作”到“管理”的跨越。","stars":48755,"starGrowth":26,"createdAt":"2026-01-02T21:45:11Z","silentRounds":0,"language":"","topics":["agents"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["Obsidian","AI Agent","知识管理","自动化","技能包"],"domainKey":"AI Agent","pushedAt":"2026-09-15T14:43:57Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7}],"aiScore":0.9,"source":"trending","url":"https://github.com/kepano/obsidian-skills","ts":"2026-09-15T14:43:57Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.09551980309711823,"copyOk":false},{"repo":"aixcoder-plugin/aiXcoder-7B","owner":"aixcoder-plugin","name":"aiXcoder-7B","desc":"","summaryCn":"多语言代码生成大模型，提供IDE插件与权重，辅助日常编程开发。","reasonCn":"这是一款专为提升编程效率而设计的代码大语言模型，核心优势在于其对多种编程语言的深度理解与生成能力。它不仅仅是一个自动补全工具，更能够辅助进行代码理解、重构以及新功能的生成，覆盖了从初级补全到复杂逻辑推理的多个场景。开发者可以通过官方的 VS Code 或 JetBrains 插件直接集成到日常工作流中，或者下载模型权重进行本地部署，从而在保持代码隐私的同时，享受 AI 带来的编码加速与错误排查便利，显著缩短软件交付周期。","stars":2271,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["代码助手","大语言模型","模型部署"],"zone":"AI","funScore":0.2,"domainTags":["代码生成","IDE插件","多语言支持","模型权重"],"domainKey":"代码生成","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/aixcoder-plugin/aiXcoder-7B","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"CodeBoarding/CodeBoarding","owner":"CodeBoarding","name":"CodeBoarding","desc":"","summaryCn":"在 AI 编程大行其道的背景下，开发者往往面临着“AI 生成的代码虽快，但难以","reasonCn":"该工具的核心亮点在于其深度集成能力。开发者可以将它作为 VS Code 扩展安装，直接在 IDE 中查看组件间的逻辑流向，这极大缩短了理解庞大遗留代码库的时间。同时，它还能在 CI/CD 流程或 Pull Request 中自动生成架构文档，使得团队在进行代码审查时，不再是盯着孤立的 diff，而是","stars":2445,"starGrowth":0,"createdAt":"2025-04-08T19:57:56Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","代码助手","可视化"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["代码可视化","架构图","AI辅助开发"],"domainKey":"代码可视化","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/CodeBoarding/CodeBoarding","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"spicetify/cli","owner":"spicetify","name":"cli","desc":"Command-line tool to customize Spotify client. Supports Windows, macOS, and Linux.","summaryCn":"让Spotify变身为你的专属音乐站，主题插件随便换","reasonCn":"这是一个开源的Spotify客户端定制工具，支持Windows、macOS和Linux平台。通过命令行即可安装主题、皮肤、扩展插件，还能修改界面布局、添加自定义菜单项。拥有活跃的社区生态，提供大量第三方主题和扩展，让你的Spotify体验焕然一新。","stars":24595,"starGrowth":5,"createdAt":"2018-12-01T19:55:18Z","silentRounds":0,"language":"JavaScript","topics":["command-line"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["Spotify","客户端定制","主题系统","脚本注入"],"domainKey":"Spotify","pushedAt":"2026-09-22T16:01:02Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/spicetify/cli","ts":"2026-09-22T16:01:02Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.015609880056922138,"copyOk":false},{"repo":"zilliztech/VectorDBBench","owner":"zilliztech","name":"VectorDBBench","desc":"Benchmark for vector databases.","summaryCn":"测各种向量数据库谁最强的工具，帮你选对数据库不踩坑","reasonCn":"VectorDBBench是专为向量数据库设计的基准测试平台，支持Milvus、Pinecone、Weaviate等主流库的对比评测。提供多维度指标：查询延迟、吞吐量、内存占用、召回率等，自动化跑测试并生成可视化报告。让开发者在搭建RAG系统或AI搜索项目时能科学选型。","stars":1179,"starGrowth":0,"createdAt":"2023-05-10T02:16:41Z","silentRounds":25,"language":"Python","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["向量数据库","性能基准","RAG选型","Milvus"],"domainKey":"向量数据库","pushedAt":"2026-09-22T07:17:43Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/zilliztech/VectorDBBench","ts":"2026-09-22T07:17:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"spmallick/learnopencv","owner":"spmallick","name":"learnopencv","desc":"Learn OpenCV  : C++ and Python Examples","summaryCn":"OpenCV图像处理与深度学习实战教程合集","reasonCn":"汇集官方文档和教程的OpenCV示例代码集合，涵盖从基础图像处理到深度学习应用的完整内容。包括图像滤波、边缘检测、特征提取、目标跟踪和神经网络等，采用Jupyter Notebook格式，理论与实战结合，注释详细且附带图解，代码均已验证可用，是系统学习OpenCV的优质资源。","stars":23159,"starGrowth":2,"createdAt":"2015-02-17T08:00:51Z","silentRounds":0,"language":"Jupyter Notebook","topics":["ml"],"aiDims":["学习资源","教程","多模态Agent"],"zone":"资源","funScore":0,"domainTags":["OpenCV","计算机视觉","教程","Jupyter Notebook","图像处理"],"domainKey":"OpenCV","pushedAt":"2026-09-20T06:56:01Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/spmallick/learnopencv","ts":"2026-09-20T06:56:01Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JannisX11/blockbench","owner":"JannisX11","name":"blockbench","desc":"Blockbench - A low poly 3D model editor","summaryCn":"低多边形3D建模神器，手残党也能轻松上手","reasonCn":"Blockbench是一款基于Web的低多边形3D模型编辑器，支持Windows、Mac和Linux。它提供直观的界面，可以快速创建、编辑和导出3D模型。项目使用JavaScript编写，无任何外部依赖，安装即用。支持导入导出多种格式，包括GLTF、OBJ、FBX等。","stars":5973,"starGrowth":2,"createdAt":"2017-06-08T12:33:46Z","silentRounds":0,"language":"JavaScript","topics":["pixel-art"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["3D Modeling","Low Poly","Pixel Art"],"domainKey":"3D Modeling","pushedAt":"2026-09-23T20:41:35Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/JannisX11/blockbench","ts":"2026-09-23T20:41:35Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"halilibrahimd27/cheat-sheet","owner":"halilibrahimd27","name":"cheat-sheet","desc":"Interactive offensive security cheat sheet — 5000+ pentest commands across 50+ categories with instant search, full CRUD, write-ups, favorites and variable fill-bar. OSCP/OSWE/OSEP/CPTS/CKS/CKA/DCA exam prep ready. Docker deployable.","summaryCn":"红队渗透测试神器，5000+命令一键搜索，安全人员必备。","reasonCn":"这是一个交互式的渗透测试命令速查表，整合了OSCP、OSWE等多个认证考试的5000多条命令。它支持全文检索、收藏、变量填充和CRUD操作，专为安全研究人员设计。项目采用JavaScript开发，提供了可视化的命令输入界面，能快速生成定制化的攻击Payload。","stars":41,"starGrowth":0,"createdAt":"2026-03-18T10:20:00Z","silentRounds":9,"language":"JavaScript","topics":["cheatsheet"],"aiDims":["安全工具","学习资源","工具调用"],"zone":"资源","funScore":0,"domainTags":["渗透测试","安全认证","命令速查"],"domainKey":"渗透测试","pushedAt":"2026-09-18T09:16:02Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/halilibrahimd27/cheat-sheet","ts":"2026-09-18T09:16:02Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"frankbria/ralph-claude-code","owner":"frankbria","name":"ralph-claude-code","desc":"Autonomous AI development loop for Claude Code with intelligent exit detection","summaryCn":"让AI自己写代码还能智能判断何时停手的自动化神器","reasonCn":"这个项目的核心技术亮点是智能退出检测机制。传统的AI编程助手需要用户每次手动输入下一个指令，而Ralph通过监控Claude Code的输出和文件变更，自动判断当前任务是否完成。如果检测到代码符合预期、测试通过或任务边界已触及，它会优雅退出；如果发现问题，它会继续迭代优化。","stars":9642,"starGrowth":0,"createdAt":"2025-08-27T16:03:45Z","silentRounds":2,"language":"Shell","topics":["ai-agent"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","Claude Code","编程助手","自动化","智能体"],"domainKey":"AI Agent","pushedAt":"2026-09-19T02:33:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/frankbria/ralph-claude-code","ts":"2026-09-19T02:33:45Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wanshuiyin/Auto-claude-code-research-in-sleep","owner":"wanshuiyin","name":"Auto-claude-code-research-in-sleep","desc":"ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works with Claude Code, Codex, OpenClaw, or any LLM agent.","summaryCn":"它利用机器学习算法来分析大量的代码和实验数据，帮助研究者快速发现潜在的","reasonCn":"它利用机器学习算法来分析大量的代码和实验数据，帮助研究者快速发现潜在的模式和规律。同时，它还提供了可视化界面，使得结果更加直观易懂。这对于需要在夜间或不方便长时间工作的情况下进行研究的研究人员来说非常有帮助。","stars":16509,"starGrowth":18,"createdAt":"2026-03-10T07:31:45Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","研究自动化","技能包","机器学习"],"domainKey":"AI Agent","pushedAt":"2026-09-18T18:05:10Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/wanshuiyin/Auto-claude-code-research-in-sleep","ts":"2026-09-18T18:05:10Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.06872513174632018,"copyOk":false},{"repo":"kubernetes-sigs/agent-sandbox","owner":"kubernetes-sigs","name":"agent-sandbox","desc":"agent-sandbox enables easy management of isolated, stateful, singleton workloads, ideal for use cases like AI agent runtimes and reinforcement learning (RL).","summaryCn":"给AI Agent装个安全牢笼，隔离运行单实例工作负载的神器","reasonCn":"该项目专为AI Agent运行时设计，解决了多Agent协作中的资源冲突与状态隔离难题。它基于Kubernetes构建，利用单实例、状态ful工作负载模型，确保每个Agent拥有独立的运行环境与持久化存储。通过Go语言实现，提供了轻量级的沙箱机制，让Agent能够安全地执行工具调用并管理自身状态，特别适合构建复杂的AI应用编排系统。","stars":3992,"starGrowth":6,"createdAt":"2025-08-12T04:55:05Z","silentRounds":0,"language":"Go","topics":["kubernetes","agent","ai","rl","sandbox"],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","Kubernetes","状态管理","沙箱隔离","云原生"],"domainKey":"AI Agent","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"rl","source":"github","weight":0.7},{"name":"sandbox","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/kubernetes-sigs/agent-sandbox","ts":"2026-09-17T04:13:51.053Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.022509018036072144,"copyOk":false},{"repo":"devsisters/DQN-tensorflow","owner":"devsisters","name":"DQN-tensorflow","desc":"","summaryCn":"这是一个基于 TensorFlow 框架复现人类水平控制深度强化学习算法的经典代码","reasonCn":"在环境依赖与前置条件方面，该项目主要支持 Python 2.7 或 Python 3.3 及以上版本，并需要安装 gym、tqdm、SciPy 或 OpenCV2 以及 TensorFlow 0.12.0 等运行库。通过简单的 pip 命令即可一键安装所需的全部依赖环境，为后续的模型训练与测试打下基础，免去了繁琐的手动配置流程。","stars":2583,"starGrowth":0,"createdAt":"2016-05-15T11:33:47Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["强化学习","深度Q网络","TensorFlow","算法实现"],"domainKey":"强化学习","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/devsisters/DQN-tensorflow","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Event-AHU/Medical_Image_Analysis","owner":"Event-AHU","name":"Medical_Image_Analysis","desc":"Foundation models based medical image analysis","summaryCn":"医学影像分析是AI落地医疗场景的重要方向。","reasonCn":"技术上，项目整合了多种预训练医学影像模型，支持图像分类、目标检测和语义分割等任务。核心亮点在于利用基础模型的通用能力，结合医学影像特点进行适配优化。相比传统方法，这种方法能更好地理解影像中的细微病变特征。","stars":241,"starGrowth":0,"createdAt":"2023-12-25T15:26:28Z","silentRounds":26,"language":"Python","topics":["llm-model"],"aiDims":["多模态Agent","AI基础设施","模型与训练"],"zone":"资源","funScore":0,"domainTags":["医学影像分析","基础模型","图像分类","目标检测","语义分割"],"domainKey":"医学影像分析","pushedAt":"2026-09-22T14:57:51Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/Event-AHU/Medical_Image_Analysis","ts":"2026-09-22T14:57:51Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"unclecode/crawl4ai","owner":"unclecode","name":"crawl4ai","desc":"🚀🤖 Crawl4AI: Open-source LLM Friendly Web Crawler &amp; Scraper. Don't be shy, join here: https://discord.gg/jP8KfhDhyN","summaryCn":"专为大模型设计的网页爬虫，爬下来的内容直接喂给AI用，省掉清洗环节","reasonCn":"这是专为大语言模型优化的开源网络爬虫工具，能将网页内容自动提取并转换为LLM友好格式。内置AI驱动的页面理解能力，自动识别正文、去除广告噪音，还支持异步批量抓取和多种数据输出格式，是构建RAG系统和AI信息获取管道的核心组件。","stars":84084,"starGrowth":21,"createdAt":"2024-05-09T09:48:50Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["工具调用","RAG与知识","AI应用"],"zone":"工具","funScore":0,"domainTags":["网页爬虫","数据清洗","RAG数据管道","LLM预处理","异步抓取"],"domainKey":"网页爬虫","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/unclecode/crawl4ai","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06374925074925075,"copyOk":false},{"repo":"tidev/titanium-sdk","owner":"tidev","name":"titanium-sdk","desc":"🚀 Native iOS and Android Apps with JavaScript & TypeScript","summaryCn":"用写网页JS的技术直接搞原生App，一套代码iOS安卓都能跑，真香。","reasonCn":"Titanium允许开发者用JavaScript和TypeScript编写真正的原生移动应用，而非只是WebView套壳。它通过动态创建平台原生API来工作，意味着应用能访问相机、文件系统等底层硬件。虽然项目较老，但它证明了跨平台开发的可行性，对于熟悉前端生态的开发者来说，学习曲线平缓，能快速产出双端应用。","stars":2802,"starGrowth":0,"createdAt":"2009-06-15T22:54:37Z","silentRounds":25,"language":"Objective-C","topics":["mobile"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["跨平台开发","移动App","JavaScript","原生编译"],"domainKey":"跨平台开发","pushedAt":"2026-09-23T21:49:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Objective-C","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/tidev/titanium-sdk","ts":"2026-09-23T21:49:01Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deforum/deforum-stable-diffusion","owner":"deforum","name":"deforum-stable-diffusion","desc":"","summaryCn":"Deforum最初是为了让编程与艺术爱好者能够无门槛接触先进的机器学习图像合成","reasonCn":"在上手使用方面，该项目设计了灵活的定制化管道，允许开发者根据特定需求对合成流程进行深度修改。由于目前该仓库已停止官方维护，使用者需要通过获取仓库代码并配置Python运行环境来启动项目。对于初学者，建议寻找基于该项目开发的第三方集成平台，或参考社区留存的开源文档进行","stars":2287,"starGrowth":2,"createdAt":"2022-10-28T23:17:50Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","模型部署"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["AI绘画","视频合成","扩散模型","图像生成"],"domainKey":"图像生成","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/deforum/deforum-stable-diffusion","ts":"2026-09-22T21:05:02.747Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"feyninc/chonkie","owner":"feyninc","name":"chonkie","desc":"🦛 CHONK docs with Chonkie ✨ — The lightweight ingestion library for fast, efficient and robust RAG pipelines","summaryCn":"专为RAG流水线打造的轻量级文档吞噬者，让大模型吃得又快又好。","reasonCn":"底层技术上，它采用Python编写，算法经过极致优化，在处理百万字级别的长文档时展现出惊人的速度和极低的内存消耗。它支持多种智能分块策略，不仅考虑字数，还考虑语义上下文的完整性，确保切出来的文本块不会断章取义，从而显著提升RAG系统的问答准确率。","stars":4766,"starGrowth":0,"createdAt":"2025-03-29T05:10:03Z","silentRounds":2,"language":"Python","topics":["rag"],"aiDims":["RAG与知识","AI基础设施","文档理解"],"zone":"AI","funScore":0.6,"domainTags":["RAG","文档摄取","向量检索","NLP"],"domainKey":"RAG","pushedAt":"2026-09-18T17:49:10Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/feyninc/chonkie","ts":"2026-09-18T17:49:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/PIKE-RAG","owner":"microsoft","name":"PIKE-RAG","desc":"","summaryCn":"PIKE-RAG 通过将特殊的知识挖掘逻辑与大语言模型的推理能力相结合，解决了传","reasonCn":"核心亮点在于其针对专业领域定制的增强生成架构。它不仅利用了外部检索的数据，更重要的是建立了一个动态的推理框架，能够引导 LLM 在回答复杂业务问题时进行多步推理，从而避免了浅层检索导致的知识丢失。这种架构对于那些要求极高精准度、严禁模型出现“幻觉”的工业应用场景来说，","stars":2482,"starGrowth":0,"createdAt":"2024-09-26T02:54:44Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["RAG与知识","文档理解"],"zone":"AI","funScore":0.2,"domainTags":["检索增强生成","工业知识推理","深度知识提取"],"domainKey":"检索增强生成","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/microsoft/PIKE-RAG","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"butaixianran/Stable-Diffusion-Webui-Civitai-Helper","owner":"butaixianran","name":"Stable-Diffusion-Webui-Civitai-Helper","desc":"","summaryCn":"这是一个在 AI 绘画社区广受欢迎的 Stable Diffusion Webui 扩展插件，主要","reasonCn":"在系统兼容性与前置要求方面，该扩展需要较新的 Stable Diffusion webui v1.6.x 及以上版本支持，并且在完成安装后需要彻底重启 Webui 程序而不仅仅是刷新界面才能正常加载。为了保证项目的长期活力与持续更新，原作者还在文档中贴心地列出了多条由社区开发者维护的优秀分支链接，","stars":2522,"starGrowth":0,"createdAt":"2023-03-07T07:46:13Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["AI应用","创意工具"],"zone":"AI","funScore":0.5,"domainTags":["Stable-Diffusion","Civitai助手","模型管理","AI绘画插件"],"domainKey":"Stable-Diffusion","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/butaixianran/Stable-Diffusion-Webui-Civitai-Helper","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Zero6992/chatGPT-discord-bot","owner":"Zero6992","name":"chatGPT-discord-bot","desc":"","summaryCn":"支持多模型及本地部署的Discord AI聊天机器人，自托管灵活配置。","reasonCn":"这是一款强大的自托管Discord机器人，让员工在Discord中随意切换OpenAI、Claude、Gemini或本地Ollama模型。它独创隔离运行时机制，允许接入Codex等CLI工具。每个用户的对话历史在不同服务器、频道甚至线程中独立保存，私密且持久，是构建私有AI助手群的完美方案。","stars":2730,"starGrowth":0,"createdAt":"2022-12-04T08:55:06Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["Discord机器人","多模型支持","本地部署","Ollama集成"],"domainKey":"Discord机器人","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/Zero6992/chatGPT-discord-bot","ts":"2026-09-21T04:10:58.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hacksider/Deep-Live-Cam","owner":"hacksider","name":"Deep-Live-Cam","desc":"real time face swap and one-click video deepfake with only a single image","summaryCn":"只需一张照片就能实现实时换脸和一键视频造假的整活神器，好玩又震撼。","reasonCn":"在底层技术上，它集成了目前最先进的人脸检测、特征提取和面部融合神经网络模型。与老一代换脸软件需要海量素材训练专属模型不同，它厉害就厉害在单图推理能力上——只要一张照片，它就能瞬间抓取面部三维特征，并在极短时间内完成精准贴合与光影融合，甚至连眨眼和微笑时的微表情都","stars":96775,"starGrowth":10,"createdAt":"2023-09-24T13:19:31Z","silentRounds":0,"language":"Python","topics":["ai"],"aiDims":["非AI-好玩","创意工具","AI应用"],"zone":"创意","funScore":1,"domainTags":["实时换脸","视频深度伪造","单图推理"],"domainKey":"实时换脸","pushedAt":"2026-09-15T11:05:04Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.98,"source":"search","url":"https://github.com/hacksider/Deep-Live-Cam","ts":"2026-09-15T11:05:04Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.030309997416688193,"copyOk":true},{"repo":"bilawalsidhu/gods-eye-view","owner":"bilawalsidhu","name":"gods-eye-view","desc":"A spy satellite simulator in your browser, except the data is real. Live open source spatial intelligence on a photorealistic 3D globe.","summaryCn":"在浏览器里看真实卫星影像，像玩间谍游戏一样浏览全球3D地图。","reasonCn":"gods‑eye‑view 是基于 WebGL 与 Three.js 构建的卫星模拟平台，实时加载公开的高分辨率遥感影像与地形数据，在浏览器中渲染光照真实、纹理细腻的 3D 球体。用户可自由切换视角、调节时间轴，模拟卫星轨道与成像过程，甚至支持自定义任务如监测热点、测算视野覆盖，提供近乎真实的空间情报体验。","stars":41280,"starGrowth":161,"createdAt":"2026-06-22T23:37:22Z","silentRounds":0,"language":"JavaScript","topics":["3d-globe","cesium","flight-tracking","geospatial","geospatial-intelligence","gis","osint","photogrammetry","satellite-tracking","spatial-intelligence","webgl","worldview"],"aiDims":["非AI-好玩","可视化"],"zone":"创意","funScore":1,"domainTags":["卫星影像可视化","WebGL渲染","三维地球浏览"],"domainKey":"卫星影像可视化","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"3d-globe","source":"github","weight":0.7},{"name":"cesium","source":"github","weight":0.7},{"name":"flight-tracking","source":"github","weight":0.7},{"name":"geospatial","source":"github","weight":0.7},{"name":"geospatial-intelligence","source":"github","weight":0.7},{"name":"gis","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"photogrammetry","source":"github","weight":0.7},{"name":"satellite-tracking","source":"github","weight":0.7},{"name":"spatial-intelligence","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"worldview","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/bilawalsidhu/gods-eye-view","ts":"2026-09-15T01:07:58.514Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3740406976744186,"copyOk":false},{"repo":"moby/moby","owner":"moby","name":"moby","desc":"The Moby Project - a collaborative project for the container ecosystem to assemble container-based systems","summaryCn":"Docker 的核心引擎，让应用打包部署像搭积木一样简单。","reasonCn":"Moby 项目是 Docker 的开源内核，由 Go 语言编写，提供了一套完整的容器生态系统。它允许开发者将应用程序及其依赖项打包成轻量级的容器镜像，实现“一次构建，到处运行”。项目包含构建、运行和管理容器的核心组件，如 containerd 和 runc，是现代云原生应用的基础设施。","stars":72128,"starGrowth":2,"createdAt":"2013-01-18T18:10:57Z","silentRounds":0,"language":"Go","topics":["docker"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Docker","容器化","DevOps","Go"],"domainKey":"Docker","pushedAt":"2026-09-23T16:26:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/moby/moby","ts":"2026-09-23T16:26:51Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"allkern/iris","owner":"allkern","name":"iris","desc":"Sony PlayStation 2 emulator for Windows, Linux and macOS","summaryCn":"在电脑上来玩PS2经典游戏，支持三大平台","reasonCn":"Iris是一个开源的PlayStation 2模拟器，支持Windows、Linux和macOS三大平台。它采用动态重编译技术模拟PS2硬件，核心由C语言编写，性能优化较好，对系统资源占用相对较低，适合怀旧玩家和模拟器爱好者体验经典主机游戏。","stars":476,"starGrowth":0,"createdAt":"2024-10-16T01:11:36Z","silentRounds":9,"language":"C","topics":["emulator"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["PS2模拟器","游戏怀旧","跨平台"],"domainKey":"PS2模拟器","pushedAt":"2026-09-23T20:08:24Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/allkern/iris","ts":"2026-09-23T20:08:24Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yuzicapp/yuzic","owner":"yuzicapp","name":"yuzic","desc":"A cross-platform, free and open source music player for Navidrome & Jellyfin","summaryCn":"专为Navidrome用户打造的高颜值开源播放器，听私人歌单更痛快","reasonCn":"从技术架构上看，yuzic使用TypeScript开发，这意味着它可以在Web端、桌面端（通过Electron或Tauri）以及移动端流畅运行。它不是简单地嵌入一个HTML5音频标签，而是实现了完整的流媒体播放逻辑，支持高音质解码、元数据提取、歌词同步（LRC格式识别）以及播放队列的智能管理。","stars":572,"starGrowth":0,"createdAt":"2024-12-18T23:10:13Z","silentRounds":2,"language":"TypeScript","topics":["jellyfin"],"aiDims":["非AI-实用","效率工具"],"zone":"创意","funScore":0.5,"domainTags":["音乐播放器","Navidrome","Jellyfin","流媒体播放"],"domainKey":"音乐播放器","pushedAt":"2026-09-23T20:07:29Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/yuzicapp/yuzic","ts":"2026-09-23T20:07:29Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"diegosouzapw/OmniRoute","owner":"diegosouzapw","name":"OmniRoute","desc":"Never stop coding. Free MIT AI gateway: one endpoint, 352 providers (150+ free), 1200+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline &amp; Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 550+ contributors","summaryCn":"一个免费开源AI网关，聚合500+模型一键接入，免费额度超香","reasonCn":"这是开源的AI网关聚合平台，一个端点接入290+厂商500+模型，其中90+完全免费。支持Kimi、Claude、GPT、Gemini、DeepSeek等主流模型，原生兼容Claude Code、Cursor等AI编程工具。MIT协议免费使用，帮你把所有AI API统一到一处，省去多次注册和切换的痛苦。","stars":69291,"starGrowth":92,"createdAt":"2026-02-13T12:38:31Z","silentRounds":0,"language":"TypeScript","topics":["a2a","ai-agents","ai-gateway","anthropic","claude","claude-code","cline","codex","copilot","cursor","deepseek","free-ai","gemini","llm-gateway","mcp","openai","openai-proxy","qwen","token-saver","kimi"],"aiDims":["AI基础设施","模型部署","AI Agent"],"zone":"AI","funScore":0,"domainTags":["AI网关","API聚合","模型路由"],"domainKey":"AI网关","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"a2a","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"cline","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"free-ai","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"llm-gateway","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"openai-proxy","source":"github","weight":0.7},{"name":"qwen","source":"github","weight":0.7},{"name":"token-saver","source":"github","weight":0.7},{"name":"kimi","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/diegosouzapw/OmniRoute","ts":"2026-09-11T22:17:29.879Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3359798415378621,"copyOk":false},{"repo":"coronalabs/corona","owner":"coronalabs","name":"corona","desc":"Solar2D Game Engine main repository (ex Corona SDK)","summaryCn":"老牌游戏引擎Solar2D，跨平台开发超简单","reasonCn":"Corona（现更名为Solar2D）是一款老牌但依然活跃的跨平台游戏引擎，以简单易用、性能优异著称。支持iOS、Android、Windows、macOS、Web等多平台发布，非常适合快速开发2D游戏和应用。对于想快速出作品的独立开发者来说，是个不错的选择。","stars":2878,"starGrowth":0,"createdAt":"2018-12-29T17:05:15Z","silentRounds":1,"language":"C++","topics":["mobile","desktop","game"],"aiDims":["游戏","创意工具","非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["游戏引擎","Solar2D","Lua脚本"],"domainKey":"游戏引擎","pushedAt":"2026-09-23T20:20:21Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/coronalabs/corona","ts":"2026-09-23T20:20:21Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"OneKeyHQ/app-monorepo","owner":"OneKeyHQ","name":"app-monorepo","desc":"Secure, open source and community driven crypto wallet runs on all platforms and trusted by millions.","summaryCn":"开源免信任加密货币钱包，手机电脑都能用，资产自己掌握","reasonCn":"OneKey团队开发的开源加密货币钱包，支持iOS、Android、macOS、Windows、Linux全平台。采用开源密钥管理架构，私钥本地生成永不上传云端，支持多链资产管理和硬件钱包联动。代码完全公开可审计，适合注重资产安全的用户。","stars":2448,"starGrowth":0,"createdAt":"2021-11-22T14:54:51Z","silentRounds":3,"language":"TypeScript","topics":["mobile"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0.5,"domainTags":["加密货币","硬件钱包","自托管","TypeScript"],"domainKey":"加密货币","pushedAt":"2026-09-23T19:47:09Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/OneKeyHQ/app-monorepo","ts":"2026-09-23T19:47:09Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cachix/devenv","owner":"cachix","name":"devenv","desc":"Fast, Declarative, Reproducible, and Composable Developer Environments using Nix","summaryCn":"告别环境配置地狱，一行命令帮你把开发环境还原到出厂设置","reasonCn":"Devenv 是基于 Nix 包管理系统的开发者环境工具，采用声明式配置管理，确保开发环境的可复现性和一致性。它允许用户在项目目录下定义依赖和配置，通过一条命令即可创建和激活专属环境，解决了“在我机器上能跑”的经典难题，极大地提升了跨平台开发效率。","stars":7672,"starGrowth":0,"createdAt":"2022-10-22T20:10:10Z","silentRounds":1,"language":"Rust","topics":["developer-tools"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Nix","开发环境","DevOps","配置管理"],"domainKey":"Nix","pushedAt":"2026-09-23T19:38:24Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/cachix/devenv","ts":"2026-09-23T19:38:24Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gruntwork-io/terratest","owner":"gruntwork-io","name":"terratest","desc":" Terratest is a Go library that makes it easier to write automated tests for your infrastructure code.","summaryCn":"把基础设施代码像写程序一样自动测试的神器","reasonCn":"Terratest 是一款基于 Go 语言的开源测试库，专门用于自动化测试基础设施代码。它允许开发者编写测试脚本，通过 HTTP、SSH 等协议验证云资源的创建和配置是否符合预期。通过 Terratest，可以将基础设施变更的验证流程集成到 CI/CD 管道中，显著降低因人为配置错误导致的生产事故风险，提升 DevOps 流程的可靠性和效率。","stars":7948,"starGrowth":0,"createdAt":"2016-03-05T00:26:17Z","silentRounds":3,"language":"Go","topics":["devops","testing"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["基础设施测试","Terraform","Go语言","DevOps自动化","云原生测试"],"domainKey":"基础设施测试","pushedAt":"2026-09-23T19:57:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"testing","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/gruntwork-io/terratest","ts":"2026-09-23T19:57:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EasyCorp/EasyAdminBundle","owner":"EasyCorp","name":"EasyAdminBundle","desc":"EasyAdmin is a fast, beautiful and modern admin generator for Symfony applications.","summaryCn":"给 Symfony 网站一键生成高大上后台的管理工具，写代码的福音","reasonCn":"这项目是 PHP 框架 Symfony 的配套神器，能自动根据数据库表结构生成界面美观的后台管理系统。开发者只需简单配置，就能拥有列表、详情、编辑、删除等全套功能，省去重复造轮子，大幅提升后端开发效率，适合需要快速交付管理界面的 PHP 团队。","stars":4303,"starGrowth":0,"createdAt":"2015-01-12T08:43:02Z","silentRounds":13,"language":"PHP","topics":["backend"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Symfony","后台生成","PHP","CRUD"],"domainKey":"Symfony","pushedAt":"2026-09-23T18:35:17Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/EasyCorp/EasyAdminBundle","ts":"2026-09-23T18:35:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"valkey-io/valkey","owner":"valkey-io","name":"valkey","desc":"A flexible distributed key-value database that is optimized for caching and other realtime workloads.","summaryCn":"专为实时场景打造的分布式键值数据库，缓存快到飞起","reasonCn":"Valkey是Redis的开源分支，专为实时缓存和分布式键值存储设计。采用C语言编写，支持集群模式、持久化和流数据结构。兼容Redis协议，迁移成本低。用于会话缓存、实时分析、消息队列等场景。性能指标：百万级QPS，亚毫秒延迟。提供Java、Python客户端库。","stars":27276,"starGrowth":2,"createdAt":"2024-03-22T00:42:17Z","silentRounds":0,"language":"C","topics":["database"],"aiDims":["AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["缓存","数据库","键值存储","Redis兼容"],"domainKey":"缓存","pushedAt":"2026-09-23T20:37:14Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/valkey-io/valkey","ts":"2026-09-23T20:37:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tixl3d/tixl","owner":"tixl3d","name":"tixl","desc":"TiXL is an open source software to create realtime motion graphics.","summaryCn":"开源实时动效创作工具，把代码变成炫酷的视觉动画，设计师和极客都爱","reasonCn":"它本质是一个基于时间轴的实时渲染引擎，你可以在里面做各种动效：粒子系统、波形变换、程序化生成图形，还能和外部设备或数据源交互。它用C#写的，界面直观但背后是代码驱动，所以既有设计的自由度，又有编程的精确控制。","stars":5111,"starGrowth":0,"createdAt":"2017-06-17T22:37:11Z","silentRounds":1,"language":"C#","topics":["design"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["实时渲染","创意编程","动效","粒子系统","交互装置"],"domainKey":"实时渲染","pushedAt":"2026-09-23T18:23:45Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/tixl3d/tixl","ts":"2026-09-23T18:23:45Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wshobson/agents","owner":"wshobson","name":"agents","desc":"Multi-harness agentic plugin marketplace for Claude Code, Codex, Cursor, OpenCode, GitHub Copilot, Google Antigravity, and Pi","summaryCn":"一个插件市场，让各类AI编程助手共享插件一套插件跑遍所有工具","reasonCn":"这是一个跨平台的AI编程助手插件市场，支持Claude Code、Cursor、GitHub Copilot等多个主流AI编程工具。通过统一的插件接口，开发者只需编写一次插件代码，就能在所有支持的AI编程助手中运行，极大提升了插件复用率和开发效率。","stars":39878,"starGrowth":7,"createdAt":"2025-07-24T23:28:14Z","silentRounds":0,"language":"Python","topics":["anthropic","agent-skills","agentic-ai","ai-agents","cursor","cursor-rules","mcp","multi-agent","opencode","claude-code","antigravity","claude","claude-skills","subagents","codex","claude-code-marketplace","claude-code-plugin","github-copilot","pi-coding-agent","coding-agents"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["AI编程助手","插件市场","多工具适配","开发工具"],"domainKey":"AI编程助手","pushedAt":"2026-09-17T04:13:51.053Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"anthropic","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"cursor-rules","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"multi-agent","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-skills","source":"github","weight":0.7},{"name":"subagents","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"claude-code-marketplace","source":"github","weight":0.7},{"name":"claude-code-plugin","source":"github","weight":0.7},{"name":"github-copilot","source":"github","weight":0.7},{"name":"pi-coding-agent","source":"github","weight":0.7},{"name":"coding-agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/wshobson/agents","ts":"2026-09-17T04:13:51.053Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0215266061487537,"copyOk":false},{"repo":"python-visualization/folium","owner":"python-visualization","name":"folium","desc":"Python Data. Leaflet.js Maps. ","summaryCn":"用Python写几行代码就能做出超炫酷交互式地图","reasonCn":"folium把Leaflet.js地图库和Python完美结合，不用写一行JavaScript，直接用Python数据就能生成带交互的网页地图。支持标记点、热力图、弹窗信息等各种效果，把枯燥的数据变成直观的可视化地图，特别适合展示地理位置信息。","stars":7403,"starGrowth":0,"createdAt":"2013-05-09T04:21:35Z","silentRounds":3,"language":"Python","topics":["data-visualization","data-science"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["数据可视化","地理信息系统","交互式地图","Python库"],"domainKey":"数据可视化","pushedAt":"2026-09-23T17:53:26Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/python-visualization/folium","ts":"2026-09-23T17:53:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KbWen/agent-virtual-office","owner":"KbWen","name":"agent-virtual-office","desc":"🏢 Watch your AI coding agents live their office life — Claude Code, Codex & Gemini show up as tiny pixel coworkers who type, take coffee breaks, get blocked, and ship. Real-time, cozy, runs right in your browser 👀✨","summaryCn":"看AI写代码的同事在办公室里摸鱼，太治愈了","reasonCn":"这是一个将AI编程代理具象化的虚拟办公室模拟器。它将Claude Code、Codex和Gemini等AI编程助手转化为像素风格的虚拟同事。项目实时展示它们的工作状态：输入代码、喝水休息、被代码逻辑卡住以及最终完成任务。基于JavaScript和像素艺术技术构建，完全运行在浏览器中。","stars":24,"starGrowth":0,"createdAt":"2026-03-16T14:30:21Z","silentRounds":7,"language":"JavaScript","topics":["pixel-art"],"aiDims":["AI应用","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["像素艺术","AI Agent","虚拟办公室"],"domainKey":"像素艺术","pushedAt":"2026-09-20T05:44:22Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/KbWen/agent-virtual-office","ts":"2026-09-20T05:44:22Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"InterviewReady/ai-engineering-resources","owner":"InterviewReady","name":"ai-engineering-resources","desc":"","summaryCn":"专为软件工程师量身打造的AI工程学转型路线图与精选论文合集。","reasonCn":"这是一个面向传统软件工程师的AI工程转型资源库，系统整理了通往AI领域所需的各类核心研究论文。内容涵盖了Tokenization、Vectorization、基础设施、核心架构以及混合专家模型等关键技术模块，帮助开发者快速建立系统的AI知识体系，无需自行大海捞针搜寻优质学习资料。","stars":2653,"starGrowth":0,"createdAt":"2025-05-13T18:17:53Z","silentRounds":1,"language":"","topics":[],"aiDims":["学习资源","教程","论文"],"zone":"资源","funScore":0.2,"domainTags":["AI工程","大语言模型","向量检索","机器学习架构"],"domainKey":"AI工程","pushedAt":"2026-09-21T19:19:33.440Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1}],"aiScore":0.71,"source":"search","url":"https://github.com/InterviewReady/ai-engineering-resources","ts":"2026-09-21T19:19:33.440Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alephpi/Texo","owner":"alephpi","name":"Texo","desc":"A minimalist SOTA LaTeX OCR model with only 20M parameters, running in browser. Full training pipeline available for self-reproduction. | 超轻量SOTA LaTeX公式识别模型，仅20M参数量，可在浏览器中运行。训练全流程代码开源，以便自学复现。","summaryCn":"你在写论文时是不是经常为了把书里的数学公式变成LaTeX代码而抓狂？","reasonCn":"核心技术上，它采用了先进的轻量级Transformer架构，通过极致的参数裁剪，将模型大小压缩到了惊人的20MB。别看它小，它在数学符号的识别率上却能和那些数百MB甚至GB级别的庞然大物硬碰硬。其训练过程非常透明，不仅开源了模型权重，还分享了处理数学图像数据集的方法，这在同类小模","stars":898,"starGrowth":0,"createdAt":"2025-07-29T10:08:28Z","silentRounds":4,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","文档理解","AI应用"],"zone":"AI","funScore":0.9,"domainTags":["LaTeX OCR","数学公式识别","WebGPU","Transformer","边缘AI"],"domainKey":"LaTeX OCR","pushedAt":"2026-09-18T15:33:12Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/alephpi/Texo","ts":"2026-09-18T15:33:12Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lioensky/VCPToolBox","owner":"lioensky","name":"VCPToolBox","desc":"VCP 部署在 AI 模型 API 与前端应用之间，是面向AGI OS开发和探索的工业级基建示范项目。通过统一指令协议、多层级持久化记忆、分布式插件引擎及多 Agent 协作框架，将原本“无状态、无记忆、无工具调用能力”的大语言模型，彻底改造成拥有永久自我意识、物理世界操作权及群体协作智能的完整智能体系统。","summaryCn":"给大模型装上永久记忆和手脚，让AI从聊天机器人变成能干活的全能助手","reasonCn":"VCPToolBox定位是AGI OS的工业级基建，核心是解决大模型无状态、无记忆、无工具能力的短板。它通过统一指令协议和分布式插件引擎，让多个AI Agent能像团队一样协作，还能持久化记忆。对于想做复杂多步任务或智能体应用的开发者来说，这是一个很硬核的底层框架。","stars":2324,"starGrowth":0,"createdAt":"2025-05-12T13:51:06Z","silentRounds":1,"language":"JavaScript","topics":["vector-db"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["Agent框架","持久化记忆","多Agent协作","AGI"],"domainKey":"Agent框架","pushedAt":"2026-09-23T12:12:55Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/lioensky/VCPToolBox","ts":"2026-09-23T12:12:55Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ClementTsang/bottom","owner":"ClementTsang","name":"bottom","desc":"Yet another cross-platform graphical process/system monitor.","summaryCn":"命令行党的神器！一眼看穿系统运行状态，CPU内存统统透明","reasonCn":"用Rust编写的跨平台系统监控终端工具，提供媲美htop的实时监控能力，但界面更加现代化美观。支持CPU、内存、磁盘、网络、电池等多维度数据可视化，支持自定义配色方案、刷新频率和组件布局，还能通过配置文件灵活调整显示内容，适合追求效率的开发者日常使用。","stars":14042,"starGrowth":3,"createdAt":"2019-08-28T23:43:30Z","silentRounds":0,"language":"Rust","topics":["monitoring"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["系统监控","Rust","终端工具","性能分析"],"domainKey":"系统监控","pushedAt":"2026-09-23T00:54:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/ClementTsang/bottom","ts":"2026-09-23T00:54:06Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"portainer/portainer","owner":"portainer","name":"portainer","desc":"Making Docker and Kubernetes management easy.","summaryCn":"给Docker和K8s装个可视化的‘遥控器’，管理容器像点外卖一样简单","reasonCn":"Portainer是用TypeScript写的Docker和Kubernetes管理平台，通过Web界面提供可视化操作。它支持容器、镜像、卷、网络等核心资源的CRUD操作，也提供内置的编排模板和监控仪表盘。对团队来说，它能降低Docker/K8s的学习门槛，解决命令行操作的繁琐问题，尤其适合中小团队快速部署和管理微服务。","stars":38543,"starGrowth":2,"createdAt":"2016-05-19T20:15:28Z","silentRounds":0,"language":"TypeScript","topics":["docker","kubernetes"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Docker管理","Kubernetes","容器运维","可视化面板"],"domainKey":"Docker管理","pushedAt":"2026-09-22T13:01:30Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/portainer/portainer","ts":"2026-09-22T13:01:30Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xMassi/webclaw","owner":"0xMassi","name":"webclaw","desc":"","summaryCn":"把混乱的网页内容转化为干净的 Markdown 和大模型专属上下文。","reasonCn":"许多网页抓取工具往往会返回被拦截的页面、登录墙或充满广告与脚本的原始 HTML，这给大语言模型和检索增强生成带来了极大的干扰。该开源项目提供了一个强大的网页提取引擎，能够将任意网址精准转换为结构化的清洁内容。它通过提供命令行、MCP 服务器以及托管 API 等多种形态，帮助开发者轻松为 AI 智能体接入实时的网页知识上下文。","stars":2359,"starGrowth":1,"createdAt":"2026-03-10T17:50:15Z","silentRounds":1,"language":"Rust","topics":[],"aiDims":["AI应用","文档理解","工具调用"],"zone":"AI","funScore":0.2,"domainTags":["网页抓取","Markdown转换","REST API","MCP服务器","AI代理"],"domainKey":"网页抓取","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/0xMassi/webclaw","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hetu-script/heavenly-tribulation","owner":"hetu-script","name":"heavenly-tribulation","desc":"Heavenly Tribulation: A RPG game mixed with cardgame-autobattler/Tycoon elements. Game logic is written in Hetu Script. 《天道奇劫》：混合了仙侠RPG+卡牌自动对战+门派经营玩法的游戏。","summaryCn":"仙侠卡牌养成大作，一边修仙一边经营门派，热血又上头。","reasonCn":"这款游戏的独特之处在于它的“仙侠”风味和“剧本”深度。它不是那种简单的数值堆砌，而是有着自己的世界观和剧情线。你扮演的修仙者，需要一步步踏上成仙之路，途中会遇到各种挑战和机遇。游戏的战斗系统采用了卡牌自动对战机制，你可以提前设置好弟子的技能组合和出牌顺序，然后看着","stars":12,"starGrowth":0,"createdAt":"2021-12-07T06:26:32Z","silentRounds":12,"language":"Dart","topics":["procedural-generation"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["仙侠游戏","卡牌对战","自动战斗","经营模拟","RPG"],"domainKey":"仙侠游戏","pushedAt":"2026-09-23T15:59:20Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/hetu-script/heavenly-tribulation","ts":"2026-09-23T15:59:20Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"openfl/openfl","owner":"openfl","name":"openfl","desc":"An open source library for creative expression on the web, desktop, mobile and consoles. Inspired by the classic Flash and AIR APIs.","summaryCn":"老牌跨平台开发框架，复刻Flash API，适合做游戏和创意动画。","reasonCn":"它是一款基于Haxe语言的跨平台图形库，核心灵感来自经典Flash和AIR技术，能在网页、桌面、移动端和游戏机上一键发布应用。项目拥有超过2000个Star，社区成熟稳定，特别适合追求高性能渲染和统一开发体验的创意工作者。","stars":2157,"starGrowth":0,"createdAt":"2013-03-19T02:50:23Z","silentRounds":6,"language":"Haxe","topics":["frontend","game-development"],"aiDims":["游戏","创意工具","非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["Haxe","跨平台开发","Flash复刻","游戏开发"],"domainKey":"Haxe","pushedAt":"2026-09-23T16:04:53Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"game-development","source":"github","weight":0.7},{"name":"Haxe","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/openfl/openfl","ts":"2026-09-23T16:04:53Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ankitpokhrel/jira-cli","owner":"ankitpokhrel","name":"jira-cli","desc":"🔥 Feature-rich interactive Jira command line.","summaryCn":"让你像敲代码一样玩转 Jira 的神器，告别网页端繁琐操作","reasonCn":"这是一个功能完备的 Jira 命令行工具，支持交互式模式。用户可以通过简单的命令创建、查询、更新工单，实现与 Jira 系统的无缝集成。项目使用 Go 语言开发，性能高效，支持多种交互模式，极大提升了敏捷开发团队的工作效率。","stars":5985,"starGrowth":2,"createdAt":"2020-11-22T08:48:39Z","silentRounds":0,"language":"Go","topics":["command-line","productivity"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Jira","命令行工具","敏捷开发","任务管理"],"domainKey":"Jira","pushedAt":"2026-09-22T12:42:52Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/ankitpokhrel/jira-cli","ts":"2026-09-22T12:42:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"homeassistant-ai/ha-mcp","owner":"homeassistant-ai","name":"ha-mcp","desc":"The Unofficial and Awesome Home Assistant MCP Server","summaryCn":"Ha-MCP是一个强大的智能家居管理服务器，它提供了一套完整的API接口，使得开","reasonCn":"Ha-MCP是一个强大的智能家居管理服务器，它提供了一套完整的API接口，使得开发者可以轻松地管理和控制家庭自动化设备。它支持多种通信协议，如Zigbee、CoAP和HTTP，并且可以与其他智能家居设备进行交互。此外，它还提供了一些高级功能，如事件触发、设备状态监控和数据记录，使得管","stars":4819,"starGrowth":3,"createdAt":"2025-09-14T13:23:45Z","silentRounds":0,"language":"Python","topics":["home-automation"],"aiDims":["Home Assistant MCP Server"],"zone":"工具","funScore":0,"domainTags":["Home Assistant","智能家居","API","自动化"],"domainKey":"Home Assistant","pushedAt":"2026-09-23T14:17:10Z","tags":[{"name":"Home Assistant MCP Server","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/homeassistant-ai/ha-mcp","ts":"2026-09-23T14:17:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xerrors/Yuxi","owner":"xerrors","name":"Yuxi","desc":"可私有部署的多租户知识智能体平台：统一 RAG、知识图谱、多智能体、MCP/Skills、沙盒与权限管理。Self-hosted knowledge agent platform for RAG, knowledge graphs and multi-agent workflows.","summaryCn":"轻量级多租户知识图谱引擎","reasonCn":"由于其轻量级和高可扩展性，Yuxi非常适合用于大规模的知识图谱构建和更新。对于需要处理大规模数据集的企业来说，它是一个非常好的选择。","stars":7171,"starGrowth":2,"createdAt":"2024-07-05T10:57:35Z","silentRounds":0,"language":"Python","topics":["rag"],"aiDims":["AI基础设施"],"zone":"AI","funScore":0,"domainTags":["知识图谱","RAG","多智能体"],"domainKey":"知识图谱","pushedAt":"2026-09-23T18:01:51Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/xerrors/Yuxi","ts":"2026-09-23T18:01:51Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cure53/DOMPurify","owner":"cure53","name":"DOMPurify","desc":"DOMPurify - a DOM-only, super-fast, uber-tolerant XSS sanitizer for HTML, MathML and SVG. DOMPurify works with a secure default, but offers a lot of configurability and hooks. Demo:","summaryCn":"前端安全的神级过滤器，防止黑客通过 HTML 注入攻击","reasonCn":"这是前端开发领域公认的最安全、最流行的 XSS（跨站脚本攻击）防护库。它能智能地清理用户输入的 HTML、SVG 和 MathML 内容，去除危险标签和事件处理器，同时保留丰富的格式化能力。无论是 React 还是 Vue，集成它只需几行代码，是构建用户生成内容（UGC）应用的必备护盾。","stars":17409,"starGrowth":0,"createdAt":"2014-02-17T21:48:14Z","silentRounds":3,"language":"JavaScript","topics":["security"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["XSS防护","HTML清理","安全过滤","JavaScript库"],"domainKey":"XSS防护","pushedAt":"2026-09-23T15:18:57Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/cure53/DOMPurify","ts":"2026-09-23T15:18:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xJacky/nginx-ui","owner":"0xJacky","name":"nginx-ui","desc":"Yet another WebUI for Nginx","summaryCn":"给Nginx装个网页版皮肤，配置服务器像玩积木一样简单","reasonCn":"这是一个专为Nginx设计的现代化Web管理界面，支持反向代理、负载均衡和HTTPS证书自动申请等功能，让复杂的配置文件管理变得直观可视。开发者可以通过图形化操作替代繁琐的命令行编辑，实时预览配置变更效果，大幅降低运维门槛。","stars":11535,"starGrowth":1,"createdAt":"2021-02-19T10:51:04Z","silentRounds":1,"language":"Go","topics":["self-hosted","docker"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["Nginx","Web服务器","配置管理","反向代理"],"domainKey":"Nginx","pushedAt":"2026-09-23T13:46:31Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/0xJacky/nginx-ui","ts":"2026-09-23T13:46:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LLMQuant/quant-mind","owner":"LLMQuant","name":"quant-mind","desc":"QuantMind is an open source agent-native knowledge extraction and retrieval framework for quantitative finance.","summaryCn":"量化金融知识处理器，将非结构化信息转化为可信结构化知识。","reasonCn":"QuantMind是一款面向量化金融领域的知识处理工具，旨在解决金融信息繁杂且非结构化的问题。它通过将论文、新闻和财报等原始资料，转化为带有类型、引用和时间戳的结构化知识，为下游的检索和推理提供可靠数据基础。项目强调知识的持久性和独立性，确保每条知识都能在特定时间点被独立查询和分析。","stars":3012,"starGrowth":0,"createdAt":"2025-03-14T02:51:22Z","silentRounds":1,"language":"Python","topics":["quantitative-finance","quantitative-research","data","knowledge","llm","pipeline","workflow","agent","context-engineering","harness-engineering"],"aiDims":["AI Agent","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["量化金融","Agent框架","知识图谱","RAG","数据分析"],"domainKey":"知识图谱","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"quantitative-finance","source":"github","weight":0.7},{"name":"quantitative-research","source":"github","weight":0.7},{"name":"data","source":"github","weight":0.7},{"name":"knowledge","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"pipeline","source":"github","weight":0.7},{"name":"workflow","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"context-engineering","source":"github","weight":0.7},{"name":"harness-engineering","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/LLMQuant/quant-mind","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"BrikerMan/Kashgari","owner":"BrikerMan","name":"Kashgari","desc":"","summaryCn":"Kashgari 的核心价值在于将复杂的自然语言处理任务标准化和模块化。","reasonCn":"在架构演进上，Kashgari 紧跟深度学习生态的步伐。其最新版本 2.0.0 正式引入了对 TensorFlow 2.0 的支持，这意味着开发者可以利用 Eager Execution 等现代特性进行更灵活的调试和开发。框架的设计哲学强调“人类友好”，其源代码结构清晰，注释丰富，且经过了严格的单元测试覆盖。","stars":2382,"starGrowth":0,"createdAt":"2019-01-19T01:53:28Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["大语言模型","NLP","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["命名实体识别","词性标注","文本分类","迁移学习","TensorFlow"],"domainKey":"命名实体识别","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"NLP","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/BrikerMan/Kashgari","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Jittor/JittorLLMs","owner":"Jittor","name":"JittorLLMs","desc":"","summaryCn":"2G内存就能跑大模型，这听起来像是画饼，但JittorLLMs确实做到了。","reasonCn":"除了低门槛，这个库还在速度和兼容性上做了不少文章。大模型加载慢是困扰本地部署的一大痛点，Jittor通过零拷贝技术，让模型加载的开销降低了40%。同时，借助元算子的自动编译优化，计算性能相比同类框架提升了20%以上。","stars":2410,"starGrowth":0,"createdAt":"2023-04-02T06:22:01Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["推理引擎","模型部署","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["大模型推理","低内存部署","国产模型适配","异构计算"],"domainKey":"大模型推理","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/Jittor/JittorLLMs","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JetBrains/compose-multiplatform","owner":"JetBrains","name":"compose-multiplatform","desc":"Compose Multiplatform, a modern UI framework for Kotlin that makes building performant and beautiful user interfaces easy and enjoyable.","summaryCn":"一套代码搞定电脑手机应用，JetBrains出品的跨平台UI神器","reasonCn":"Compose Multiplatform是JetBrains推出的跨平台UI框架，允许开发者使用Kotlin编写代码，一次编译即可运行在桌面、Web和移动设备上。它复用Android Compose的设计理念，提供了现代化的声明式UI体验，极大提升了跨端开发效率。","stars":19381,"starGrowth":1,"createdAt":"2020-09-07T10:40:49Z","silentRounds":1,"language":"Kotlin","topics":["desktop"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.7,"domainTags":["Kotlin","跨平台UI","Compose","多端开发"],"domainKey":"Kotlin","pushedAt":"2026-09-23T12:21:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/JetBrains/compose-multiplatform","ts":"2026-09-23T12:21:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Keralots/AnimatedPixelClock","owner":"Keralots","name":"AnimatedPixelClock","desc":"Animated retro-arcade clock on a 128x64 HUB75 RGB LED matrix (ESP32-S3), with optional PC stats monitoring","summaryCn":"用ESP32-S3控制128x64像素屏，打造复古街机风格的动态时钟。","reasonCn":"这是一个基于ESP32-S3开发板的创意硬件项目，通过编程在LED矩阵屏上绘制动态像素时钟。它不仅显示时间，还融合了街机游戏的视觉风格，甚至能监测PC性能数据。它完美结合了复古美学与现代技术，是极客们打造桌面装饰品的绝佳选择。","stars":74,"starGrowth":0,"createdAt":"2026-07-02T11:37:09Z","silentRounds":8,"language":"C","topics":["pixel-art"],"aiDims":["创意工具","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["ESP32-S3","LED矩阵","像素艺术","复古风格","硬件DIY"],"domainKey":"ESP32-S3","pushedAt":"2026-09-19T17:25:58Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Keralots/AnimatedPixelClock","ts":"2026-09-19T17:25:58Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fengdu78/machine_learning_beginner","owner":"fengdu78","name":"machine_learning_beginner","desc":"","summaryCn":"该仓库拥有一个包含“强烈推荐的TensorFlow、Pytorch和Keras样例资源”的深度","reasonCn":"项目内容覆盖了从两天入门Python到Sklearn经典案例实战的完整链路，实现了理论与实战的同步推进。无论你是想通过吴恩达的课程笔记夯实统计学基础，还是想直接下载项目代码上手复现《深度学习》原书中的示例，这里都能提供相应的配套文件夹，将抽象的算法逻辑通过Python代码展现得直","stars":2358,"starGrowth":0,"createdAt":"2019-01-09T08:15:12Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["机器学习","深度学习","Python编程","人工智能教程"],"domainKey":"机器学习","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/fengdu78/machine_learning_beginner","ts":"2026-09-21T23:17:36.060Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Revolutionary-Games/Thrive","owner":"Revolutionary-Games","name":"Thrive","desc":"The main repository for the development of the evolution game Thrive. ","summaryCn":"从单细胞一路进化到星际文明，自己当造物主的游戏","reasonCn":"基于C#开发的进化模拟游戏，涵盖从微观单细胞到宏观星际文明的完整演化过程。支持细胞个体和整个物种两种视角，结合程序化生成、物理模拟和AI驱动的生态环境系统，让每个生物体根据基因和环境实时变化。游戏内置编辑器允许自定义物种形态、行为模式和世界规则，画面呈现精美的2D风格配合动态音效，沉浸感很强。","stars":3714,"starGrowth":0,"createdAt":"2012-08-30T15:31:36Z","silentRounds":1,"language":"C#","topics":["game","simulation"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["进化模拟","策略游戏","细胞模拟","程序化生成","生物演化"],"domainKey":"进化模拟","pushedAt":"2026-09-23T14:16:49Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/Revolutionary-Games/Thrive","ts":"2026-09-23T14:16:49Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NVIDIA/TensorRT-LLM","owner":"NVIDIA","name":"TensorRT-LLM","desc":"TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way.","summaryCn":"英伟达针对大语言模型和视觉生成模型推出了这款官方开源的推理加速框架，它","reasonCn":"核心亮点在于其极致的性能压榨能力。通过结合 GEMM 量化、注意力机制量化以及 Skip Softmax 等先进技术，它能有效降低显存占用并成倍提升视频生成与文本推理的执行效率。同时，它还不断跟进前沿负载需求，支持智能体（Agentic）服务的评估与大规模集群扩展，展现出极强的工业级生产","stars":14698,"starGrowth":5,"createdAt":"2023-08-16T17:14:27Z","silentRounds":0,"language":"Python","topics":["blackwell","cuda","moe","pytorch","llm-serving"],"aiDims":["推理引擎","AI基础设施","大语言模型"],"zone":"AI","funScore":0,"domainTags":["大语言模型","推理加速","GPU优化","AI部署"],"domainKey":"大语言模型","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"blackwell","source":"github","weight":0.7},{"name":"cuda","source":"github","weight":0.7},{"name":"moe","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"llm-serving","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"trending","url":"https://github.com/NVIDIA/TensorRT-LLM","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.016020547013199073,"copyOk":false},{"repo":"google-parfait/tensorflow-federated","owner":"google-parfait","name":"tensorflow-federated","desc":"","summaryCn":"在训练移动端键盘预测模型时，传统的云端训练方式要求上传用户所有的打字记","reasonCn":"TFF 的核心价值在于提供了一个标准化的实验平台，开发者无需自行搭建分布式通信协议。框架内置了多种联邦学习算法，从基础的 FedAvg 到更复杂的个性化训练策略，都可以通过拖拽式的 Python API 进行组合。这意味着研究者可以快速验证“如果数据分散在不同设备上，模型收敛性会怎样变","stars":2452,"starGrowth":0,"createdAt":"2018-12-12T23:15:35Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["模型与训练","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["联邦学习","隐私计算","移动端预测","分布式训练","机器学习框架"],"domainKey":"联邦学习","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/google-parfait/tensorflow-federated","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CSAILVision/places365","owner":"CSAILVision","name":"places365","desc":"","summaryCn":"麻省理工发布的场景识别模型库，能判断照片属于哪类地方。","reasonCn":"这是MIT计算机视觉实验室开源的场景识别模型库。它能够识别一张照片属于哪一类物理场所，比如厨房、海滩、办公室等365种场景。项目提供了在大型数据集上预训练好的多种卷积神经网络模型，可以直接用于分类任务，也可以作为特征提取器迁移到其他视觉项目中。","stars":2089,"starGrowth":2,"createdAt":"2016-05-12T16:31:05Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0.26,"domainTags":["场景识别","图像分类","计算机视觉","卷积神经网络"],"domainKey":"场景识别","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/CSAILVision/places365","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yuanyang749/softie-webgpu","owner":"yuanyang749","name":"softie-webgpu","desc":"使用GPT6 Astra 制作的一个可爱的解压小应用：Softie","summaryCn":"AI辅助生成的WebGPU解压小游戏，软萌互动超治愈。","reasonCn":"基于WebGPU技术构建的浏览器端创意解压应用，采用GPT辅助编程生成核心逻辑。JavaScript实现高性能图形渲染，无需安装大型插件。在浏览器中即可流畅运行生成式视觉粒子交互，操作轻盈，界面软萌，主打放松与视觉娱乐，科技感与趣味性结合。","stars":23,"starGrowth":0,"createdAt":"2026-09-06T15:54:56Z","silentRounds":11,"language":"JavaScript","topics":["creative-coding"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["WebGPU","创意编程","解压","互动应用","数字宠物"],"domainKey":"WebGPU","pushedAt":"2026-09-22T12:02:15Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/yuanyang749/softie-webgpu","ts":"2026-09-22T12:02:15Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pnpm/pnpm","owner":"pnpm","name":"pnpm","desc":"Fast, disk space efficient package manager","summaryCn":"装依赖贼快还不占硬盘的包管理器，程序员的神器","reasonCn":"这是一个用Rust编写的超快包管理器，解决了传统工具内存占用高、安装慢的痛点。它采用硬链接和符号链接技术，让同一版本的依赖包在所有项目中只存在一份，极大节省磁盘空间。它拥有严格的依赖隔离机制，彻底解决了Node.js生态常见的幽灵依赖问题。","stars":36612,"starGrowth":4,"createdAt":"2016-01-28T07:40:43Z","silentRounds":0,"language":"Rust","topics":["npm","dependency-manager","package-manager","node","nodejs","install","modules","javascript"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["包管理器","Node.js","前端工程","依赖管理"],"domainKey":"包管理器","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"npm","source":"github","weight":0.7},{"name":"dependency-manager","source":"github","weight":0.7},{"name":"package-manager","source":"github","weight":0.7},{"name":"node","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"install","source":"github","weight":0.7},{"name":"modules","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/pnpm/pnpm","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"langroid/langroid","owner":"langroid","name":"langroid","desc":"Harness LLMs with Multi-Agent Programming","summaryCn":"有时候让一个大模型去完成极其复杂的任务，比如写一整套包含前后端的软件或","reasonCn":"技术底层上，它采用了面向对象的编程范式，把每个Agent看作是一个独立的计算实体，通过直观的Python API进行消息传递和状态同步。它内置了强大的RAG（检索增强生成）引擎，能够高效处理海量私有文档。和同类框架相比，它的代码结构更加纯粹和模块化，类型提示完善，调试起来非常直","stars":4104,"starGrowth":1,"createdAt":"2023-04-16T20:47:28Z","silentRounds":19,"language":"Python","topics":["rag"],"aiDims":["AI Agent","RAG与知识","大语言模型"],"zone":"AI","funScore":0,"domainTags":["AI Agent","多智能体","RAG","Python"],"domainKey":"AI Agent","pushedAt":"2026-09-17T23:56:01Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/langroid/langroid","ts":"2026-09-17T23:56:01Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"h2oai/h2o-llmstudio","owner":"h2oai","name":"h2o-llmstudio","desc":"H2O LLM Studio - a framework and no-code GUI for fine-tuning LLMs. Documentation: https://docs.h2o.ai/h2o-llmstudio/","summaryCn":"如果微调大模型像做饭一样，那这个项目就是你的全自动智能电饭煲。","reasonCn":"核心架构采用了模块化的微调管道设计。它不仅支持LoRA等主流微调技术，还内置了丰富的评估工具，可以在训练的同时实时监控模型性能指标，比如准确率、损失函数变化等。与命令行工具相比，这种可视化交互方式让训练过程变得透明且可控，大大降低了试错成本，特别适合需要快速验证多","stars":5182,"starGrowth":1,"createdAt":"2023-04-17T17:56:33Z","silentRounds":12,"language":"Python","topics":["chatbot"],"aiDims":["微调","模型部署","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["大模型微调","LLM","无代码工具","MLOps"],"domainKey":"大模型微调","pushedAt":"2026-09-18T22:06:07Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/h2oai/h2o-llmstudio","ts":"2026-09-18T22:06:07Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TinyLLaVA/TinyLLaVA_Factory","owner":"TinyLLaVA","name":"TinyLLaVA_Factory","desc":"A Framework of Small-scale Large Multimodal Models","summaryCn":"轻量版多模态大模型框架，省显卡也能玩图文对话","reasonCn":"TinyLLaVA_Factory 基于 PyTorch 与 transformers，提供可微调的轻量多模态模型，实现图像特征提取、文本生成和跨模态检索。配套高效数据管线和统一训练脚本，用户仅需几行代码即可在普通 GPU 上完成训练或微调，支持图文问答、视觉指令等实际场景。","stars":1008,"starGrowth":0,"createdAt":"2024-02-21T00:28:14Z","silentRounds":3,"language":"Python","topics":["transformers"],"aiDims":["多模态Agent","模型与训练","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["多模态模型","图像理解","视觉语言","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-18T06:00:50Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/TinyLLaVA/TinyLLaVA_Factory","ts":"2026-09-18T06:00:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"frontman-ai/frontman","owner":"frontman-ai","name":"frontman","desc":"The AI agent that lives in your framework/browser","summaryCn":"嵌进浏览器帮你跑腿的AI小助理，自动操作网页太香了","reasonCn":"这是一个专为前端开发者设计的AI Agent，可以嵌入到你的项目框架或浏览器中运行。它具备网页自动化操作能力，能理解页面结构并执行点击、填写、滚动等交互任务。支持ReScript语言开发，与TypeScript生态兼容，适合需要AI辅助前端调试或自动化的场景。","stars":705,"starGrowth":1,"createdAt":"2025-09-29T08:38:03Z","silentRounds":3,"language":"ReScript","topics":["frontend"],"aiDims":["AI Agent","AI应用","开发者工具"],"zone":"AI","funScore":0.7,"domainTags":["AI代理","浏览器自动化","前端开发"],"domainKey":"AI代理","pushedAt":"2026-09-21T05:13:55Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"ReScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/frontman-ai/frontman","ts":"2026-09-21T05:13:55Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chatboxai/chatbox","owner":"chatboxai","name":"chatbox","desc":"Powerful AI Client","summaryCn":"简单来说，这个工具就像是一个全能的AI聊天遥控器。","reasonCn":"技术上，它完全基于Web前端技术栈构建，跨平台能力极佳，完美支持Windows、macOS和Linux。它将所有的聊天历史和配置都安全地保存在本地，不用担心隐私泄露。它支持主流的各大模型API转发，自带强大的Markdown和代码高亮渲染，甚至还支持画图、文件上传解析和联网搜索等扩展功能，极","stars":41836,"starGrowth":2,"createdAt":"2023-03-06T12:22:15Z","silentRounds":0,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI应用","效率工具"],"zone":"AI","funScore":1,"domainTags":["AI聊天","多模型聚合","桌面客户端","跨平台"],"domainKey":"AI聊天","pushedAt":"2026-09-16T13:37:52Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/chatboxai/chatbox","ts":"2026-09-16T13:37:52Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"encoredev/encore","owner":"encoredev","name":"encore","desc":"The infrastructure platform for the intelligence era","summaryCn":"用Go写云原生应用的新一代平台，自带部署和服务治理一把梭","reasonCn":"Encore是一个用Go语言编写的云原生应用开发平台，它将应用框架、服务发现和部署基础设施融为一体。开发者只需关注业务逻辑，Encore自动处理服务间通信、API路由、资源管理和云端部署，支持一键部署到AWS等云平台，大幅简化云原生应用的开发运维复杂度。","stars":12378,"starGrowth":1,"createdAt":"2021-02-17T16:33:24Z","silentRounds":1,"language":"Go","topics":["devops","database","backend"],"aiDims":["非AI-实用","开发者工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["云原生开发","服务发现","应用部署"],"domainKey":"云原生开发","pushedAt":"2026-09-23T14:18:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/encoredev/encore","ts":"2026-09-23T14:18:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jeedom/core","owner":"jeedom","name":"core","desc":" Software for home automation","summaryCn":"许多人在搭建智能家居时，往往会被厂商的云服务限制，一旦断网家里就瘫痪了","reasonCn":"从架构上看，它采用了基于PHP的后端系统，提供了极其丰富的插件生态（Plugin Market），通过不同协议的插件支持Z-Wave、Zigbee、WiFi等多种连接方式。它的独特之处在于强大的场景设计器（Scenario Builder），你可以通过拖拽或简单的脚本逻辑，设置复杂的联动：比如“当光线暗且门口","stars":415,"starGrowth":0,"createdAt":"2014-11-04T09:25:36Z","silentRounds":18,"language":"PHP","topics":["home-automation"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["智能家居","IoT","设备联动","本地自动化"],"domainKey":"智能家居","pushedAt":"2026-09-23T10:06:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/jeedom/core","ts":"2026-09-23T10:06:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sveltejs/svelte","owner":"sveltejs","name":"svelte","desc":"web development for the rest of us","summaryCn":"写前端代码像搭积木，比React简单十倍","reasonCn":"Svelte是一个革命性的前端框架，通过编译时转换将响应式状态管理逻辑注入到浏览器中，避免了React等框架运行时的虚拟DOM开销。它采用声明式语法，让开发者只需关注数据变化，无需手动操作DOM，大幅降低了前端开发的复杂度。","stars":88171,"starGrowth":3,"createdAt":"2016-11-20T18:13:05Z","silentRounds":0,"language":"JavaScript","topics":["template","ui","compiler"],"aiDims":["开发者工具","效率工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["前端框架","JavaScript","编译时框架","Web开发"],"domainKey":"前端框架","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"template","source":"github","weight":0.7},{"name":"ui","source":"github","weight":0.7},{"name":"compiler","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/sveltejs/svelte","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sb-ai-lab/RePlay","owner":"sb-ai-lab","name":"RePlay","desc":"A Comprehensive Framework for Building End-to-End Recommendation Systems with State-of-the-Art Models","summaryCn":"像搭积木一样构建推荐系统的开源框架，让大模型落地更简单","reasonCn":"该框架的核心亮点在于其高度的模块化和SOTA模型的集成。它内置了多种先进的推荐模型架构，从传统的矩阵分解到基于Transformer的深度模型一应俱全，并且持续更新最前沿的论文复现。开发者可以通过简单的配置文件，像搭积木一样组合不同的模型组件，而不需要从零开始写代码。","stars":413,"starGrowth":0,"createdAt":"2022-04-19T13:34:50Z","silentRounds":18,"language":"Python","topics":["transformers"],"aiDims":["大语言模型","模型与训练","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["推荐系统","Transformer","矩阵分解","深度学习","模型训练"],"domainKey":"推荐系统","pushedAt":"2026-09-16T12:43:10Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/sb-ai-lab/RePlay","ts":"2026-09-16T12:43:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"anomalyco/opencode","owner":"anomalyco","name":"opencode","desc":"The open source coding agent.","summaryCn":"开源版Cursor，AI帮你写代码改bug提PR一条龙","reasonCn":"技术上，Opencode的核心是一个可编程的AI Agent，它不是简单的ChatGPT套壳，而是能主动读取文件、执行命令、理解项目结构，然后根据上下文做出决策。它支持接入多种大模型，包括Claude、GPT-4、以及开源的Llama等，你可以按需切换。","stars":209405,"starGrowth":50,"createdAt":"2025-04-30T20:08:00Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0,"domainTags":["AI编程","Agent","代码生成","IDE集成"],"domainKey":"Agent","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/anomalyco/opencode","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.15071631527422935,"copyOk":true},{"repo":"google-coral/coralnpu","owner":"google-coral","name":"coralnpu","desc":"","summaryCn":"该项目代表了当前AI硬件加速领域的顶尖开源水平，其核心价值在于提供了一种","reasonCn":"在实际的应用场景中，硬件工程师可以将这一开源IP集成到自己的System-on-Chips设计中，通过软硬件协同优化的方式，在极其有限的电力预算内发挥出卓越的深度学习推理能力。由于这是一项底层的硬件IP工程，它并没有传统软件那种简单的点击运行安装方式，而是需要通过标准硬件工程流程","stars":2575,"starGrowth":1,"createdAt":"2025-10-02T22:32:37Z","silentRounds":1,"language":"Emacs Lisp","topics":[],"aiDims":["模型部署"],"zone":"工具","funScore":0.2,"domainTags":["NPU","RISC-V","硬件加速","AI芯片设计"],"domainKey":"NPU","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"Emacs Lisp","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/google-coral/coralnpu","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yusufkaraaslan/Skill_Seekers","owner":"yusufkaraaslan","name":"Skill_Seekers","desc":"Convert documentation websites, GitHub repositories, and PDFs into Claude AI skills with automatic conflict detection","summaryCn":"把文档和代码仓库一键变成AI能理解的技能配置，还能自动检测冲突。","reasonCn":"这个项目能把任何文档网站、GitHub仓库或PDF文件自动转换成Claude AI可用的技能模块。它内置了冲突检测机制，防止多个技能配置互相覆盖。对开发者来说，意味着可以把自己的知识文档快速变成AI agent能调用的工具，无需手写配置文件。支持批量转换和自动化流程。","stars":15020,"starGrowth":3,"createdAt":"2025-10-17T14:43:48Z","silentRounds":0,"language":"Python","topics":["automation"],"aiDims":["文档理解","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI技能包","文档转换","Claude"],"domainKey":"AI技能包","pushedAt":"2026-09-20T16:14:08Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/yusufkaraaslan/Skill_Seekers","ts":"2026-09-20T16:14:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AWeirdDev/flights","owner":"AWeirdDev","name":"flights","desc":"Fast, robust Google Flights scraper (API) for Python. (Probably)","summaryCn":"用Python写个爬虫，自动抓取谷歌机票信息，比人工快多了。","reasonCn":"这是一个基于Python开发的谷歌机票爬虫库，核心功能是通过编写脚本自动访问谷歌航班搜索页面并提取航班信息。它使用了强大的异步网络请求库，能够高效地抓取大量数据，并将结果以JSON格式返回，方便后续处理。项目支持多种过滤条件，用户可以轻松筛选出符合需求的航班。","stars":2067,"starGrowth":0,"createdAt":"2024-05-12T14:45:12Z","silentRounds":1,"language":"Python","topics":["travel"],"aiDims":["工具调用","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["机票爬虫","航班数据","异步网络请求","JSON解析"],"domainKey":"机票爬虫","pushedAt":"2026-09-21T13:48:46Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"travel","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/AWeirdDev/flights","ts":"2026-09-21T13:48:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aigc3d/LHM","owner":"aigc3d","name":"LHM","desc":"","summaryCn":"LHM++ 是该项目最新的开源版本，其最显著的技术突破在于极高的能效比：它支","reasonCn":"该项目隶属于阿里巴巴通义实验室，已被 ICCV 2025 录用，显示出其在计算机视觉领域的学术地位。除了最新的 LHM++ 外，项目还持续迭代了多种优化版本，例如发布于 2025 年 4 月的记忆节省版和 LHM-MINI。这些版本分别允许用户在 14GB 和 16GB 显存的显卡上完整运行处理流水线，为用","stars":2675,"starGrowth":0,"createdAt":"2025-03-13T18:48:56Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型部署","多模态Agent","推理引擎"],"zone":"AI","funScore":0.2,"domainTags":["3D高斯泼溅","视频理解","PyTorch","图形渲染","计算机视觉"],"domainKey":"3D高斯泼溅","pushedAt":"2026-09-21T12:25:34.711Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/aigc3d/LHM","ts":"2026-09-21T12:25:34.711Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"study8677/awesome-architecture","owner":"study8677","name":"awesome-architecture","desc":"🧭 Architecture-first system design: 26 bilingual tutorials, 25 architecture templates, and 6 end-to-end cases covering distributed systems, AI-native systems, RAG, coding Agents, and production trade-offs.","summaryCn":"26套架构教程加25个模板，手把手教你设计AI系统","reasonCn":"涵盖分布式系统、AI原生系统、RAG和编码Agent等方向，提供26个双语教程、25个架构模板和6个端到端案例，从入门到生产全覆盖。虽然本质是学习资源而非AI工具，但内容紧密围绕AI系统架构设计，对开发者搭建AI应用很有帮助，star数中等热度，实用性较强。","stars":2349,"starGrowth":0,"createdAt":"2026-05-23T02:44:43Z","silentRounds":2,"language":"Vue","topics":["backend"],"aiDims":["学习资源","教程","AI应用"],"zone":"资源","funScore":0,"domainTags":["系统架构","AI系统设计","教程","RAG","分布式系统"],"domainKey":"系统架构","pushedAt":"2026-09-21T03:20:49Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/study8677/awesome-architecture","ts":"2026-09-21T03:20:49Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/lucene","owner":"apache","name":"lucene","desc":"Apache Lucene open-source search software","summaryCn":"用它给App和网站加搜索功能，比AI还实用","reasonCn":"Apache Lucene是老牌Java搜索引擎库，用倒排索引实现毫秒级全文检索。Elasticsearch底层就是它。在AI领域可用于知识库检索、RAG系统的精确匹配层，与向量检索互补。成熟稳定，生产环境验证充分，是构建搜索相关AI应用的重要基础组件。","stars":3564,"starGrowth":0,"createdAt":"2021-02-23T17:16:56Z","silentRounds":3,"language":"Java","topics":["backend"],"aiDims":["AI搜索","RAG与知识","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["搜索引擎","倒排索引","Java"],"domainKey":"搜索引擎","pushedAt":"2026-09-23T08:36:19Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/apache/lucene","ts":"2026-09-23T08:36:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cloudreve/cloudreve","owner":"cloudreve","name":"cloudreve","desc":"🌩 Self-hosted file management and sharing system, supports multiple storage providers","summaryCn":"自建网盘神器，支持多平台存储，数据完全掌握在自己手里。","reasonCn":"这是一个完全开源的云存储分享系统，支持本地、阿里云OSS、S3等多种存储后端。它具备完整的文件管理功能，包括多用户、权限控制、在线预览、拖拽上传等。作为自建方案，它能替代各类付费网盘，确保数据隐私安全，同时提供类似网盘的分享功能。","stars":28740,"starGrowth":1,"createdAt":"2018-02-04T04:56:38Z","silentRounds":3,"language":"Go","topics":["file-sharing","file","cloud","file-manager","cloud-storage","cloudreve","golang"],"aiDims":["非AI-实用","工具调用"],"zone":"工具","funScore":0,"domainTags":["网盘","文件管理","自托管","多存储后端","Go"],"domainKey":"网盘","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"file-sharing","source":"github","weight":0.7},{"name":"file","source":"github","weight":0.7},{"name":"cloud","source":"github","weight":0.7},{"name":"file-manager","source":"github","weight":0.7},{"name":"cloud-storage","source":"github","weight":0.7},{"name":"cloudreve","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/cloudreve/cloudreve","ts":"2026-09-21T04:10:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"williamngan/pts","owner":"williamngan","name":"pts","desc":"A library for visualization and creative-coding","summaryCn":"创意编程神器，把数据变成炫酷的动态视觉艺术。","reasonCn":"它的核心技术在于强大的几何计算和灵活的渲染管线。支持多种数据输入格式，并内置了丰富的预设效果，比如粒子系统、流体模拟、分形生成等。你可以通过简单的参数调整，创造出千变万化的视觉体验。 安装非常简单，只需通过 npm 安装 pts 包，然后在代码中引入即可。","stars":5347,"starGrowth":0,"createdAt":"2017-03-24T14:09:43Z","silentRounds":4,"language":"TypeScript","topics":["design","creative-coding","generative-art"],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":0.5,"domainTags":["创意编程","可视化","生成艺术"],"domainKey":"创意编程","pushedAt":"2026-09-19T07:53:31Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"creative-coding","source":"github","weight":0.7},{"name":"generative-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/williamngan/pts","ts":"2026-09-19T07:53:31Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"datalevin/datalevin","owner":"datalevin","name":"datalevin","desc":"A simple, fast and versatile Datalog database","summaryCn":"向量数据库加Datalog查询，AI检索新玩法","reasonCn":"Datalevin是一个基于Datalog的向量数据库，将传统数据库的逻辑查询能力与向量相似性搜索相结合。核心亮点是Datalog查询语言支持递归推理和复杂关系查询，而向量搜索能力让语义检索成为可能。两者结合后可以同时做逻辑推理和语义匹配，这在现有向量数据库市场中很少见。","stars":1479,"starGrowth":0,"createdAt":"2020-06-08T23:53:52Z","silentRounds":9,"language":"Clojure","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","知识图谱"],"zone":"工具","funScore":0,"domainTags":["向量数据库","Datalog查询","知识图谱","语义检索"],"domainKey":"向量数据库","pushedAt":"2026-09-23T13:51:05Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Clojure","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/datalevin/datalevin","ts":"2026-09-23T13:51:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kijai/ComfyUI-SUPIR","owner":"kijai","name":"ComfyUI-SUPIR","desc":"","summaryCn":"SUPIR是一个基于扩散模型的图像超分辨率工具，能把模糊低清图片恢复成细节丰","reasonCn":"项目经历了几次重要更新。早期版本只能加载固定的SDXL模型，后来改进了加载方式，允许用户在超分流程中叠加LoRA——这给想要对特定风格或内容做定向增强的用户提供了很大的灵活性。同时作者还发布了经过剪枝处理的safetensors格式模型文件，体积更小、加载更快，对显存有限的用户更友","stars":2310,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","模型部署"],"zone":"工具","funScore":0.38,"domainTags":["图像放大","AI修图","ComfyUI","图像超分辨率"],"domainKey":"图像放大","pushedAt":"2026-09-22T20:36:00.388Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/kijai/ComfyUI-SUPIR","ts":"2026-09-22T20:36:00.388Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lintsinghua/DeepAudit","owner":"lintsinghua","name":"DeepAudit","desc":"DeepAudit：人人拥有的 AI 黑客战队，让漏洞挖掘触手可及。国内首个开源的代码漏洞挖掘多智能体系统。小白一键部署运行，自主协作审计 + 自动化沙箱 PoC 验证。支持 Ollama 私有部署 ，一键生成报告。支持中转站。​让安全不再昂贵，让审计不再复杂。","summaryCn":"让人人都能轻松使用的国内首个开源自动化代码漏洞挖掘战队","reasonCn":"技术架构上，它是国内首个专注于代码漏洞挖掘的多智能体系统。不同智能体各司其职，有的负责静态代码分析，有的负责漏洞推理，最厉害的是它配备了自动化沙箱环境，能够自主编写并运行PoC进行验证，极大地降低了误报率。","stars":7066,"starGrowth":0,"createdAt":"2025-09-19T15:53:48Z","silentRounds":1,"language":"Python","topics":["developer-tools"],"aiDims":["安全工具","AI Agent","开发者工具"],"zone":"AI","funScore":0,"domainTags":["代码审计","漏洞挖掘","安全"],"domainKey":"代码审计","pushedAt":"2026-09-16T08:31:47Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/lintsinghua/DeepAudit","ts":"2026-09-16T08:31:47Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"IridiumIO/CompactGUI","owner":"IridiumIO","name":"CompactGUI","desc":"Reduce the space taken up by games and programs on disk by using native Windows APIs","summaryCn":"游戏硬盘总不够？一键无损压缩腾空间，不删文件也能装更多","reasonCn":"CompactGUI利用Windows原生API对已安装的游戏和应用进行无损压缩，支持ZIP、TAR、7z等多种格式。压缩率可达20%-80%，无需重新下载安装，压缩后的文件依然可直接运行。界面基于WPF，操作直观，右键即可一键压缩或还原，是Windows玩家释放磁盘空间的实用利器。","stars":8888,"starGrowth":0,"createdAt":"2017-10-12T08:23:14Z","silentRounds":1,"language":"Visual Basic .NET","topics":["game"],"aiDims":["非AI-实用","游戏","效率工具"],"zone":"工具","funScore":0,"domainTags":["Windows","磁盘空间管理","游戏优化","无损压缩","压缩工具"],"domainKey":"Windows","pushedAt":"2026-09-23T03:20:39Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Visual Basic .NET","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/IridiumIO/CompactGUI","ts":"2026-09-23T03:20:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"vllm-project/production-stack","owner":"vllm-project","name":"production-stack","desc":"","summaryCn":"该项目从 Lambda Labs 到 AWS EKS 再到 Google GCP 等主流云服务商的实战部","reasonCn":"该架构方案不仅是一份静态指南，更是为了应对高并发、低延迟的实际生产需求。它通过预设好的环境配置与优化路径，让用户无需从零摸索复杂的 Kubernetes 调度策略或云端扩容设置。项目团队每两周举行一次社区会议，确保架构能够实时跟进 vLLM 最新的推理技术演进，保持部署方案的前","stars":2628,"starGrowth":3,"createdAt":"2025-01-21T23:09:11Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型部署","推理引擎"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","云原生部署","模型推理"],"domainKey":"大语言模型","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/vllm-project/production-stack","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"anchore/syft","owner":"anchore","name":"syft","desc":"CLI tool and library for generating a Software Bill of Materials from container images and filesystems","summaryCn":"给Docker镜像做体检，精准列出所有组件和漏洞，安全审计的必备神器。","reasonCn":"这是一个用于生成容器镜像软件物料清单（SBOM）的CLI工具，能精确识别镜像中的所有文件、库和组件。它不仅列出组件版本，还能与漏洞数据库对比，自动标记已知的安全隐患。支持Docker、Podman等多种运行时和文件系统格式，是DevSecOps流程中不可或缺的一环。","stars":9599,"starGrowth":0,"createdAt":"2020-05-07T18:19:29Z","silentRounds":1,"language":"Go","topics":["containers","docker","go","golang","static-analysis","tool","oci","sbom","spdx","cyclonedx","hacktoberfest"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["容器安全","SBOM","漏洞扫描","DevSecOps"],"domainKey":"容器安全","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"containers","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"static-analysis","source":"github","weight":0.7},{"name":"tool","source":"github","weight":0.7},{"name":"oci","source":"github","weight":0.7},{"name":"sbom","source":"github","weight":0.7},{"name":"spdx","source":"github","weight":0.7},{"name":"cyclonedx","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/anchore/syft","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nats-io/nats-server","owner":"nats-io","name":"nats-server","desc":"High-Performance server for NATS.io, the cloud and edge native messaging system.","summaryCn":"云原生消息队列之王，让AI服务之间高效通讯","reasonCn":"它的核心优势在于“高性能”和“弹性伸缩”。基于Go语言开发，单机就能轻松处理每秒几十万的消息，而且支持集群部署，数据可以自动同步，保证高可用。它支持多种通信模式：发布订阅适合广播场景，请求响应适合RPC调用，队列组适合负载均衡。","stars":20757,"starGrowth":1,"createdAt":"2012-10-29T16:12:24Z","silentRounds":2,"language":"Go","topics":["go","golang","messaging","message-bus","message-queue","cloud-native","microservices-architecture","nats-server","distributed-systems","cloud","cloud-computing","connected-vehicle","edge","edge-ai","edge-computing"],"aiDims":["AI基础设施"],"zone":"工具","funScore":0,"domainTags":["消息队列","NATS","微服务","通信","高性能"],"domainKey":"消息队列","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"messaging","source":"github","weight":0.7},{"name":"message-bus","source":"github","weight":0.7},{"name":"message-queue","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"microservices-architecture","source":"github","weight":0.7},{"name":"nats-server","source":"github","weight":0.7},{"name":"distributed-systems","source":"github","weight":0.7},{"name":"cloud","source":"github","weight":0.7},{"name":"cloud-computing","source":"github","weight":0.7},{"name":"connected-vehicle","source":"github","weight":0.7},{"name":"edge","source":"github","weight":0.7},{"name":"edge-ai","source":"github","weight":0.7},{"name":"edge-computing","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/nats-io/nats-server","ts":"2026-09-18T10:23:42.801Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liaohch3/claude-tap","owner":"liaohch3","name":"claude-tap","desc":"","summaryCn":"这个项目支持多达十几种Agent客户端的接入，无论是Claude Code还是主流的Ki","reasonCn":"在上手使用上，你只需将本地运行的CLI工具流量指向该代理即可。它不改变原有的调用方式，而是作为一名透明的旁观者，将原本枯燥的HTTP报文转化为直观的trace数据流。无论你是在排查工具调用参数的格式错误，还是想深入分析Agent处理大型代码库的路径规划，它都能为你提供近乎上帝视","stars":3231,"starGrowth":5,"createdAt":"2026-02-15T12:51:26Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["工具调用","AI Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI编程","Agent调试","网络代理"],"domainKey":"AI编程","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/liaohch3/claude-tap","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02357103064066852,"copyOk":false},{"repo":"castorini/pyserini","owner":"castorini","name":"pyserini","desc":"","summaryCn":"可复现的信息检索工具包，支持稀疏和密集向量检索两种方式。","reasonCn":"这是一个用于信息检索研究的Python工具包，主打可复现性。它同时支持基于Lucene的稀疏检索和基于Faiss的密集向量检索，可以方便地跑通BM25、DPR等经典模型。项目自带标准测试集和评估脚本，你输入一条命令就能重现论文里的实验结果，非常适合做检索方向的研究对比。","stars":2166,"starGrowth":4,"createdAt":"2019-11-01T23:36:34Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["RAG与知识","AI基础设施"],"zone":"AI","funScore":0.26,"domainTags":["信息检索","搜索引擎","密集检索","稀疏检索","RAG"],"domainKey":"RAG","pushedAt":"2026-09-22T22:51:50.275Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.59,"source":"search","url":"https://github.com/castorini/pyserini","ts":"2026-09-22T22:51:50.275Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wfxr/forgit","owner":"wfxr","name":"forgit","desc":":zzz: A utility tool powered by fzf for using git interactively.","summaryCn":"一款基于Fzf的Git交互工具，让你通过命令行更高效地使用Git。","reasonCn":"这款工具利用了Fzf强大的命令行搜索功能，让用户在处理复杂的Git操作时更加直观和高效。","stars":5083,"starGrowth":1,"createdAt":"2017-11-15T15:27:38Z","silentRounds":2,"language":"Shell","topics":["git"],"aiDims":["Git"],"zone":"工具","funScore":0,"domainTags":["Git","命令行","版本控制"],"domainKey":"Git","pushedAt":"2026-09-17T01:48:13Z","tags":[{"name":"Git","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/wfxr/forgit","ts":"2026-09-17T01:48:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cmang/gifterm","owner":"cmang","name":"gifterm","desc":"View animated .GIF files in a text console. Renders as ASCII art and/or Unicode art. Extended colors. ANSI and Durdraw file export. Linux/Mac/Windows","summaryCn":"在终端里把GIF画成ASCII艺术，复古又赛博朋克。","reasonCn":"技术上，它利用了Python的图像处理能力将GIF解码，然后通过终端的字符网格进行映射渲染。它不仅支持纯黑白字符，还扩展了彩色显示，让ASCII画也能五彩斑斓。对于开发者来说，它还能把渲染好的画面导出为ANSI或Durdraw格式的文件，方便在其他绘图工具里二次编辑。","stars":54,"starGrowth":0,"createdAt":"2015-09-18T11:14:29Z","silentRounds":12,"language":"Python","topics":["ascii-art"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"工具","funScore":0.8,"domainTags":["ASCII艺术","终端图形","GIF渲染","极客工具","字符动画"],"domainKey":"ASCII艺术","pushedAt":"2026-09-16T21:11:34Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ascii-art","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/cmang/gifterm","ts":"2026-09-16T21:11:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FasterDecoding/Medusa","owner":"FasterDecoding","name":"Medusa","desc":"","summaryCn":"Medusa是一个旨在简化大语言模型生成加速流程的开源框架。","reasonCn":"该项目的核心亮点在于其易用性和通用性。在Medusa-2版本中，项目推出了针对全模型训练的配方，同时也提供了自蒸馏功能，允许用户将Medusa添加到任何已经微调过的LLM中。这意味着开发者无需重新训练整个模型，只需在现有模型上添加Medusa组件，就能获得显著的性能提升。","stars":2773,"starGrowth":0,"createdAt":"2023-09-10T06:14:07Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["推理引擎","大语言模型","模型部署"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","推理加速","多解码头","开源框架"],"domainKey":"大语言模型","pushedAt":"2026-09-20T18:36:02.638Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/FasterDecoding/Medusa","ts":"2026-09-20T18:36:02.638Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rustfs/rustfs","owner":"rustfs","name":"rustfs","desc":"RustFS is an open-source, S3-compatible high-performance object storage system supporting migration and coexistence with other S3-compatible platforms such as MinIO and Ceph.","summaryCn":"Rust写的对象存储，速度比MinIO快两倍多，兼容S3协议","reasonCn":"这是一个用Rust编写的高性能对象存储系统，兼容Amazon S3协议。它的亮点是在4KB小对象场景下比MinIO快2.3倍，同时还支持与现有S3平台的迁移和共存。如果你在做AI相关的存储服务或者需要高性能对象存储，这个项目值得了解。","stars":33658,"starGrowth":24,"createdAt":"2023-11-23T13:45:10Z","silentRounds":0,"language":"Rust","topics":["bigdata","cloud-native","filesystem","object-storage","rust","s3","minio","amazon-s3","objectstorage","ai-native","multi-cloud","ai-storage"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["对象存储","S3兼容","Rust性能优化","边缘计算"],"domainKey":"对象存储","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"bigdata","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"filesystem","source":"github","weight":0.7},{"name":"object-storage","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"s3","source":"github","weight":0.7},{"name":"minio","source":"github","weight":0.7},{"name":"amazon-s3","source":"github","weight":0.7},{"name":"objectstorage","source":"github","weight":0.7},{"name":"ai-native","source":"github","weight":0.7},{"name":"multi-cloud","source":"github","weight":0.7},{"name":"ai-storage","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/rustfs/rustfs","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07413916453740567,"copyOk":false},{"repo":"PickleBoxer/dev-chatgpt-prompts","owner":"PickleBoxer","name":"dev-chatgpt-prompts","desc":"","summaryCn":"该资源库的核心价值在于提供了一种“提示词链”的思维模式，通过将复杂的开发","reasonCn":"你可以在仓库中直接查阅这些分类明确的提示词内容，涵盖了从简单的代码优化请求到复杂的项目架构建议，所有内容均以文本形式开源供人免费参考。它不需要任何安装流程，只需要下载或收藏仓库，便可在需要时随时查阅并复制其中的技巧。","stars":2272,"starGrowth":0,"silentRounds":0,"language":"","topics":[],"aiDims":["提示工程","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["提示工程","开发者资源","ChatGPT技巧"],"domainKey":"提示工程","pushedAt":"2026-09-22T21:05:02.747Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.62,"source":"search","url":"https://github.com/PickleBoxer/dev-chatgpt-prompts","ts":"2026-09-22T21:05:02.747Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"veloren/veloren","owner":"veloren","name":"veloren","desc":"[mirror of https://gitlab.com/veloren/veloren] An open world, open source voxel RPG inspired by Dwarf Fortress and Cube World. This repository is a mirror. Please submit all PRs and issues on our GitLab page.","summaryCn":"不用花一分钱，开源体素RPG带你穿越矮人地牢冒险","reasonCn":"基于Rust开发的开放世界体素RPG游戏，灵感来自矮人要塞和Cube World，支持多人联机。游戏从1998年就开始开发，历经二十多年社区持续贡献，目前已支持Linux、Windows和Mac平台。虽然画面风格偏可爱，但游戏深度和自由度相当出色，完全免费开源，是独立游戏开发的优秀参考项目。","stars":7582,"starGrowth":1,"createdAt":"2018-05-29T00:49:57Z","silentRounds":1,"language":"Rust","topics":["game"],"aiDims":["游戏"],"zone":"创意","funScore":0.6,"domainTags":["体素RPG","开放世界","多人联机","Rust游戏"],"domainKey":"体素RPG","pushedAt":"2026-09-22T22:13:24Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/veloren/veloren","ts":"2026-09-22T22:13:24Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"spicylobstergames/shotcaller-godot","owner":"spicylobstergames","name":"shotcaller-godot","desc":"2D top-down 1v1 MOBA with autonomous units. Design doc below 👇 ","summaryCn":"像素风2D俯视视角MOBA，单机版1v1对战，AI小兵自行进攻防守。","reasonCn":"ShotCaller是用Godot与GDScript实现的2D顶视MOBA，玩家在同一屏幕上进行1对1对战，场上每支阵营拥有自主行动的AI单位，能够根据脚本自行寻找目标、移动、攻击并协同防守。项目提供完整的像素艺术资源、技能系统与路径寻找模块，设计文档详细阐述了单位行为树、冷却机制及平衡调参方法，源码 MIT 许可，支持自定义英雄属性和插件，可加入局域网多人对战，适合作为案例构建策略游戏。","stars":300,"starGrowth":0,"createdAt":"2022-01-18T16:59:25Z","silentRounds":12,"language":"GDScript","topics":["pixel-art"],"aiDims":["游戏","AI Agent","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["Godot","MOBA 游戏","像素艺术"],"domainKey":"Godot","pushedAt":"2026-09-20T18:25:14Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/spicylobstergames/shotcaller-godot","ts":"2026-09-20T18:25:14Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"plankanban/planka","owner":"plankanban","name":"planka","desc":"PLANKA is the Kanban-style project mastering tool for everyone","summaryCn":"开源看板神器，像Trello那样简单好用，却完全掌控在自己手里。","reasonCn":"在功能上，Planka不仅限于基础的拖拽。它支持多看板管理、标签过滤、子任务以及附件上传。用户还可以自定义视图布局，比如甘特图或日历视图，以适应不同的工作流程。所有数据都保存在本地数据库（SQLite）中，这意味着你不需要配置复杂的数据库服务，开箱即用，非常适合私有化部署","stars":12575,"starGrowth":1,"createdAt":"2019-08-30T20:35:36Z","silentRounds":2,"language":"JavaScript","topics":["self-hosted","docker"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["看板管理","项目协作","Trello替代"],"domainKey":"看板管理","pushedAt":"2026-09-17T01:52:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/plankanban/planka","ts":"2026-09-17T01:52:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVlabs/cuda-oxide","owner":"NVlabs","name":"cuda-oxide","desc":"cuda-oxide is a Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign language bindings, just Rust.","summaryCn":"在 GPU 开发领域，传统方式往往需要学习 CUDA C/C++，或使用 Python 的 Num","reasonCn":"在 GPU 开发领域，传统方式往往需要学习 CUDA C/C++，或使用 Python 的 Numba、CuPy 等封装层，学习成本高且易出错。cuda-oxide 直接把纯 Rust 代码编译成 PTX，省去中间语言的桥接，让 Rust 开发者能够像写普通函数一样写 GPU kernel，既保持了语言的安全性，又能利用 CUDA 的强大","stars":3566,"starGrowth":5,"createdAt":"2026-04-22T17:03:09Z","silentRounds":0,"language":"Rust","topics":["compiler-backend","async","cuda","gpu","high-performance-computing","nvidia","programming-languages","rust","heterogeneous-computing"],"aiDims":["AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Rust","GPU编程","CUDA","编译器","高性能计算"],"domainKey":"Rust","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"compiler-backend","source":"github","weight":0.7},{"name":"async","source":"github","weight":0.7},{"name":"cuda","source":"github","weight":0.7},{"name":"gpu","source":"github","weight":0.7},{"name":"high-performance-computing","source":"github","weight":0.7},{"name":"nvidia","source":"github","weight":0.7},{"name":"programming-languages","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"heterogeneous-computing","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/NVlabs/cuda-oxide","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.023047672462142454,"copyOk":false},{"repo":"Provenance-Emu/Provenance","owner":"Provenance-Emu","name":"Provenance","desc":"iOS & tvOS multi-emulator frontend, supporting various Atari, Bandai, NEC, Nintendo, Sega, SNK and Sony console systems… Get Started:  https://wiki.provenance-emu.com  | ","summaryCn":"一部iPhone搞定N台经典游戏机，情怀拉满","reasonCn":"Provenance是一款iOS和tvOS平台的开源多模拟器前端，整合了Atari、世嘉、任天堂、索尼等多家经典游戏机的模拟核心。界面精美，支持蓝牙手柄，能流畅运行大量复古游戏。适合怀旧玩家和收藏经典游戏的朋友，一键安装即可体验童年。","stars":6377,"starGrowth":0,"createdAt":"2013-08-27T20:34:36Z","silentRounds":3,"language":"C","topics":["game","emulator"],"aiDims":["游戏","非AI-好玩","可视化"],"zone":"创意","funScore":1,"domainTags":["iOS模拟器","复古游戏","怀旧","游戏前端","手柄支持"],"domainKey":"iOS模拟器","pushedAt":"2026-09-21T07:47:55Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"emulator","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/Provenance-Emu/Provenance","ts":"2026-09-21T07:47:55Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"huggingface/agents-course","owner":"huggingface","name":"agents-course","desc":"This repository contains the Hugging Face Agents Course. ","summaryCn":"想要快速掌握 AI Agent，却不知道从哪儿入手？","reasonCn":"想要快速掌握 AI Agent，却不知道从哪儿入手？Hugging Face 官方推出的 agents-course 正是为此而生。它把抽象的智能体概念拆解成易懂的模块，用实际代码示例演示如何让模型自行调用工具、检索文档并进行多轮对话。\\n· 课程结构分为七大章节：从 Agent 基础、Prompt 设计、工具函数","stars":32778,"starGrowth":12,"createdAt":"2025-01-16T19:50:22Z","silentRounds":0,"language":"MDX","topics":["agents"],"aiDims":["AI Agent","教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["AI Agent","教程","机器学习","课程"],"domainKey":"AI Agent","pushedAt":"2026-09-15T00:53:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"MDX","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/huggingface/agents-course","ts":"2026-09-15T00:53:42Z","category":"learning","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.03709829763866007,"copyOk":false},{"repo":"streamyfin/streamyfin","owner":"streamyfin","name":"streamyfin","desc":"A modern Jellyfin client built with Expo","summaryCn":"StreamYin是一个基于ExoTv平台的流媒体播放器，它提供了一种简单的方式来观","reasonCn":"StreamYin是一个基于ExoTv平台的流媒体播放器，它提供了一种简单的方式来观看各种视频内容。它支持多种视频格式，并且具有内置的视频解码器，可以播放高清视频内容。这使得用户无需安装额外的视频播放器，就可以享受流畅的视频播放体验。","stars":5206,"starGrowth":2,"createdAt":"2024-07-31T21:20:53Z","silentRounds":0,"language":"TypeScript","topics":["jellyfin"],"aiDims":["Jellyfin"],"zone":"工具","funScore":0,"domainTags":["Jellyfin","流媒体","视频播放","影音"],"domainKey":"Jellyfin","pushedAt":"2026-09-21T21:29:55Z","tags":[{"name":"Jellyfin","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/streamyfin/streamyfin","ts":"2026-09-21T21:29:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"generative-computing/granite-switch","owner":"generative-computing","name":"granite-switch","desc":"Granite Switch — Build AI models like you build software","summaryCn":"开发AI模型经常面临代码像一团乱麻的窘境，尤其是当你尝试替换模型架构中的","reasonCn":"它的核心技术亮点在于高度的解耦设计，底层深度兼容Transformers库，但上层封装了更优雅的Switch机制。这种设计允许你在同一个训练管道中无缝切换不同的层级实现，并能自动处理配置的同步，极大地提升了实验的可重现性与迭代开发效率。","stars":96,"starGrowth":0,"createdAt":"2026-05-04T19:20:26Z","silentRounds":18,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","AI基础设施","大语言模型"],"zone":"AI","funScore":0,"domainTags":["AI模型","深度学习","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-19T22:55:58Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/generative-computing/granite-switch","ts":"2026-09-19T22:55:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kserve/kserve","owner":"kserve","name":"kserve","desc":"Standardized Distributed Generative and Predictive AI Inference Platform for Scalable, Multi-Framework Deployment on Kubernetes","summaryCn":"在传统部署环境里，模型上线常常要写大量脚本、手动配置负载均衡，容易出现","reasonCn":"在传统部署环境里，模型上线常常要写大量脚本、手动配置负载均衡，容易出现环境不一致、扩容慢等痛点。KServe 把这些繁杂步骤封装成 Kubernetes 原生的 CRD（自定义资源定义），只需要写一个 YAML 文件描述模型位置、运行时框架和资源需求，平台就会自动创建 Deployment、Service、","stars":5989,"starGrowth":1,"createdAt":"2019-03-27T21:14:14Z","silentRounds":1,"language":"Go","topics":["knative","machine-learning","model-interpretability","model-serving","istio","kubeflow","artificial-intelligence","tensorflow","pytorch","xgboost","kubernetes","k8s","service-mesh","kserve","hacktoberfest","mlops","genai","llm-inference","cncf","vllm"],"aiDims":["AI基础设施","推理引擎","模型部署"],"zone":"AI","funScore":0,"domainTags":["模型部署","Kubernetes","推理服务","MLOps"],"domainKey":"模型部署","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"knative","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"model-interpretability","source":"github","weight":0.7},{"name":"model-serving","source":"github","weight":0.7},{"name":"istio","source":"github","weight":0.7},{"name":"kubeflow","source":"github","weight":0.7},{"name":"artificial-intelligence","source":"github","weight":0.7},{"name":"tensorflow","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"xgboost","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"k8s","source":"github","weight":0.7},{"name":"service-mesh","source":"github","weight":0.7},{"name":"kserve","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"mlops","source":"github","weight":0.7},{"name":"genai","source":"github","weight":0.7},{"name":"llm-inference","source":"github","weight":0.7},{"name":"cncf","source":"github","weight":0.7},{"name":"vllm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.94,"source":"trending","url":"https://github.com/kserve/kserve","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"snape/RVO2","owner":"snape","name":"RVO2","desc":"Optimal Reciprocal Collision Avoidance (C++)","summaryCn":"多人游戏NPC智能避让、机器人编队不撞车的核心算法库","reasonCn":"RVO2实现了最优互惠碰撞避免算法，专为多智能体动态避障设计。核心是速度障碍法，让数百个虚拟角色或机器人实时计算避让路径，互不碰撞到达各自目标。在机器人协同导航、游戏NPC行为控制、人群模拟等领域应用广泛，是群体智能和运动规划的经典开源实现。","stars":966,"starGrowth":0,"createdAt":"2013-06-25T03:11:32Z","silentRounds":27,"language":"C++","topics":["collision-avoidance","cpp","motion-planning","robotics","velocity-obstacles","virtual-agents"],"aiDims":["AI Agent","游戏","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["碰撞检测","路径规划","机器人仿真"],"domainKey":"碰撞检测","pushedAt":"2026-09-22T18:04:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"collision-avoidance","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"motion-planning","source":"github","weight":0.7},{"name":"robotics","source":"github","weight":0.7},{"name":"velocity-obstacles","source":"github","weight":0.7},{"name":"virtual-agents","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.62,"source":"bigbro","url":"https://github.com/snape/RVO2","ts":"2026-09-22T18:04:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"derailed/k9s","owner":"derailed","name":"k9s","desc":"🐶 Kubernetes CLI To Manage Your Clusters In Style!","summaryCn":"Kubernetes的炫酷命令行工具，像玩RPG游戏一样管理集群。","reasonCn":"这是Kubernetes的终端管理工具，界面炫酷，操作高效。支持键盘快捷键快速导航、查看Pod/Service/Deployment等资源，还有实时日志监控和故障排查功能。相比kubectl，更直观易用。适合每天与K8s打交道的运维和开发同学，大幅提升集群管理效率。","stars":34649,"starGrowth":1,"createdAt":"2019-01-25T18:46:02Z","silentRounds":1,"language":"Go","topics":["kubernetes"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["Kubernetes","CLI","终端","运维工具","资源监控"],"domainKey":"Kubernetes","pushedAt":"2026-09-22T18:35:00Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/derailed/k9s","ts":"2026-09-22T18:35:00Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DioxusLabs/dioxus","owner":"DioxusLabs","name":"dioxus","desc":"Fullstack app framework for web, desktop, and mobile.","summaryCn":"用Rust写UI跨端应用，一套代码同时搞定桌面网页移动","reasonCn":"Dioxus是一款基于Rust语言的全栈应用框架，采用类似React的声明式UI开发模式。它允许开发者用单一代码库同时构建Web、桌面端及移动端应用，通过成熟的跨平台编译方案显著提升开发效率，适合追求高性能与代码复用率的现代应用开发者。","stars":39213,"starGrowth":3,"createdAt":"2021-01-15T01:57:26Z","silentRounds":0,"language":"Rust","topics":["desktop"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Rust","跨平台","UI框架","全栈"],"domainKey":"Rust","pushedAt":"2026-09-21T12:33:02Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/DioxusLabs/dioxus","ts":"2026-09-21T12:33:02Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Tencent-Hunyuan/Hunyuan3D-1","owner":"Tencent-Hunyuan","name":"Hunyuan3D-1","desc":"","summaryCn":"腾讯开源的3D生成模型，输入文字或图片即可创建立体模型。","reasonCn":"这是腾讯混元团队推出的统一3D生成框架，支持文本描述或参考图片直接生成可用的3D模型资产。项目已迭代到2.1版本，号称生产级可用，还提供在线试玩网站。对游戏开发、电商展示、影视特效等需要3D素材的行业尤其有用，无需专业建模功底。","stars":3486,"starGrowth":2,"createdAt":"2024-10-31T09:04:57Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","AI应用"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["3D生成","文本转3D","图像转3D","多视图扩散模型"],"domainKey":"3D生成","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/Tencent-Hunyuan/Hunyuan3D-1","ts":"2026-09-19T18:31:48.187Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"facebookresearch/tribev2","owner":"facebookresearch","name":"tribev2","desc":"","summaryCn":"融合多模态感知与大模型的神经科学大脑编码预测工具。","reasonCn":"在实际使用时，研究人员可以非常方便地从 HuggingFace 平台加载预训练好的模型权重，并直接输入一段视频来预测大脑的响应情况。需要说明的是，这些预测结果是针对标准平均受试者在约两万个顶点的 fsaverage5 皮层网格上计算得出的，并且为了补偿血流动力学的延迟，模型在时间上进行","stars":3231,"starGrowth":0,"createdAt":"2026-03-24T10:28:01Z","silentRounds":3,"language":"Jupyter Notebook","topics":[],"aiDims":["大语言模型","多模态Agent","AI基础设施"],"zone":"AI","funScore":0.44,"domainTags":["神经科学","多模态","fMRI","Transformer"],"domainKey":"多模态","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/facebookresearch/tribev2","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"dsh0416/godot-cef","owner":"dsh0416","name":"godot-cef","desc":"A cross-platform GPU-accelerated CEF-based (Chromium Embedded Framework) WebView Extension for Godot 4","summaryCn":"给Godot游戏引擎装上Chromium内核，网页技术玩转游戏开发。","reasonCn":"这是一个基于Rust开发的Godot 4插件，它利用CEF技术将Chromium的强大Web渲染能力嵌入到游戏引擎中。这让开发者可以直接在游戏里加载复杂的HTML5内容，或者把Web应用直接变成游戏的一部分，极大地扩展了Godot的表现力。","stars":222,"starGrowth":0,"createdAt":"2025-12-31T12:25:54Z","silentRounds":4,"language":"Rust","topics":["godot"],"aiDims":["游戏","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["Godot","WebView","Chromium","Rust","游戏开发"],"domainKey":"Godot","pushedAt":"2026-09-19T17:48:43Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/dsh0416/godot-cef","ts":"2026-09-19T17:48:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"project-baize/baize-chatbot","owner":"project-baize","name":"baize-chatbot","desc":"","summaryCn":"这是一个在开源大模型发展早期备受关注的对话机器人项目，其核心特色在于探","reasonCn":"为了让这些开源模型能够适配更多应用场景，项目开源了不同参数量级的模型权重，涵盖了7B、13B甚至更大的30B版本。同时，开发团队贴心地提供了专门的转换脚本，能够将通过LoRA微调得到的模型权重无缝合并到标准的Hugging Face模型架构中，从而确保它能够在所有支持Hugging Face生态","stars":3145,"starGrowth":0,"createdAt":"2023-03-31T17:56:03Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["大语言模型","微调","AI应用"],"zone":"AI","funScore":0.38,"domainTags":["大语言模型","LoRA微调","对话机器人","ChatGPT"],"domainKey":"大语言模型","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/project-baize/baize-chatbot","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lightdash/lightdash","owner":"lightdash","name":"lightdash","desc":"Agentic BI. Analytics at the speed of code ⚡️","summaryCn":"用AI对话就能分析数据出报表，不用写SQL也能做商业智能","reasonCn":"核心能力是它的Agentic架构，本质上是把LLM当作数据分析的代理，你提问，它去查数据、写查询、生成可视化，全程自动化。它支持多种数据库连接，包括Snowflake、BigQuery、PostgreSQL等主流数据源，同时也支持传统的BI建模方式，既有AI的灵活性也有传统BI的严谨性。","stars":6152,"starGrowth":2,"createdAt":"2021-03-19T09:44:40Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["AI应用","非AI-实用","可视化"],"zone":"工具","funScore":0,"domainTags":["Agentic BI","数据分析","数据可视化","自托管","数据工程"],"domainKey":"Agentic BI","pushedAt":"2026-09-23T21:17:31Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/lightdash/lightdash","ts":"2026-09-23T21:17:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fluxcd/flux2","owner":"fluxcd","name":"flux2","desc":"Open and extensible continuous delivery solution for Kubernetes. Powered by GitOps Toolkit.","summaryCn":"让K8s应用像Git一样自动部署的自动化神器","reasonCn":"基于GitOps理念构建的Kubernetes持续交付工具，通过声明式配置管理集群状态。核心是GitOps Toolkit，利用Git作为单一事实来源，实现应用配置与集群环境的实时同步，无需手动操作。支持多集群管理、镜像自动更新和故障自愈。","stars":8421,"starGrowth":2,"createdAt":"2020-04-24T09:38:21Z","silentRounds":0,"language":"Go","topics":["gitops-toolkit","gitops","continuous-delivery","kubernetes","kustomize","helm"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["GitOps","Kubernetes","持续交付","CI/CD"],"domainKey":"GitOps","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"gitops-toolkit","source":"github","weight":0.7},{"name":"gitops","source":"github","weight":0.7},{"name":"continuous-delivery","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"kustomize","source":"github","weight":0.7},{"name":"helm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/fluxcd/flux2","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Warzone2100/warzone2100","owner":"Warzone2100","name":"warzone2100","desc":"Command the forces of The Project in a battle to rebuild the world after mankind has been nearly destroyed by nuclear missiles. A 100% free and open source real-time strategy game for Windows, macOS, Linux, BSD+","summaryCn":"核战后重建世界的经典开源RTS游戏，策略对战超上瘾","reasonCn":"Warzone2100是开源即时战略游戏，继承自1999年的经典名作。游戏设定在核战后世界，玩家需要指挥部队重建文明。支持单人战役和多人对战，拥有完整的游戏引擎和编辑器。跨平台支持Windows、macOS、Linux。适合RTS爱好者和怀旧玩家，是一款完全免费的高质量策略游戏。","stars":3964,"starGrowth":1,"createdAt":"2010-11-16T04:36:26Z","silentRounds":1,"language":"C++","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["即时战略","开源游戏","多人对战"],"domainKey":"即时战略","pushedAt":"2026-09-22T16:37:46Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/Warzone2100/warzone2100","ts":"2026-09-22T16:37:46Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MARKTECHPOST-AI-MEDIA-INC/AI-Agents-Projects-Tutorials","owner":"MARKTECHPOST-AI-MEDIA-INC","name":"AI-Agents-Projects-Tutorials","desc":"","summaryCn":"你可以在此项目中发现多个关于构建“智能代理”的具体教程，这些教程不仅限于","reasonCn":"项目的一大特色是其对“边缘计算”与“自我进化”能力的关注。通过OpenSpace等架构，教程引导用户尝试构建能够重复利用技能、具备上下文记忆的AI工作流，这对于追求降低API使用成本、提升任务执行稳定性的开发者非常有参考价值。","stars":2909,"starGrowth":1,"createdAt":"2025-05-15T09:00:54Z","silentRounds":3,"language":"Jupyter Notebook","topics":[],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI代理","金融分析","自动化工作流","编程辅助"],"domainKey":"AI代理","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/MARKTECHPOST-AI-MEDIA-INC/AI-Agents-Projects-Tutorials","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"golang/go","owner":"golang","name":"go","desc":"The Go programming language","summaryCn":"Google出的高性能编程语言，让开发变得又快又稳","reasonCn":"它的内存管理采用了自动垃圾回收机制，开发者不需要手动分配和释放内存，这大大降低了内存泄漏的风险。Go语言的编译速度非常快，能实现热编译，这对于开发效率的提升是巨大的。此外，Go语言对网络编程和系统调用有很好的支持，非常适合编写网络服务器、命令行工具和微服务组件。","stars":138953,"starGrowth":7,"createdAt":"2014-08-19T04:33:40Z","silentRounds":0,"language":"Go","topics":["programming-language","language","go","golang"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["编程语言","编译器","并发编程","后端开发","系统编程"],"domainKey":"编程语言","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"programming-language","source":"github","weight":0.7},{"name":"language","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/golang/go","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02115113023828201,"copyOk":true},{"repo":"clusterzx/paperless-ai","owner":"clusterzx","name":"paperless-ai","desc":"An automated document analyzer for Paperless-ngx using OpenAI API, Ollama, Deepseek-r1, Azure and all OpenAI API compatible Services to automatically analyze and tag your documents.","summaryCn":"发票合同丢进去就能自动分类打标签的小助手","reasonCn":"它给开源文档归档系统装上智能内核，读取扫描页文字后判断文件类型、日期、金额与来源，再按模板生成标签并写入目录。可接云端接口、本地推理或兼容模型，也支持自定义字段与提示词，把杂乱票据、合同、账单整理成可搜索的档案库，并保留人工复核通道，方便日后追踪。","stars":5950,"starGrowth":0,"createdAt":"2024-12-01T16:41:07Z","silentRounds":1,"language":"JavaScript","topics":["automation"],"aiDims":["AI应用","文档理解"],"zone":"AI","funScore":0.6,"domainTags":["文档管理","自动打标","Paperless-ngx","插件","发票识别"],"domainKey":"文档管理","pushedAt":"2026-09-19T00:23:27Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/clusterzx/paperless-ai","ts":"2026-09-19T00:23:27Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"IlyaRice/RAG-Challenge-2","owner":"IlyaRice","name":"RAG-Challenge-2","desc":"","summaryCn":"该项目是 RAG Challenge 竞赛的获胜解决方案，专门解决从复杂公司年报中提取","reasonCn":"为了应对更复杂的场景，如多公司对比分析，项目实现了专门的查询路由机制。这意味着系统能够识别用户意图，将简单查询和复杂对比查询分流处理，从而在保持速度的同时保证深度分析的准确性。虽然作者自称代码风格“scrappy”（粗糙实用），但这正是竞赛代码的真实写照——专注于效果最大","stars":2446,"starGrowth":0,"createdAt":"2025-03-19T04:48:45Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["RAG与知识","文档理解","提示工程"],"zone":"AI","funScore":0.2,"domainTags":["RAG 检索","PDF 解析","年度报告问答"],"domainKey":"RAG 检索","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/IlyaRice/RAG-Challenge-2","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"riffusion/riffusion-app-hobby","owner":"riffusion","name":"riffusion-app-hobby","desc":"","summaryCn":"Riffusion 曾首创了一种极具创意的玩法：不训练专用的音乐模型，而是把音频","reasonCn":"从技术栈视角看，这里是 Next.js 与重计算 AI 任务对接的典型案例。参考文档显示，它由 Next.js、React 和三.js 堆叠而成，其中 /pages/api 目录专门用于映射 API 路由。这种分离式设计（前端负责实时交互，后端负责繁重的扩散计算）是构建此类 AI 应用的经典架构，为开发者提供了","stars":2673,"starGrowth":0,"createdAt":"2022-11-20T19:20:04Z","silentRounds":3,"language":"TypeScript","topics":[],"aiDims":["大语言模型","模型与训练","多模态Agent"],"zone":"创意","funScore":0.44,"domainTags":["AI 音乐生成","Stable Diffusion","Web 应用"],"domainKey":"AI 音乐生成","pushedAt":"2026-09-21T12:25:34.711Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/riffusion/riffusion-app-hobby","ts":"2026-09-21T12:25:34.711Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"saharmor/dalle-playground","owner":"saharmor","name":"dalle-playground","desc":"","summaryCn":"生成式图像模型更新极快，普通用户往往难以获取最新的Demo环境。","reasonCn":"该项目的最大亮点在于其极简的部署路径。用户无需配置复杂的Python环境或下载庞大的模型权重，只需在 Google Colab 中运行一个简单的Notebook即可启动后端服务。随后，复制生成的URL，即可在浏览器中访问这个前端界面，实现了从代码到可视化工具的无缝衔接。","stars":2740,"starGrowth":0,"createdAt":"2021-09-13T10:35:50Z","silentRounds":3,"language":"JavaScript","topics":[],"aiDims":["创意工具","AI应用"],"zone":"创意","funScore":0.5,"domainTags":["Stable Diffusion","文生图","Web应用"],"domainKey":"Stable Diffusion","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/saharmor/dalle-playground","ts":"2026-09-21T04:10:58.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DataDog/datadog-agent","owner":"DataDog","name":"datadog-agent","desc":"Main repository for Datadog Agent","summaryCn":"全球大厂都在用的监控侦探，给服务器装上百双眼睛，任何异常都逃不掉","reasonCn":"这是Datadog官方的主机和容器代理，用Go写成，轻量且资源占用低。它负责从你的基础设施里采集指标、日志和分布式追踪数据，并加密发送给Datadog平台。支持Kubernetes、Docker等多种环境自动发现，是构建可观测性体系的基石组件。","stars":3737,"starGrowth":0,"createdAt":"2016-01-19T17:40:41Z","silentRounds":2,"language":"Go","topics":["monitoring"],"aiDims":["非AI-实用","开发者工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["监控","Kubernetes","日志采集","可观测性","基础设施"],"domainKey":"监控","pushedAt":"2026-09-23T21:48:11Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/DataDog/datadog-agent","ts":"2026-09-23T21:48:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huytieu/COG-second-brain","owner":"huytieu","name":"COG-second-brain","desc":"Self-evolving second brain with 33 AI skills, 10 agents, and people CRM. Closed-loop harness: a V-model verification lifecycle where the worker never grades its own homework. Plus paired anti-slop design skills for marketing and product UI. Works with Claude Code, Cursor, Kiro, Gemini CLI, Codex.","summaryCn":"你是否也觉得大脑里存的笔记越来越多，却从未真正用起来？","reasonCn":"核心架构采用了闭环验证机制，这类似于软件工程中的V模型开发生命周期。一个Agent负责产出，另一个Agent负责监督审核，杜绝了AI胡言乱语的可能。这种设计不仅提升了逻辑推理的可靠性，还通过整合个人CRM功能，让你的知识与人际互动产生了深度联系。","stars":1212,"starGrowth":1,"createdAt":"2025-10-15T03:50:52Z","silentRounds":1,"language":"HTML","topics":["knowledge-management"],"aiDims":["AI Agent","知识管理"],"zone":"AI","funScore":0.7,"domainTags":["知识管理","AI Agent","个人助手","对话系统"],"domainKey":"AI Agent","pushedAt":"2026-09-18T05:15:09Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识管理","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/huytieu/COG-second-brain","ts":"2026-09-18T05:15:09Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cosmicstack-labs/mercury-agent","owner":"cosmicstack-labs","name":"mercury-agent","desc":"Soul-driven AI agent with permission-hardened tools, token budgets, and multi-channel access. Runs 24/7 from CLI or Telegram.","summaryCn":"带着钱包和安全锁的灵魂级AI，全天候在电报和命令行陪你聊天。","reasonCn":"在技术架构上，它采用了TypeScript开发，强调权限的硬化隔离，所有外部工具调用都经过严格的沙盒和权限审查，防止恶意注入。它完美打通了Telegram和CLI双端，支持全天候持续在线运行，能够处理复杂的长文本记忆和多模态上下文交互。","stars":3168,"starGrowth":1,"createdAt":"2026-04-20T12:46:31Z","silentRounds":1,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","安全工具"],"zone":"AI","funScore":1,"domainTags":["AI Agent","Telegram","多模态交互","自动化"],"domainKey":"AI Agent","pushedAt":"2026-09-16T17:33:29Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.89,"source":"search","url":"https://github.com/cosmicstack-labs/mercury-agent","ts":"2026-09-16T17:33:29Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"neo4j/neo4j","owner":"neo4j","name":"neo4j","desc":"Graphs for Everyone","summaryCn":"图数据库界的老大，关系数据存储和查询的王者","reasonCn":"最流行的开源图数据库Neo4j，用Cypher查询语言操作节点和关系。原生图存储引擎，查询性能远超传统关系型数据库。支持ACID事务、高性能图算法和向量索引。社区版免费，企业版功能更强大。广泛应用于社交网络、推荐系统、风控等领域。","stars":17252,"starGrowth":0,"createdAt":"2012-11-12T08:46:15Z","silentRounds":1,"language":"Java","topics":["database"],"aiDims":["知识图谱","AI基础设施","向量数据库"],"zone":"工具","funScore":0,"domainTags":["图数据库","Cypher查询","知识图谱","关系存储","图算法"],"domainKey":"图数据库","pushedAt":"2026-09-22T10:36:59Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/neo4j/neo4j","ts":"2026-09-22T10:36:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenXRay/xray-16","owner":"OpenXRay","name":"xray-16","desc":"Improved version of the X-Ray Engine, the game engine used in the world-famous S.T.A.L.K.E.R. game series by GSC Game World. Join OpenXRay! ;)","summaryCn":"经典潜行者游戏引擎的重制改进版，持续开发中","reasonCn":"OpenXRay是基于X-Ray引擎的开源重制版，用于运行S.T.A.L.K.E.R.系列经典游戏。X-Ray引擎是乌克兰工作室GSC Game World专为潜行者系列开发的渲染引擎，OpenXRay在其基础上进行大量改进和持续开发。支持现代操作系统，修复原版漏洞，提升图形兼容性，让这款经典末日生存游戏在新硬件上也能流畅运行。","stars":3566,"starGrowth":1,"createdAt":"2014-09-05T04:28:23Z","silentRounds":1,"language":"C++","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["游戏引擎","S.T.A.L.K.E.R.","开源重制","C++","游戏开发"],"domainKey":"游戏引擎","pushedAt":"2026-09-22T13:03:31Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/OpenXRay/xray-16","ts":"2026-09-22T13:03:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"googleapis/python-genai","owner":"googleapis","name":"python-genai","desc":"","summaryCn":"谷歌官方的 Python Gen AI SDK 为开发者提供了一个标准化的接口，用于在 Py","reasonCn":"项目核心亮点在于其官方正统性与持续更新的特性。它不仅支持基础的模型内容生成，还针对未来版本规划了自动函数调用等高级功能的优化，确保开发者能够用上最新、最稳妥的技术规范。通过统一的模块划分，开发者可以极大地减少样板代码，专注于核心业务逻辑的实现。","stars":3987,"starGrowth":2,"createdAt":"2024-12-06T23:16:52Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","推理引擎","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["Python SDK","Gemini","生成式AI","Google API","大模型开发"],"domainKey":"Python SDK","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/googleapis/python-genai","ts":"2026-09-18T20:05:54.376Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OnionBrowser/OnionBrowser","owner":"OnionBrowser","name":"OnionBrowser","desc":"An open-source, privacy-enhancing web browser for iOS, utilizing the Tor anonymity network","summaryCn":"iPhone专用隐私浏览器，上洋葱网不露脸。","reasonCn":"技术上，它完全基于Swift开发，利用Tor的混合网络协议（包括Tor SOCKS代理）来路由流量。它内置了Tor控制器的集成，可以动态调整路由节点，甚至支持创建私人Tor网络。相比Firefox的Tor插件，它更原生，运行更流畅，而且针对iOS的权限做了优化，不会过度索取系统权限。","stars":2678,"starGrowth":1,"createdAt":"2012-02-24T08:50:51Z","silentRounds":10,"language":"Swift","topics":["mobile"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0,"domainTags":["隐私浏览","Tor网络","iOS","匿名上网"],"domainKey":"隐私浏览","pushedAt":"2026-09-22T12:38:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/OnionBrowser/OnionBrowser","ts":"2026-09-22T12:38:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tensorflow/probability","owner":"tensorflow","name":"probability","desc":"Probabilistic reasoning and statistical analysis in TensorFlow","summaryCn":"TensorFlow 官方概率计算库，让机器学会“不确定性”。","reasonCn":"这是Google TensorFlow 生态中的核心库之一，专注于概率编程和统计建模。它提供了贝叶斯神经网络、随机梯度下降等多种算法，帮助开发者构建能处理不确定性和随机性的机器学习模型，是金融风控、自动驾驶等领域的必备工具。","stars":4433,"starGrowth":0,"createdAt":"2017-10-23T23:50:54Z","silentRounds":3,"language":"Jupyter Notebook","topics":["data-science"],"aiDims":["AI基础设施","模型与训练","数据科学"],"zone":"工具","funScore":0,"domainTags":["概率编程","贝叶斯推断","TensorFlow","统计建模"],"domainKey":"概率编程","pushedAt":"2026-09-15T20:21:21Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"数据科学","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/tensorflow/probability","ts":"2026-09-15T20:21:21Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"f/prompts.chat","owner":"f","name":"prompts.chat","desc":"f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.","summaryCn":"收集了全球最实用的ChatGPT提示词大全，直接复制粘贴就能变大神","reasonCn":"这是一个开源的提示词（Prompt）合集网站，收录了来自社区的几千个高质量提示词，涵盖写作、代码、营销、学习等各个领域。你可以直接复制使用，也可以学习别人的思路来优化自己的提示词，是Prompt Engineering入门的最佳教材。","stars":171036,"starGrowth":56,"createdAt":"2022-12-05T13:54:13Z","silentRounds":0,"language":"HTML","topics":["ml","llm","ai"],"aiDims":["提示工程","学习资源","AI写作"],"zone":"资源","funScore":0,"domainTags":["Prompt工程","AI提示词","开源资源"],"domainKey":"Prompt工程","pushedAt":"2026-09-09T10:27:05Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/f/prompts.chat","ts":"2026-09-09T10:27:05Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.16898224935101383,"copyOk":false},{"repo":"catboost/catboost","owner":"catboost","name":"catboost","desc":"A fast, scalable, high performance Gradient Boosting on Decision Trees library, used for ranking, classification, regression and other machine learning tasks for Python, R, Java, C++. Supports computation on CPU and GPU.","summaryCn":"Yandex出的梯度提升树神器，处理表格数据效果拔群，多种语言都支持","reasonCn":"这个库的核心优势在于它原生支持类别特征（categorical features）的自动处理，不需要手动做one-hot编码，大幅简化了特征工程流程。同时它还引入了有序 boosting（ordered boosting）等技术来缓解预测偏移问题，在多个机器学习竞赛和实际业务场景中都取得了顶级效果。","stars":9113,"starGrowth":0,"createdAt":"2017-07-18T05:29:04Z","silentRounds":2,"language":"C++","topics":["tutorial","data-science"],"aiDims":["模型与训练","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["梯度提升","机器学习","表格数据","分类回归"],"domainKey":"梯度提升","pushedAt":"2026-09-23T21:29:25Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/catboost/catboost","ts":"2026-09-23T21:29:25Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"minio/minio","owner":"minio","name":"minio","desc":"MinIO is a high-performance, S3 compatible object store, open sourced under GNU AGPLv3 license.","summaryCn":"像云存储一样好用的自建文件服务器，开箱即用","reasonCn":"MinIO 最厉害的地方在于它的性能和兼容性。它直接兼容 AWS S3 的接口，这意味着你在 S3 上写的代码，拿到 MinIO 上基本不用改就能跑，这对很多基于云存储开发的应用来说是巨大的福音。它采用 Go 语言编写，天生支持多线程并发，读写速度非常快，而且单个节点就能跑起来，数据可以自","stars":61357,"starGrowth":0,"createdAt":"2015-01-14T19:23:58Z","silentRounds":7,"language":"Go","topics":["go","storage","cloud","s3","objectstorage","cloudstorage","amazon-s3","cloudnative","k8s","kubernetes","multi-cloud","multi-cloud-kubernetes"],"aiDims":["AI基础设施","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["S3兼容","文件存储","自建网盘","分布式存储","高性能"],"domainKey":"S3兼容","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"storage","source":"github","weight":0.7},{"name":"cloud","source":"github","weight":0.7},{"name":"s3","source":"github","weight":0.7},{"name":"objectstorage","source":"github","weight":0.7},{"name":"cloudstorage","source":"github","weight":0.7},{"name":"amazon-s3","source":"github","weight":0.7},{"name":"cloudnative","source":"github","weight":0.7},{"name":"k8s","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"multi-cloud","source":"github","weight":0.7},{"name":"multi-cloud-kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/minio/minio","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TriliumNext/Trilium","owner":"TriliumNext","name":"Trilium","desc":"Build your personal knowledge base with Trilium Notes","summaryCn":"自托管笔记神器，知识图谱一拉就清晰，打造你的私人大脑","reasonCn":"Trilium Notes是一个完全自托管的个人知识库工具，支持无限层级笔记组织、全文检索和可视化知识图谱。采用Electron框架开发，数据存储在SQLite中，支持端到端加密同步。相比Obsidian，它内置了强大的图可视化；相比Notion，它更注重本地隐私和完全控制权。","stars":37940,"starGrowth":3,"createdAt":"2017-05-23T00:30:20Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted","note-taking","knowledge-management"],"aiDims":["RAG与知识","知识图谱","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["笔记应用","知识图谱","自托管","SQLite"],"domainKey":"笔记应用","pushedAt":"2026-09-23T19:58:12Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"note-taking","source":"github","weight":0.7},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/TriliumNext/Trilium","ts":"2026-09-23T19:58:12Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mufeedvh/code2prompt","owner":"mufeedvh","name":"code2prompt","desc":"A CLI tool to convert your codebase into a single LLM prompt with source tree, prompt templating, and token counting.","summaryCn":"一键把你的代码库打包成AI能读懂的提示词，让LLM秒变你的专属代码顾问","reasonCn":"这个工具能把整个代码库一键转成LLM能看懂的prompt，支持源树结构展示、自定义prompt模板和token计数功能。开发者只需一个命令，就能把复杂的代码结构、文件层次和代码内容打包成AI友好的格式，让大模型瞬间理解你的项目架构和逻辑，大幅提升与AI协作开发代码的效率。","stars":7694,"starGrowth":3,"createdAt":"2024-03-09T12:42:06Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["代码助手","提示工程","开发者工具"],"zone":"工具","funScore":0,"domainTags":["代码转提示词","LLM Prompt","命令行工具"],"domainKey":"代码转提示词","pushedAt":"2026-09-18T18:23:08Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/mufeedvh/code2prompt","ts":"2026-09-18T18:23:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Zackriya-Solutions/meetily","owner":"Zackriya-Solutions","name":"meetily","desc":"Privacy first, AI meeting assistant with 4x faster Parakeet/Whisper live transcription, speaker diarization, and Ollama summarization built on Rust. 100% local processing. no cloud required. Meetily (Meetly Ai - https://meetily.ai) is the #1 Self-hosted, Open-source Ai meeting note taker for macOS &amp; Windows. Understand How to write meeting minutes","summaryCn":"本地AI会议助手，转录速度快四倍隐私保护拉满无需云端上传数据","reasonCn":"基于Rust开发的本地AI会议助手，使用Parakeet/Whisper进行实时语音转文字，支持说话人分离和Ollama本地总结。100%本地处理无需云端，保护隐私。速度比开源方案快四倍，准确率高，适合会议记录场景，数据完全不出本地。","stars":31035,"starGrowth":6,"createdAt":"2024-12-26T12:52:14Z","silentRounds":0,"language":"Rust","topics":["self-hosted","ai","llm"],"aiDims":["AI应用","AI Agent"],"zone":"AI","funScore":0,"domainTags":["会议转录","说话人分离","本地大模型"],"domainKey":"会议转录","pushedAt":"2026-09-15T18:32:32Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/Zackriya-Solutions/meetily","ts":"2026-09-15T18:32:32Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01857999033349444,"copyOk":true},{"repo":"microsoft/genaiscript","owner":"microsoft","name":"genaiscript","desc":"","summaryCn":"GenAIScript通过引入一种创新的编程范式，将AI提示词（Prompt）视作代码的一","reasonCn":"该项目的一大亮点是其极佳的开发体验。它深度集成了Visual Studio Code，开发者在编写脚本时能获得实时补全与调试支持，这种“所写即所得”的体验大幅降低了AI应用的门槛。此外，它对多种主流模型（包括OpenAI、Azure OpenAI和Anthropic等）的内置支持，意味着开发者无需为不同模型编","stars":2884,"starGrowth":1,"createdAt":"2023-08-17T15:57:21Z","silentRounds":4,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手","提示工程"],"zone":"AI","funScore":0.5,"domainTags":["JavaScript编程","LLM脚本化","自动化提示","VSCode集成"],"domainKey":"JavaScript编程","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/microsoft/genaiscript","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"apache/casbin-gateway","owner":"apache","name":"casbin-gateway","desc":"Casbin AI & MCP security gateway for HTTP, online demo: https://door.caswaf.com","summaryCn":"Apache出品的AI安全网关，给大模型接口穿上防弹衣","reasonCn":"技术上，它的核心是一个基于Casbin的权限引擎。Casbin是一个强大的开源访问控制库，支持RBAC、ABAC等多种权限模型。gateway在此基础上增加了AI场景特有的能力，比如模型级别的权限控制、请求频率限制、敏感信息过滤等。","stars":636,"starGrowth":0,"createdAt":"2017-05-13T02:42:27Z","silentRounds":7,"language":"Go","topics":["llm-model"],"aiDims":["AI基础设施","安全工具","AI Agent"],"zone":"AI","funScore":0,"domainTags":["AI安全","API网关","权限控制","MCP","Apache"],"domainKey":"AI安全","pushedAt":"2026-09-19T15:10:14Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/apache/casbin-gateway","ts":"2026-09-19T15:10:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"plasmicapp/plasmic","owner":"plasmicapp","name":"plasmic","desc":"Visual builder for React. Build apps, websites, and content. Integrate with your codebase.","summaryCn":"可视化搭建React应用，设计师也能写代码","reasonCn":"这是一个可视化的React应用构建平台，允许设计师或开发者通过拖拽界面元素来生成高质量的React代码。它与主流框架深度集成，支持复杂的交互逻辑，让非技术人员也能参与前端开发，同时保留开发者的代码控制权，实现了设计与开发的无缝协作。","stars":7034,"starGrowth":1,"createdAt":"2021-02-02T19:37:59Z","silentRounds":1,"language":"TypeScript","topics":["design","frontend"],"aiDims":["AI应用","非AI-实用","创意工具"],"zone":"工具","funScore":0,"domainTags":["可视化开发","React框架","低代码平台","前端工程化","网页构建"],"domainKey":"可视化开发","pushedAt":"2026-09-23T17:44:26Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/plasmicapp/plasmic","ts":"2026-09-23T17:44:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ItzCrazyKns/Vane","owner":"ItzCrazyKns","name":"Vane","desc":"Vane is an AI-powered answering engine.","summaryCn":"像装了个超级大脑，能读懂你所有文件还能随时提问","reasonCn":"Vane 是一个基于 TypeScript 的 RAG（检索增强生成）引擎，专为构建本地知识库问答系统设计。它支持多种向量数据库和模型后端，能够自动解析文档并构建向量索引，实现精准的知识检索与生成。项目结构清晰，插件化架构使其易于扩展，适合个人开发者快速搭建私有知识库助手。","stars":36900,"starGrowth":4,"createdAt":"2024-04-09T10:51:32Z","silentRounds":0,"language":"TypeScript","topics":["llm","rag","ml"],"aiDims":["AI Agent","AI搜索","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["RAG","向量检索","知识库","LLM"],"domainKey":"RAG","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/ItzCrazyKns/Vane","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"allenai/open-instruct","owner":"allenai","name":"open-instruct","desc":"","summaryCn":"Allen Institute for AI 推出的 open-instruct 并非简单的训练脚本集合，而","reasonCn":"项目的一大特色是数据的统一处理。它将来自不同来源的指令数据集整合为统一格式，解决了多源数据清洗与预处理的繁琐工作。对于希望复现 TÜLU 3 训练逻辑的研究者而言，直接使用其提供的代码库可大幅减少配置环境的精力，尤其是在处理 DPO 和 RLVR 这类复杂算法时，其稳定性优于零散","stars":3875,"starGrowth":0,"createdAt":"2023-06-09T16:50:38Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["模型与训练","微调","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["指令微调","强化学习","TÜLU 3","语言模型"],"domainKey":"指令微调","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/allenai/open-instruct","ts":"2026-09-18T20:41:13.636Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"btspoony/mstar-harness","owner":"btspoony","name":"mstar-harness","desc":"An omni-plugin for harness engineering workflows with multi-agents, programmatic gates and skills.","summaryCn":"让多个AI机器人组团打工，自动完成复杂工程任务的超级插件。","reasonCn":"在技术实现上，它是一个功能强大的全能型插件系统，基于TypeScript开发，具备极佳的扩展性和类型安全。它引入了独特的程序化门禁（Programmatic Gates）概念，也就是说AI在推进任务时，必须通过自动化测试或预设的规则校验才能进入下一步，这就有效避免了AI盲目胡说八道或者把代码","stars":61,"starGrowth":0,"createdAt":"2026-03-06T16:47:43Z","silentRounds":15,"language":"TypeScript","topics":["knowledge-management"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["多Agent系统","AI编程助手","工程自动化","工作流引擎"],"domainKey":"多Agent系统","pushedAt":"2026-09-18T11:26:59Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/btspoony/mstar-harness","ts":"2026-09-18T11:26:59Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"eosphoros-ai/DB-GPT","owner":"eosphoros-ai","name":"DB-GPT","desc":"open-source agentic AI data assistant for the next generation of AI + Data products.","summaryCn":"跟AI聊聊天就能查数据库，让普通人也能秒变数据分析师","reasonCn":"DB-GPT是开源的AI数据助手框架，支持自然语言与数据库直接对话。能将口语化问题自动转为SQL查询，支持RAG检索增强生成和多模型接入。具备Agent能力，可调用工具执行复杂任务，支持MySQL、PostgreSQL等多种数据库，提供可视化Web界面，适用于数据分析、BI报表、业务查询等场景。","stars":20033,"starGrowth":1,"createdAt":"2023-04-13T14:52:43Z","silentRounds":1,"language":"Python","topics":["rag","security","database","agents"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["自然语言转SQL","数据库问答","AI数据助手","RAG应用","多模型接入"],"domainKey":"自然语言转SQL","pushedAt":"2026-09-16T06:51:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/eosphoros-ai/DB-GPT","ts":"2026-09-16T06:51:17Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Paitesanshi/LLM-Agent-Survey","owner":"Paitesanshi","name":"LLM-Agent-Survey","desc":"","summaryCn":"大语言模型的兴起促使研究者将其作为自主智能体的核心控制器，但现有的研究","reasonCn":"在核心内容上，该调查首先深入探讨了AI智能体的基本组成部分。值得注意的是，它并不局限于代码实现，而是从认知架构的角度出发，详细定义了包含一个画像模块、一个记忆模块、一个规划模块和一个行动模块的标准结构。","stars":2909,"starGrowth":0,"createdAt":"2023-08-03T09:19:42Z","silentRounds":3,"language":"","topics":[],"aiDims":["AI Agent","大语言模型","论文"],"zone":"资源","funScore":0.2,"domainTags":["LLM自主智能体","智能体架构","学术论文","系统综述"],"domainKey":"LLM自主智能体","pushedAt":"2026-09-20T15:45:52.959Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1}],"aiScore":0.71,"source":"tier","url":"https://github.com/Paitesanshi/LLM-Agent-Survey","ts":"2026-09-20T15:45:52.959Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"shareAI-lab/claw0","owner":"shareAI-lab","name":"claw0","desc":"","summaryCn":"许多开发者在接触 AI 智能体时，往往只能写出调用一次大模型接口的简单脚本","reasonCn":"项目的核心亮点在于其独特的渐进式教学设计。通过精心划分的十个部分，它将复杂的 AI 网关拆解为易于消化的模块。你不仅能学到如何处理基本的交互循环，还能深入掌握高阶的设计模式与并发控制。当学完整套课程后，你将拥有足够的底气和技术储备去阅读和理解诸如 OpenClaw 这样复杂","stars":3364,"starGrowth":0,"createdAt":"2026-02-24T10:16:38Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","推理引擎","开发者工具"],"zone":"AI","funScore":0.2,"domainTags":["AI Agent","网关","Python教程","架构设计"],"domainKey":"AI Agent","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/shareAI-lab/claw0","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"faiscadev/fakecloud","owner":"faiscadev","name":"fakecloud","desc":"Free, open-source AWS emulator. LocalStack alternative: 105 services, 7,404 operations, true 100% Smithy conformance (248,557/248,557 variants pass). No account, no auth token, no paid tier.","summaryCn":"零配置模拟AWS全栈，105项服务7400操作本地完美复刻","reasonCn":"这个项目最大的亮点在于它的规范遵从度。很多云模拟器只能模拟部分服务的部分功能，但fakecloud声称支持105项AWS服务和7404个操作，并且所有248557个Smithy协议变体全部通过测试。这意味着它的模拟精度远高于市面上大多数同类工具。","stars":556,"starGrowth":0,"createdAt":"2026-04-04T13:45:31Z","silentRounds":1,"language":"Rust","topics":["emulator"],"aiDims":["开发者工具","AI基础设施","非AI-实用"],"zone":"工具","funScore":0.7,"domainTags":["AWS 模拟器","云服务测试","Smithy 协议","本地开发","本地 AWS"],"domainKey":"AWS 模拟器","pushedAt":"2026-09-21T13:26:29Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/faiscadev/fakecloud","ts":"2026-09-21T13:26:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"google-deepmind/dnc","owner":"google-deepmind","name":"dnc","desc":"","summaryCn":"DNC实现：循环网络结合可微外存，支持记忆增强模型实验。","reasonCn":"Differentiable Neural Computer (DNC) 是一种结合循环网络与可微记忆的架构，能够提升模型的长期记忆能力。该开源实现基于Python，方便研究者在实验中快速构建和训练 DNC 模型，并严格要求引用原始Nature论文以保证实验可重复性与学术严谨性。","stars":2532,"starGrowth":0,"createdAt":"2017-03-28T13:24:38Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0.2,"domainTags":["循环神经网络","可微记忆","Python","机器学习"],"domainKey":"循环神经网络","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/google-deepmind/dnc","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"anthropics/skills","owner":"anthropics","name":"skills","desc":"Public repository for Agent Skills","summaryCn":"Anthropic官方开源的海量AI技能大全库","reasonCn":"这是Anthropic官方开源的Agent技能大宝库，收录了大量预构建的Agent Skills。每个Skill都是独立的功能模块，可以让AI Agent执行特定任务，比如操作文件、调用API、控制浏览器等。star数超过16万，是目前最流行的Agent技能集合之一。","stars":177652,"starGrowth":37,"createdAt":"2025-09-22T15:53:31Z","silentRounds":0,"language":"Python","topics":["agent-skills"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["Agent","技能包","Claude","AI工具","自动化"],"domainKey":"Agent","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/anthropics/skills","ts":"2026-09-06T13:43:54.327Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.11162481705806858,"copyOk":false},{"repo":"ax-llm/ax","owner":"ax-llm","name":"ax","desc":"","summaryCn":"一个编程模型搞定六种语言的LLM应用开发，TypeScript优先。","reasonCn":"Ax把自己定位成DSPy的跨语言版本，让你用一套编程模型就能在TypeScript、Python、Java、C++、Go和Rust里开发大模型应用。它支持字符串DSL、流式f()构建器或标准Schema验证器（Zod、Valibot、ArkType），还内置了多Provider映射和Agent运行时的编译生成。对多语言团队来说，一套业务逻辑到处跑，确实省事。","stars":2945,"starGrowth":2,"createdAt":"2023-02-23T19:21:40Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.32,"domainTags":["大语言模型","TypeScript","多语言支持","Agent框架","结构化生成"],"domainKey":"大语言模型","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/ax-llm/ax","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"matthew-kissinger/kiln","owner":"matthew-kissinger","name":"kiln","desc":"Text-to-3D as code: language models write JavaScript programs that build game-ready GLBs, look at their own renders, and revise. Ships as an MCP server plus skills for any harness.","summaryCn":"用文字生成3D游戏模型，像写代码一样简单","reasonCn":"它的核心工作流程非常聪明。首先，你可以用自然语言描述你想要的3D模型，比如“一个可爱的卡通机器人”。AI会根据你的描述，生成一个初步的GLB模型文件。但是，这还不是最终结果。你可以让AI“看”到自己生成的模型，然后根据观察结果，对代码进行修改和优化，生成一个更好的版本。","stars":104,"starGrowth":0,"createdAt":"2026-06-23T17:26:16Z","silentRounds":2,"language":"TypeScript","topics":["procedural-generation"],"aiDims":["创意工具","AI应用","工具调用"],"zone":"AI","funScore":1,"domainTags":["文本生成3D","MCP服务器","程序化建模","AI生成内容","游戏资产"],"domainKey":"文本生成3D","pushedAt":"2026-09-17T05:11:34Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/matthew-kissinger/kiln","ts":"2026-09-17T05:11:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/TypeScript","owner":"microsoft","name":"TypeScript","desc":"TypeScript is a superset of JavaScript that compiles to clean JavaScript output.","summaryCn":"给JavaScript加类型系统的超牛工具，写大项目必备。","reasonCn":"在AI领域，很多AI工具和框架底层就是用TypeScript编写的，比如VS Code和Vercel的AI SDK。它强大的类型推断和智能提示功能，能帮助开发者快速理解复杂代码，减少上手时间。对于团队协作，TypeScript的接口定义文档化特性，让前后端对接变得非常顺畅。","stars":111163,"starGrowth":1,"createdAt":"2014-06-17T15:28:39Z","silentRounds":1,"language":"Go","topics":["typescript","javascript","language","typechecker"],"aiDims":["开发者工具","效率工具","AI搜索"],"zone":"工具","funScore":0,"domainTags":["类型系统","JavaScript","编译"],"domainKey":"类型系统","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"typescript","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"language","source":"github","weight":0.7},{"name":"typechecker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/microsoft/TypeScript","ts":"2026-09-19T04:31:19.347Z","category":"tool","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"cilame/v_jstools","owner":"cilame","name":"v_jstools","desc":"","summaryCn":"该项目最新升级至manifest V3标准，全面提升了对现代Chrome浏览器的适配能力","reasonCn":"工具内置了专门用于生成JS运行环境的代码包，能够辅助用户快速构建仿真运行环境，进而深入分析代码的混淆机理。对于那些难以直接阅读的压缩代码，v_jstools通过提供深层的分析工具集，将原本琐碎的调试过程结构化。这一功能极大地提高了逆向工程师在应对各类防调试机制时的效率，使","stars":3011,"starGrowth":0,"createdAt":"2021-09-25T14:18:21Z","silentRounds":4,"language":"JavaScript","topics":[],"aiDims":["AI应用","开发者工具","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["JS逆向","浏览器插件","代码分析","自动化调试"],"domainKey":"JS逆向","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/cilame/v_jstools","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"guoqincode/Open-AnimateAnyone","owner":"guoqincode","name":"Open-AnimateAnyone","desc":"","summaryCn":"该项目通过集成Magic-Animate和AnimateDiff这两个主流模型框架，为开发者提","reasonCn":"由于目前项目主要依赖TikTok和UBC等小规模数据集进行训练，开发者在实际使用中可能会发现其在复杂场景下的生成质量与官方论文展示的效果存在一定差距。这种开源模式的初衷是降低研究门槛，通过社区协作的方式不断优化模型权重。","stars":2919,"starGrowth":0,"createdAt":"2023-12-07T16:23:46Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["模型与训练","多模态Agent","非AI-好玩"],"zone":"AI","funScore":0.5,"domainTags":["视频生成","人物动作迁移","图像动画化"],"domainKey":"视频生成","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/guoqincode/Open-AnimateAnyone","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"shaobeichen/dsh-pocket","owner":"shaobeichen","name":"dsh-pocket","desc":"把 DeepSeek Harness 装进你的口袋：电脑上跑 dsh web，手机扫码即同步访问（局域网 + 公网，实时同屏）Put DeepSeek Harness in your pocket: run dsh web on your computer and access it synchronously by scanning a QR code on your phone (LAN + public network, real‑time screen mirroring)","summaryCn":"把电脑上的 AI 界面瞬间同步到手机里的神操作","reasonCn":"技术实现上，它并不复杂但极其巧妙。它在本地构建了一个 Web 代理层，利用 JavaScript 实现了高效的请求转发和界面同步。通过集成局域网穿透或简单的 IP 映射，它将原本只能在 localhost 访问的本地服务转化为可扫码访问的链接。","stars":1292,"starGrowth":1,"createdAt":"2026-08-15T09:36:29Z","silentRounds":1,"language":"JavaScript","topics":["mobile"],"aiDims":["效率工具","开发者工具"],"zone":"AI","funScore":1,"domainTags":["AI","DeepSeek","局域网穿透","移动端","远程访问"],"domainKey":"AI","pushedAt":"2026-09-16T16:08:58Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/shaobeichen/dsh-pocket","ts":"2026-09-16T16:08:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FudanNLP/fastNLP","owner":"FudanNLP","name":"fastNLP","desc":"","summaryCn":"复旦NLP团队开发的轻量级NLP工具包，简化数据处理和模型训练流程。","reasonCn":"fastNLP是复旦大学NLP团队开发的轻量级自然语言处理工具包，目标是减少NLP项目中的工程代码。支持通过apply函数避免数据处理循环、使用多进程加速，无需改代码即可实现fp16、多卡训练和ZeRO优化，兼容多种深度学习框架后端。","stars":3139,"starGrowth":0,"createdAt":"2018-03-07T13:30:20Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["NLP","自然语言处理","PyTorch"],"domainKey":"NLP","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/FudanNLP/fastNLP","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"SeanNaren/deepspeech.pytorch","owner":"SeanNaren","name":"deepspeech.pytorch","desc":"","summaryCn":"Deepspeech.pytorch 是一个专注于语音识别技术的研究型项目，它在 PyTorch ","reasonCn":"该项目的特色在于对推理流程的灵活支持。除了基础的神经网络解码外，它允许用户可选地引入 KenLM 语言模型。这意味着系统可以结合声学模型和语言模型的统计规律，利用上下文信息修正可能的识别错误，从而在复杂语音环境下获得更高的文本准确率，这对于提升最终用户的使用体验至关重","stars":2136,"starGrowth":0,"silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","多模态Agent"],"zone":"AI","funScore":0.2,"domainTags":["语音识别","DeepSpeech2","PyTorch"],"domainKey":"语音识别","pushedAt":"2026-09-23T18:23:13.861Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/SeanNaren/deepspeech.pytorch","ts":"2026-09-23T18:23:13.861Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dat999zx/knowl","owner":"dat999zx","name":"knowl","desc":"Persistent memory for Claude Code, Cursor and Codex. Facts are typed, sourced, and retired when they change. Local SQLite, no API keys.","summaryCn":"如果你在使用 Claude、Cursor 或 Codex 等大型语言模型时，总会遇到模型记忆","reasonCn":"核心技术是事件驱动的知识图谱模型，支持标签、来源标识、事实版本控制；当事实更新时，knowl 会自动触发相应的缓存刷新，保持知识库与实际数据同步。插件化架构让你可以轻松接入第三方源，如 RSS、JSON 或自定义 API，甚至支持自定义校验逻辑。","stars":98,"starGrowth":0,"createdAt":"2026-06-24T12:45:15Z","silentRounds":1,"language":"TypeScript","topics":["knowledge-management"],"aiDims":["RAG与知识","知识图谱"],"zone":"AI","funScore":0,"domainTags":["Claude","持久记忆","知识管理"],"domainKey":"Claude","pushedAt":"2026-09-21T06:47:16Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/dat999zx/knowl","ts":"2026-09-21T06:47:16Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"supElement/ComfyUI_Element_easy","owner":"supElement","name":"ComfyUI_Element_easy","desc":"一些方便使用的节点。Element Multi REF, Element Load and Edit Video，Minimax_H3-LatentUpscaler, Smart merge images,LoadImage_Preview,Element_SigmaGraph,Element ImageCurve,Element HueSat,Element HueBright,Element HueHue,black_white_color,...","summaryCn":"给ComfyUI加了一堆好用的节点，视频图片处理像搭积木一样简单。","reasonCn":"这是一个专为ComfyUI设计的节点扩展库，提供了Element多路复用、视频加载与编辑、LatentUpscaler以及Hue/Sat/Brightness调整等实用功能。它极大地丰富了ComfyUI的节点生态，让用户能更便捷地实现复杂的图像处理、视频生成与风格化任务，显著降低了使用门槛。","stars":11,"starGrowth":0,"createdAt":"2025-12-04T08:37:43Z","silentRounds":9,"language":"JavaScript","topics":["video-editor"],"aiDims":["创意工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["ComfyUI","节点扩展","图像处理","视频编辑"],"domainKey":"ComfyUI","pushedAt":"2026-09-17T03:03:30Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/supElement/ComfyUI_Element_easy","ts":"2026-09-17T03:03:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"martinpacesa/BindCraft","owner":"martinpacesa","name":"BindCraft","desc":"User friendly and accurate binder design pipeline","summaryCn":"在3D打印的世界里，设计环节往往比打印本身更让人头疼。","reasonCn":"在3D打印的世界里，设计环节往往比打印本身更让人头疼。想要打印一个精美的镂空结构或复杂几何体，没有几年的建模经验根本玩不转。BindCraft就是为了解决这个问题而生的。它就像一个懂你的智能助手，你只需要输入一些简单的参数，比如你想要的形状、大小、孔洞数量，它就能在几秒钟","stars":1228,"starGrowth":0,"createdAt":"2024-09-18T10:45:09Z","silentRounds":1,"language":"Python","topics":["design"],"aiDims":["创意工具","设计"],"zone":"创意","funScore":0,"domainTags":["3D打印","参数化建模","模型生成","STL导出"],"domainKey":"3D打印","pushedAt":"2026-09-21T09:13:05Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"设计","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/martinpacesa/BindCraft","ts":"2026-09-21T09:13:05Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ruiwuniu/Aivatar-Demo","owner":"ruiwuniu","name":"Aivatar-Demo","desc":"Aivatar is a local-first desktop companion for AI coding agents. It turns Codex, Claude Code, opencode, or a custom local agent bridge into a lively pixel-room companion with selectable characters, upgraded room materials, live status, growth, bits, and room customization.","summaryCn":"让你的 AI 编程助手变成像素房间里的萌系小精灵陪伴你","reasonCn":"Aivatar 是一个本地优先的桌面伴侣应用，将 Codex、Claude Code 等 AI 编程代理的交互界面转化为像素风格的虚拟房间。你在代码编辑器里执行命令时，房间里的像素小精灵会做出对应反应，让编程过程更有陪伴感。支持多种可切换角色，界面充满复古游戏风格。","stars":13,"starGrowth":0,"createdAt":"2026-06-11T09:13:43Z","silentRounds":16,"language":"TypeScript","topics":["pixel-art","virtual-pet"],"aiDims":["AI Agent","非AI-好玩","创意工具"],"zone":"AI","funScore":1,"domainTags":["AI编程助手","虚拟宠物","像素艺术","桌面伴侣","命令行交互"],"domainKey":"AI编程助手","pushedAt":"2026-09-20T08:11:02Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"virtual-pet","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/ruiwuniu/Aivatar-Demo","ts":"2026-09-20T08:11:02Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"THU-MAIC/OpenMAIC","owner":"THU-MAIC","name":"OpenMAIC","desc":"Open Multi-Agent Interactive Classroom — Get an immersive, multi-agent learning experience in just one click","summaryCn":"一键开启多人沉浸式课堂，AI老师学生齐上阵","reasonCn":"OpenMAIC 是一个基于 TypeScript 开发的开源多智能体交互课堂系统，旨在提供沉浸式的多人学习体验。它通过多智能体协作模拟真实的课堂互动场景，支持师生多角色互动、课程内容生成与实时问答。项目基于先进的 AI Agent 技术栈构建，强调多模态交互与流畅的用户体验，目前已获得超过两万颗星，成为 AI 教育与多智能体应用领域的热门项目。","stars":38602,"starGrowth":31,"createdAt":"2026-03-11T10:13:16Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","多模态Agent","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["多智能体协作","虚拟课堂","AI教育","角色扮演"],"domainKey":"多智能体协作","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/THU-MAIC/OpenMAIC","ts":"2026-09-10T22:16:01.544Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.11449104191492668,"copyOk":false},{"repo":"KKKKhazix/khazix-skills","owner":"KKKKhazix","name":"khazix-skills","desc":"数字生命卡兹克开源的 AI Skills 合集 | Agent Skills: leader（帮你定义目标）, neat-freak 洁癖, hv-analysis, khazix-writer & more — Claude Code, Codex & 40+ agents","summaryCn":"给AI打工仔装上40多种技能包，干活效率直接拉满","reasonCn":"这个项目整理了一套AI Agent技能合集，包含leader帮你定义目标、neat-freak做代码洁癖检查、hv-analysis做技术分析等四十多种开箱即用的Skill。支持Claude Code、Codex等主流Agent框架，直接克隆就能给AI工具人加技能，让写代码、分析、写作都变得更聪明高效。","stars":20878,"starGrowth":1,"createdAt":"2026-04-06T07:37:02Z","silentRounds":1,"language":"Python","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","技能包","开发工具"],"domainKey":"AI Agent","pushedAt":"2026-09-16T03:30:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/KKKKhazix/khazix-skills","ts":"2026-09-16T03:30:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Cp0204/SmartStrm","owner":"Cp0204","name":"SmartStrm","desc":"媒体库 STRM 文件生成工具。和 Emby 优雅配合，媒体资源丝滑入库，支持 302 直链播放，支持同步删除。 配合 Quark-Auto-Save, CloudSaver, Emby 力求即存即看。🥳","summaryCn":"Emby媒体库自动化神器，自动生成STRM直链。","reasonCn":"核心功能非常强大。它支持 302 跳转直链，这通常意味着更高的稳定性和更低的带宽消耗。它还能和 Quark-Auto-Save、CloudSaver 等工具配合，实现自动同步和清理。比如，你删除了 Emby 里的条目，SmartStrm 可以自动帮你清理对应的云盘文件，保持两边的同步。","stars":758,"starGrowth":1,"createdAt":"2025-08-07T10:59:40Z","silentRounds":3,"language":"Shell","topics":["jellyfin"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Media Library","STRM","Cloud Storage"],"domainKey":"Media Library","pushedAt":"2026-09-17T15:38:59Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Cp0204/SmartStrm","ts":"2026-09-17T15:38:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hiyouga/LlamaFactory","owner":"hiyouga","name":"LlamaFactory","desc":"Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)","summaryCn":"大模型微调界的瑞士军刀，不用写代码也能把通用AI调教成行业专家","reasonCn":"LlamaFactory是一个统一的efficient微调框架，支持100+主流LLM和VLM的指令微调、预训练和奖励建模。它集成了QLoRA、LoRA、全参微调等多种策略，提供Web UI和命令行双模式，极大降低了大模型本地化定制的门槛，是开发者和企业落地的利器。","stars":74979,"starGrowth":5,"createdAt":"2023-05-28T10:09:12Z","silentRounds":0,"language":"Python","topics":["llm","ai","transformers"],"aiDims":["微调","模型与训练","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["大模型微调","QLoRA","LLM训练","多模态微调","参数高效微调"],"domainKey":"大模型微调","pushedAt":"2026-09-14T08:09:48Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/hiyouga/LlamaFactory","ts":"2026-09-14T08:09:48Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015200056015684391,"copyOk":false},{"repo":"FurkanGozukara/Stable-Diffusion","owner":"FurkanGozukara","name":"Stable-Diffusion","desc":"FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tutorials, Guides, Lectures, Courses, ComfyUI, Google Colab, RunPod, Kaggle, NoteBooks, ControlNet, TTS, Voice Cloning, AI, AI News, ML, ML News, News, Tech, Tech News, Kohya, Midjourney, RunPod","summaryCn":"保姆级的AI绘画与大模型训练全套实战教程与工具集","reasonCn":"技术上，它涵盖了从经典Stable Diffusion到最新FLUX、SD3的核心架构解析。不仅教你怎么用Automatic1111和ComfyUI，更深入到DreamBooth、LoRA训练的每一个核心参数。作者通过大量的Jupyter Notebook和图文教程，把复杂的数学原理和炼丹技巧变成了大白话。","stars":2766,"starGrowth":1,"createdAt":"2023-05-01T18:36:11Z","silentRounds":21,"language":"Jupyter Notebook","topics":["learning"],"aiDims":["模型与训练","微调","创意工具"],"zone":"资源","funScore":0,"domainTags":["AI 绘画","模型训练","深度学习","教程"],"domainKey":"AI 绘画","pushedAt":"2026-09-16T02:23:15Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/FurkanGozukara/Stable-Diffusion","ts":"2026-09-16T02:23:15Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fqscfqj/Y2A-Auto","owner":"fqscfqj","name":"Y2A-Auto","desc":"","summaryCn":"在跨国内容分发日益频繁的今天，手动下载 YouTube 视频、转换格式、翻译字幕","reasonCn":"该项目的核心亮点在于其精细化的处理流程。在视频下载后，系统会自动进行语音识别并生成字幕，随后利用 AI 进行精准翻译与后处理。特别值得注意的是，它集成了阿里云 Green 进行内容安全审核，确保搬运的内容符合平台规范，避免了因违规而被限流或封号的风险。","stars":3376,"starGrowth":0,"createdAt":"2025-05-10T11:07:45Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["视频搬运","自动化流水线","内容审核"],"domainKey":"视频搬运","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/fqscfqj/Y2A-Auto","ts":"2026-09-19T18:31:48.187Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tirth8205/code-review-graph","owner":"tirth8205","name":"code-review-graph","desc":"Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.","summaryCn":"本地代码智能图谱，让AI工具只读它该读的代码，大幅提升review效率","reasonCn":"它构建了一个本地的代码知识图谱，用图数据库的方式存储代码的结构和关系。通过MCP协议，AI工具可以精准查询到它真正需要的代码片段，而不是读取整个仓库。这意味着在做代码审查或修改功能时，AI只需要处理它真正相关的代码，上下文占用大幅减少。","stars":31714,"starGrowth":3,"createdAt":"2026-02-26T17:25:13Z","silentRounds":0,"language":"Python","topics":["llm"],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"AI","funScore":1,"domainTags":["代码图谱","MCP","代码审查","AI工具"],"domainKey":"代码图谱","pushedAt":"2026-09-19T04:31:19.347Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/tirth8205/code-review-graph","ts":"2026-09-19T04:31:19.347Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wonderwhy-er/DesktopCommanderMCP","owner":"wonderwhy-er","name":"DesktopCommanderMCP","desc":"This is MCP server for Claude that gives it terminal control, file system search and diff file editing capabilities","summaryCn":"传统的AI助手往往被困在提示框里，而Desktop Commander MCP让AI拥有了“手脚","reasonCn":"其核心亮点在于成本优化的架构设计。用户不需要为这个Agent单独购买昂贵的API Token，而是连接已有的宿主客户端，如Claude、GPT-4.5或Gemini 2.0的订阅。它作为一个MCP服务器运行，复用现有的认证和计费体系，极大地降低了使用门槛和边际成本。","stars":9708,"starGrowth":4,"createdAt":"2024-12-04T08:26:08Z","silentRounds":0,"language":"TypeScript","topics":["agent","ai","code-analysis","code-generation","mcp","terminal-ai","terminal-automation","vibe-coding","gemini-cli-extension"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":1,"domainTags":["MCP","终端控制","Agent","自动化"],"domainKey":"Agent","pushedAt":"2026-09-18T10:23:42.801Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"code-analysis","source":"github","weight":0.7},{"name":"code-generation","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"terminal-ai","source":"github","weight":0.7},{"name":"terminal-automation","source":"github","weight":0.7},{"name":"vibe-coding","source":"github","weight":0.7},{"name":"gemini-cli-extension","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.77,"source":"trending","url":"https://github.com/wonderwhy-er/DesktopCommanderMCP","ts":"2026-09-18T10:23:42.801Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/huggingface.js","owner":"huggingface","name":"huggingface.js","desc":"","summaryCn":"该项目通过一系列模块化的库，解决了开发者在前端或后端JS环境中调用复杂AI","reasonCn":"除了基础的推理功能，项目还紧跟技术趋势，引入了对模型上下文协议（MCP）的支持，允许开发 Agent 应用时能够更顺畅地串联上下文信息。此外，针对性能要求极高的场景，它还包含了对 GGUF 和 DDUF 等高效模型格式的解析能力，使得在远程环境下处理模型文件变得异常高效。","stars":2530,"starGrowth":0,"createdAt":"2023-02-06T18:37:09Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["开发者工具","代码助手"],"zone":"工具","funScore":0.2,"domainTags":["JavaScript","TypeScript","API交互","机器学习"],"domainKey":"JavaScript","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/huggingface/huggingface.js","ts":"2026-09-21T23:17:36.060Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Zettlr/Zettlr","owner":"Zettlr","name":"Zettlr","desc":"Your One-Stop Publication Workbench","summaryCn":"写论文的终极神器，Markdown加文献管理一气呵成","reasonCn":"Zettlr是一款专为学术写作和知识管理设计的桌面编辑器，支持Markdown编辑、Zettelkasten笔记法、LaTeX公式和BibTeX文献引用。它把写作、引用管理、PDF标注和最终出版流程整合在一个应用中，让你不用在多个软件之间来回切换。跨平台支持Windows、Mac和Linux，完全免费开源。对于写论文的学生、做知识管理的研究者以及任何喜欢Markdown的写作者来说，它是提升写作效率的利器。虽然不直接涉及AI大模型，但它集成了AI辅助写作功能，可以智能补全和润色文本。","stars":13565,"starGrowth":2,"createdAt":"2017-12-26T18:34:19Z","silentRounds":0,"language":"TypeScript","topics":["desktop","productivity"],"aiDims":["AI写作","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Markdown编辑器","学术写作","知识管理","文献引用"],"domainKey":"Markdown编辑器","pushedAt":"2026-09-22T02:36:38Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/Zettlr/Zettlr","ts":"2026-09-22T02:36:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ashishpatel26/LLM-Finetuning","owner":"ashishpatel26","name":"LLM-Finetuning","desc":"","summaryCn":"该项目通过集成 Hugging Face 的 transformers 库，将大语言模型微调的门槛","reasonCn":"项目中包含了详细的微调实验流程表，通过 Notebook 的形式一步步指引用户操作。开发者可以从中学习如何配置模型加载、如何准备训练数据集以及如何管理微调过程中的评估指标。这些代码不仅提供了功能性实现，更像是为初学者量身打造的教学手册，展示了如何高效驾驭大型模型的核心手","stars":2985,"starGrowth":0,"createdAt":"2023-06-08T05:11:00Z","silentRounds":4,"language":"Jupyter Notebook","topics":[],"aiDims":["微调","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","LoRA","微调","深度学习"],"domainKey":"大语言模型","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/ashishpatel26/LLM-Finetuning","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"adongwanai/AgentGuide","owner":"adongwanai","name":"AgentGuide","desc":"https://adongwanai.github.io/AgentGuide | AI Agent开发指南 | LangGraph实战 | 高级RAG | 转行大模型 | 大模型面试 | 算法工程师 | 面试题库 | 强化学习｜数据合成","summaryCn":"从LangGraph到面试题库，一站式拿下AI Agent开发全攻略","reasonCn":"这个仓库是一个系统化的AI Agent开发学习指南，深入讲解LangGraph框架实战和高级RAG架构设计。内容覆盖从入门到进阶的完整知识体系，并整理了大模型公司算法工程师面试题库、强化学习和数据合成等前沿 topics，非常适合准备转行或提升技能的开发者。","stars":9913,"starGrowth":0,"createdAt":"2025-11-03T10:46:38Z","silentRounds":1,"language":"HTML","topics":["ai-agent","rag"],"aiDims":["AI Agent","教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["AI Agent","LangGraph","RAG","面试题库","技术教程"],"domainKey":"AI Agent","pushedAt":"2026-09-15T06:38:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/adongwanai/AgentGuide","ts":"2026-09-15T06:38:22Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mgba-emu/mgba","owner":"mgba-emu","name":"mgba","desc":"mGBA Game Boy Advance Emulator","summaryCn":"想象一下你翻箱倒柜找出了落满灰尘的GBA，却发现屏幕漏液或按键失灵，mGBA就","reasonCn":"技术上，mGBA采用了极度精简的代码架构，通过对ARM7TDMI指令集的高度优化，实现了超越大多数同类项目的执行效率。它不仅能够准确模拟原版硬件的渲染周期，还加入了针对现代显示技术的画面平滑算法和滤镜支持。与传统的VisualBoyAdvance相比，它彻底解决了部分游戏在特定场景下的卡","stars":7426,"starGrowth":0,"createdAt":"2014-12-09T21:37:23Z","silentRounds":1,"language":"C","topics":["emulator"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["Game Boy Advance","游戏模拟","复古游戏"],"domainKey":"Game Boy Advance","pushedAt":"2026-09-19T22:55:26Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/mgba-emu/mgba","ts":"2026-09-19T22:55:26Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Toni-SM/skrl","owner":"Toni-SM","name":"skrl","desc":"Modular Reinforcement Learning (RL) library (implemented in PyTorch, JAX, and NVIDIA Warp) with support for Gymnasium/Gym, NVIDIA Isaac Lab, MuJoCo Playground and other environments","summaryCn":"skrl是一个现代化的模块化强化学习库，支持PyTorch、JAX和NVIDIA Warp三种后","reasonCn":"技术上，它的设计哲学是「解耦」——算法、网络、环境、经验回放这些组件都可以独立替换。比如你可以在一个项目里用JAX做GPU加速训练，在另一个项目里切回PyTorch做研究对比。它原生支持Gymnasium环境标准，也能对接MuJoCo物理仿真、NVIDIA Isaac Lab机器人仿真等高级场景。","stars":1098,"starGrowth":0,"createdAt":"2021-10-18T07:53:11Z","silentRounds":11,"language":"Python","topics":["robotics"],"aiDims":["模型与训练","AI基础设施","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["强化学习","机器人","PyTorch","RL库"],"domainKey":"强化学习","pushedAt":"2026-09-17T21:39:00Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Toni-SM/skrl","ts":"2026-09-17T21:39:00Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"marcelm/cutadapt","owner":"marcelm","name":"cutadapt","desc":"Cutadapt removes adapter sequences from sequencing reads","summaryCn":"生物信息学必备工具，帮你快速清洗测序数据","reasonCn":"它的核心能力非常强大。不仅能处理单端和双端测序数据，还能同时匹配多个接头序列，甚至能自动识别序列中的 3' 或 5' 端的接头。更高级的功能是它支持复杂的修剪模式，比如只保留序列中特定的区域，或者将序列的 3' 端修剪到特定的碱基位置。","stars":587,"starGrowth":0,"createdAt":"2012-06-06T20:26:25Z","silentRounds":12,"language":"Python","topics":["bioinformatics"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["生物信息学","测序数据","接头修剪"],"domainKey":"生物信息学","pushedAt":"2026-09-21T00:34:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"bioinformatics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/marcelm/cutadapt","ts":"2026-09-21T00:34:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Deepractice/PromptX","owner":"Deepractice","name":"PromptX","desc":"","summaryCn":"PromptX 的核心在于解决通用大模型缺乏垂直领域深度知识和长期记忆的问题。","reasonCn":"与传统提示词工具不同，PromptX 提供了“AI Role Creation Platform”（AI 角色创建平台）、“Intelligent Tool Development Platform”（智能工具开发平台）和“Cognitive Memory System”（认知记忆系统）三位一体的能力。","stars":3689,"starGrowth":1,"createdAt":"2025-05-13T04:17:28Z","silentRounds":3,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","提示工程"],"zone":"AI","funScore":0.2,"domainTags":["AI Agent","MCP协议","角色创建","认知记忆","Claude"],"domainKey":"AI Agent","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/Deepractice/PromptX","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JunjoSick/bookforge","owner":"JunjoSick","name":"bookforge","desc":"EPUB-first AI book translation tool supporting OpenAI-compatible APIs (DeepSeek, OpenRouter)","summaryCn":"用AI一键翻译整本书，支持多种API接口，一键生成EPUB","reasonCn":"它的技术优势在于对EPUB格式的深度支持。普通的翻译工具可能只支持TXT文本，翻译完还得自己排版。而Bookforge处理的是真正的电子书文件，它能保留原文的目录、排版和样式，翻译后的内容无缝融入，就像原作者写的一样。","stars":12,"starGrowth":0,"createdAt":"2026-05-01T22:35:11Z","silentRounds":12,"language":"Rust","topics":["ebook"],"aiDims":["AI写作","工具调用","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["电子书翻译","AI翻译","EPUB"],"domainKey":"电子书翻译","pushedAt":"2026-09-21T04:22:15Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/JunjoSick/bookforge","ts":"2026-09-21T04:22:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"seek-oss/playroom","owner":"seek-oss","name":"playroom","desc":"Design with JSX, powered by your own component library.","summaryCn":"写代码的同时实时预览组件效果，前端设计师的创意工作台","reasonCn":"由Seek OSS团队开发的组件实时预览工具，基于React和TypeScript构建。支持在设计时实时看到组件在不同状态、主题和尺寸下的表现，与现有组件库无缝集成。开发体验流畅，帮助设计师和前端工程师高效协作，是组件驱动开发的理想辅助工具。","stars":4584,"starGrowth":0,"createdAt":"2018-10-29T07:14:53Z","silentRounds":26,"language":"TypeScript","topics":["design"],"aiDims":["创意工具","可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["组件预览","React","设计系统","实时渲染"],"domainKey":"组件预览","pushedAt":"2026-09-22T01:15:48Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/seek-oss/playroom","ts":"2026-09-22T01:15:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ML-GSAI/LLaDA","owner":"ML-GSAI","name":"LLaDA","desc":"","summaryCn":"在主流大语言模型几乎完全由 Transformer 自回归（AR）架构主导的今天，ML-","reasonCn":"该项目的核心优势在于其极致的推理效率。LLaDA-MoE-7B-A1B-Instruct 在推理时仅激活约 1B 参数，但其表现不仅超越了前代的 8B 稠密模型 LLaDA 1.5，还与 Qwen2.5-3B-Instruct 处于同一水平。这意味着在保持高精度的同时大幅降低了计算资源消耗。","stars":3987,"starGrowth":1,"createdAt":"2025-02-13T19:28:36Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练","多模态Agent"],"zone":"AI","funScore":0.2,"domainTags":["扩散模型","LLaDA","MoE架构","视觉语言模型"],"domainKey":"扩散模型","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/ML-GSAI/LLaDA","ts":"2026-09-18T20:05:54.376Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LibreTranslate/LibreTranslate","owner":"LibreTranslate","name":"LibreTranslate","desc":"Free and Open Source Machine Translation API. Self-hosted, offline capable and easy to setup.","summaryCn":"想象一下，你有一堆敏感的商业合同或私密笔记需要翻译，但又不敢把它们传到","reasonCn":"技术上，它利用Argos Translate作为后端引擎，底层采用深度的Transformer神经网络架构。与竞品相比，它最大的优势在于轻量级且极易部署，通过Docker容器即可完成一键安装。由于它完全解耦了第三方依赖，你不需要再为每条翻译支付API费用，且能在完全脱机（Offline）的状态下处理成","stars":16710,"starGrowth":3,"createdAt":"2020-12-19T19:19:34Z","silentRounds":0,"language":"Python","topics":["api"],"aiDims":["AI应用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["机器翻译","自托管服务","离线翻译","API 服务"],"domainKey":"机器翻译","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/LibreTranslate/LibreTranslate","ts":"2026-09-15T01:07:58.514Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vulhub/vulhub","owner":"vulhub","name":"vulhub","desc":"Pre-Built Vulnerable Environments Based on Docker-Compose","summaryCn":"漏洞复现神器，几十秒拉起一个靶场，安全入门必看","reasonCn":"它的技术基础是Docker和Docker-Compose。每个漏洞都被封装成一个独立的镜像，包含了操作系统、Web服务、配置文件和漏洞代码。你可以通过简单的JSON配置文件来启动这些环境，比如“启动Struts2漏洞环境”、“启动MySQL弱口令环境”。","stars":21264,"starGrowth":0,"createdAt":"2017-04-09T10:13:13Z","silentRounds":1,"language":"Dockerfile","topics":["docker"],"aiDims":["安全工具","非AI-实用","学习资源"],"zone":"工具","funScore":0,"domainTags":["漏洞复现","Docker","渗透测试","CTF","网络安全"],"domainKey":"漏洞复现","pushedAt":"2026-09-18T17:56:42Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"Dockerfile","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/vulhub/vulhub","ts":"2026-09-18T17:56:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fudan-generative-vision/hallo2","owner":"fudan-generative-vision","name":"hallo2","desc":"","summaryCn":"Hallo2 的核心突破在于“长时长”与“高保真”的兼得。","reasonCn":"该模型由复旦大学、百度及南京大学联合研发，论文已被顶级会议 ICLR 2025 正式录用。技术层面，它属于音频驱动人像图像动画（Audio-driven Portrait Image Animation）领域，通过先进的生成对抗网络或扩散模型机制，精确捕捉音频中的韵律并映射到面部肌肉运动。","stars":3740,"starGrowth":0,"createdAt":"2024-10-16T09:10:54Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["多模态Agent","模型与训练","AI应用"],"zone":"AI","funScore":0.38,"domainTags":["人像动画","音频驱动","ICLR2025","高分辨率生成","4K视频"],"domainKey":"人像动画","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/fudan-generative-vision/hallo2","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bigscience-workshop/promptsource","owner":"bigscience-workshop","name":"promptsource","desc":"","summaryCn":"PromptSource 的核心设计理念是将每一个数据集实例通过特定的函数映射为自然","reasonCn":"该项目通过提供一个交互式的框架，解决了提示模板在不同模型与任务间难以复用和分享的痛点。用户可以利用它定义复杂的模板函数，将繁琐的数据处理逻辑抽象出来，让模型训练过程更加简洁高效。这种规范化的处理方式，使得开发者能够把精力集中在如何设计更具启发性的提示词上，而不","stars":3033,"starGrowth":1,"createdAt":"2021-05-19T15:26:25Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["提示工程","大语言模型","微调"],"zone":"AI","funScore":0.2,"domainTags":["提示模板","数据集处理","自然语言处理","机器学习工作流"],"domainKey":"提示模板","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/bigscience-workshop/promptsource","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"atlassian/atlassian-mcp-server","owner":"atlassian","name":"atlassian-mcp-server","desc":"Official remote MCP server for Atlassian. Securely connect Jira, Confluence, Jira Service Management, Bitbucket, and Compass to Claude, ChatGPT, Cursor, VS Code, and other AI tools using OAuth 2.1 or API tokens.","summaryCn":"很多企业想在AI助手里直接查看Jira工单或更新Confluence文档，但直接连接会","reasonCn":"它的安全性是最大的卖点。所有数据交互都通过OAuth 2.1进行认证授权，这意味着AI工具不需要知道你的密码，也不需要拥有管理员权限，它只能执行你预先授权的操作。它支持连接Jira、Confluence、Jira Service Management、Bitbucket和Compass等多个核心产品。","stars":1059,"starGrowth":0,"createdAt":"2025-08-01T22:31:35Z","silentRounds":1,"language":"JavaScript","topics":["ai","ai-agents","atlassian","bitbucket","chatgpt","claude","compass","confluence","copilot","cursor","developer-tools","jira","jira-service-management","llm","mcp","mcp-server","model-context-protocol","oauth","rovo","vscode"],"aiDims":["工具调用","AI基础设施","AI应用"],"zone":"AI","funScore":0,"domainTags":["MCP","Atlassian","AI 工具集成"],"domainKey":"MCP","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"atlassian","source":"github","weight":0.7},{"name":"bitbucket","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"compass","source":"github","weight":0.7},{"name":"confluence","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"jira","source":"github","weight":0.7},{"name":"jira-service-management","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"model-context-protocol","source":"github","weight":0.7},{"name":"oauth","source":"github","weight":0.7},{"name":"rovo","source":"github","weight":0.7},{"name":"vscode","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/atlassian/atlassian-mcp-server","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dexmal/dexbotic","owner":"dexmal","name":"dexbotic","desc":"Dexbotic: Open-Source Vision-Language-Action Toolbox","summaryCn":"让AI机器人不仅能看懂画面还能动手操作的开源多模态控制箱。","reasonCn":"技术架构上，它采用了最前沿的视觉-语言-动作（VLA）三位一体设计，能够无缝对接各大开源视觉大模型。它内置了丰富的仿真环境接口，支持主流的物理引擎，允许开发者在虚拟世界里反复训练和测试机器人的抓取、导航等复杂动作，大大降低了硬件损耗和研发成本。","stars":3203,"starGrowth":0,"createdAt":"2025-10-17T03:23:00Z","silentRounds":3,"language":"Python","topics":["simulation"],"aiDims":["多模态Agent","AI Agent","开发者工具"],"zone":"AI","funScore":0.7,"domainTags":["具身智能","机器人仿真","视觉语言模型","VLA"],"domainKey":"具身智能","pushedAt":"2026-09-16T11:54:57Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/dexmal/dexbotic","ts":"2026-09-16T11:54:57Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LiberCoders/FeatureBench","owner":"LiberCoders","name":"FeatureBench","desc":"[ICLR 2026] Official Implementation of \"FeatureBench: Benchmarking Agentic Coding for Complex Feature Development\" ","summaryCn":"给AI写代码能力打分的基准测试，看谁更懂复杂开发","reasonCn":"FeatureBench 是一个针对 AI 编码代理能力的基准测试框架，旨在评估大模型在复杂功能开发任务上的表现。它不仅关注代码生成的正确性，还深入考察了代理在长上下文理解、多文件修改、工具调用和逻辑推理方面的综合能力。","stars":95,"starGrowth":1,"createdAt":"2026-01-20T09:35:46Z","silentRounds":25,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","代码助手","大语言模型"],"zone":"资源","funScore":0.6,"domainTags":["AI Benchmark","Agent Evaluation","代码生成","基准测试","评测框架"],"domainKey":"AI Benchmark","pushedAt":"2026-09-15T08:56:16Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/LiberCoders/FeatureBench","ts":"2026-09-15T08:56:16Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"minitap-ai/mobile-use","owner":"minitap-ai","name":"mobile-use","desc":"AI agents can now use real Android and iOS apps, just like a human.","summaryCn":"让AI像真人一样刷手机，自动操作各类App，解放你的双手","reasonCn":"这个项目让AI智能体能像人类一样操作Android和iOS手机应用。通过视觉识别屏幕内容，理解界面元素，模拟真实用户的手势点击、滑动、输入等操作。支持主流手机应用，从社交、购物到出行APP都能操控。基于大语言模型理解任务意图，结合手机操控工具实现端到端自动化，无需ROOT或越狱即可在真实设备上运行。","stars":3161,"starGrowth":0,"createdAt":"2025-08-16T21:49:49Z","silentRounds":1,"language":"Python","topics":["mobile"],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["AI Agent","移动端自动化","视觉识别","Android","iOS"],"domainKey":"AI Agent","pushedAt":"2026-09-14T17:13:50Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/minitap-ai/mobile-use","ts":"2026-09-14T17:13:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Developer-Y/cs-video-courses","owner":"Developer-Y","name":"cs-video-courses","desc":"List of Computer Science courses with video lectures.","summaryCn":"全网最全的CS课程视频合集，从入门到精通一站式找齐","reasonCn":"这个项目整理收录了全球顶尖高校计算机科学各方向的优质视频课程，涵盖机器学习、深度学习、系统设计、算法等核心领域。8万star证明其权威性与实用性，按类别清晰分类，省去四处搜课的时间，是系统学习AI和计算机科学的宝藏资源库。","stars":83552,"starGrowth":1,"createdAt":"2016-10-21T17:02:11Z","silentRounds":1,"language":"","topics":["ml","security"],"aiDims":["学习资源","教程","模型与训练"],"zone":"资源","funScore":0,"domainTags":["计算机科学","课程合集","MIT","Stanford","视频教程"],"domainKey":"计算机科学","pushedAt":"2026-09-21T16:48:31Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7}],"aiScore":0.62,"source":"search","url":"https://github.com/Developer-Y/cs-video-courses","ts":"2026-09-21T16:48:31Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/smollm","owner":"huggingface","name":"smollm","desc":"","summaryCn":"在人工智能追求参数规模军备竞赛的当下，Hugging Face 的 SmolModels 项目反","reasonCn":"SmolLM3 是该家族中的明星产品，其最引人注目的特点是以 3B 的参数量实现了超越同级甚至部分更大模型的惊人性能。根据官方基准测试，它在多项任务上击败了 Llama 3.2 3B 和 Qwen2.5 3B，甚至在某些指标上与 4B 级别的模型相提并论。","stars":3904,"starGrowth":0,"createdAt":"2024-11-04T13:01:54Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["轻量级LLM","端侧AI","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/huggingface/smollm","ts":"2026-09-18T20:41:13.636Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"vercel-labs/knowledge-agent-template","owner":"vercel-labs","name":"knowledge-agent-template","desc":"Open source file-system and knowledge based agent template. Build AI agents that stay up to date with your knowledge base","summaryCn":"帮你把个人资料库喂给AI，让它秒懂你的所有文档","reasonCn":"这个项目最厉害的地方在于它把AI和文件系统结合得很紧密。它不是简单的“读文件再回答”，而是构建了一个完整的Agent框架。AI不仅仅是聊天，还能使用工具去执行具体的文件操作，比如搜索、读取、修改文件，甚至调用外部API。","stars":1015,"starGrowth":0,"createdAt":"2026-01-20T15:02:20Z","silentRounds":3,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["文件系统","知识库","Agent","TypeScript","本地部署"],"domainKey":"Agent","pushedAt":"2026-09-15T08:10:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/vercel-labs/knowledge-agent-template","ts":"2026-09-15T08:10:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jundot/omlx","owner":"jundot","name":"omlx","desc":"LLM inference server with continuous batching &amp; SSD caching for Apple Silicon — managed from the macOS menu bar","summaryCn":"在Mac菜单栏就能一键管理大模型，本地推理又快又稳","reasonCn":"技术上，omlx深度利用了Apple Silicon的统一内存架构和神经引擎，通过连续批处理技术，让系统可以同时处理多个请求，大幅提升了吞吐量。SSD缓存机制则将模型加载到内存中，避免了反复从硬盘读取的延迟，确保推理速度飞快。","stars":22080,"starGrowth":16,"createdAt":"2026-02-13T14:13:27Z","silentRounds":0,"language":"Python","topics":["apple-silicon","inference-server","llm","macos","mlx","openai-api"],"aiDims":["推理引擎","大语言模型","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["LLM推理","大模型","Apple Silicon","本地部署"],"domainKey":"LLM推理","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"apple-silicon","source":"github","weight":0.7},{"name":"inference-server","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"mlx","source":"github","weight":0.7},{"name":"openai-api","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/jundot/omlx","ts":"2026-09-12T22:00:03.438Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.06020869565217391,"copyOk":false},{"repo":"zexadev/lapisnote","owner":"zexadev","name":"lapisnote","desc":"简洁高效的本地笔记应用，内置 AI 写作助手 | Local-first note-taking app with AI assistant","summaryCn":"如果你是一个对隐私极其敏感，又不想错过AI时代生产力红利的人，这个应用就","reasonCn":"它的核心技术架构采用了模块化的TypeScript设计，保证了应用的响应速度和扩展性。在AI集成方面，它不仅支持接入主流模型API进行在线增强，还考虑了本地轻量模型的交互适配，能够根据网络环境智能切换任务逻辑。同时，该应用高度重视数据的本地化管理，所有笔记内容以标准格式存放在","stars":16,"starGrowth":0,"createdAt":"2025-12-02T04:35:58Z","silentRounds":16,"language":"TypeScript","topics":["writing"],"aiDims":["AI写作","效率工具"],"zone":"工具","funScore":0,"domainTags":["本地笔记","AI写作助手","Markdown","隐私优先"],"domainKey":"本地笔记","pushedAt":"2026-09-16T12:49:31Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/zexadev/lapisnote","ts":"2026-09-16T12:49:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DarkflameUniverse/DarkflameServer","owner":"DarkflameUniverse","name":"DarkflameServer","desc":"The main repository for the Darkflame Universe Server Emulator project.","summaryCn":"硬核还原经典网游《天际传说》的服务端模拟器。","reasonCn":"这个项目的最大难点在于“逆向工程”。因为原版游戏没有公开的服务器源码，开发者需要通过抓包、反编译、协议分析等手段，把游戏里的每一条指令、每一个数据包都搞清楚，然后自己重新写一套逻辑。这就像是在黑盒子里复原一台精密机器，难度极高。","stars":755,"starGrowth":0,"createdAt":"2021-12-04T19:09:48Z","silentRounds":10,"language":"C++","topics":["emulator"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0.7,"domainTags":["服务器模拟器","逆向工程","C++","MMO"],"domainKey":"服务器模拟器","pushedAt":"2026-09-18T12:25:08Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/DarkflameUniverse/DarkflameServer","ts":"2026-09-18T12:25:08Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Cassette-Editor/oh-my-cassette","owner":"Cassette-Editor","name":"oh-my-cassette","desc":"你的随身 AI 剪辑搭档 | Pocket AI co-editor for video montage — AI video editing plugin & MCP server for Claude Code, Codex, Hermes & OpenCode","summaryCn":"你是不是也觉得剪视频最痛苦的不是创意，而是没完没了的素材整理和拖拽？","reasonCn":"核心亮点在于它支持MCP（模型上下文协议）架构。简单说，这让AI不仅能看懂视频内容，还能直接下达剪辑指令，比如“帮我剪掉所有沉默片段”或者“把这段视频对齐到BGM的鼓点上”，它都能直接调用剪辑引擎帮你完成。这种“命令式剪辑”大大颠覆了传统非线性编辑的模式。","stars":152,"starGrowth":0,"createdAt":"2026-06-10T14:17:55Z","silentRounds":5,"language":"Python","topics":["video-editor"],"aiDims":["创意工具","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["视频剪辑","MCP服务器","自动化剪辑","创意工作流","AI辅助"],"domainKey":"视频剪辑","pushedAt":"2026-09-15T17:50:13Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Cassette-Editor/oh-my-cassette","ts":"2026-09-15T17:50:13Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tmewett/BrogueCE","owner":"tmewett","name":"BrogueCE","desc":"Brogue: Community Edition - a community-lead fork of the much-loved minimalist roguelike game","summaryCn":"经典地牢冒险游戏续命版，随机关卡让人欲罢不能","reasonCn":"这是广受喜爱的极简风格Roguelike游戏《Brogue》的社区活跃分支。它修复了原版陈旧问题，支持高分辨率显示与现代操作系统，拥有深度随机生成的地图和物品。代码基于C语言，架构清晰，适合想体验硬核回合制探险或研究游戏AISourceCode的开发者。","stars":1471,"starGrowth":1,"createdAt":"2019-09-02T16:27:37Z","silentRounds":2,"language":"C","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["Roguelike","地牢冒险","程序生成","ASCII游戏","C语言"],"domainKey":"Roguelike","pushedAt":"2026-09-20T15:01:43Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/tmewett/BrogueCE","ts":"2026-09-20T15:01:43Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Eigenwise/eigenwise-toolshed","owner":"Eigenwise","name":"eigenwise-toolshed","desc":"Six Claude Code plugins for the work that keeps coming back: repo maps, conditional rules, ticketed parallel work, extra subscription models, local usage metrics, and guided setup.","summaryCn":"很多开发者在使用 AI 编程时发现，虽然 AI 写代码很快，但很难让它处理长期","reasonCn":"这套插件的核心亮点在于它解决了 AI 编程中的“上下文管理”和“任务拆解”痛点。比如，当开发者说“修复这个 Bug”时，普通 AI 可能会直接改代码，而有了这个插件，AI 会先自动生成“仓库地图”，理解整个代码库的架构，然后制定详细的修复计划，甚至可以在后台并行运行多个测试用例来验证","stars":269,"starGrowth":0,"createdAt":"2026-06-13T08:55:58Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["AI Agent","工具调用","效率工具"],"zone":"AI","funScore":0,"domainTags":["Claude Code","插件","任务拆解"],"domainKey":"Claude Code","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/Eigenwise/eigenwise-toolshed","ts":"2026-09-15T01:07:58.514Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenHUTB/hutb","owner":"OpenHUTB","name":"hutb","desc":"人车模拟器（Human-vehicle Simulator）","summaryCn":"跨平台人车模拟器，实时渲染车辆与行人交互，助力交通研究与教学。","reasonCn":"OpenHUTB 的 hutb 项目旨在为交通研究和教学提供一个可视化、可交互的人车仿真环境。传统的交通仿真往往只能在二维平面或简化模型上运行，而 hutb 通过高性能的 C++ 引擎结合 OpenGL 渲染，实现了三维场景下的真实车辆动力学和行人行为交互，让使用者能够直观看到车辆加速、转向、","stars":616,"starGrowth":0,"createdAt":"2024-05-05T13:22:42Z","silentRounds":1,"language":"C++","topics":["simulation"],"aiDims":["非AI-实用","可视化"],"zone":"工具","funScore":0.7,"domainTags":["交通仿真","人车模拟","OpenGL","三维可视化"],"domainKey":"交通仿真","pushedAt":"2026-09-21T13:30:15Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.61,"source":"search","url":"https://github.com/OpenHUTB/hutb","ts":"2026-09-21T13:30:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Machine-Learning-Tokyo/Interactive_Tools","owner":"Machine-Learning-Tokyo","name":"Interactive_Tools","desc":"","summaryCn":"传统机器学习教程往往堆满密集的公式和难以理解的参数黑盒，让初学者在入门","reasonCn":"在深度学习模块中，它提供了一系列极具针对性的专业可视化工具。例如专门用于拆解 Transformer 架构的交互解释器，能让用户清晰地追踪注意力权重的分布；还有针对 BERT 等预训练模型的 BERTViz 工具，将复杂的词向量映射关系转化为直观的可视化图谱。","stars":2886,"starGrowth":0,"createdAt":"2019-04-29T23:24:38Z","silentRounds":4,"language":"","topics":[],"aiDims":["学习资源","可视化"],"zone":"资源","funScore":0.5599999999999999,"domainTags":["Transformer 可视化","BERT 解释工具","GAN 交互演示","概率分布图表","贝叶斯推断"],"domainKey":"Transformer 可视化","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1}],"aiScore":0.68,"source":"search","url":"https://github.com/Machine-Learning-Tokyo/Interactive_Tools","ts":"2026-09-19T22:19:04.205Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"strakam/generals-bots","owner":"strakam","name":"generals-bots","desc":"Develop your agent for generals.io! ","summaryCn":"用Python给即时战略游戏造一个能自己打比赛的AI机器人","reasonCn":"专为 generals.io 网页版即时战略游戏打造的开源 bot 开发框架，基于 pygame 实现渲染与交互。提供完整 API 让开发者编写 AI 策略逻辑，包括资源管理、兵力调度、进攻防守等模块。GitHub 116 star，社区虽小众但稳定，适合想动手实践游戏 AI 的程序员。","stars":120,"starGrowth":0,"createdAt":"2024-06-05T10:44:06Z","silentRounds":10,"language":"Python","topics":["pygame"],"aiDims":["AI Agent","游戏","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["即时战略游戏","Bot开发","pygame","游戏AI"],"domainKey":"即时战略游戏","pushedAt":"2026-09-21T12:51:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"pygame","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/strakam/generals-bots","ts":"2026-09-21T12:51:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"junegunn/fzf","owner":"junegunn","name":"fzf","desc":"🌸 A command-line fuzzy finder","summaryCn":"敲一个键就能模糊搜索，终端里找文件快如闪电的神器","reasonCn":"fzf是命令行模糊查找工具，让你在终端输入几个字母就能实时匹配并显示所有相关结果。它支持文件查找、历史命令搜索、进程查看等各种场景，可直接结合管道使用。性能极为出色，即使面对成千上万条目录也能毫秒级响应，是开发者提升效率的必备神器。","stars":83199,"starGrowth":5,"createdAt":"2013-10-23T16:04:23Z","silentRounds":0,"language":"Go","topics":["fzf","go","bash","zsh","fish","vim","neovim","cli","unix","tmux"],"aiDims":["效率工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["命令行工具","模糊搜索","文件查找","终端交互","效率工具"],"domainKey":"命令行工具","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"fzf","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"bash","source":"github","weight":0.7},{"name":"zsh","source":"github","weight":0.7},{"name":"fish","source":"github","weight":0.7},{"name":"vim","source":"github","weight":0.7},{"name":"neovim","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"unix","source":"github","weight":0.7},{"name":"tmux","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/junegunn/fzf","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01518029062849313,"copyOk":true},{"repo":"serengil/deepface","owner":"serengil","name":"deepface","desc":"A Lightweight Face Recognition and Facial Attribute Analysis (Age, Gender, Emotion and Race) Library for Python","summaryCn":"人脸识别加情绪分析，一键识别年龄性别情绪种族","reasonCn":"集成多个主流人脸识别模型，支持年龄、性别、情绪、种族分析。API简洁，几行代码就能完成人脸检测和属性识别。底层支持VGG-Face、FaceNet、OpenFace等多个模型架构，可以直接在图片上画出人脸识别框并标注属性信息，适合快速开发各类人脸相关应用。","stars":23476,"starGrowth":2,"createdAt":"2020-02-08T20:42:28Z","silentRounds":0,"language":"Python","topics":["ml"],"aiDims":["AI应用","多模态Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["人脸识别","Python库","深度学习","情感分析","计算机视觉"],"domainKey":"人脸识别","pushedAt":"2026-09-16T12:15:47Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/serengil/deepface","ts":"2026-09-16T12:15:47Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hwchase17/langchain-hub","owner":"hwchase17","name":"langchain-hub","desc":"","summaryCn":"LangChainHub 的设计初衷是成为连接开发者与复杂 LLM 应用的桥梁，它通过集","reasonCn":"该项目的运作方式高度依赖社区的协同效应，通过 GitHub 的 Pull Request 流程，开发者可以贡献自己调试好的提示词或链路模板。这种设计不仅保证了资源的多样性，还通过开源协作的方式不断优化提示词工程的质量。对于开发者而言，它不仅是一个工具库，更是学习如何构建生产级 AI 逻","stars":3413,"starGrowth":0,"createdAt":"2023-01-11T03:22:40Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["提示工程","工具调用"],"zone":"AI","funScore":0.2,"domainTags":["提示词工程","大语言模型","链式架构"],"domainKey":"大语言模型","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/hwchase17/langchain-hub","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fathah/hermes-desktop","owner":"fathah","name":"hermes-desktop","desc":"Desktop Companion for Hermes Agent","summaryCn":"电脑桌面上的AI小助手，随时帮你处理各种日常任务","reasonCn":"基于Electron构建的桌面端AI Agent伴侣应用，为Hermes Agent提供图形化操作界面。支持本地文件读写、代码执行、浏览器控制等多种工具调用能力，可接入OpenAI、Anthropic等主流模型。界面简洁现代，支持多窗口布局和自定义主题，适合希望在本地运行AI助手又不想写代码的用户。","stars":14279,"starGrowth":1,"createdAt":"2026-04-02T15:26:49Z","silentRounds":1,"language":"TypeScript","topics":["ai-agent","chatbot"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["AI助手","桌面应用","Agent","Electron"],"domainKey":"Agent","pushedAt":"2026-09-16T05:56:49Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/fathah/hermes-desktop","ts":"2026-09-16T05:56:49Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"freestylefly/awesome-gpt-image-2","owner":"freestylefly","name":"awesome-gpt-image-2","desc":"Prompt as Code | GPT Image 2 / 2.5 提示词与案例库，530+ 个案例、20+ 套工业级模板与可复用 Skills，新增 2.5 同提示词对比专区，附完整提示词与生成记录，持续更新。","summaryCn":"把提示词当成代码来写，解锁GPT-Image2的工业级创作能力","reasonCn":"这个项目把GPT-Image2的提示词工程做到了极致，逆向分析了470多个优秀案例，提炼出20多套可复用的工业级模板。它不仅仅是收集prompt，而是构建了一套Prompt as Code体系，让设计师和开发者能像写代码一样精准控制AI绘图输出，大幅降低试错成本，是视觉创作领域的实用工具库。","stars":33277,"starGrowth":19,"createdAt":"2026-04-25T09:19:13Z","silentRounds":0,"language":"JavaScript","topics":["agents"],"aiDims":["提示工程","AI应用","创意工具"],"zone":"资源","funScore":0,"domainTags":["提示词工程","AI绘图","案例库"],"domainKey":"提示词工程","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/freestylefly/awesome-gpt-image-2","ts":"2026-09-12T22:00:03.438Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07045547375063857,"copyOk":true},{"repo":"polakowo/vectorbt","owner":"polakowo","name":"vectorbt","desc":"The backtesting engine that gives you an unfair advantage. Run thousands of trading ideas before others finish one.","summaryCn":"量化交易的回测引擎，用一行代码跑几千个交易策略","reasonCn":"核心技术是用向量化计算替代逐笔模拟。传统回测是逐条处理历史数据，vectorbt直接对整个矩阵进行操作，充分利用CPU的SIMD指令。这意味着你可以同时在几秒钟内测试数千个参数组合，找到最优策略。 安装：pip install vectorbt。","stars":9156,"starGrowth":2,"createdAt":"2017-11-14T11:56:32Z","silentRounds":0,"language":"Python","topics":["data-visualization","trading","finance"],"aiDims":["AI应用","可视化"],"zone":"工具","funScore":0,"domainTags":["量化交易","回测引擎","向量化计算","金融策略"],"domainKey":"量化交易","pushedAt":"2026-09-17T20:38:03Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"trading","source":"github","weight":0.7},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/polakowo/vectorbt","ts":"2026-09-17T20:38:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yfeng95/GAN","owner":"yfeng95","name":"GAN","desc":"","summaryCn":"GAN 项目是一个专为学习生成对抗网络（GAN）而准备的代码仓库，所有实现均基","reasonCn":"项目目录包括 data.py 用于数据预处理，nets.py 包含 Generator 与 Discriminator 的定义，此外还有用于保存训练数据和生成样本的文件夹。用户只需按照文档说明安装 TensorFlow1.0+，运行训练脚本即可在本地生成手写数字图片，观察生成质量随训练轮数的提升。","stars":3099,"starGrowth":0,"createdAt":"2017-03-27T12:29:40Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["模型与训练","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["GAN","TensorFlow","MNIST","生成模型"],"domainKey":"GAN","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/yfeng95/GAN","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"markusschanta/awesome-jupyter","owner":"markusschanta","name":"awesome-jupyter","desc":"A curated list of awesome Jupyter projects, libraries and resources","summaryCn":"Jupyter星球宝藏地图，从教程到工具一键打包","reasonCn":"这是一份精心整理的Jupyter生态大清单，汇聚了全球开发者贡献的笔记模板、交互式可视化组件、AI数据处理库以及高效使用技巧。无论是刚入门的数据科学新手，还是追求极致体验的资深分析师，都能在这里找到提升生产力的利器，是离线查阅的首选指南。","stars":4676,"starGrowth":1,"createdAt":"2017-08-27T22:00:02Z","silentRounds":1,"language":"","topics":["data-visualization","frontend","data-science"],"aiDims":["学习资源","AI应用","可视化"],"zone":"资源","funScore":0.6,"domainTags":["Jupyter","资源导航","数据科学","可视化","精选清单"],"domainKey":"Jupyter","pushedAt":"2026-09-22T23:57:48Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7}],"aiScore":0.55,"source":"search","url":"https://github.com/markusschanta/awesome-jupyter","ts":"2026-09-22T23:57:48Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pi-hole/PADD","owner":"pi-hole","name":"PADD","desc":"PADD (formerly Chronometer2) is a more expansive version of the original chronometer.sh that is included with Pi-Hole. PADD provides in-depth information about your Pi-hole.","summaryCn":"把树莓派小屏幕变成酷炫的去广告监控仪表盘的神器","reasonCn":"PADD是专门为树莓派全局广告拦截系统打造的终端仪表盘面板，采用纯粹的Shell脚本编写，绝无任何额外繁重的依赖组件。它能够实时提取并优雅地呈现内部域名解析查询总量、广告过滤比例、实时屏蔽名单规模以及树莓派系统的物理内存和处理器负载状态，极度适合部署在微型屏幕上作为极客桌面显示终端。","stars":1553,"starGrowth":0,"createdAt":"2017-03-11T02:54:53Z","silentRounds":13,"language":"Shell","topics":["raspberry-pi"],"aiDims":["非AI-好玩","非AI-实用","可视化"],"zone":"创意","funScore":1,"domainTags":["树莓派","广告拦截","仪表盘"],"domainKey":"树莓派","pushedAt":"2026-09-19T10:05:19Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/pi-hole/PADD","ts":"2026-09-19T10:05:19Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"huggingface/huggingface_hub","owner":"huggingface","name":"huggingface_hub","desc":"","summaryCn":"只要运行一个简单的独立安装器，开发者就能迅速构建起连接 Hugging Face 社","reasonCn":"该库提供了完善的认证机制，通过配置安全令牌，用户能够轻松实现对自己项目的版本控制与私有模型管理。对于那些渴望在开源社区贡献力量的人来说，它不仅是一个下载工具，更是一个高效的发布平台，让你编写的算法、整理的数据集或精心调试的 Demo 能够迅速触达全球同行，通过这种方","stars":3923,"starGrowth":0,"createdAt":"2020-12-22T10:20:28Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型部署","开发者工具"],"zone":"AI","funScore":0.2,"domainTags":["机器学习模型","数据集管理","开源协作平台"],"domainKey":"机器学习模型","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/huggingface/huggingface_hub","ts":"2026-09-18T20:41:13.636Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"QwenLM/Qwen3.8","owner":"QwenLM","name":"Qwen3.8","desc":"","summaryCn":"Qwen3.8 的发布标志着开源大语言模型在能力边界上的进一步突破，其最显著的","reasonCn":"在技术领域，Qwen3.8 在代码生成、专业文档处理和科学研究辅助方面表现得尤为出色。它能够更好地理解上下文中的隐含意图，并在长对话或长文档保持逻辑一致性。基于 Qwen3.5 的架构改进，该模型在处理大规模上下文窗口时效率更高，延迟更低。","stars":4161,"starGrowth":0,"createdAt":"2025-09-11T05:32:39Z","silentRounds":1,"language":"","topics":[],"aiDims":["大语言模型","AI Agent"],"zone":"AI","funScore":0.2,"domainTags":["通义千问","大语言模型","Agent 任务"],"domainKey":"大语言模型","pushedAt":"2026-09-18T10:52:55.418Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1}],"aiScore":0.74,"source":"tier","url":"https://github.com/QwenLM/Qwen3.8","ts":"2026-09-18T10:52:55.418Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"torinmb/mediapipe-touchdesigner","owner":"torinmb","name":"mediapipe-touchdesigner","desc":"","summaryCn":"在创意编码和实时视觉艺术领域，TouchDesigner 是行业标准，但它缺乏直接调","reasonCn":"该插件的一个显著特点是其“自包含”和“无需安装”的特性。这意味着它直接在 GPU 上加速运行，用户只需下载 release.zip 并打开 .toe 文件即可开始使用，极大地降低了环境配置的门槛。官方警告用户在拖拽组件时务必选择“启用外部 .tox”，以防止文件体积过大，体现了其对独立部署的优化","stars":2765,"starGrowth":0,"createdAt":"2023-08-02T13:51:17Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":0.68,"domainTags":["MediaPipe","TouchDesigner","计算机视觉"],"domainKey":"MediaPipe","pushedAt":"2026-09-21T04:10:58.868Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/torinmb/mediapipe-touchdesigner","ts":"2026-09-21T04:10:58.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nihui/waifu2x-ncnn-vulkan","owner":"nihui","name":"waifu2x-ncnn-vulkan","desc":"","summaryCn":"这款工具的核心亮点在于极致的性能优化与硬件兼容性。","reasonCn":"在分发形式上，该项目对新手极为友好。官方在 Releases 页面直接提供了 Windows、Linux 和 macOS 的全平台可执行文件，并且压缩包内已经预置了所需的所有二进制文件和模型权重。用户下载解压后，完全不需要去 Python 官网配置环境，也不需要手动下载几百兆的模型文件，真正做到了“","stars":3474,"starGrowth":0,"createdAt":"2019-04-03T14:14:34Z","silentRounds":3,"language":"C++","topics":[],"aiDims":["推理引擎","AI应用","图像增强"],"zone":"AI","funScore":0.5,"domainTags":["图像超分辨率","视频补帧","Vulkan加速","AI降噪"],"domainKey":"图像超分辨率","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"图像增强","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/nihui/waifu2x-ncnn-vulkan","ts":"2026-09-19T18:31:48.187Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"anthropics/claude-plugins-community","owner":"anthropics","name":"claude-plugins-community","desc":"","summaryCn":"Claude Plugins Community 是一个开放的插件生态池，汇集了众多由社区开发者","reasonCn":"该项目解决了AI模型“孤岛化”的问题，让模型不再仅仅局限于文字处理，而是通过插件机制深度介入生产工作流。不论是需要自动化执行终端命令，还是希望让Claude直接管理你的笔记系统，这里都可能藏着你需要的功能雏形。","stars":4387,"starGrowth":11,"createdAt":"2026-03-20T02:20:30Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["工具调用","AI Agent"],"zone":"工具","funScore":0.2,"domainTags":["Claude","插件开发","集成工具"],"domainKey":"Claude","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/anthropics/claude-plugins-community","ts":"2026-09-18T20:05:54.376Z","category":"tool","momentum":["daily"],"fromOfficial":true,"heatScore":0.048626669705949394,"copyOk":false},{"repo":"Lightning-AI/litgpt","owner":"Lightning-AI","name":"litgpt","desc":"20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.","summaryCn":"你是不是觉得ChatGPT很好用，但又想让它学会你们公司的专业知识？","reasonCn":"它的厉害之处在于极高的性能优化，利用Flash Attention等尖端技术，极大减少了GPU内存开销，让你即使在普通的消费级显卡（如RTX系列）上也能跑得动精细化的微调任务。此外，它采用模块化设计，你可以随意替换模型的层结构、修改优化器参数，这种灵活性让它成了研究人员调试模型参数","stars":13679,"starGrowth":1,"createdAt":"2023-05-04T17:46:11Z","silentRounds":1,"language":"Python","topics":["deep-learning"],"aiDims":["大语言模型","微调","模型部署"],"zone":"AI","funScore":0.7,"domainTags":["大模型微调","性能优化","Flash Attention","PyTorch框架"],"domainKey":"大模型微调","pushedAt":"2026-09-14T10:24:06Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/Lightning-AI/litgpt","ts":"2026-09-14T10:24:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"netbirdio/netbird","owner":"netbirdio","name":"netbird","desc":"Connect your devices into a secure WireGuard®-based overlay network with SSO, MFA and granular access controls.","summaryCn":"把你的设备连成网，像打局域网游戏一样简单安全","reasonCn":"技术上，Netbird 使用 WireGuard 的轻量级内核，保证了极高的加密效率和低延迟。它最大的亮点是采用了基于 OpenID Connect 的 SSO 机制，这意味着你可以直接用公司现有的 Active Directory 或 SAML 提供商来管理用户登录，不用再记一堆复杂的 VPN 账号密码。","stars":29440,"starGrowth":3,"createdAt":"2021-04-14T12:27:20Z","silentRounds":0,"language":"Go","topics":["wireguard","wireguard-vpn","vpn","nat-traversal","mesh-networks","mesh","golang","wiretrustee","zero-trust-network-access","netbird"],"aiDims":["安全工具","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["VPN","WireGuard","远程访问"],"domainKey":"VPN","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"wireguard","source":"github","weight":0.7},{"name":"wireguard-vpn","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"nat-traversal","source":"github","weight":0.7},{"name":"mesh-networks","source":"github","weight":0.7},{"name":"mesh","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"wiretrustee","source":"github","weight":0.7},{"name":"zero-trust-network-access","source":"github","weight":0.7},{"name":"netbird","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/netbirdio/netbird","ts":"2026-09-20T04:13:28.323Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"getoffmystagestevewilkos-spec/velo-print-rigs","owner":"getoffmystagestevewilkos-spec","name":"velo-print-rigs","desc":"3D Printable Cycling Gear 2026: Wheel Blocks, Tablet Stands & AirTag Holders","summaryCn":"专为骑行爱好者打造的3D打印好物合集，车轮固定块和平板支架自己动手做。","reasonCn":"技术上，这个项目采用开源的3D打印设计规范，所有的源文件和STL格式都经过精心优化。这些模型在设计时充分考虑了骑行过程中的受力情况和震动环境，通过合理的壁厚设计和填充结构，在保证轻量化的同时具备极高的机械强度，普通家用FDM打印机就能完美搞定。","stars":56,"starGrowth":0,"createdAt":"2026-08-25T08:57:21Z","silentRounds":13,"language":"HTML","topics":["3d-printing"],"aiDims":["非AI-好玩","创意工具","非AI-实用"],"zone":"创意","funScore":0,"domainTags":["3D打印","骑行","硬件","配件"],"domainKey":"3D打印","pushedAt":"2026-09-16T21:27:42Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/getoffmystagestevewilkos-spec/velo-print-rigs","ts":"2026-09-16T21:27:42Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"visresearch/WordAgent","owner":"visresearch","name":"WordAgent","desc":"An AI agent-powered writing assistance system (Codex style) that enables AI-assisted content creation via WPS and Microsoft Word add-ins. 基于AI智能体的写作辅助系统，通过WPS、Microsoft Word加载项，实现AI辅助的文字创作","summaryCn":"在Word和WPS里直接召唤AI帮你写文案改稿子","reasonCn":"将AI智能体嵌入WPS和Microsoft Word加载项，无需切换界面即可实现AI辅助写作。支持续写、改写、纠错等功能，像Codex一样理解上下文并给出建议。集成国内用户常用的办公软件，解决频繁切换工具的不便，对内容创作者、学生党、职场写作者尤其友好。","stars":152,"starGrowth":0,"createdAt":"2026-03-10T11:17:15Z","silentRounds":1,"language":"Python","topics":["writing"],"aiDims":["AI写作","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["写作辅助","AI Agent","办公软件集成","提示词工程"],"domainKey":"AI Agent","pushedAt":"2026-09-17T04:55:00Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/visresearch/WordAgent","ts":"2026-09-17T04:55:00Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"awslabs/agentcore-samples","owner":"awslabs","name":"agentcore-samples","desc":"","summaryCn":"一套可插拔的AI代理部署方案，支持任意框架和模型，实现安全规模化运行。","reasonCn":"这是亚马逊官方提供的AgentCore示例仓库。它解决了AI代理从开发到生产环境的关键痛点——框架锁定和模型锁定。无论你用LangGraph还是CrewAI构建代理，都能通过这套基础设施统一部署、运维和扩展，同时兼顾企业级的安全与合规要求。Python SDK和CLI让上手变得直观。","stars":3395,"starGrowth":1,"createdAt":"2025-07-03T21:05:14Z","silentRounds":1,"language":"Python","topics":["agent","agentic-ai","agents","authentication","bedrock","core","gateway","identity-management","memory-management","production-code","runtime"],"aiDims":["AI Agent","模型部署","AI基础设施"],"zone":"AI","funScore":0.26,"domainTags":["AI Agent","模型部署","多框架支持","云服务","安全运维"],"domainKey":"AI Agent","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"authentication","source":"github","weight":0.7},{"name":"bedrock","source":"github","weight":0.7},{"name":"core","source":"github","weight":0.7},{"name":"gateway","source":"github","weight":0.7},{"name":"identity-management","source":"github","weight":0.7},{"name":"memory-management","source":"github","weight":0.7},{"name":"production-code","source":"github","weight":0.7},{"name":"runtime","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/awslabs/agentcore-samples","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nasa/apod-api","owner":"nasa","name":"apod-api","desc":"Astronomy Picture of the Day API service","summaryCn":"NASA每日天文美图API，每天给你送上宇宙浪漫。","reasonCn":"NASA每天都会发布一张最新的天文图片，这个开源项目提供了一个简单的API接口，开发者可以轻松调用NASA每日天文图（APOD）的数据。它返回JSON格式的图片链接、标题、解释等元数据，方便集成到任何网站、App或桌面工具中。","stars":1086,"starGrowth":0,"createdAt":"2015-12-01T21:33:12Z","silentRounds":9,"language":"Python","topics":["astronomy"],"aiDims":["非AI-实用","可视化","开发者工具"],"zone":"资源","funScore":0,"domainTags":["NASA","天文摄影","RESTful API"],"domainKey":"NASA","pushedAt":"2026-09-18T01:30:34Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"astronomy","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/nasa/apod-api","ts":"2026-09-18T01:30:34Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"52North/WeatherRoutingTool","owner":"52North","name":"WeatherRoutingTool","desc":"A tool to perform optimization of ship routes based on fuel consumption in different weather conditions","summaryCn":"在全球航运行业，燃料成本占据船舶运营的大头。","reasonCn":"核心技术：\n· 采用Python编写，利用Pandas、NumPy进行海量数据处理；\n· 燃料消耗模型基于船舶动力学，考虑速度、风速、海流等因素；\n· 采用A*或Dijkstra等图算法在网格地图上搜索最短燃料成本路径；\n· 输出结果可直接导入船舶导航系统或生成KML文件供Google Earth等查看。","stars":65,"starGrowth":0,"createdAt":"2024-01-11T10:38:03Z","silentRounds":13,"language":"Python","topics":["weather"],"aiDims":["非AI-实用","工具调用"],"zone":"工具","funScore":0,"domainTags":["船舶航线","燃油优化","气象数据"],"domainKey":"船舶航线","pushedAt":"2026-09-21T11:45:26Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"weather","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/52North/WeatherRoutingTool","ts":"2026-09-21T11:45:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PaddlePaddle/FastDeploy","owner":"PaddlePaddle","name":"FastDeploy","desc":"","summaryCn":"FastDeploy的核心价值在于它极大地简化了模型在异构硬件上的落地过程。","reasonCn":"想要使用FastDeploy，开发者可以参考其详尽的文档与安装指引，从基础的推理演示开始上手，迅速打通模型服务的闭环。它不仅仅是一个库，更是一套包含了性能调优、服务化部署和生态兼容的综合体系，特别是在处理MoE（混合专家模型）和多模态大模型时，它表现出的前缀缓存性能优化能够","stars":3719,"starGrowth":0,"createdAt":"2022-06-27T10:22:10Z","silentRounds":2,"language":"Python","topics":["serving","ernie","llm","inference","llm-serving","openai","vllm","ernie-45","ernie-45-vl"],"aiDims":["模型部署","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["模型部署","飞桨","多模态模型","推理加速"],"domainKey":"模型部署","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"serving","source":"github","weight":0.7},{"name":"ernie","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"llm-serving","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"vllm","source":"github","weight":0.7},{"name":"ernie-45","source":"github","weight":0.7},{"name":"ernie-45-vl","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/PaddlePaddle/FastDeploy","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ultralytics/yolov5","owner":"ultralytics","name":"yolov5","desc":"Ultralytics YOLOv5 in PyTorch for object detection, instance segmentation, classification, training, and export.","summaryCn":"视觉识别领域的王者，让机器一眼认出所有物体","reasonCn":"技术核心在于其单一阶段（One-Stage）的检测架构。不同于传统方法先找候选区域再分类，YOLOv5 直接在一次前向传递中就预测出物体的类别和位置。它在 PyTorch 框架下进行了极致优化，支持自动锚框计算、先进的数据增强策略以及多种模型尺寸（从 Nano 到 XLarge），用户可以根据设备","stars":58066,"starGrowth":0,"createdAt":"2020-05-18T03:45:11Z","silentRounds":1,"language":"Python","topics":["deep-learning"],"aiDims":["模型与训练","多模态Agent","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["物体检测","深度学习","PyTorch","计算机视觉","YOLO"],"domainKey":"物体检测","pushedAt":"2026-09-16T10:18:54Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ultralytics/yolov5","ts":"2026-09-16T10:18:54Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aden-hive/hive","owner":"aden-hive","name":"hive","desc":"Multi-Agent Harness for Production AI","summaryCn":"让多个AI Agent组队协作、自动完成复杂任务的开源框架","reasonCn":"核心亮点在于它的编排引擎。不同于单个Agent只能处理单一任务，Hive支持多个Agent之间的消息传递、任务委派和结果汇总。你可以定义复杂的工作流，比如先让一个Agent收集信息，再让另一个Agent分析，最后由第三个Agent输出报告。","stars":11065,"starGrowth":2,"createdAt":"2026-01-12T00:04:22Z","silentRounds":0,"language":"Python","topics":["automation","self-hosted"],"aiDims":["AI Agent"],"zone":"AI","funScore":0.5,"domainTags":["多Agent","AI编排","工作流框架"],"domainKey":"多Agent","pushedAt":"2026-09-14T00:10:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/aden-hive/hive","ts":"2026-09-14T00:10:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"blader/humanizer","owner":"blader","name":"humanizer","desc":"Agent skill that removes signs of AI-generated writing from text","summaryCn":"给AI生成的文字加点“人味”，像人写的不再像机器","reasonCn":"这是一个专门用来检测和移除AI生成文本特征的Python脚本。它通过重写句式、调整语气和增加自然语言的不规则性，让机器产出的内容看起来更像真人撰写。虽然它不生成文本，但通过反向处理，能让AI内容更难被检测出AI痕迹，适合需要降低查重率或通过AI检测的场景。","stars":51403,"starGrowth":64,"createdAt":"2026-01-18T05:30:15Z","silentRounds":0,"language":"Python","topics":["agent-skills","ai-writing","claude-code","codex","cursor","prompt-engineering","writing-tools"],"aiDims":["AI写作","非AI-实用","非AI-好玩"],"zone":"AI","funScore":0,"domainTags":["文本改写","AI检测","风格迁移","Python","Agent技能"],"domainKey":"文本改写","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"ai-writing","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"prompt-engineering","source":"github","weight":0.7},{"name":"writing-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/blader/humanizer","ts":"2026-09-07T22:38:09.240Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.23488222866369665,"copyOk":false},{"repo":"dogxii/iFace","owner":"dogxii","name":"iFace","desc":"Agent/前端/Golang/Java 面试题库 -  智能刷题工具，自定义 AI 接口，实现高效辅助学习","summaryCn":"基于AI的智能面试刷题助手，自定义接口还能一键生成代码，刷题效率翻倍","reasonCn":"这是一个面向前端、Agent及Golang/Java开发者的智能面试题库工具。它集成了大语言模型的问答能力，支持自定义AI接口，能够根据用户的输入生成面试题、解析代码并提供解题思路。项目支持多种题型，包括算法题、系统设计题等，旨在通过AI辅助让开发者更高效地准备面试。它采用了模块化设计，用户可以自由配置大模型参数，以适应不同的学习场景和难度需求。","stars":149,"starGrowth":0,"createdAt":"2026-03-08T16:56:04Z","silentRounds":1,"language":"TypeScript","topics":["interview"],"aiDims":["AI应用","工具调用","AI Agent"],"zone":"资源","funScore":0,"domainTags":["面试题库","代码练习","AI 辅助","刷题"],"domainKey":"面试题库","pushedAt":"2026-09-14T13:16:43Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/dogxii/iFace","ts":"2026-09-14T13:16:43Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"judahpaul16/gpt-home","owner":"judahpaul16","name":"gpt-home","desc":"ChatGPT at home! A better alternative to commercial smart home assistants, built on the Raspberry Pi using LiteLLM and LangGraph.","summaryCn":"把树莓派变成拥有大模型灵魂的智能家居管家，告别智障语音助手。","reasonCn":"技术实现上，它巧妙地组合了树莓派作为硬件载体，利用LiteLLM来统一管理各种大模型接口，并用LangGraph来构建复杂的智能体工作流。这样一来，大模型就不再只是陪聊的玩具，而是拥有了感知环境、调用工具和执行多步任务的能力，真正把大语言模型变成了家里的中枢大脑。","stars":649,"starGrowth":0,"createdAt":"2023-09-05T17:17:22Z","silentRounds":16,"language":"Python","topics":["home-automation"],"aiDims":["AI Agent","AI应用","非AI-实用"],"zone":"AI","funScore":0.6,"domainTags":["智能家居","树莓派","LangGraph","LLM"],"domainKey":"智能家居","pushedAt":"2026-09-13T23:14:52Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/judahpaul16/gpt-home","ts":"2026-09-13T23:14:52Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"FlagAI-Open/FlagAI","owner":"FlagAI-Open","name":"FlagAI","desc":"","summaryCn":"该工具包的核心诉求是“快速与通用”，它通过整合多模态模型的训练逻辑，让用","reasonCn":"在易用性上，FlagAI的表现尤为突出。它不仅提供了开箱即用的预训练模型API，还贴心地适配了多种常用下游任务。这种设计让开发者无需从零开始构建模型训练代码，通过调用预置的API接口即可快速完成模型参数的调整。这对于希望将大模型能力快速注入自身业务产品的开发者而言，是一套","stars":3869,"starGrowth":0,"createdAt":"2022-05-16T06:43:12Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["大语言模型","微调","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["大模型训练","微调","多模态","模型部署"],"domainKey":"微调","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/FlagAI-Open/FlagAI","ts":"2026-09-18T20:41:13.636Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"frappe/books","owner":"frappe","name":"books","desc":"Free Accounting Software","summaryCn":"开源免费的企业记账软件，从小微商户到专业会计都能用","reasonCn":"基于Frappe框架开发的桌面端会计软件，支持多币种、多公司账目管理，内置发票、采购、库存、税务报表等完整财务模块。采用TypeScript编写，支持本地部署，数据完全自控，适合中小型企业日常财务核算和报表生成，无需订阅付费服务。","stars":4979,"starGrowth":3,"createdAt":"2018-01-08T12:29:11Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["会计软件","财务管理","桌面应用","多币种"],"domainKey":"会计软件","pushedAt":"2026-09-20T16:14:54Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/frappe/books","ts":"2026-09-20T16:14:54Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"thinkwee/AwesomeOPD","owner":"thinkwee","name":"AwesomeOPD","desc":"Awesome List for On-Policy Distillation","summaryCn":"政策内蒸馏精选资料库，大模型训练者的宝藏地图","reasonCn":"这是一个专注于On-Policy Distillation（政策内蒸馏）的精选资源列表，收录了该领域的重要论文、开源项目和实现代码。对于从事大语言模型训练和蒸馏的AI研究者来说，这个项目整理了最新的研究进展和实践方案，帮助你快速掌握政策内蒸馏的核心技术和前沿动态。","stars":868,"starGrowth":0,"createdAt":"2026-04-27T22:30:47Z","silentRounds":3,"language":"","topics":["llm-model"],"aiDims":["模型与训练","微调","大语言模型"],"zone":"资源","funScore":0,"domainTags":["知识蒸馏","论文合集","LLM训练","学术资源"],"domainKey":"知识蒸馏","pushedAt":"2026-09-18T17:12:52Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/thinkwee/AwesomeOPD","ts":"2026-09-18T17:12:52Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"asukaminato0721/telegram-summary-bot","owner":"asukaminato0721","name":"telegram-summary-bot","desc":"Summarize group chat with AI, LLM && query group chat, FREE to deploy your own, support img, link meta info, reply to, auto fold result, 支持中文检索.","summaryCn":"群聊神器用AI秒变摘要，再也不用手滑漏看消息","reasonCn":"这款机器人接入大语言模型，专门针对Telegram群组聊天内容进行自动摘要和检索。支持图片、链接元数据提取，还能 replies 自动折叠结果。最大的亮点是支持中文检索，解决了多语言群聊的信息过载痛点，部署简单且免费，是社恐和忙碌打工人的福音。","stars":200,"starGrowth":0,"createdAt":"2024-11-17T22:16:03Z","silentRounds":24,"language":"TypeScript","topics":["llm-model"],"aiDims":["AI应用","AI Agent"],"zone":"AI","funScore":0,"domainTags":["Telegram","摘要","LLM"],"domainKey":"Telegram","pushedAt":"2026-09-18T05:03:57Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/asukaminato0721/telegram-summary-bot","ts":"2026-09-18T05:03:57Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jonas1ara/73","owner":"jonas1ara","name":"73","desc":"The 73 most useful LeetCode problems implemented in C# and F#","summaryCn":"准备算法面试或者想提升C#编程水平？这73道题就是你的通关秘籍。","reasonCn":"这个项目的设计初衷是“少而精”。与其刷几百道题，不如把这73道题彻底吃透，理解背后的算法原理和编程技巧。代码风格规范，符合C#最佳实践，同时展示了F#这种函数式语言在算法问题上的独特魅力。对于面试党来说，这里不仅有代码，还有对题目的深度解析，帮助你应对面试中的追问。","stars":147,"starGrowth":0,"createdAt":"2022-12-06T19:35:23Z","silentRounds":16,"language":"C#","topics":["algorithms"],"aiDims":["代码助手","学习资源"],"zone":"资源","funScore":0,"domainTags":["LeetCode","算法面试","C#实现"],"domainKey":"LeetCode","pushedAt":"2026-09-18T06:08:08Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"algorithms","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/jonas1ara/73","ts":"2026-09-18T06:08:08Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CopilotC-Nvim/CopilotChat.nvim","owner":"CopilotC-Nvim","name":"CopilotChat.nvim","desc":"","summaryCn":"对于追求极致编辑器的 Neovim 忠实用户，这款插件相当于给原本的纯文本空间","reasonCn":"除了依赖 GitHub 账号，项目还支持 Ollama 和 Mistral.ai 等自定义提供商，这对于关注数据隐私的公司或个人开发者至关重要。这意味着你可以在完全断网或内网环境中，加载一个本地微调的 Llama 或 CodeLlama 模型，通过该插件实现代码解释和重构功能，真正做到“代码不出门”。","stars":3652,"starGrowth":0,"createdAt":"2023-12-12T00:59:30Z","silentRounds":2,"language":"Lua","topics":["copilot","copilot-chat","github","github-copilot","neovim-plugin","hacktoberfest"],"aiDims":["代码助手","AI Agent"],"zone":"AI","funScore":0.2,"domainTags":["Neovim","代码生成","Copilot"],"domainKey":"Neovim","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"copilot","source":"github","weight":0.7},{"name":"copilot-chat","source":"github","weight":0.7},{"name":"github","source":"github","weight":0.7},{"name":"github-copilot","source":"github","weight":0.7},{"name":"neovim-plugin","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Lua","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/CopilotC-Nvim/CopilotChat.nvim","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NVlabs/SegFormer","owner":"NVlabs","name":"SegFormer","desc":"","summaryCn":"该项目解决的是图像中像素级语义分割的问题，即让计算机精准识别出图片里的","reasonCn":"该项目的核心代码基于 PyTorch 实现，并且与 MMSegmentation v0.13.0 版本深度集成。开发者可以直接使用预训练模型进行推理，也可以参考仓库中的配置文件对 B0 到 B5 的任意变体进行微调。对于初次接触该项目的用户，官方提供了详细的在线 Demo 和 Intro Video，直观展示了模型在不","stars":3661,"starGrowth":0,"createdAt":"2021-06-11T17:22:07Z","silentRounds":1,"language":"Python","topics":["semantic-segmentation","transformer","ade20k","cityscapes"],"aiDims":["AI应用","文档理解"],"zone":"AI","funScore":0.2,"domainTags":["语义分割","Transformer","计算机视觉","图像处理"],"domainKey":"语义分割","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"semantic-segmentation","source":"github","weight":0.7},{"name":"transformer","source":"github","weight":0.7},{"name":"ade20k","source":"github","weight":0.7},{"name":"cityscapes","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/NVlabs/SegFormer","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"brightmart/albert_zh","owner":"brightmart","name":"albert_zh","desc":"","summaryCn":"ALBERT中文项目是针对大规模中文语料预训练的优秀范例，它通过对模型结构的","reasonCn":"该项目不仅提供了适用于TensorFlow、PyTorch和Keras等主流框架的预训练模型，还配套了CLUE工具包，极大简化了开发者的工作流。你只需编写极简的代码，便可在短短几分钟内定制一个专门用于零样本学习的API。对于那些需要快速验证业务模型效果或者需要同时在多个基准模型间进行性能对","stars":3983,"starGrowth":0,"createdAt":"2019-09-26T16:45:30Z","silentRounds":5,"language":"Python","topics":["albert","bert","roberta","xlnet","tensorflow","pytorch","pre-trained-model","chinese-corpus","pre-trained"],"aiDims":["大语言模型","模型与训练","文档理解"],"zone":"AI","funScore":0.2,"domainTags":["中文预训练模型","ALBERT","自然语言处理","深度学习"],"domainKey":"中文预训练模型","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"albert","source":"github","weight":0.7},{"name":"bert","source":"github","weight":0.7},{"name":"roberta","source":"github","weight":0.7},{"name":"xlnet","source":"github","weight":0.7},{"name":"tensorflow","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"pre-trained-model","source":"github","weight":0.7},{"name":"chinese-corpus","source":"github","weight":0.7},{"name":"pre-trained","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/brightmart/albert_zh","ts":"2026-09-18T20:05:54.376Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sourcebox/mi-plaits-dsp-rs","owner":"sourcebox","name":"mi-plaits-dsp-rs","desc":"Rust port of Mutable Instruments Plaits module DSP code.","summaryCn":"用Rust实现的可编程合成模块，玩转黑科技音色合成。","reasonCn":"mi-plaits-dsp-rs 将 Mutable Instruments 的 Plaits 合成模块 DSP 代码完整移植到 Rust，保留原始算法的音色模型与参数控制，利用 Rust 的零成本抽象与安全内存管理，实现高性能、跨平台的实时音频合成库，适合电子乐器、模块化合成器以及音频插件开发者使用。","stars":63,"starGrowth":0,"createdAt":"2022-05-29T17:02:19Z","silentRounds":13,"language":"Rust","topics":["synthesizer"],"aiDims":["非AI-好玩","创意工具"],"zone":"工具","funScore":0,"domainTags":["音频合成","Rust","DSP"],"domainKey":"音频合成","pushedAt":"2026-09-21T07:49:53Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"synthesizer","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/sourcebox/mi-plaits-dsp-rs","ts":"2026-09-21T07:49:53Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mshumer/gpt-llm-trainer","owner":"mshumer","name":"gpt-llm-trainer","desc":"","summaryCn":"输入任务描述即可自动生成数据并微调专属大模型的 AI 工具。","reasonCn":"这是一个致力于消除大模型训练复杂性的实验性流水线项目，旨在让普通人也能轻松完成特定任务模型的训练。用户只需要简单输入对某个任务的自然语言描述，系统就能自动从零生成对应的数据集、调整为正确的格式，并直接为你微调出高性能的 LLaMA 2 或 GPT-3.5 模型。","stars":4176,"starGrowth":0,"createdAt":"2023-08-09T15:34:24Z","silentRounds":5,"language":"Jupyter Notebook","topics":[],"aiDims":["微调","大语言模型","AI应用"],"zone":"AI","funScore":0.44,"domainTags":["模型微调","数据生成","LLaMA","GPT"],"domainKey":"模型微调","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/mshumer/gpt-llm-trainer","ts":"2026-09-18T00:01:25.926Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"redis/redis","owner":"redis","name":"redis","desc":"For developers, who are building real-time data-driven applications, Redis is the preferred, fastest, and most feature-rich cache, data structure server, and document and vector query engine.","summaryCn":"内存数据库界的瑞士军刀，AI应用必备的高速缓存和向量检索","reasonCn":"Redis是业界最流行的内存数据存储引擎，读写速度可达百万级QPS。它不仅是缓存工具，还支持字符串、列表、集合等多种数据结构，以及发布订阅、Lua脚本、事务等高级功能。2.8版本后新增的Redis Stack模块支持向量检索和JSON存储，直接成为AI应用的热门基础设施。","stars":76449,"starGrowth":4,"createdAt":"2009-03-21T22:32:25Z","silentRounds":0,"language":"C","topics":["database"],"aiDims":["AI基础设施","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["内存数据库","缓存","数据结构","Redis","性能优化"],"domainKey":"内存数据库","pushedAt":"2026-09-23T11:56:29Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/redis/redis","ts":"2026-09-23T11:56:29Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Kosinkadink/ComfyUI-AnimateDiff-Evolved","owner":"Kosinkadink","name":"ComfyUI-AnimateDiff-Evolved","desc":"","summaryCn":"AnimateDiff-Evolved的核心价值在于它将AI动画生成的控制权提升到了一个新的","reasonCn":"为了处理复杂的合成任务，项目还整合了ControlNet的支持，允许用户通过结构图引导AI的动作表现，这意味着你可以精确控制视频里的角色或物体位移。结合VideoHelperSuite节点包，用户可以像在非编软件里一样对视频片段进行拆分、堆叠和循环处理，将原本抽象的AI采样逻辑转变为具体的","stars":3542,"starGrowth":0,"createdAt":"2023-08-26T16:24:00Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["视频生成","动画制作","图像采样"],"domainKey":"视频生成","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/Kosinkadink/ComfyUI-AnimateDiff-Evolved","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"continue-revolution/sd-webui-animatediff","owner":"continue-revolution","name":"sd-webui-animatediff","desc":"","summaryCn":"让 SD 画图一键变视频，操作跟画图一样。","reasonCn":"这是给 Stable Diffusion WebUI 加的视频生成插件，能把静态图片变成流畅动画。它运行时直接往模型里插入运动模块，不用重新加载权重文件，配合 ControlNet 能精准控制画面变化。激活插件后，生成 GIF 的操作和平时画图完全一样，门槛很低。","stars":3426,"starGrowth":0,"createdAt":"2023-07-18T02:59:45Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["推理引擎","多模态Agent"],"zone":"AI","funScore":0.62,"domainTags":["Stable Diffusion","视频生成","AnimateDiff","AI视频"],"domainKey":"视频生成","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/continue-revolution/sd-webui-animatediff","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"diasurgical/DevilutionX","owner":"diasurgical","name":"DevilutionX","desc":"Diablo build for modern operating systems","summaryCn":"经典暗黑破坏神重制版，现代电脑也能流畅玩","reasonCn":"把1996年经典暗黑破坏神原样复刻到现代系统，保留原汁原味但去掉老毛病。用C++重写核心代码，原生支持Windows、macOS、Linux，还能跑在现代手机和掌机上。支持高清纹理和宽屏分辨率，键鼠手柄都兼容，9000多star说明需求旺盛。","stars":9761,"starGrowth":1,"createdAt":"2018-08-02T17:19:01Z","silentRounds":2,"language":"C++","topics":["game"],"aiDims":["游戏"],"zone":"创意","funScore":1,"domainTags":["暗黑破坏神","复古游戏","游戏复刻","C++"],"domainKey":"暗黑破坏神","pushedAt":"2026-09-20T17:58:47Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/diasurgical/DevilutionX","ts":"2026-09-20T17:58:47Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"campusx-official/100-days-of-machine-learning","owner":"campusx-official","name":"100-days-of-machine-learning","desc":"","summaryCn":"系统化的 100 天机器学习笔记本合集，帮助快速入门实战。","reasonCn":"仓库结构清晰，根目录下按天数划分文件夹，每个文件夹包含 .ipynb 笔记本和必要的数据文件。使用者只需在本地或云端（如 Google Colab）打开对应 Notebook，即可运行代码并观察输出，无需额外配置复杂的环境。项目提供了简要的环境说明，推荐使用 Anaconda 配置 Python 环境，并列","stars":3169,"starGrowth":1,"createdAt":"2021-03-30T14:22:03Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.5599999999999999,"domainTags":["机器学习","Jupyter Notebook","教程"],"domainKey":"机器学习","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.65,"source":"tier","url":"https://github.com/campusx-official/100-days-of-machine-learning","ts":"2026-09-19T22:19:04.205Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"k2-fsa/OmniVoice","owner":"k2-fsa","name":"OmniVoice","desc":"High-Quality Voice Cloning TTS for 600+ Languages","summaryCn":"支持六百多种语言的高质量声音克隆与语音合成开源黑科技","reasonCn":"技术层面上，它采用了最先进的神经网络架构，摆脱了传统声音合成需要海量录音室数据的束缚。通过优秀的零样本声音克隆技术，它能够精准提取说话人的音色、音调和呼吸感特征。同时，其庞大的语种支持得益于先进的多语言共享表征空间，使得低资源语言也能获得极高清晰度和自然度的合","stars":13738,"starGrowth":10,"createdAt":"2026-03-31T12:00:45Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","创意工具","多模态Agent"],"zone":"AI","funScore":1,"domainTags":["语音合成","声音克隆","TTS","深度学习","音频AI"],"domainKey":"语音合成","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/k2-fsa/OmniVoice","ts":"2026-09-12T22:00:03.438Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03862046877274712,"copyOk":false},{"repo":"vercel-labs/skills","owner":"vercel-labs","name":"skills","desc":"The open agent skills tool - npx skills","summaryCn":"AI技能插件商店，一行命令给你的Agent装上搜索代码执行等各种超能力","reasonCn":"Vercel Labs推出的开源Agent技能工具，通过npx命令快速安装各类AI Agent技能插件。支持搜索、代码执行、文件操作等常用能力，技能可复用可分享。技术栈TypeScript，CLI和API双接口，适合开发者快速扩展Agent能力边界。","stars":32274,"starGrowth":14,"createdAt":"2026-01-14T21:31:18Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":1,"domainTags":["MCP","AI技能","Agent","代码执行"],"domainKey":"Agent","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/vercel-labs/skills","ts":"2026-09-12T22:00:03.438Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.05196162855549359,"copyOk":true},{"repo":"sansan0/TrendRadar","owner":"sansan0","name":"TrendRadar","desc":"⭐AI-driven public opinion & trend monitor with multi-platform aggregation, RSS, and smart alerts.🎯 告别信息过载，你的 AI 舆情监控助手与热点筛选工具！聚合多平台热点 +  RSS 订阅，支持关键词精准筛选。AI 智能筛选新闻 + AI 翻译 +  AI 分析简报直推手机，也支持接入 MCP 架构，赋能 AI 自然语言对话分析、情感洞察与趋势预测等。支持 Docker ，数据本地/云端自持。集成微信/飞书/钉钉/Telegram/邮件/ntfy/bark/slack 等渠道智能推送。","summaryCn":"AI帮你盯着全网热点，自动过滤噪音，精彩新闻直接推送到手。","reasonCn":"TrendRadar是一个基于大语言模型的舆情监控与趋势发现工具。它聚合多平台热点数据和RSS源，利用AI进行智能筛选、翻译和分析，生成简报直推用户。支持关键词精准过滤，接入MCP架构，有效解决信息过载问题，是高效获取资讯的实用利器。","stars":62480,"starGrowth":2,"createdAt":"2025-04-28T11:42:16Z","silentRounds":0,"language":"Python","topics":["llm","ai"],"aiDims":["AI应用","AI Agent","效率工具"],"zone":"AI","funScore":0.7,"domainTags":["舆情监控","新闻聚合","AI分析"],"domainKey":"舆情监控","pushedAt":"2026-09-13T06:38:25Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/sansan0/TrendRadar","ts":"2026-09-13T06:38:25Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nguyenphutrong/quotio","owner":"nguyenphutrong","name":"quotio","desc":"Stop juggling AI accounts. Quotio is a beautiful native macOS menu bar app that unifies your Claude, Gemini, OpenAI, Qwen, and Antigravity subscriptions – with real-time quota tracking and smart auto-failover for AI coding tools like Claude Code, OpenCode, and Droid.","summaryCn":"统一管理多个AI账户的macOS菜单栏应用","reasonCn":"Quotio是一个专为macOS设计的应用程序，旨在帮助用户统一管理和监控他们的Claude、Gemini、OpenAI、Qwen、Antigravity等订阅服务。通过Quotio，用户可以在一个统一的界面中查看和管理所有相关的订阅，包括调整配额、跟踪使用情况等。这款应用特别适合需要同时管理多个AI账户的用户，提高了用户的工作效率。","stars":4867,"starGrowth":0,"createdAt":"2025-12-24T14:42:00Z","silentRounds":1,"language":"Swift","topics":["developer-tools"],"aiDims":["开发者工具"],"zone":"工具","funScore":0,"domainTags":["订阅管理","账户聚合","配额追踪","菜单栏应用","macOS"],"domainKey":"订阅管理","pushedAt":"2026-09-17T17:14:23Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/nguyenphutrong/quotio","ts":"2026-09-17T17:14:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"magnitudedev/magnitude","owner":"magnitudedev","name":"magnitude","desc":"Open source inference server that runs the best local models for your hardware, plugged into the agent you already use. Works with Pi, OpenCode, Hermes, OpenClaw, Codex, Claude Code, Oh My Pi, and Cline.","summaryCn":"一个能让你在本地跑大模型的神器，还能和现有工具无缝连接","reasonCn":"它的核心厉害之处在于“轻量”和“生态”。它基于高效的推理引擎，能针对不同硬件（CPU、集成显卡甚至树莓派）做深度优化，支持多种模型格式。更重要的是，它把复杂的推理过程封装成了简单的HTTP API或本地调用方式，让开发者能像调用函数一样使用大模型。","stars":4847,"starGrowth":28,"createdAt":"2026-06-12T09:06:26Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["本地推理","模型服务","HTTP API"],"domainKey":"本地推理","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/magnitudedev/magnitude","ts":"2026-09-07T02:44:20.540Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.12159636888797193,"copyOk":true},{"repo":"pinecone-io/examples","owner":"pinecone-io","name":"examples","desc":"Jupyter Notebooks to help you get hands-on with Pinecone vector databases","summaryCn":"基于Pinecone的向量数据库入门指南，AI应用快速上手","reasonCn":"这个项目是Pinecone官方收集的Jupyter Notebook示例集合，专注于向量数据库的实战应用。它涵盖了文本嵌入、语义搜索、相似度检索等核心功能，提供可直接运行的代码。通过丰富的教程案例，帮助开发者快速掌握如何用向量数据库为AI应用添加记忆和搜索能力，适合从入门到进阶的学习者。","stars":3043,"starGrowth":0,"createdAt":"2021-03-09T19:42:10Z","silentRounds":26,"language":"Jupyter Notebook","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI应用"],"zone":"资源","funScore":0,"domainTags":["向量数据库","embedding","语义搜索","RAG教程","Jupyter"],"domainKey":"向量数据库","pushedAt":"2026-09-19T01:49:05Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/pinecone-io/examples","ts":"2026-09-19T01:49:05Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"opencode-ai/opencode","owner":"opencode-ai","name":"opencode","desc":"A powerful AI coding agent. Built for the terminal.","summaryCn":"像极客特工一样在终端里指挥 AI 写代码","reasonCn":"这是一个基于 Go 语言构建的 AI 编码智能体，专为终端用户设计。它不仅能自动完成代码编写、重构和调试任务，还能像人类程序员一样进行上下文理解和多步推理。通过自然语言指令即可驱动，支持集成多种模型，让开发者从繁琐的打字中解放出来。","stars":13762,"starGrowth":1,"createdAt":"2025-03-16T23:38:55Z","silentRounds":1,"language":"Go","topics":["ai","claude","code","llm","openai"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["AI 编程","Agent","终端工具","代码生成"],"domainKey":"Agent","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"code","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/opencode-ai/opencode","ts":"2026-09-12T22:00:03.438Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DataTalksClub/ai-dev-tools-zoomcamp","owner":"DataTalksClub","name":"ai-dev-tools-zoomcamp","desc":"A free, hands-on course on using AI developer tools to build, test, deploy, extend, and audit software without losing engineering discipline. The 2026 cohort starts August 31.","summaryCn":"教你如何规范使用AI开发工具来写代码、测试和部署的免费实战训练营。","reasonCn":"内容设计上，它摒弃了枯燥的纯理论说教，全部采用真实世界的动手项目驱动。课程涵盖了主流AI编码助手（如Cursor、GitHub Copilot等）的高级用法、上下文管理技巧、测试用例生成策略以及代码安全审计规范。它不仅教你怎么让AI写代码，更教你如何审查AI写出的Bug，避免被AI带进沟里。","stars":1639,"starGrowth":2,"createdAt":"2025-05-27T16:43:27Z","silentRounds":0,"language":"Python","topics":["course"],"aiDims":["学习资源","教程","开发者工具"],"zone":"资源","funScore":0.5,"domainTags":["AI课程","实战训练营","AI编程"],"domainKey":"AI课程","pushedAt":"2026-09-15T12:22:20Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"course","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/DataTalksClub/ai-dev-tools-zoomcamp","ts":"2026-09-15T12:22:20Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MohamedRejeb/Calf","owner":"MohamedRejeb","name":"Calf","desc":"Calf is a library that allows you to easily create adaptive UIs and access platform specific APIs with Compose Multiplatform (Adaptive UI, File Picker, WebView, Permissions...).","summaryCn":"Compose跨平台神器，一行代码搞定文件选择和权限申请","reasonCn":"它是JetBrains Compose Multiplatform的UI辅助库，用Kotlin编写。核心价值在于提供一套适配iOS、Android、Windows、macOS和Linux的统一UI组件，开发者无需重复造轮子就能轻松调用系统级功能，大幅降低多端开发成本。","stars":1712,"starGrowth":0,"createdAt":"2023-08-05T07:56:40Z","silentRounds":2,"language":"Kotlin","topics":["desktop"],"aiDims":["开发工具","跨平台"],"zone":"工具","funScore":0,"domainTags":["Compose Multiplatform","跨平台UI","文件选择器","权限管理","桌面开发"],"domainKey":"Compose Multiplatform","pushedAt":"2026-09-20T04:57:14Z","tags":[{"name":"开发工具","source":"llm","weight":1},{"name":"跨平台","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/MohamedRejeb/Calf","ts":"2026-09-20T04:57:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aws-samples/generative-ai-use-cases","owner":"aws-samples","name":"generative-ai-use-cases","desc":"Application implementation with business use cases for safely utilizing generative AI in business operations","summaryCn":"很多公司老板想搞AI，但技术团队往往不知道怎么把ChatGPT这类东西用到公司业","reasonCn":"技术上，它集成了AWS的强大云端生态，如Amazon Bedrock和OpenSearch。它的架构高度模块化，采用TypeScript构建，能够轻松调用各类主流模型，并深度整合了RAG（检索增强生成）技术，让模型能准确理解企业内部的私有文档，彻底告别AI幻觉带来的业务风险。","stars":1388,"starGrowth":0,"createdAt":"2023-08-17T02:08:52Z","silentRounds":3,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI应用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["生成式AI","企业应用","RAG"],"domainKey":"RAG","pushedAt":"2026-09-13T22:00:38Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/aws-samples/generative-ai-use-cases","ts":"2026-09-13T22:00:38Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gotenberg/gotenberg","owner":"gotenberg","name":"gotenberg","desc":"A developer-friendly API for converting many document formats into PDF files, and more!","summaryCn":"把各种文档一键转成PDF的神级API，程序员必备","reasonCn":"Gotenberg 是一个开发者友好的文档转换 API，基于 Go 语言构建，支持将多种格式文件（如 HTML、Office 文档、图片等）高效转换为 PDF。它通过简单的 HTTP 请求即可处理复杂的格式转换，内置了 LibreOffice、Chromium 等强大组件，能够完美保留文档格式和排版。项目在 GitHub 上拥有 1.2 万星，以其稳定、快速和易用性，成为众多开发者在构建涉及文档处理的应用时的首选工具。","stars":13146,"starGrowth":0,"createdAt":"2018-03-23T14:05:59Z","silentRounds":1,"language":"Go","topics":["docker","api"],"aiDims":["文档理解","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["PDF转换","文档处理","API服务"],"domainKey":"PDF转换","pushedAt":"2026-09-18T07:27:47Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/gotenberg/gotenberg","ts":"2026-09-18T07:27:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lukasmasuch/best-of-ml-python","owner":"lukasmasuch","name":"best-of-ml-python","desc":"🏆 A ranked list of awesome machine learning Python libraries. Updated weekly.","summaryCn":"Python机器学习库的活地图，每周更新，从入门到专家一站收录。","reasonCn":"这是一个精心整理的机器学习Python库排行榜，按热度、质量、功能多维度排序。每周自动更新，涵盖NLP、计算机视觉、强化学习、数据科学等全领域。对想快速了解Python AI生态的开发者来说，是省去检索时间的宝藏清单。","stars":23816,"starGrowth":2,"createdAt":"2020-11-29T19:41:36Z","silentRounds":0,"language":"","topics":["ml"],"aiDims":["AI基础设施","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["机器学习","Python生态","技术导航","资源清单"],"domainKey":"机器学习","pushedAt":"2026-09-17T14:16:12Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/lukasmasuch/best-of-ml-python","ts":"2026-09-17T14:16:12Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"msitarzewski/agency-agents","owner":"msitarzewski","name":"agency-agents","desc":"A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.","summaryCn":"一个AI公司，什么角色都有，前端到社区运营全靠智能体搞定","reasonCn":"这是msitarzewski打造的多智能体协作框架，每个agent都有专属人设和专业能力，涵盖前端开发、Reddit社区运营、创意注入、现实校验等多个领域。基于Shell脚本实现，通过组合不同agent可完成复杂工作流，还能按需注入幽默感或做事实核查。","stars":154188,"starGrowth":54,"createdAt":"2025-10-13T12:12:29Z","silentRounds":0,"language":"Shell","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.9,"domainTags":["多智能体协作","AI Agent","Shell脚本","虚拟团队","Prompt工程"],"domainKey":"AI Agent","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/msitarzewski/agency-agents","ts":"2026-08-14T01:23:41.092Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.1956607985057203,"copyOk":false},{"repo":"reactive-resume/app","owner":"reactive-resume","name":"app","desc":"A one-of-a-kind resume builder that keeps your privacy in mind. Completely secure, customizable, portable, open-source and free forever. Try it out today!","summaryCn":"主打绝对隐私且免费的在线简历制作神器，让你的求职履历瞬间高级起来。","reasonCn":"技术架构上，它采用了TypeScript、React和TailwindCSS开发，不仅界面极其丝滑，还深度整合了现代Web技术。它最大的亮点在于极致的隐私保护——你的所有数据都保存在本地或你自己的服务器上，绝对不会被第三方平台拿去偷偷训练AI。","stars":43331,"starGrowth":12,"createdAt":"2020-03-25T14:08:22Z","silentRounds":0,"language":"TypeScript","topics":["ai"],"aiDims":["AI写作","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["Resume","Privacy","TypeScript","React","SaaS"],"domainKey":"Resume","pushedAt":"2026-09-11T10:54:59Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/reactive-resume/app","ts":"2026-09-11T10:54:59Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.036830813966905906,"copyOk":true},{"repo":"killop/anything_about_game","owner":"killop","name":"anything_about_game","desc":"A wonderful list of Game Development resources.","summaryCn":"游戏开发资源大合集，小白到大佬都能找到所需教程和工具链接","reasonCn":"这是一个精心整理的游戏开发资源清单，涵盖了从零基础入门到高级技巧的各类内容。不管你是想学Unity、Unreal还是Godot，这里都能找到对应的教程、插件和文档链接。虽然不直接涉及AI技术，但对想入坑游戏开发的开发者非常实用，资源分类清晰，节省了大量搜索时间。","stars":4128,"starGrowth":0,"createdAt":"2020-02-26T07:52:04Z","silentRounds":1,"language":"","topics":["game","ai-agent"],"aiDims":["游戏","非AI-好玩","学习资源"],"zone":"资源","funScore":0,"domainTags":["游戏开发","教程","资源索引","Unity","Godot"],"domainKey":"游戏开发","pushedAt":"2026-09-21T02:28:35Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7}],"aiScore":0.6,"source":"search","url":"https://github.com/killop/anything_about_game","ts":"2026-09-21T02:28:35Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bghira/SimpleTuner","owner":"bghira","name":"SimpleTuner","desc":"","summaryCn":"这是一个注重代码简洁与易懂性、适合学术交流的开源模型微调与训练项目。","reasonCn":"SimpleTuner 旨在让代码更易于理解，并作为共享的学术练习项目接受社区贡献。它默认不会向任何第三方发送数据，所有数据上报或推送功能都需要手动配置开启，保障了训练过程中的隐私与安全。如果你在学习或研究模型训练过程中遇到疑问，还可以通过 Discord 上的研究小组与开发者社区取得联系，获取进一步的交流与支持。","stars":2927,"starGrowth":1,"createdAt":"2023-05-28T18:01:02Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["模型与训练","微调"],"zone":"AI","funScore":0.2,"domainTags":["模型训练","微调","学术研究"],"domainKey":"模型训练","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/bghira/SimpleTuner","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mksglu/context-mode","owner":"mksglu","name":"context-mode","desc":"Context window optimization for AI coding agents. Sandboxes tool output (98% reduction), persists session memory, and enforces routing across 17 platforms via MCP + hooks.","summaryCn":"给AI代码助手装个“外挂”，让它干活更稳、记得住上下文","reasonCn":"Context Mode 是一个专为 AI 编程 Agent 设计的上下文管理优化工具。它通过 MCP 协议连接 17 个主流开发平台，将工具输出自动存入沙箱，并持久化会话记忆。这套机制不仅能大幅减少冗余信息，还能智能路由任务，让 AI Agent 在复杂开发场景下更专注、更高效地执行任务。","stars":23952,"starGrowth":11,"createdAt":"2026-02-23T05:56:28Z","silentRounds":0,"language":"TypeScript","topics":["claude","claude-code","claude-code-plugins","mcp","skills","codex","copilot","opencode","antigravity","kiro","openclaw","claude-code-hooks","claude-code-skill","codex-cli","cursor-plugin","mcp-server","mcp-tools","pi-agent","zed-extension","context-mode"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["AI编程助手","MCP协议","上下文管理","工具路由","沙箱"],"domainKey":"AI编程助手","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"claude-code-plugins","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"kiro","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"claude-code-hooks","source":"github","weight":0.7},{"name":"claude-code-skill","source":"github","weight":0.7},{"name":"codex-cli","source":"github","weight":0.7},{"name":"cursor-plugin","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"mcp-tools","source":"github","weight":0.7},{"name":"pi-agent","source":"github","weight":0.7},{"name":"zed-extension","source":"github","weight":0.7},{"name":"context-mode","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/mksglu/context-mode","ts":"2026-09-11T22:17:29.879Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.04125330661322645,"copyOk":false},{"repo":"01-ai/Yi","owner":"01-ai","name":"Yi","desc":"","summaryCn":"零一万物推出的高性能开源双语大模型，支持灵活部署与微调。","reasonCn":"Yi系列是由零一万物开发的下一代开源双语大模型，旨在提供卓越的语言理解与生成能力。该项目涵盖了基础模型与对话优化模型，能够满足开发者在科研或生产环境中的多种需求。通过提供详细的生态系统文档和模型部署方案，它降低了高性能AI模型的上手门槛，让开发者可以利用pip或docker等工具快速集成，是构建智能应用的核心引擎。","stars":7833,"starGrowth":0,"createdAt":"2023-11-03T16:08:37Z","silentRounds":5,"language":"Jupyter Notebook","topics":["large-language-models"],"aiDims":["大语言模型","模型与训练","模型部署"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","双语模型","模型部署","微调"],"domainKey":"大语言模型","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"large-language-models","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/01-ai/Yi","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"testtimescaling/testtimescaling.github.io","owner":"testtimescaling","name":"testtimescaling.github.io","desc":"\"what, how, where, and how well? a survey on test-time scaling in large language models\" repository","summaryCn":"AI大模型的临时开窍秘籍，全面梳理测试时如何多花算力让模型变聪明","reasonCn":"这是一篇关于大语言模型测试时缩放（test-time scaling）的系统性综述论文，探讨了如何在模型推理阶段通过动态分配计算资源来提升输出质量。文章涵盖思维链推理、自一致性采样、检索增强等多种技术路径，并对比了不同策略在准确性、延迟、成本之间的权衡。作为前沿方向的文献综述，适合研究人员和工程师快速把握领域全景，理解测试时缩放的核心思路与技术流派。","stars":112,"starGrowth":0,"createdAt":"2025-04-08T02:50:44Z","silentRounds":25,"language":"HTML","topics":["llm-model"],"aiDims":["大语言模型","学习资源","论文"],"zone":"资源","funScore":0,"domainTags":["大语言模型","测试时缩放","综述论文"],"domainKey":"大语言模型","pushedAt":"2026-09-21T02:31:22Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/testtimescaling/testtimescaling.github.io","ts":"2026-09-21T02:31:22Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/deepseek-harness","owner":"deepseek-ai","name":"deepseek-harness","desc":"DeepSeek Harness: Everything is a Plugin.","summaryCn":"给大模型插上各种插件，让它变强变聪明的瑞士军刀","reasonCn":"这是一个专为 DeepSeek 等大模型设计的强大插件化框架，核心思想是“万物皆插件”。它通过 Cordis 插件系统，让大模型能够像搭积木一样灵活调用外部工具，极大地扩展了模型的能力边界。该项目在 GitHub 上获得了超过 4.9k 的 Star，是构建复杂 AI 应用的基础设施级项目。","stars":233386,"starGrowth":113,"createdAt":"2026-08-13T11:56:32Z","silentRounds":0,"language":"TypeScript","topics":["cordis","dsh","dsh-plugin"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["插件系统","大模型","工具调用","DeepSeek","Agent框架"],"domainKey":"插件系统","pushedAt":"2026-08-17T12:01:58Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"cordis","source":"github","weight":0.7},{"name":"dsh","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"bigbro","url":"https://github.com/deepseek-ai/deepseek-harness","ts":"2026-08-17T12:01:58Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.4521787939293702,"copyOk":false},{"repo":"charmbracelet/crush","owner":"charmbracelet","name":"crush","desc":"Glamourous agentic coding for all 💘","summaryCn":"在终端里跟AI一起写代码，像有个程序员搭档随时陪你干活","reasonCn":"crush是charmbracelet推出的AI编程助手，在终端中以对话形式帮助开发者完成编码任务。它支持读取项目上下文、理解代码结构，用自然语言即可生成、修改和调试代码。设计美学出色，交互流畅，让终端编程体验焕然一新。","stars":28248,"starGrowth":3,"createdAt":"2025-05-21T12:14:57Z","silentRounds":0,"language":"Go","topics":["agentic-ai","ai","llms","ravishing"],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0.9,"domainTags":["AI编程","终端交互","代码助手","Rust"],"domainKey":"代码助手","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"ravishing","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/charmbracelet/crush","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chuspeeism/dashi-ppt-skill","owner":"chuspeeism","name":"dashi-ppt-skill","desc":"An AI-agent skill that generates browser-editable presentations from multiple visual themes, exportable to HTML, PDF, and PPTX.","summaryCn":"只要一句指令，AI 自动帮你生成可编辑的 PPT，支持多主题导出。","reasonCn":"该项目提供基于JavaScript的AIAgent能力，读取多张图片或文字素材，利用大模型生成符合视觉主题的演示稿，可直接在浏览器中编辑，支持导出为HTML、PDF与PPTX，兼容主流办公软件，省去排版和设计时间，并可通过API接入其他工作流，自动化报告生成。","stars":8620,"starGrowth":2,"createdAt":"2026-06-10T07:46:43Z","silentRounds":0,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":1,"domainTags":["AI Agent","演示文稿","PPT生成"],"domainKey":"AI Agent","pushedAt":"2026-09-12T14:13:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/chuspeeism/dashi-ppt-skill","ts":"2026-09-12T14:13:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"safepay/ha_google_weather","owner":"safepay","name":"ha_google_weather","desc":"A Home Assistant integration for the Google Weather API","summaryCn":"把谷歌天气搬进你的智能家居中控，看天气再也不用切App了","reasonCn":"这是一个Home Assistant的集成组件，能直接调用谷歌天气API获取精准的天气数据。通过它，你可以在智能家居控制面板上实时显示温度、湿度、空气质量等信息，甚至能联动控制家里的灯光和空调，让天气数据真正服务于生活。","stars":75,"starGrowth":0,"createdAt":"2025-11-25T11:59:52Z","silentRounds":13,"language":"Python","topics":["weather"],"aiDims":["非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["智能家居","天气API","Home Assistant"],"domainKey":"智能家居","pushedAt":"2026-09-16T00:57:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"weather","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/safepay/ha_google_weather","ts":"2026-09-16T00:57:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"roboticforce/sugar","owner":"roboticforce","name":"sugar","desc":"Persistent memory for AI coding agents. Local-first, cross-session context, global knowledge, and optional autonomous task execution.","summaryCn":"让AI写代码记住上下文，跨会话持续作业的本地记忆库","reasonCn":"Sugar为AI编码助手提供持久化记忆，支持本地优先、跨会话上下文保存，并可接入全局知识库或自主任务执行，让AI在多轮编程时保持状态，避免重复上下文注入，提升开发效率。它基于SQLite实现轻量持久，兼容LangChain接口，可在VS Code等环境使用。","stars":94,"starGrowth":0,"createdAt":"2025-08-21T18:26:58Z","silentRounds":13,"language":"Python","topics":["knowledge-management"],"aiDims":["AI Agent","RAG与知识","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI记忆","本地存储","上下文管理"],"domainKey":"AI记忆","pushedAt":"2026-09-16T18:09:03Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/roboticforce/sugar","ts":"2026-09-16T18:09:03Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fudan-generative-vision/champ","owner":"fudan-generative-vision","name":"champ","desc":"","summaryCn":"基于3D参数化引导的人体图像动画生成模型，实现动作可控且外观一致。","reasonCn":"Champ 是一款用于生成可控且一致性人体图像动画的 AI 模型。它利用 3D 参数化引导技术，让静态人物图像能够自然地跟随指定动作进行动画化，同时保持人物身份特征不改变。项目由复旦大学等机构研发，发布于 ECCV 2024，目前已开源训练代码及样本数据，适合研究者探索人体动作合成与虚拟形象制作。","stars":4261,"starGrowth":0,"createdAt":"2024-03-17T04:49:43Z","silentRounds":4,"language":"Python","topics":["human-animation","video-generation","image-animatioln"],"aiDims":["AI应用","模型与训练","非AI-实用"],"zone":"AI","funScore":0.2,"domainTags":["人体动画","3D人体模型","图像生成","ECCV论文"],"domainKey":"图像生成","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"human-animation","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"image-animatioln","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/fudan-generative-vision/champ","ts":"2026-09-18T00:01:25.926Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"1rgs/jsonformer","owner":"1rgs","name":"jsonformer","desc":"","summaryCn":"该项目专注于解决大语言模型生成规范结构化 JSON 数据的难题。","reasonCn":"让大语言模型稳定输出符合特定模式的结构化 JSON 数据一直是个巨大挑战。当前主流方法极易出错，而该项目创新性地通过仅生成内容令牌并填充固定令牌的方式，确保了生成结果在语法上的绝对正确，为开发者提供了一种更加可靠的结构化文本生成途径。","stars":4939,"starGrowth":0,"createdAt":"2023-04-29T23:25:54Z","silentRounds":5,"language":"Jupyter Notebook","topics":[],"aiDims":["大语言模型","推理引擎","工具调用"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","结构化数据","JSON格式","提示工程"],"domainKey":"大语言模型","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/1rgs/jsonformer","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"localsend/localsend","owner":"localsend","name":"localsend","desc":"An open-source cross-platform alternative to AirDrop","summaryCn":"一款开源免费的跨平台文件传输神器，不用网也能像AirDrop一样互传。","reasonCn":"LocalSend 是一款开源的跨平台局域网文件传输工具，旨在替代 AirDrop 等方案。它支持 Windows、macOS、Linux 和 iOS/Android，无需配置服务器，利用本地网络即可实现极速、安全的点对点传输。其核心优势在于零配置、无需互联网、支持大文件、隐私保护以及完全开源。","stars":92466,"starGrowth":16,"createdAt":"2022-12-16T00:46:07Z","silentRounds":0,"language":"Dart","topics":["flutter","file-sharing","dart","flutter-apps"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["文件传输","局域网","AirDrop替代","跨平台","隐私保护"],"domainKey":"文件传输","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"flutter","source":"github","weight":0.7},{"name":"file-sharing","source":"github","weight":0.7},{"name":"dart","source":"github","weight":0.7},{"name":"flutter-apps","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/localsend/localsend","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.048519109726818506,"copyOk":false},{"repo":"jordan-gibbs/hyperresearch","owner":"jordan-gibbs","name":"hyperresearch","desc":"Agent-driven research knowledge base. Agents collect, search, and synthesize web research into a persistent, searchable wiki.","summaryCn":"让AI小助手自动搜网、摘要、写入可搜索维基的研究利器","reasonCn":"该项目采用基于大模型的自治代理，自动爬取互联网公开资源，利用摘要与检索技术把信息结构化后存入持久化的 Markdown 维基。用户可以通过自然语言查询，系统会调用内部代理进行多轮检索、过滤与合成，返回可追溯的来源和编辑建议，实现类似科研助理的全流程知识管理。","stars":3498,"starGrowth":5,"createdAt":"2026-04-09T20:46:46Z","silentRounds":0,"language":"Python","topics":["agents","agentskills","claude-code","deep-research","deep-research-agent"],"aiDims":["AI Agent","RAG与知识","AI搜索"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","知识管理","网页爬取","向量检索"],"domainKey":"AI Agent","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"agentskills","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"deep-research","source":"github","weight":0.7},{"name":"deep-research-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/jordan-gibbs/hyperresearch","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.023145797598627785,"copyOk":false},{"repo":"Alishahryar1/free-claude-code","owner":"Alishahryar1","name":"free-claude-code","desc":"Use Claude Code, Codex, Pi, and OpenCode and more for free (1.3B+ free tokens) from your terminal, app, IDE, or phone like OpenClaw (voice supported + ToS friendly)","summaryCn":"在终端或IDE里免费调用Claude等大模型写代码，语音也能玩。","reasonCn":"该项目提供跨平台的CLI与GUI，能够直接在终端、IDE甚至手机上调用Claude Code、Codex、Pi、OpenCode等多家大模型，赠送130亿免费token，支持语音输入并遵循服务条款友好，兼容OpenClaw等前端，极大降低AI编码门槛，适合开发者随时获得代码补全、重构与单元测试建议。","stars":55687,"starGrowth":23,"createdAt":"2026-01-28T19:02:30Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["代码助手","AI应用","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["AI编程助手","大模型聚合","代码生成"],"domainKey":"AI编程助手","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/Alishahryar1/free-claude-code","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.08428688203710022,"copyOk":false},{"repo":"ZengLiangYi/ChatCrystal","owner":"ZengLiangYi","name":"ChatCrystal","desc":"Local-first AI PKM for coding conversations: import Claude Code/Cursor/Codex, distill notes, semantic search, tag graph, MCP memory.","summaryCn":"很多程序员都有过这种苦恼：刚让AI帮你解决了复杂的bug，过了几天代码逻辑又","reasonCn":"技术上，它采取了“本地优先”的设计原则。核心亮点在于它的语义搜索与知识图谱系统，利用Embedding向量化技术，系统能自动理解你代码对话中的上下文逻辑，而不仅仅是简单的关键词搜索。通过MCP（Model Context Protocol）协议，它能够与外部AI模型深度联动，自动打标签、整理知识脉","stars":58,"starGrowth":0,"createdAt":"2026-03-27T04:02:57Z","silentRounds":15,"language":"TypeScript","topics":["knowledge-management"],"aiDims":["AI Agent","知识图谱","RAG与知识"],"zone":"AI","funScore":0.7,"domainTags":["AI知识管理","语义搜索","代码对话记录"],"domainKey":"AI知识管理","pushedAt":"2026-09-14T01:19:43Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/ZengLiangYi/ChatCrystal","ts":"2026-09-14T01:19:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"honojs/hono","owner":"honojs","name":"hono","desc":"Web framework built on Web Standards","summaryCn":"专为AI应用打造的极速框架，几行代码就能出API","reasonCn":"Hono是一个基于Web标准的轻量级Web框架，支持Cloudflare Workers、Deno、Bun等多种运行时。它内置类型安全的路由中间件系统，可以与Zod等校验库无缝集成。对于构建AI应用来说，它可以快速搭建LLM推理API、Agent工具端点，性能比Express强3倍以上，而且对边缘计算环境友好。","stars":32302,"starGrowth":0,"createdAt":"2021-12-14T20:05:30Z","silentRounds":1,"language":"TypeScript","topics":["web-framework"],"aiDims":["开发者工具","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Web框架","边缘计算","TypeScript"],"domainKey":"Web框架","pushedAt":"2026-09-22T06:02:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/honojs/hono","ts":"2026-09-22T06:02:51Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"awslabs/aidlc-workflows","owner":"awslabs","name":"aidlc-workflows","desc":"AI-Driven Life Cycle (AI-DLC) adaptive workflow steering rules for AI coding agents","summaryCn":"AWS推出的AI编码助手智能调度中枢，让编程机器人按需切换工作模式","reasonCn":"AWS Labs开源的AI驱动生命周期自适应工作流引擎，专为AI coding agents设计。内置多种steering规则，根据任务类型自动切换工作流模式，支持代码生成、重构、测试、文档等全流程。TypeScript实现，可轻松集成到现有Agent框架中。","stars":4764,"starGrowth":3,"createdAt":"2025-11-13T22:05:08Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","工作流引擎","代码生成","AWS","TypeScript"],"domainKey":"AI Agent","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/awslabs/aidlc-workflows","ts":"2026-09-11T22:17:29.879Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"agamrossen/VolAnti","owner":"agamrossen","name":"VolAnti","desc":"Open-source acoustic drone detection. It hears the propellers, not the radio, so it works against fibre-optic FPV aircraft that emit no signal at all.","summaryCn":"反制FPV无人机神器，不用无线电也能听声辨位。","reasonCn":"它的核心技术亮点在于“听”而不是“看”。绝大多数反无人机设备靠拦截无线电信号，但光纤FPV无人机完全没有任何无线电发射，VolAnti 通过高灵敏度麦克风阵列拾取空气传播的振动声波，再通过信号处理算法进行特征提取，从而精准定位声源。","stars":383,"starGrowth":0,"createdAt":"2026-09-05T15:51:21Z","silentRounds":1,"language":"C","topics":["3d-printing"],"aiDims":["非AI-实用","安全工具"],"zone":"工具","funScore":0.6,"domainTags":["声学探测","反无人机","FPV无人机","3D打印"],"domainKey":"声学探测","pushedAt":"2026-09-20T19:44:28Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/agamrossen/VolAnti","ts":"2026-09-20T19:44:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ai-glimpse/toyllm","owner":"ai-glimpse","name":"toyllm","desc":"ToyLLM: Learning LLM from Scratch","summaryCn":"从零手写迷你版大语言模型，深度学习最佳实践教程","reasonCn":"这是一个教学性质的开源项目，带你从最基础的神经网络开始，逐步构建一个完整的小型语言模型。项目使用PyTorch实现，涵盖分词、Embedding、Transformer架构、注意力机制、训练循环等核心概念。代码结构清晰、注释详细，每个步骤都有数学原理解释。虽然模型规模很小无法用于生产，但它是理解大语言模型工作原理的绝佳学习资源，非常适合AI初学者建立完整的知识体系。","stars":26,"starGrowth":0,"createdAt":"2024-03-17T02:01:05Z","silentRounds":24,"language":"Python","topics":["llm-model"],"aiDims":["模型与训练","大语言模型","教程"],"zone":"资源","funScore":0,"domainTags":["Transformer","PyTorch","深度学习","教学","模型原理"],"domainKey":"Transformer","pushedAt":"2026-09-14T19:44:27Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ai-glimpse/toyllm","ts":"2026-09-14T19:44:27Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wackojice/pwnagotchi-neuromancer","owner":"wackojice","name":"pwnagotchi-neuromancer","desc":"Neuromancer theme for pwnagotchi: cyberpunk pixel-art faces, Gibson-flavoured voice, and an ICE BROKEN screen on every handshake","summaryCn":"赛博朋克风的黑客机器人，表情包和语音全都有","reasonCn":"在功能实现上，这个项目深度定制了Pwnagotchi的UI和交互系统。它通过Python脚本替换了默认的图像渲染器，加载了精心设计的像素画资源，让机器人的喜怒哀乐更加鲜明。同时，它集成了语音合成模块，用合成器风格的声音朗读状态信息，甚至模仿电影《神经漫游者》中的经典台词，极大地","stars":18,"starGrowth":0,"createdAt":"2026-09-06T02:52:20Z","silentRounds":7,"language":"Python","topics":["pixel-art"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0.7,"domainTags":["Pwnagotchi","赛博朋克","WiFi","黑客","像素画"],"domainKey":"Pwnagotchi","pushedAt":"2026-09-17T03:31:36Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/wackojice/pwnagotchi-neuromancer","ts":"2026-09-17T03:31:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"guangzhengli/ChatFiles","owner":"guangzhengli","name":"ChatFiles","desc":"","summaryCn":"这是一个支持上传文件并与文档进行实时智能对话的开源知识库问答应用。","reasonCn":"该项目基于 LangchainJS 和向量数据库技术打造，允许用户轻松上传各类本地文件并向 GPT-3.5 提问。它非常适合需要从大量文档中快速提取关键信息的技术人员或普通用户，能有效省去人工翻阅长篇大论的繁琐过程，大幅提升日常查阅资料的效率。","stars":3337,"starGrowth":0,"createdAt":"2023-03-28T11:20:09Z","silentRounds":4,"language":"TypeScript","topics":["chatbot","chatgpt","chatgpt-api","chatpdf","chatfile"],"aiDims":["RAG与知识","文档理解","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["文件对话","向量数据库","LangChain","Supabase"],"domainKey":"向量数据库","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"chatgpt-api","source":"github","weight":0.7},{"name":"chatpdf","source":"github","weight":0.7},{"name":"chatfile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/guangzhengli/ChatFiles","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"enescingoz/awesome-n8n-templates","owner":"enescingoz","name":"awesome-n8n-templates","desc":"280+ free n8n automation templates — ready-to-use workflows for Gmail, Telegram, Slack, Discord, WhatsApp, Google Drive, Notion, OpenAI, and more. AI agents, RAG   chatbots, email automation, social media, DevOps, and document processing. The largest open-source n8n template collection.","summaryCn":"280+免费自动化模板，一键复制就能搞定AI办公","reasonCn":"收录280余个n8n工作流模板，覆盖AI Agent构建、RAG聊天机器人、邮件自动化、多平台消息同步等实用场景。支持Gmail、Telegram、Slack、Discord、Notion、OpenAI等主流工具，用户无需从零开发，直接导入模板即可运行。n8n支持自托管，保障数据隐私安全。适合想用低代码方式快速搭建AI应用的用户。","stars":25535,"starGrowth":8,"createdAt":"2025-05-08T18:34:01Z","silentRounds":0,"language":"","topics":["self-hosted","automation","awesome"],"aiDims":["AI Agent","工具调用","效率工具"],"zone":"资源","funScore":0.6,"domainTags":["n8n","自动化工作流","模板聚合","低代码"],"domainKey":"n8n","pushedAt":"2026-09-12T13:00:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.88,"source":"search","url":"https://github.com/enescingoz/awesome-n8n-templates","ts":"2026-09-12T13:00:32Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.024939886430389663,"copyOk":true},{"repo":"RyanCodrai/turbovec","owner":"RyanCodrai","name":"turbovec","desc":"A vector index built on TurboQuant, written in Rust with Python bindings","summaryCn":"用Rust写的高性能向量索引，让RAG检索快得飞起","reasonCn":"这是一个基于TurboQuant构建的高性能向量索引库，用Rust编写并提供了Python绑定。它专注于解决RAG（检索增强生成）场景下的向量检索性能瓶颈，通过底层优化实现毫秒级响应。相比传统向量数据库，它在内存占用和吞吐量上有显著优势，特别适合需要处理海量数据的知识库场景。","stars":17226,"starGrowth":2,"createdAt":"2026-03-26T10:32:44Z","silentRounds":0,"language":"Rust","topics":["rag"],"aiDims":["向量数据库","RAG与知识","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["向量检索","RAG","Rust","嵌入式索引"],"domainKey":"RAG","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/RyanCodrai/turbovec","ts":"2026-09-15T01:07:58.514Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NaiboWang/EasySpider","owner":"NaiboWang","name":"EasySpider","desc":"A visual no-code/code-free web crawler/spider易采集：一个可视化浏览器自动化测试/数据采集/网页爬虫软件，可以无代码图形化的设计和执行爬虫任务。别名：ServiceWrapper面向Web应用的智能化服务封装系统。","summaryCn":"拖拖拽拽就能采集网页数据的爬虫神器，不用写一行代码","reasonCn":"这是一个可视化网页爬虫工具，通过图形界面拖拽操作就能完成数据采集任务，无需编写任何代码。支持Chrome浏览器插件和独立桌面客户端两种方式，可以模拟人类浏览器行为进行页面滚动、点击、表单填写等操作，自动提取网页中的数据并导出为CSV、JSON等格式。特别适合非技术背景的用户快速采集电商数据、新闻资讯、招聘信息等，大幅降低数据采集的技术门槛。","stars":44583,"starGrowth":3,"createdAt":"2020-07-18T02:41:27Z","silentRounds":0,"language":"JavaScript","topics":["frontend","robotics"],"aiDims":["非AI-实用","工具调用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["网页爬虫","数据采集","浏览器自动化","无代码工具","可视化设计"],"domainKey":"网页爬虫","pushedAt":"2026-09-17T12:09:24Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"robotics","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/NaiboWang/EasySpider","ts":"2026-09-17T12:09:24Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"datawhalechina/self-llm","owner":"datawhalechina","name":"self-llm","desc":"《开源大模型食用指南》针对中国宝宝量身打造的基于Linux环境快速微调（全参数/Lora）、部署国内外开源大模型（LLM）/多模态大模型（MLLM）教程","summaryCn":"想象一下，你想自己训练一个能像私人助理一样说话的大模型，但对着铺天盖地","reasonCn":"项目核心在于其对“微调”这一高门槛技术的降维打击。它不仅涵盖了全参数微调和LoRA（低秩自适应）等主流技术路线，还贴心地为不同性能的显卡准备了方案。与市面上那些只有理论的教程不同，它基于Linux环境，通过Jupyter Notebook的形式，将代码段与文字解释一一对应，确保了每一个操","stars":32302,"starGrowth":1,"createdAt":"2023-11-16T02:31:29Z","silentRounds":1,"language":"Jupyter Notebook","topics":["llm"],"aiDims":["教程","微调","大语言模型"],"zone":"资源","funScore":0,"domainTags":["大模型","微调","教程","Python"],"domainKey":"大模型","pushedAt":"2026-09-12T02:51:21Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/datawhalechina/self-llm","ts":"2026-09-12T02:51:21Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0x90d/videoduplicatefinder","owner":"0x90d","name":"videoduplicatefinder","desc":"","summaryCn":"基于感知哈希与AI匹配，彻底清理硬盘中各种变体重复视频的跨平台神器。","reasonCn":"它在扫描时通过音频指纹定位短片是否内嵌于长视频中，即便被裁剪或压缩也能精准识别。内置神经哈希模块能捕捉被镜像、缩放或遮盖的视觉副本。不仅支持快速全量扫描，还特别优化了增量复检机制。它深刻重塑了数字内容资产的治理流程，极大减少了存储空间的冗余浪费。无论个人整理家庭相册还是机构清洗素材库，都能获得极佳的体验。","stars":3707,"starGrowth":1,"createdAt":"2019-01-08T02:27:21Z","silentRounds":1,"language":"C#","topics":[],"aiDims":["AI应用","非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["视频去重","图像识别","文件管理","FFmpeg集成"],"domainKey":"视频去重","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/0x90d/videoduplicatefinder","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"patchy631/ai-engineering-hub","owner":"patchy631","name":"ai-engineering-hub","desc":"In-depth tutorials on LLMs, RAGs and real-world AI agent applications.","summaryCn":"从大模型到智能体，手把手教你打造真正能落地的AI应用","reasonCn":"该项目提供LLM、RAG和AI Agent的深度教程，涵盖RAG系统构建、多模态AI应用、AI Agent编排等实战内容。每个教程都配有Jupyter Notebook代码示例，读者可直接运行并修改。项目覆盖LangChain、LlamaIndex、Hugging Face等主流框架，帮助开发者系统掌握大模型应用开发的核心技能。","stars":37955,"starGrowth":9,"createdAt":"2024-10-21T10:43:24Z","silentRounds":0,"language":"Jupyter Notebook","topics":["rag","ml","ai","agents"],"aiDims":["AI应用","RAG与知识","教程"],"zone":"资源","funScore":0,"domainTags":["LLM教程","RAG系统","AI智能体"],"domainKey":"LLM教程","pushedAt":"2026-09-10T21:32:11Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/patchy631/ai-engineering-hub","ts":"2026-09-10T21:32:11Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02771136872612304,"copyOk":false},{"repo":"AnswerDotAI/RAGatouille","owner":"AnswerDotAI","name":"RAGatouille","desc":"","summaryCn":"把先进检索模型接入任意 RAG 流水线，并支持训练与使用","reasonCn":"很多 RAG 项目把检索当作固定部件，却忽略了它可能拖慢整体表现。该项目关注这个容易被忽略的环节，提供可插拔的研究级检索方案，让用户按任务比较、选择或训练更贴合的模型，降低整条链路像炼金术一样凭经验调参的难度，也更容易把论文里的方法落到生产场景。","stars":3960,"starGrowth":0,"createdAt":"2023-12-29T16:26:42Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["RAG与知识","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["RAG流水线","检索模型","稠密检索","嵌入模型","OpenAI text-ada-002"],"domainKey":"RAG流水线","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/AnswerDotAI/RAGatouille","ts":"2026-09-18T20:05:54.376Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"angelos-p/llm-from-scratch","owner":"angelos-p","name":"llm-from-scratch","desc":"","summaryCn":"零基础手搓小型LLM的工作坊，一小时在笔记本跑通训练。","reasonCn":"项目专注于从零实现GPT训练流程，不依赖预置模型。通过将目标缩小至千万参数级，降低了硬件门槛，让普通笔记本也能在短时间内部署并训练出一个能生成文本的雏形。适合想深挖Transformer底层原理、不满足于调用API的开发者。","stars":3413,"starGrowth":0,"createdAt":"2026-04-04T19:00:12Z","silentRounds":1,"language":"","topics":[],"aiDims":["大语言模型","教程","模型与训练"],"zone":"资源","funScore":0.2,"domainTags":["LLM训练","PyTorch","GPT架构","动手实验"],"domainKey":"LLM训练","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1}],"aiScore":0.68,"source":"search","url":"https://github.com/angelos-p/llm-from-scratch","ts":"2026-09-18T22:32:16.581Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"colbymchenry/codegraph","owner":"colbymchenry","name":"codegraph","desc":"Pre-indexed code knowledge graph, auto syncs on code changes, for Claude Code, Codex, Gemini, Cursor, OpenCode, AntiGravity, Kiro, and Hermes Agent — fewer tokens, fewer tool calls, 100% local","summaryCn":"让AI编程助手一秒读懂你代码库的结构化秘籍","reasonCn":"这个项目为你的代码库构建了一个预先索引的知识图谱，能够自动同步代码变更。支持Cursor、Claude Code、Gemini等多个主流AI编程工具，显著减少token消耗和工具调用次数，所有数据100%本地运行，完全无需上传代码到云端。","stars":71811,"starGrowth":98,"createdAt":"2026-01-18T21:45:37Z","silentRounds":0,"language":"C","topics":[],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0,"domainTags":["代码知识图谱","AI辅助编程","本地索引"],"domainKey":"代码知识图谱","pushedAt":"2026-08-08T02:56:05Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.9,"source":"bigbro","url":"https://github.com/colbymchenry/codegraph","ts":"2026-08-08T02:56:05Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3577128963529264,"copyOk":false},{"repo":"entropy-research/Devon","owner":"entropy-research","name":"Devon","desc":"","summaryCn":"开源 AI 结对编程代理，可自主执行开发任务，支持主流大模型接口接入。","reasonCn":"Devon 是一个开源的 AI 编程智能体，旨在像人类伙伴一样协助开发。它通过连接 Anthropic 或 OpenAI 的 API 接口，能够理解并执行具体的编程任务。项目强调社区驱动的团队协作模式，虽然目前对 Windows 支持尚在发展，但已具备基本的自动化开发能力，适合希望提效的开发者尝试。","stars":3456,"starGrowth":0,"createdAt":"2024-03-15T04:52:31Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0.38,"domainTags":["编程代理","自动化开发","AI 队友"],"domainKey":"编程代理","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/entropy-research/Devon","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"KimYx0207/AI-Coding-Guide-Zh","owner":"KimYx0207","name":"AI-Coding-Guide-Zh","desc":"Claude Code + OpenClaw + Codex + WorkBuddy 中文教程 | 50篇完整教程 + 1张速查卡 | 80万+内容量 | 1500+实操示例 | AI Coding / Agent 四线学习路径（编程+助手+Agent+办公）","summaryCn":"这个项目解决了中文开发者学习AI编程工具的最大痛点资料分散且不系统。","reasonCn":"内容结构上教程按照编程助手、Agent、办公应用四个阶段递进从最基础的代码辅助到高级的自主agent覆盖了不同水平的学习者。每篇教程都有大量实操示例不是空谈理论而是手把手教你怎么用。50篇教程加起来超过80万字几乎是本mini书了。","stars":6091,"starGrowth":1,"createdAt":"2026-02-11T05:34:29Z","silentRounds":1,"language":"","topics":["ai-agent"],"aiDims":["教程","学习资源","AI Agent"],"zone":"资源","funScore":0,"domainTags":["大模型编程教程","开发者指南","学习路径","技术文档","速查手册"],"domainKey":"大模型编程教程","pushedAt":"2026-09-14T10:26:42Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7}],"aiScore":0.85,"source":"search","url":"https://github.com/KimYx0207/AI-Coding-Guide-Zh","ts":"2026-09-14T10:26:42Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LifetimeLabsDev/PrivacyNotes.app","owner":"LifetimeLabsDev","name":"PrivacyNotes.app","desc":"End-to-end encrypted notes, tasks, files, passwords, journal and bookmarks. Keyed by a 12-word BIP-39 phrase: no email, no password, no account to leak. The full client source for web, desktop and mobile is published, along with the encryption (XChaCha20-Poly1305, HKDF-SHA256, Ed25519) and the threat model, so the privacy claims are falsifiable.","summaryCn":"完全基于12个助记词加密的无密码隐私笔记，无需注册邮箱和账号。","reasonCn":"技术架构上，它采用了严格的端到端加密设计。所有的敏感数据在离开你的本地设备之前就已经完成了高强度加密，服务器上只存储无法被解密的乱码。由于直接采用了区块链常用的BIP-39助记词机制，用户只需要记住这12个单词就能在网页端、桌面端和移动端无缝同步和找回所有数据，真正实","stars":450,"starGrowth":5,"createdAt":"2026-05-03T07:49:36Z","silentRounds":0,"language":"TypeScript","topics":["note-taking"],"aiDims":["安全工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["端到端加密","密码管理","笔记应用","隐私保护"],"domainKey":"端到端加密","pushedAt":"2026-09-15T09:49:09Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/LifetimeLabsDev/PrivacyNotes.app","ts":"2026-09-15T09:49:09Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.057999999999999996,"copyOk":true},{"repo":"shareAI-lab/learn-claude-code","owner":"shareAI-lab","name":"learn-claude-code","desc":"Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1","summaryCn":"从零造了一个Claude Code，纯脚本就能用的AI编程助手","reasonCn":"从零用Python实现了一个类Claude Code的AI智能体，支持bash操作、文件系统读写、代码编辑等核心能力，通过调用大模型API实现智能编程辅助。不仅可以用，还能帮你彻底搞懂AI智能体的底层原理，是学习Agent架构的最佳实践项目。","stars":77448,"starGrowth":9,"createdAt":"2025-06-29T15:34:15Z","silentRounds":0,"language":"Python","topics":["ai-agent","llm"],"aiDims":["AI Agent","工具调用","教程"],"zone":"AI","funScore":0.7,"domainTags":["AI Agent","编程助手","Python"],"domainKey":"AI Agent","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/shareAI-lab/learn-claude-code","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.027348621010226216,"copyOk":false},{"repo":"apache/datafusion","owner":"apache","name":"datafusion","desc":"Apache DataFusion SQL Query Engine","summaryCn":"用写SQL的方式查询本地文件，不用写代码","reasonCn":"Apache官方推出的Rust语言SQL查询引擎，原生支持Parquet、JSON、CSV等多种数据格式，具备查询优化、并发执行、内存计算等核心能力。可作为AI数据处理管道的基础引擎，也适合各类数据分析场景，性能对标Arrow生态数据库技术。","stars":9350,"starGrowth":0,"createdAt":"2021-04-17T15:40:23Z","silentRounds":1,"language":"Rust","topics":["arrow","big-data","dataframe","datafusion","olap","python","query-engine","rust","sql"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["SQL查询","Rust","Arrow","数据分析"],"domainKey":"SQL查询","pushedAt":"2026-09-22T04:07:30.369Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"arrow","source":"github","weight":0.7},{"name":"big-data","source":"github","weight":0.7},{"name":"dataframe","source":"github","weight":0.7},{"name":"datafusion","source":"github","weight":0.7},{"name":"olap","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"query-engine","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"sql","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/apache/datafusion","ts":"2026-09-22T04:07:30.369Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"facebookresearch/Mask2Former","owner":"facebookresearch","name":"Mask2Former","desc":"","summaryCn":"该模型的核心亮点在于其通用性，它将以往需要分开构建的全景、实例和语义分","reasonCn":"在功能表现上，除了支持多种经典的分割数据集（如COCO、ADE20K），该项目还扩展到了视频实例分割领域，这使得它能够处理连续帧视频中的物体追踪与分割问题，性能表现极具竞争力。对于需要对视频流进行实时分析的开发者来说，这一功能的引入无疑拓宽了该架构在安防、自动驾驶等领域","stars":3415,"starGrowth":0,"createdAt":"2021-11-24T16:00:44Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["多模态Agent","文档理解"],"zone":"AI","funScore":0.5,"domainTags":["图像分割","实例分割","全景分割","计算机视觉"],"domainKey":"图像分割","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/facebookresearch/Mask2Former","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"esokolov/ml-course-hse","owner":"esokolov","name":"ml-course-hse","desc":"","summaryCn":"高经院机器学习在线课程，含 NLP 与 LLM 案例笔记","reasonCn":"这是莫斯科高等经济学校（HSE）开源的机器学习课程资料。它提供从基础理论到自然语言处理（NLP）和大语言模型（LLM）应用的完整学习路径。内容以 Jupyter Notebook 为主，包含大量可执行的代码示例和详细讲解，适合希望系统性掌握 AI 知识的开发者与学生，无需部署环境即可离线阅读与练习。","stars":3915,"starGrowth":0,"createdAt":"2016-09-04T15:36:26Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","大语言模型","教程"],"zone":"资源","funScore":0.2,"domainTags":["机器学习","自然语言处理","Jupyter Notebook"],"domainKey":"机器学习","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/esokolov/ml-course-hse","ts":"2026-09-18T20:41:13.636Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"R-D-BioTech-Alaska/Qelm","owner":"R-D-BioTech-Alaska","name":"Qelm","desc":"Qelm - Quantum Enhanced Language Model","summaryCn":"量子计算加持语言模型，让AI推理进入量子时代","reasonCn":"探索量子计算与语言模型结合的前沿项目，尝试用量子叠加态和量子纠缠增强传统神经网络的语言理解能力。项目基于Python实现，融合量子电路设计思路。虽然目前处于早期实验阶段，Star数不多，但代表了AI与量子计算交叉领域的一次大胆尝试，值得持续关注。","stars":27,"starGrowth":0,"createdAt":"2024-12-08T09:47:05Z","silentRounds":26,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","模型与训练"],"zone":"创意","funScore":1,"domainTags":["量子计算","语言模型","实验","前沿"],"domainKey":"量子计算","pushedAt":"2026-09-23T04:48:32Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/R-D-BioTech-Alaska/Qelm","ts":"2026-09-23T04:48:32Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"FreedomIntelligence/LLMZoo","owner":"FreedomIntelligence","name":"LLMZoo","desc":"","summaryCn":"LLM Zoo 项目的核心目标是降低构建多语言大语言模型的门槛，通过开放数据、","reasonCn":"该项目的独特价值在于其完整性，涵盖了从原始数据到最终评估的全链路。除了模型文件外，它还提供了一份技术监督报告（Tech Report）和详细的评估基准结果，例如包含 MOSS 等对比案例。这种“数据+模型+基准”的组合拳，使得开发者在训练自己的多语言助手时，能够有明确的参照系来衡量","stars":2934,"starGrowth":0,"createdAt":"2023-04-01T12:08:23Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练","微调"],"zone":"AI","funScore":0.2,"domainTags":["指令遵循数据","多语言LLM","评估基准"],"domainKey":"指令遵循数据","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/FreedomIntelligence/LLMZoo","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PrimeIntellect-ai/prime-agent","owner":"PrimeIntellect-ai","name":"prime-agent","desc":"A self-improving RLM agent for coding workflows and long-running autonomous tasks.","summaryCn":"写代码最怕什么？重复踩同一个坑。这个项目是一个能自我改进的AI Agent，专","reasonCn":"技术核心是RLM（可能是某种强化学习或记忆机制），Agent能进行多轮自主任务执行、代码审查和迭代优化。支持长时间运行的工作流，适合自动化测试、代码重构等持续性任务。相比Claude Code、Cursor等，它更强调自主性和自我进化能力。","stars":21175,"starGrowth":9,"createdAt":"2026-05-08T01:42:41Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":1,"domainTags":["AI-Agent","Coding-Assistant","Self-Improvement"],"domainKey":"AI-Agent","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/PrimeIntellect-ai/prime-agent","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03393010625737898,"copyOk":false},{"repo":"flick9000/winscript","owner":"flick9000","name":"winscript","desc":"Open-source tool to build your Windows script from scratch. It includes debloat, privacy, performance & app installing scripts.","summaryCn":"一键优化Windows系统，卸载垃圾软件还能装软件，小白也能用。","reasonCn":"这是一个基于CSS构建的Windows自动化脚本工具，集成了系统清理、隐私保护、性能优化和软件安装功能。用户可以通过编写或运行预设脚本，批量卸载预装应用、禁用遥测、关闭后台服务，并快速安装常用开发工具。脚本采用声明式配置，支持模块化组合，旨在降低Windows系统维护门槛。","stars":2895,"starGrowth":0,"createdAt":"2024-05-29T19:55:23Z","silentRounds":1,"language":"CSS","topics":["privacy"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Windows优化","系统清理","隐私保护","脚本工具"],"domainKey":"Windows优化","pushedAt":"2026-09-13T20:09:23Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/flick9000/winscript","ts":"2026-09-13T20:09:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deepseek-ai/DeepSeek-V2","owner":"deepseek-ai","name":"DeepSeek-V2","desc":"","summaryCn":"DeepSeek-V2的核心创新在于其混合专家（MoE）架构在236B总参数中实现了仅21","reasonCn":"从工程效率来看，该项目在成本控制和推理速度上取得了显著突破。依据官方数据，相比前代DeepSeek 67B模型，V2节省了42.5%的训练成本，并将KV缓存大小减少了93.3%。这种内存优化直接转化为更快的最大生成吞吐量，提升了5.76倍。","stars":5040,"starGrowth":0,"createdAt":"2024-04-22T06:55:47Z","silentRounds":5,"language":"","topics":[],"aiDims":["大语言模型","模型部署","推理引擎"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","MoE架构","高效推理","成本控制"],"domainKey":"大语言模型","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1}],"aiScore":0.77,"source":"search","url":"https://github.com/deepseek-ai/DeepSeek-V2","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Mesh-LLM/mesh-llm","owner":"Mesh-LLM","name":"mesh-llm","desc":"Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat.","summaryCn":"在单机资源受限的情况下，想要跑 70B 级别的大语言模型几乎是不可能的任务。","reasonCn":"在单机资源受限的情况下，想要跑 70B 级别的大语言模型几乎是不可能的任务。mesh-llm 把这类模型拆分成多个子图层，分别部署在不同的计算节点上，再通过高效的 RPC 通道进行梯度聚合和推理调度，让算力碎片化的机器也能共同完成一次完整的推理。","stars":3437,"starGrowth":7,"createdAt":"2026-02-11T03:18:27Z","silentRounds":0,"language":"Rust","topics":["agents","ai","decentralized","distributed","llm"],"aiDims":["AI基础设施","大语言模型","AI Agent"],"zone":"AI","funScore":0.6,"domainTags":["大语言模型","分布式推理","Rust","算力共享","Agent"],"domainKey":"大语言模型","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"decentralized","source":"github","weight":0.7},{"name":"distributed","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.89,"source":"trending","url":"https://github.com/Mesh-LLM/mesh-llm","ts":"2026-09-11T22:17:29.879Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.032531975560081465,"copyOk":false},{"repo":"openai/parameter-golf","owner":"openai","name":"parameter-golf","desc":"","summaryCn":"这项挑战的核心在于如何在极度有限的资源条件下榨取模型的最大价值。","reasonCn":"为了能够完成这项挑战，参与者通常需要精通各种压缩算法，如量化感知训练 (QAT) 或使用 BitNets 等前沿技术。项目深受 NanoGPT Speedrunning 的启发，评估指标是通过压缩在 FineWeb 验证集上的 bits per byte 表现，这要求训练出的模型在极小的体积下依然拥有出色的泛化能力，是对","stars":5190,"starGrowth":0,"createdAt":"2026-02-09T06:22:45Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["模型与训练","AI基础设施","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["模型压缩","参数限制","语言模型训练"],"domainKey":"模型压缩","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/openai/parameter-golf","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"dragen1860/Deep-Learning-with-PyTorch-Tutorials","owner":"dragen1860","name":"Deep-Learning-with-PyTorch-Tutorials","desc":"","summaryCn":"深度学习往往存在着理论与实践脱节的鸿沟，这个开源项目正是为了解决这一痛","reasonCn":"在技术栈选型上，该项目具有鲜明的时代特征和实用性，它明确要求用户安装 Anaconda 环境并配置特定版本的 CUDA。虽然对于熟悉新版工具链的开发者来说这听起来略显旧式，但对于初学者而言，这种精确的版本锁定避免了环境依赖冲突的噩梦，保证了代码的一致性和可复现性，极大地降低了","stars":3141,"starGrowth":0,"createdAt":"2018-11-26T04:21:46Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["PyTorch实战","视频课程配套","深度学习笔记"],"domainKey":"PyTorch实战","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/dragen1860/Deep-Learning-with-PyTorch-Tutorials","ts":"2026-09-19T22:19:04.205Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"joeltelling/print-farm-manager","owner":"joeltelling","name":"print-farm-manager","desc":"Self-hosted print farm manager for multi-brand 3D printer fleets — Prusa, Bambu, Elegoo, Klipper. Live fleet monitoring, automated job dispatch, no cloud.","summaryCn":"局域网内的3D打印机牧场总管，多品牌设备集中监控不求云端。","reasonCn":"这是一个专为3D打印发烧友和工作室打造的自托管打印农场管理系统。它能够无缝兼容Prusa、Bambu、Elegoo以及Klipper等市面上主流的多品牌打印机，提供实时状态监控与自动化任务调度功能，且完全不依赖任何第三方云服务。","stars":199,"starGrowth":0,"createdAt":"2026-03-30T21:58:55Z","silentRounds":3,"language":"JavaScript","topics":["3d-printing"],"aiDims":["非AI-实用","开发者工具","可视化"],"zone":"工具","funScore":0,"domainTags":["3D打印农场","自托管管理","打印机监控","局域网调度","创客工具"],"domainKey":"3D打印农场","pushedAt":"2026-09-13T15:49:07Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/joeltelling/print-farm-manager","ts":"2026-09-13T15:49:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lavague-ai/LaVague","owner":"lavague-ai","name":"LaVague","desc":"","summaryCn":"在网页自动化测试和数据抓取日益繁琐的今天，许多开发者面临着编写和维护脆","reasonCn":"该项目的核心亮点在于其独特的双层架构设计。世界模型负责审视当前网页的最新状态，并结合宏观目标规划出下一步的抽象指令；而行动引擎则充当翻译官的角色，把这些抽象指令精确编译成能够直接调度的操作代码。这种分工让智能体在面对经常变动的网页结构时，具备了更强的适应能力和","stars":6389,"starGrowth":2,"createdAt":"2024-02-26T23:40:23Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.38,"domainTags":["大模型框架","网页自动化","Web Agent"],"domainKey":"大模型框架","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/lavague-ai/LaVague","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mindmuxai/brain.md","owner":"mindmuxai","name":"brain.md","desc":"A persistent, file-based memory layer for coding agents — give Claude Code, Codex & others a project brain (durable decisions, requirements, constraints) via a zero-dependency CLI.","summaryCn":"我们在用Claude Code或Cursor写代码时，最痛苦的就是AI总是忘记项目的全局约","reasonCn":"它没有复杂的数据库依赖，就是一个零依赖的CLI工具，通过简单的命令就能快速更新项目需求、添加禁用项或者记录关键的技术决策。这种“基于文件”的设计极其巧妙，因为Markdown文件本身就是代码仓库的一部分，不仅AI能读，人类开发者也能随时查看、修改，实现了人机记忆的同步，让项目","stars":552,"starGrowth":0,"createdAt":"2026-06-18T06:45:01Z","silentRounds":3,"language":"JavaScript","topics":["knowledge-management"],"aiDims":["AI Agent","知识管理","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI Agent","记忆层","CLI工具","Claude Code","项目大脑"],"domainKey":"AI Agent","pushedAt":"2026-09-11T11:50:05Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识管理","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/mindmuxai/brain.md","ts":"2026-09-11T11:50:05Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Beckschen/TransUNet","owner":"Beckschen","name":"TransUNet","desc":"","summaryCn":"用Transformer做医学图像分割的经典模型，2D和3D都支持。","reasonCn":"TransUNet是医学图像分割领域的标杆模型，用Transformer增强编码器，突破了传统U-Net的局部感受野限制。2024年论文正式发表在Medical Image Analysis期刊上，还推出了3D版本，在BTCV数据集上达到88.11%的Dice分数，超过nn-UNet，并在BraTs 2021挑战赛中拿下第一。医学影像研究者可以拿它作为基线模型。","stars":3238,"starGrowth":0,"createdAt":"2021-02-08T06:12:54Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":0.26,"domainTags":["医学图像分割","Transformer","深度学习","3D图像","ViT"],"domainKey":"医学图像分割","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/Beckschen/TransUNet","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"livekit/livekit","owner":"livekit","name":"livekit","desc":"End-to-end realtime stack for connecting humans and AI","summaryCn":"一键搭建音视频实时通信，让AI听得见看得见能说话","reasonCn":"开源音视频实时通信平台，支持WebRTC，轻松实现多人会议、直播和远程协作。提供SDK和API，可快速集成到各类应用中。对于AI开发者，它是构建具备视听交互能力的智能体（Agent）的理想基础设施，让AI真正融入实时场景。","stars":21043,"starGrowth":17,"createdAt":"2020-09-30T06:49:46Z","silentRounds":0,"language":"Go","topics":["golang","webrtc","sfu","media-server","video","voice","voice-ai"],"aiDims":["AI基础设施","多模态Agent","AI应用"],"zone":"工具","funScore":0,"domainTags":["实时音视频通信","WebRTC","音视频流媒体"],"domainKey":"实时音视频通信","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"golang","source":"github","weight":0.7},{"name":"webrtc","source":"github","weight":0.7},{"name":"sfu","source":"github","weight":0.7},{"name":"media-server","source":"github","weight":0.7},{"name":"video","source":"github","weight":0.7},{"name":"voice","source":"github","weight":0.7},{"name":"voice-ai","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/livekit/livekit","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.05342360880102647,"copyOk":false},{"repo":"raeq/ibook2epub","owner":"raeq","name":"ibook2epub","desc":"Export your Apple Books (iBooks) library as standard EPUB files. Batch-converts the .epub package folders macOS keeps in iCloud into spec-valid epubs. Pure-Python CLI, zero dependencies, safe to re-run, optional Kindle-safe naming.","summaryCn":"一键把苹果云端的 iBooks 导出成标准 EPUB 文件。","reasonCn":"这是一个纯Python编写的命令行工具，专门用于批量处理苹果设备上的电子书数据。它能够读取iCloud同步的.epub文件包，将其重新封装为符合标准的EPUB3文件。工具完全零依赖，利用Python内置模块进行文件解析和转换，安全可靠且无需安装任何外部库。","stars":11,"starGrowth":0,"createdAt":"2023-05-08T14:28:58Z","silentRounds":11,"language":"Python","topics":["ebook"],"aiDims":["非AI-实用","工具调用"],"zone":"工具","funScore":0,"domainTags":["EPUB转换","苹果生态","电子书"],"domainKey":"EPUB转换","pushedAt":"2026-09-13T15:00:50Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ebook","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/raeq/ibook2epub","ts":"2026-09-13T15:00:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"advaitpaliwal/feynman","owner":"advaitpaliwal","name":"feynman","desc":"The open source AI research agent.","summaryCn":"普通的 AI 助手（比如 ChatGPT）通常只能被动回答问题，或者一次性改完代码","reasonCn":"它的核心能力在于“反思”和“迭代”。当你给它一个任务，比如“修复这个 Bug”时，它会先分析代码，提出一个计划，然后尝试修改。如果修改失败，它会自动回滚，分析错误原因，调整策略，再次尝试。这种“试错-学习-再试”的闭环机制，让它越来越接近人类的工程师。","stars":9750,"starGrowth":9,"createdAt":"2026-03-19T18:53:05Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["AI代理","代码分析","MCP协议"],"domainKey":"AI代理","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/advaitpaliwal/feynman","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03572307692307692,"copyOk":false},{"repo":"Gentleman-Programming/engram","owner":"Gentleman-Programming","name":"engram","desc":"Persistent memory system for AI coding agents. Agent-agnostic Go binary with SQLite + FTS5, MCP server, HTTP API, CLI, and TUI.","summaryCn":"给AI代理配个记忆硬盘，SQLite+FTS5全文搜索永不忘事","reasonCn":"这是一个为AI编码代理设计的持久化记忆系统，用Go编写，底层使用SQLite配合FTS5全文搜索引擎。它支持MCP服务器协议、HTTP API和CLI/TUI接口，能让AI代理跨会话记住上下文和关键信息。对AI应用开发者来说，这是解决记忆缺失痛点的实用基础设施。","stars":6758,"starGrowth":16,"createdAt":"2026-02-16T13:38:41Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["持久化记忆","SQLite搜索","MCP协议","AI代理","上下文管理"],"domainKey":"持久化记忆","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/Gentleman-Programming/engram","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.06612323172536252,"copyOk":false},{"repo":"coreyhaines31/marketingskills","owner":"coreyhaines31","name":"marketingskills","desc":"Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering.","summaryCn":"AI营销助手，让Claude Code帮你搞定SEO、文案和转化优化","reasonCn":"这是一个专为Claude Code和AI代理设计的营销技能库，涵盖CRO转化优化、文案写作、搜索引擎优化、数据分析和增长工程等多个营销领域。通过将专业营销方法论转化为可复用的提示词和工具链，让AI代理能够像资深营销专家一样思考和工作，帮助企业快速提升获客效率和品牌影响力。","stars":51192,"starGrowth":111,"createdAt":"2026-01-15T19:45:23Z","silentRounds":0,"language":"JavaScript","topics":["claude","codex","marketing"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI技能包","营销","SEO优化"],"domainKey":"AI技能包","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"marketing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/coreyhaines31/marketingskills","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.3678059071729957,"copyOk":false},{"repo":"karpathy/makemore","owner":"karpathy","name":"makemore","desc":"","summaryCn":"该项目专注于用最直观的代码揭示语言模型的运作机制，它不追求臃肿庞大的功","reasonCn":"在技术实现上，它提供了一个从最基础的二元模型一直到现代 GPT 所采用的完整 Transformer 架构的选择范围。这种渐进式的设计让学习者能够一步步跟进，理解不同复杂度模型在文本生成上的表现差异。整个代码库保持了极高的克制，没有引入成百上千个让人眼花缭乱的开关和配置参数，非","stars":4286,"starGrowth":2,"createdAt":"2022-06-09T19:29:36Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练","微调"],"zone":"AI","funScore":0.38,"domainTags":["语言模型","自回归模型","文本生成","深度学习"],"domainKey":"语言模型","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/karpathy/makemore","ts":"2026-09-18T00:01:25.926Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"symbolica-dev/numerica","owner":"symbolica-dev","name":"numerica","desc":"Numerica is an open-source mathematics library for Rust, that provides high-performance number types, such as error-tracking floats and finite field elements.","summaryCn":"Rust 写的高性能数学库，专门处理误差和有限域运算。","reasonCn":"Numerica 是一个面向 Rust 开发者的开源数学库，专注于提供高精度的数值计算类型。它特别解决了传统浮点数在科学计算中常见的精度丢失问题，提供了带误差跟踪的浮点数和有限域元素，非常适合密码学、密码分析以及需要高可靠性的数值计算场景。","stars":54,"starGrowth":0,"createdAt":"2025-10-18T16:08:06Z","silentRounds":11,"language":"Rust","topics":["mathematics"],"aiDims":["非AI-实用","数学库","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Rust","数学库","数值计算"],"domainKey":"Rust","pushedAt":"2026-09-15T08:58:13Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"数学库","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/symbolica-dev/numerica","ts":"2026-09-15T08:58:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Anakin-Inc/anakin","owner":"Anakin-Inc","name":"anakin","desc":"","summaryCn":"项目通过 HTTP 请求、防检测浏览器和外部 API 三级处理链路，实现了对互联网","reasonCn":"该项目的核心亮点在于其架构的灵活性，作为一个自托管的开源工具，它没有云服务商的供应商锁定，开发者完全可以将其部署在自己的服务器上，通过单条命令即可启动。这种自托管模式不仅保证了数据的安全性和私有化，还允许用户自由地定义抓取策略。","stars":4404,"starGrowth":0,"createdAt":"2026-03-17T04:51:39Z","silentRounds":2,"language":"Go","topics":["ai","anti-detect","golang","markdown","open-source","playwright","scraper","self-hosted","web-crawler","web-scraping"],"aiDims":["AI应用","RAG与知识"],"zone":"AI","funScore":0.5,"domainTags":["网页爬取","数据处理","自动化API"],"domainKey":"网页爬取","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"anti-detect","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"markdown","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"playwright","source":"github","weight":0.7},{"name":"scraper","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"web-crawler","source":"github","weight":0.7},{"name":"web-scraping","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/Anakin-Inc/anakin","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mockoon/mockoon","owner":"mockoon","name":"mockoon","desc":"Mockoon is the easiest and quickest way to run mock APIs locally. No remote deployment, no account required, open source.","summaryCn":"离线跑Mock API的神器，本地快速调试前端后端联调","reasonCn":"Mockoon让开发者无需部署云端服务，在本地即可快速创建并运行功能完整的Mock API。它提供图形化界面，支持REST和GraphQL，无需注册账户即可开始使用。项目基于Electron构建跨平台桌面应用，内置环境变量和动态响应功能，非常适合前后端分离开发、API测试以及无需真实后端的原型验证场景。","stars":8417,"starGrowth":0,"createdAt":"2017-07-17T13:19:16Z","silentRounds":3,"language":"TypeScript","topics":["desktop","api"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["API Mock","本地服务","前端调试","Electron"],"domainKey":"API Mock","pushedAt":"2026-09-20T07:49:55Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/mockoon/mockoon","ts":"2026-09-20T07:49:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yilewang/llm-for-zotero","owner":"yilewang","name":"llm-for-zotero","desc":"An open-source research agent system for your Zotero library.","summaryCn":"给Zotero装上大脑，让文献管理像跟AI聊天一样简单。","reasonCn":"技术上，它是一个典型的AI Agent架构。前端是Zotero插件，负责界面交互和数据读取；后端是LLM，负责理解和推理；中间通过工具调用机制连接Zotero的API。这意味着它能执行增删改查等复杂操作，而不仅仅是读取文本。它支持RAG（检索增强生成）技术，先在本地向量数据库中检索相关文献","stars":3118,"starGrowth":1,"createdAt":"2026-01-29T00:13:12Z","silentRounds":1,"language":"TypeScript","topics":["academic-paper","literature-analysis","llm-agent","zotero-plugin","zotero","codex","research-tool","rag","mcp"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["Zotero","RAG","文献管理"],"domainKey":"RAG","pushedAt":"2026-09-10T17:23:15.587Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"academic-paper","source":"github","weight":0.7},{"name":"literature-analysis","source":"github","weight":0.7},{"name":"llm-agent","source":"github","weight":0.7},{"name":"zotero-plugin","source":"github","weight":0.7},{"name":"zotero","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"research-tool","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/yilewang/llm-for-zotero","ts":"2026-09-10T17:23:15.587Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"novicezk/midjourney-proxy","owner":"novicezk","name":"midjourney-proxy","desc":"","summaryCn":"项目通过模拟 Discord 通信协议，将繁琐的 MidJourney 操作转化为标准的 RE","reasonCn":"除了基础绘图，该工具还支持包括焦点移动（Pan）、图片变焦（Zoom）及局部重绘（Vary Region）在内的一系列高级操作。特别是针对人脸替换需求，它直接集成了 InsightFace 机器人支持，能够满足商业场景中对人物形象的一致性要求。","stars":5360,"starGrowth":0,"createdAt":"2023-04-24T13:43:45Z","silentRounds":3,"language":"Java","topics":["midjourney","midjourney-api"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["MidJourney","AI绘画","API代理"],"domainKey":"MidJourney","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"midjourney","source":"github","weight":0.7},{"name":"midjourney-api","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/novicezk/midjourney-proxy","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"org2AI/ORG2","owner":"org2AI","name":"ORG2","desc":"","summaryCn":"ORG2 是一款针对 AI 编程 Agent 的本地优先记录系统，核心痛点在于解决现有","reasonCn":"在技术实现上，该项目采用 Rust 和 Tauri 构建，确保了极小的资源占用（磁盘占用低于 100MB）和高效的本地执行性能。这种架构选择使得 ORG2 无需依赖云端服务即可处理大量的会话数据，非常适合对数据隐私敏感或网络环境复杂的开发场景。","stars":2631,"starGrowth":0,"createdAt":"2026-06-01T12:52:57Z","silentRounds":2,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"工具","funScore":0.2,"domainTags":["Agent会话记录","代码溯源","本地优先软件"],"domainKey":"Agent会话记录","pushedAt":"2026-09-21T19:19:33.440Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/org2AI/ORG2","ts":"2026-09-21T19:19:33.440Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jdtoscano94/NABLA-SciML","owner":"jdtoscano94","name":"NABLA-SciML","desc":"Physics Informed Machine Learning Tutorials (Pytorch and Jax)","summaryCn":"把物理公式直接喂给AI，教你用代码写物理学的神仙教程。","reasonCn":"这是一个物理信息机器学习的系列教程，使用PyTorch和JAX两种主流框架。它结合了微分方程数值解法与深度学习，教你如何将物理定律作为约束条件融入神经网络训练过程。教程内容覆盖从基础到进阶的多个课题，通过Jupyter Notebook形式提供完整代码，帮助开发者理解如何构建科学计算AI模型。","stars":686,"starGrowth":0,"createdAt":"2022-04-04T00:29:04Z","silentRounds":10,"language":"Jupyter Notebook","topics":["tutorial"],"aiDims":["教程","模型与训练","AI应用"],"zone":"资源","funScore":0,"domainTags":["PyTorch","JAX","科学计算"],"domainKey":"PyTorch","pushedAt":"2026-09-12T05:31:33Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/jdtoscano94/NABLA-SciML","ts":"2026-09-12T05:31:33Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mshumer/gpt-prompt-engineer","owner":"mshumer","name":"gpt-prompt-engineer","desc":"","summaryCn":"该项目将提示词优化过程从繁琐的“试错”转变为“实验”。","reasonCn":"针对不同版本的模型，项目也保持了快速跟进，特别是对 Claude 3 Opus 的支持，使得提示词的生成质量和推理深度得到了质的飞跃。对于那些在 LLM 应用中苦苦挣扎于 Prompt 稳定性的人来说，这套框架不仅是一个提效器，更像是一个拥有敏锐“直觉”的提示词实验室，帮你把抽象的需求转化","stars":9680,"starGrowth":0,"createdAt":"2023-07-04T03:47:09Z","silentRounds":7,"language":"Jupyter Notebook","topics":[],"aiDims":["提示工程","AI应用","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["提示词生成","自动化测试","模型优化"],"domainKey":"提示词生成","pushedAt":"2026-09-15T20:33:40.833Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/mshumer/gpt-prompt-engineer","ts":"2026-09-15T20:33:40.833Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AIGODLIKE/AIGODLIKE-ComfyUI-Translation","owner":"AIGODLIKE","name":"AIGODLIKE-ComfyUI-Translation","desc":"","summaryCn":"ComfyUI汉化补丁曾服务数十万人，现已停止维护，词条将并入官方。","reasonCn":"这是ComfyUI生态中曾最流行的汉化插件，把界面、菜单、右键和节点名称翻译成中文，让不懂英文的玩家也能用上AI绘画工具。作者宣布停止维护，因为官方编辑器已内置翻译功能，并计划把现有词条迁移到官方核心，避免已有视频和教程因翻译不一致产生理解偏差。","stars":2459,"starGrowth":0,"createdAt":"2023-08-15T10:13:45Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.2,"domainTags":["ComfyUI","多语言翻译","本地化","AI绘画"],"domainKey":"ComfyUI","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/AIGODLIKE/AIGODLIKE-ComfyUI-Translation","ts":"2026-09-21T23:17:36.060Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"QwenAudio/Fun-ASR","owner":"QwenAudio","name":"Fun-ASR","desc":"Fun-ASR speech recognition models, with native Hugging Face Transformers support for Fun-ASR-Nano and separate FunASR, vLLM and llama.cpp deployment paths.","summaryCn":"Fun-ASR是一个实用的语音识别工具，专门为了在资源受限的环境中提供高效准确","reasonCn":"Fun-ASR是一个实用的语音识别工具，专门为了在资源受限的环境中提供高效准确的语音转文本服务而设计。它利用了最新的深度学习技术，能够识别多种语言和方言，即使在没有高性能GPU的情况下也能保持较高的识别准确率。","stars":1551,"starGrowth":0,"createdAt":"2025-12-15T01:46:48Z","silentRounds":1,"language":"C","topics":["transformers"],"aiDims":["AI Agent"],"zone":"AI","funScore":0,"domainTags":["语音识别","ASR","HuggingFace","模型","开源AI"],"domainKey":"语音识别","pushedAt":"2026-09-10T11:04:44Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/QwenAudio/Fun-ASR","ts":"2026-09-10T11:04:44Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google-deepmind/alphafold3","owner":"google-deepmind","name":"alphafold3","desc":"","summaryCn":"精准预测蛋白质与配体等生物分子三维结构的科研神器。","reasonCn":"要真正跑起来，需要先向Google申请模型参数，这是唯一合法的获取渠道，且使用要遵守Google的条款。如果你只想体验效果，不必折腾本地环境，直接访问alphafoldserver.com就能用，只是支持配体和共价修饰的种类会少一些，但完成基础的结构预测绰绰有余。","stars":8580,"starGrowth":0,"createdAt":"2024-11-11T10:38:11Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":1,"domainTags":["蛋白质结构预测","深度学习","计算生物学","药物研发"],"domainKey":"蛋白质结构预测","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/google-deepmind/alphafold3","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alirezadir/Production-Level-Deep-Learning","owner":"alirezadir","name":"Production-Level-Deep-Learning","desc":"","summaryCn":"一份面向生产环境的深度学习系统工程指南，涵盖部署难点与面试技巧。","reasonCn":"该项目旨在解决模型训练好却无法在生产环境落地的痛点，总结了构建大规模深度学习系统的工程经验。它梳理了从数据管道到服务监控的各个环节，并额外提供了机器学习面试准备资料，适合希望将 AI 项目真正推向实战的数据科学家和后端工程师阅读参考。","stars":4662,"starGrowth":0,"createdAt":"2019-11-19T20:11:21Z","silentRounds":6,"language":"","topics":[],"aiDims":["学习资源","教程","模型部署"],"zone":"资源","funScore":0.2,"domainTags":["深度学习","系统工程","机器学习面试"],"domainKey":"深度学习","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1}],"aiScore":0.74,"source":"search","url":"https://github.com/alirezadir/Production-Level-Deep-Learning","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"windmill-labs/windmill","owner":"windmill-labs","name":"windmill","desc":"Open-source developer platform to power your entire infra and turn scripts into webhooks, workflows and UIs. Fastest workflow engine (13x vs Airflow). Open-source alternative to Retool and Temporal.","summaryCn":"把写脚本变成搭积木，最快13倍的自动化工作流引擎","reasonCn":"Windmill 是一个开源的开发者平台，旨在将脚本、SQL 和 Python 代码快速转化为 Web 应用和工作流。它内置了强大的工作流引擎，性能比 Airflow 快 13 倍，同时提供了类似 Retool 的 UI 拖拽功能，让开发者能用最少的代码构建企业级应用。","stars":18006,"starGrowth":2,"createdAt":"2022-05-05T02:21:17Z","silentRounds":0,"language":"Rust","topics":["low-code","open-source","platform","python","typescript","postgresql","self-hostable"],"aiDims":["开发者工具","AI应用","效率工具"],"zone":"工具","funScore":0,"domainTags":["工作流引擎","开发者平台","自动化","Docker","Retool替代品"],"domainKey":"工作流引擎","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"low-code","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"platform","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"self-hostable","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/windmill-labs/windmill","ts":"2026-09-11T22:17:29.879Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"daveebbelaar/ai-cookbook","owner":"daveebbelaar","name":"ai-cookbook","desc":"","summaryCn":"面向AI工程师的实用代码手册，收录可直接复用的系统构建教程与实战片段。","reasonCn":"ai-cookbook 是一本专为实战派 AI 开发者准备的“食谱”。它不纠结于理论推导，而是聚焦于如何把 AI 模型真正落地到生产环境。书中包含大量经过验证的 Python 代码片段，覆盖从数据预处理到模型部署的全流程。对于厌倦了在文档和报错信息之间反复横跳的工程师来说，这里提供了即插即用的解决方案，让你能更快搭建出真正能跑的 AI 系统，而非停留在 Demo 阶段。","stars":4569,"starGrowth":1,"createdAt":"2025-01-30T15:14:29Z","silentRounds":1,"language":"Python","topics":["agents","ai","anthropic","llm","openai","python"],"aiDims":["教程","AI应用","开发者工具"],"zone":"资源","funScore":0.2,"domainTags":["AI开发","实践教程","代码片段","系统工程"],"domainKey":"AI开发","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/daveebbelaar/ai-cookbook","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"letta-ai/letta-code","owner":"letta-ai","name":"letta-code","desc":"Stateful agents that are like people, with memory, identity, and the ability to learn and adapt","summaryCn":"拥有记忆身份的智能体，能像真人一样学习适应","reasonCn":"Letta‑Code 基于 TypeScript 开发，可为每个 Agent 持久化对话记忆与身份属性，支持自定义工具调用与在线学习，用户可在浏览器或服务器上部署，实现跨会话的个性化交互和任务适配，兼容 OpenAI 接口标准，便于与现有 LLM 集成。","stars":3406,"starGrowth":11,"createdAt":"2025-10-25T04:17:51Z","silentRounds":0,"language":"TypeScript","topics":["agent-memory","ai","claude","codex","continual-learning","letta","memgpt","stateful-agents"],"aiDims":["AI Agent","工具调用","AI写作"],"zone":"AI","funScore":0.6,"domainTags":["AI智能体","记忆持久化","对话系统"],"domainKey":"AI智能体","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"agent-memory","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"continual-learning","source":"github","weight":0.7},{"name":"letta","source":"github","weight":0.7},{"name":"memgpt","source":"github","weight":0.7},{"name":"stateful-agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/letta-ai/letta-code","ts":"2026-09-10T22:16:01.544Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.05122654139753376,"copyOk":false},{"repo":"eole-nlp/eole","owner":"eole-nlp","name":"eole","desc":"Open language modeling toolkit based on PyTorch","summaryCn":"在如今大模型漫天飞的时代，大家都在用别人训练好的模型。","reasonCn":"它建立在成熟的PyTorch生态之上，核心代码高度模块化。它完美支持目前主流的各种Transformer变体架构，不仅在传统的机器翻译领域表现神勇，在现代的文本生成、大语言模型微调任务中同样游刃有余。它内置了极其完善的数据预处理流水线和分布式训练支持，能够帮你在多张显卡上平稳跑","stars":194,"starGrowth":0,"createdAt":"2024-06-03T15:08:00Z","silentRounds":16,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","大语言模型","开发者工具"],"zone":"AI","funScore":0,"domainTags":["语言模型","PyTorch","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-13T18:40:10Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/eole-nlp/eole","ts":"2026-09-13T18:40:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AutoHotkey/AutoHotkey","owner":"AutoHotkey","name":"AutoHotkey","desc":"AutoHotkey - macro-creation and automation-oriented scripting utility for Windows.","summaryCn":"让电脑自动干活的脚本神器，Windows必备","reasonCn":"AutoHotkey是Windows平台上一款开源的宏录制与自动化脚本工具，拥有超过1.2万星。它通过热键和脚本语言实现复杂的键盘鼠标自动化操作，能模拟点击、输入文本、窗口管理，甚至编写迷你程序。作为Windows自动化领域的鼻祖级项目，它支持C++编写，生态极其丰富，是效率提升的利器。","stars":13179,"starGrowth":5,"createdAt":"2009-11-25T11:08:21Z","silentRounds":0,"language":"C++","topics":["automation"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["自动化","脚本","Windows","热键","宏"],"domainKey":"自动化","pushedAt":"2026-09-12T05:31:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/AutoHotkey/AutoHotkey","ts":"2026-09-12T05:31:32Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.016138174368313225,"copyOk":false},{"repo":"Anil-matcha/awesome-generative-ai-apps","owner":"Anil-matcha","name":"awesome-generative-ai-apps","desc":"50+ open-source generative AI apps you can clone, deploy, and monetize — image generators, video tools, virtual try-ons, AI SaaS templates, and platform integrations. One-click Vercel deploy on every template.","summaryCn":"精选五十多个拿来即用的生成式人工智能应用，支持一键部署快速变现。","reasonCn":"该项目汇总了超过五十款优秀的开源生成式人工智能全栈应用，涵盖图像生成、视频制作、虚拟试衣工具、模板以及自动化平台集成。所有应用均提供代码仓库与前端界面配置，并支持单键部署到云端平台，帮助开发者快速掌握前后端技术栈，低成本打造可盈利的创新产品。","stars":3300,"starGrowth":2,"createdAt":"2023-11-10T07:55:31Z","silentRounds":0,"language":"JavaScript","topics":["nextjs","open-source","ai-apps","ai-image-generator","ai-saas","ai-tools","ai-video-generator","awesome","awesome-list","generative-ai","generative-ai-apps","image-generation","nextjs-boilerplate","open-source-ai","saas-boilerplate","saas-template","text-to-image","text-to-video","video-generation","seedance2"],"aiDims":["AI应用","创意工具","学习资源"],"zone":"资源","funScore":0,"domainTags":["生成式AI","开源应用","一键部署","独立开发"],"domainKey":"生成式AI","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"nextjs","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"ai-apps","source":"github","weight":0.7},{"name":"ai-image-generator","source":"github","weight":0.7},{"name":"ai-saas","source":"github","weight":0.7},{"name":"ai-tools","source":"github","weight":0.7},{"name":"ai-video-generator","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"awesome-list","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"generative-ai-apps","source":"github","weight":0.7},{"name":"image-generation","source":"github","weight":0.7},{"name":"nextjs-boilerplate","source":"github","weight":0.7},{"name":"open-source-ai","source":"github","weight":0.7},{"name":"saas-boilerplate","source":"github","weight":0.7},{"name":"saas-template","source":"github","weight":0.7},{"name":"text-to-image","source":"github","weight":0.7},{"name":"text-to-video","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"seedance2","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Anil-matcha/awesome-generative-ai-apps","ts":"2026-09-10T22:16:01.544Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"magenta/magenta-realtime","owner":"magenta","name":"magenta-realtime","desc":"Magenta RealTime 2: An Open-Weights Live Music Model","summaryCn":"你有没有想过，当你随手弹奏一段琴声，电脑里居然传出了另一段完美的合奏？","reasonCn":"核心亮点在于它的低延迟推理架构。为了保证音乐的实时感，模型不仅在算法上进行了针对性的性能优化，还对数据流的处理方式做了精简，使得音频输出能够紧跟演奏者的节奏。它采用了开放权重设计，这意味着你可以深入了解模型内部的参数结构，甚至根据自己的乐感需求对其进行微调，这","stars":1802,"starGrowth":0,"createdAt":"2025-06-12T17:09:21Z","silentRounds":3,"language":"Python","topics":["music"],"aiDims":["创意工具","可视化","AI基础设施"],"zone":"AI","funScore":0.8,"domainTags":["音乐生成","AI音乐","实时推理"],"domainKey":"音乐生成","pushedAt":"2026-09-13T09:32:25Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.86,"source":"search","url":"https://github.com/magenta/magenta-realtime","ts":"2026-09-13T09:32:25Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yifanfeng97/Hyper-Extract","owner":"yifanfeng97","name":"Hyper-Extract","desc":"Hypergraph is more powerful. Transform unstructured text into structured knowledge with LLMs. Graphs, hypergraphs, and spatio-temporal extractions — with one command.","summaryCn":"把杂乱的文字瞬间变成知识图谱，让LLM真正读懂文档","reasonCn":"技术上，它超越了简单的三元组（主谓宾）提取，引入了‘超图’（Hypergraph）概念。这意味着它能处理一个关系涉及多个实体的情况，而不仅仅是点对点。它利用LLM的推理能力对文本进行深层语义分析，将非结构化数据转化为结构化的时空关系图谱。","stars":3993,"starGrowth":0,"createdAt":"2026-01-07T09:44:41Z","silentRounds":1,"language":"Python","topics":["rag"],"aiDims":["知识图谱","RAG与知识","文档理解"],"zone":"AI","funScore":0,"domainTags":["知识图谱构建","超图提取","LLM应用","GraphRAG","非结构化数据处理"],"domainKey":"知识图谱构建","pushedAt":"2026-09-14T01:08:51Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/yifanfeng97/Hyper-Extract","ts":"2026-09-14T01:08:51Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ogkalu2/comic-translate","owner":"ogkalu2","name":"comic-translate","desc":"AI comic and manga translator app/browser extension for automatically translating comics, manga, manhwa, BDs, fumetti, and more in multiple languages and formats (Images, PDF, EPUB, CBR, CBZ etc).","summaryCn":"让海外漫画秒变中文的AI神器，看漫再也不用求汉化组","reasonCn":"技术层面上，它走的是一套完整的多模态流水线。首先利用高精度的OCR技术精准定位漫画中的文本区域，随后通过图像处理算法实现文字擦除（Inpainting），确保背景不被破坏。翻译环节则支持对接多种主流翻译API，通过上下文优化让译文更自然。","stars":2948,"starGrowth":3,"createdAt":"2024-01-25T08:45:17Z","silentRounds":0,"language":"Python","topics":["anime"],"aiDims":["AI应用","多模态Agent","文档理解"],"zone":"AI","funScore":0,"domainTags":["漫画翻译","OCR","图像处理","AI翻译","电子书"],"domainKey":"漫画翻译","pushedAt":"2026-09-11T01:22:10Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/ogkalu2/comic-translate","ts":"2026-09-11T01:22:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"labmlai/annotated_deep_learning_paper_implementations","owner":"labmlai","name":"annotated_deep_learning_paper_implementations","desc":"","summaryCn":"这个开源项目专注于用简洁易懂的 PyTorch 代码来复现各种神经网络和前沿深度","reasonCn":"项目核心亮点在于其独特的双栏阅读体验。在官方网站上，左侧是可直接运行的 Python 代码，右侧则是图文并茂的逐行原理解析与数学公式对照。这种形式避免了传统教程中代码与文档分离的痛点，让读者能够一眼看清每一行代码究竟对应着算法的哪一个具体步骤，极大地提升了阅读效率与学","stars":67496,"starGrowth":15,"createdAt":"2020-08-25T02:29:34Z","silentRounds":0,"language":"Python","topics":["deep-learning","deep-learning-tutorial","pytorch","gan","transformers","reinforcement-learning","optimizers","neural-networks","transformer","machine-learning","attention","literate-programming","lora"],"aiDims":["学习资源","论文","模型与训练"],"zone":"资源","funScore":0.6,"domainTags":["深度学习","PyTorch","论文复现","数学公式","教程"],"domainKey":"深度学习","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"deep-learning-tutorial","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"gan","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"reinforcement-learning","source":"github","weight":0.7},{"name":"optimizers","source":"github","weight":0.7},{"name":"neural-networks","source":"github","weight":0.7},{"name":"transformer","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"attention","source":"github","weight":0.7},{"name":"literate-programming","source":"github","weight":0.7},{"name":"lora","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/labmlai/annotated_deep_learning_paper_implementations","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.045666706175180746,"copyOk":false},{"repo":"meta-pytorch/gpt-fast","owner":"meta-pytorch","name":"gpt-fast","desc":"","summaryCn":"项目初衷不仅是为了实现一个高性能推理器，更是作为一份高性能模型部署的「","reasonCn":"对于想要深入了解 Transformer 底层优化的开发者，该项目提供了极其清晰的实现参考。代码的每一行都经过精心简化，旨在演示如何通过原生的 PyTorch 张量操作来实现复杂的算子并行计算。无论是 Nvidia 的 GPU 加速还是 AMD 的 ROCm 环境，项目都做了良好的支持，这为跨硬件平台的模","stars":6255,"starGrowth":0,"createdAt":"2023-10-17T05:30:32Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["推理引擎","模型部署"],"zone":"AI","funScore":0.5,"domainTags":["LLaMA","大模型加速","PyTorch推理"],"domainKey":"LLaMA","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/meta-pytorch/gpt-fast","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AgriciDaniel/claude-obsidian","owner":"AgriciDaniel","name":"claude-obsidian","desc":"Self-organizing AI second brain for Obsidian + Claude Code. Drop any source and Claude reads, links, and files it into one connected knowledge graph of plain Markdown you own. AI note-taking, personal knowledge management (PKM), and an open-source Notion alternative. Based on Karpathy's LLM Wiki pattern.","summaryCn":"Obsidian里的第二大脑，把零散笔记自动整理成知识网络，AI帮你把书读薄。","reasonCn":"它的核心逻辑是“理解与连接”。当你往库里扔一堆文章或PDF时，Claude-obsidian 会自动解析这些Markdown内容。它不只是提取关键词，而是真正理解文章的主旨、论点，甚至引用了哪些笔记。然后，它会利用Obsidian的内部链接机制，自动在笔记之间建立连接。","stars":15136,"starGrowth":13,"createdAt":"2026-04-07T08:59:37Z","silentRounds":0,"language":"Python","topics":["note-taking","knowledge-management"],"aiDims":["知识图谱","AI应用","文档理解"],"zone":"AI","funScore":0,"domainTags":["笔记管理","AI 助手","Obsidian","知识图谱"],"domainKey":"知识图谱","pushedAt":"2026-09-10T17:44:23Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.89,"source":"trending","url":"https://github.com/AgriciDaniel/claude-obsidian","ts":"2026-09-10T17:44:23Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.04989196617336152,"copyOk":false},{"repo":"coleam00/ai-agents-masterclass","owner":"coleam00","name":"ai-agents-masterclass","desc":"","summaryCn":"在人工智能技术飞速迭代的当下，单纯调用大模型 API 已不足以释放其全部价值","reasonCn":"仓库的核心价值在于其对“交互能力”的具象化展示。正如项目自述所言，AI Agent 本质上是具备外部交互能力的大语言模型。因此，课程涵盖了工具调用、API 集成以及多步骤任务规划等关键环节。学习者无需在复杂的框架文档中迷失，而是跟随视频节奏，一行行复制并运行代码，亲眼见证一个","stars":3464,"starGrowth":1,"createdAt":"2024-06-23T20:17:36Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","教程"],"zone":"资源","funScore":0.2,"domainTags":["AI代理","编程教程","YouTube课程"],"domainKey":"AI代理","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/coleam00/ai-agents-masterclass","ts":"2026-09-18T22:32:16.581Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mono0926/LicensePlist","owner":"mono0926","name":"LicensePlist","desc":"A license list generator of all your dependencies for iOS applications","summaryCn":"iOS 依赖许可证生成神器，一键搞定合规性","reasonCn":"一款高效的 iOS 应用许可证列表生成器，能够自动扫描项目中的所有第三方依赖库，并生成符合 Apple App Store 审核要求的许可证文本文件。支持 CocoaPods、SPM、Carthage 等主流依赖管理工具，内置数千种库的许可证信息。","stars":2534,"starGrowth":0,"createdAt":"2017-04-28T06:53:36Z","silentRounds":12,"language":"Swift","topics":["productivity"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["iOS 开发","许可证管理","CocoaPods","合规性","依赖扫描"],"domainKey":"iOS 开发","pushedAt":"2026-09-09T02:46:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/mono0926/LicensePlist","ts":"2026-09-09T02:46:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"elementalsouls/Claude-BugHunter","owner":"elementalsouls","name":"Claude-BugHunter","desc":"","summaryCn":"给Claude助手的红队技能包，内置上千种漏洞模式。","reasonCn":"这是一个专为Claude编程助手设计的红队攻击技能包，主要解决漏洞挖掘和外部红队工作中需要大量专业知识与经验的问题。它包含83个技能和15条斜杠命令，覆盖24类核心漏洞，内置681个已披露漏洞模式。项目支持企业级身份与基础设施攻击矩阵，并内置Burp集成，帮助安全研究员和红队成员更高效地发现和利用漏洞。","stars":4612,"starGrowth":1,"createdAt":"2026-05-05T11:22:32Z","silentRounds":1,"language":"Python","topics":["ai-security","anthropic","application-security","bug-bounty","bugbounty","bugcrowd","claude","claude-code","claude-skills","ethical-hacking","hackerone","offensive-security","pentesting","red-team","security-tools","web-security"],"aiDims":["AI应用","工具调用","AI Agent"],"zone":"AI","funScore":0.32,"domainTags":["安全攻防","漏洞挖掘","红队","渗透测试","Burp"],"domainKey":"安全攻防","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-security","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"application-security","source":"github","weight":0.7},{"name":"bug-bounty","source":"github","weight":0.7},{"name":"bugbounty","source":"github","weight":0.7},{"name":"bugcrowd","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"claude-skills","source":"github","weight":0.7},{"name":"ethical-hacking","source":"github","weight":0.7},{"name":"hackerone","source":"github","weight":0.7},{"name":"offensive-security","source":"github","weight":0.7},{"name":"pentesting","source":"github","weight":0.7},{"name":"red-team","source":"github","weight":0.7},{"name":"security-tools","source":"github","weight":0.7},{"name":"web-security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/elementalsouls/Claude-BugHunter","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"HaD0Yun/Doyunha-Gopeak","owner":"HaD0Yun","name":"Doyunha-Gopeak","desc":"GoPeak — The most comprehensive MCP server for Godot Engine. 95+ tools: scene management, GDScript LSP, DAP debugger, screenshot capture, input injection, ClassDB introspection, CC0 asset library. npx gopeak","summaryCn":"让AI直接操控Godot引擎，打造最聪明的游戏开发助手","reasonCn":"这是一个专为Godot游戏引擎开发的MCP服务器，它拥有95多个强大的工具。AI不仅能通过GDScript LSP深度理解代码逻辑，还能像人类玩家一样注入输入、截取屏幕截图，甚至进行断点调试。它彻底打通了AI与游戏引擎之间的隔阂，让自动化开发变得前所未有的简单和高效。","stars":259,"starGrowth":0,"createdAt":"2026-01-11T09:16:06Z","silentRounds":8,"language":"TypeScript","topics":["godot"],"aiDims":["AI Agent","工具调用","游戏"],"zone":"AI","funScore":1,"domainTags":["MCP","Godot Engine","游戏开发","AI Agent"],"domainKey":"AI Agent","pushedAt":"2026-09-12T13:23:34Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/HaD0Yun/Doyunha-Gopeak","ts":"2026-09-12T13:23:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SakuraLLM/SakuraLLM","owner":"SakuraLLM","name":"SakuraLLM","desc":"","summaryCn":"开源可控的日中翻译模型，专为轻小说与游戏文本优化，禁止商用。","reasonCn":"SakuraLLM 是一个专注于 ACGN（动画、漫画、游戏、小说）领域的日中翻译大语言模型。它在通用双语语料的基础上，针对轻小说、视觉小说（Galgame）等特定类型的文本进行了深度的继续预训练和微调。这使得它在处理充满圈内术语、双关语和特殊语气的文本时，比通用翻译引擎更具“神韵”。该项目完全开源，支持离线自部署，允许用户在不依赖云服务的情况下私有化运行，但严格限制仅用于非商业用途的学习和交流。","stars":4783,"starGrowth":0,"createdAt":"2023-08-23T03:37:56Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["大语言模型","提示工程"],"zone":"AI","funScore":0.2,"domainTags":["日文翻译","轻小说","Galgame"],"domainKey":"日文翻译","pushedAt":"2026-09-17T11:19:48.210Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/SakuraLLM/SakuraLLM","ts":"2026-09-17T11:19:48.210Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kohunglee/openworld-js","owner":"kohunglee","name":"openworld-js","desc":"A 40kb+ webgl library, with concise code and extensive world sandbox! World engine prioritizes first-person view!The world is your viewpoint. ","summaryCn":"轻量WebGL世界沙盒库，专为第一人称视角和创意编程设计的迷你引擎。","reasonCn":"OpenWorld.js是一个仅40kb+的WebGL库，专注于构建第一人称视角的三维世界沙盒。其代码精简，强调“世界即视角”的设计理念，提供扩展的世界引擎功能。与Three.js等重型库相比，它更适合创意编码场景，加载极速，交互直观。适合开发者快速构建沉浸式的3D网页体验，如虚拟现实原型、艺术装置或互动小说，体现了极简主义在3D开发中的应用。","stars":35,"starGrowth":0,"createdAt":"2025-06-29T04:33:54Z","silentRounds":10,"language":"JavaScript","topics":["creative-coding"],"aiDims":["可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["WebGL","创意编程","3D引擎","第一人称视角","沙盒"],"domainKey":"WebGL","pushedAt":"2026-09-12T16:21:36Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.86,"source":"search","url":"https://github.com/kohunglee/openworld-js","ts":"2026-09-12T16:21:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ZSeven-W/dsh-openpencil","owner":"ZSeven-W","name":"dsh-openpencil","desc":"The DeepSeek Harness plugin for OpenPencil — preview, inspect, and edit real .op documents inside a conversation.","summaryCn":"第一段：想象一下，你正和同事在聊天工具里讨论 UI 细节，忽然需要看一眼最","reasonCn":"第一段：想象一下，你正和同事在聊天工具里讨论 UI 细节，忽然需要看一眼最新的 OpenPencil 设计稿，却得打开另一个软件切换窗口。dsh-openpencil 把这一步省掉，直接在聊天框里把 .op 文件渲染出来，让设计稿随时可视。","stars":177,"starGrowth":1,"createdAt":"2026-08-12T16:05:11Z","silentRounds":2,"language":"TypeScript","topics":["design"],"aiDims":["AI应用","创意工具","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI插件","设计预览","聊天工具集成"],"domainKey":"AI插件","pushedAt":"2026-09-17T09:39:03Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/ZSeven-W/dsh-openpencil","ts":"2026-09-17T09:39:03Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NVIDIA/garak","owner":"NVIDIA","name":"garak","desc":"the LLM vulnerability scanner","summaryCn":"给大模型做体检的CT机，一秒查出漏洞和偏见。","reasonCn":"Garak的工作原理非常聪明。它通过发送精心构造的“攻击”提示词来测试模型的防御能力。如果模型回复了攻击指令，或者输出了不该说的话，就说明存在漏洞。此外，它还能分析模型的训练数据，检查是否包含版权材料或有毒内容。","stars":9332,"starGrowth":1,"createdAt":"2023-05-10T18:52:16Z","silentRounds":1,"language":"Python","topics":["ai","llm-evaluation","llm-security","security-scanners","vulnerability-assessment"],"aiDims":["安全工具","大语言模型","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["LLM安全","漏洞扫描","提示词攻击","模型体检","AI安全"],"domainKey":"LLM安全","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"llm-evaluation","source":"github","weight":0.7},{"name":"llm-security","source":"github","weight":0.7},{"name":"security-scanners","source":"github","weight":0.7},{"name":"vulnerability-assessment","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/NVIDIA/garak","ts":"2026-09-11T22:17:29.879Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"joey-zhou/xiaozhi-esp32-server-java","owner":"joey-zhou","name":"xiaozhi-esp32-server-java","desc":"小智ESP32的Java企业级管理平台，提供设备监控、音色定制、角色切换和对话记录管理的前后端及服务端一体化解决方案","summaryCn":"Xiaozhi-ESP32 Server提供了一个易于使用的界面，使管理员能够轻松地查看和","reasonCn":"Xiaozhi-ESP32 Server提供了一个易于使用的界面，使管理员能够轻松地查看和管理他们的小智ESP32设备。它支持实时监控设备状态、远程控制设备等功能，大大提高了设备的管理效率。此外，Xiaozhi-ESP32 Server还提供了一些实用的功能，如自动化任务调度、故障预警等，帮助管理员更好地","stars":1357,"starGrowth":0,"createdAt":"2025-02-27T07:35:36Z","silentRounds":1,"language":"Java","topics":["esp32"],"aiDims":["Java企业级管理"],"zone":"工具","funScore":0,"domainTags":["ESP32","物联网","设备管理","Java"],"domainKey":"ESP32","pushedAt":"2026-09-20T06:30:11Z","tags":[{"name":"Java企业级管理","source":"llm","weight":1},{"name":"esp32","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/joey-zhou/xiaozhi-esp32-server-java","ts":"2026-09-20T06:30:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aieditor-team/AiEditor","owner":"aieditor-team","name":"AiEditor","desc":"AiEditor is a next-generation rich text editor for AI.","summaryCn":"让大模型帮你写富文本的新一代编辑器，内容创作效率翻倍","reasonCn":"技术层面基于 TypeScript 开发，使用现代前端架构，编辑器核心支持富文本格式，包括标题、列表、代码块、表格等常见元素。AI 能力方面支持智能续写、风格改写、语法纠错、摘要生成等。架构上采用了模块化设计，AI 功能作为插件集成，方便扩展。","stars":1776,"starGrowth":0,"createdAt":"2023-11-10T09:04:14Z","silentRounds":14,"language":"TypeScript","topics":["editor"],"aiDims":["AI应用","AI写作","编辑器"],"zone":"AI","funScore":0.6,"domainTags":["富文本编辑器","AI写作","TypeScript"],"domainKey":"AI写作","pushedAt":"2026-09-13T18:13:27Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"编辑器","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/aieditor-team/AiEditor","ts":"2026-09-13T18:13:27Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alondmnt/joplin-mcp","owner":"alondmnt","name":"joplin-mcp","desc":"MCP server for the Joplin note taking app","summaryCn":"让AI大模型直接读写你本地笔记软件的神奇桥梁工具","reasonCn":"技术上，它严格遵循了Anthropic提出的MCP规范，用Python编写了一个轻量级的数据网关。它通过Joplin开放的API接口，将复杂的笔记增删改查操作封装成了标准化工具（Tools），供支持MCP的AI客户端（如Claude Desktop）调用。","stars":173,"starGrowth":0,"createdAt":"2025-07-06T14:50:17Z","silentRounds":16,"language":"Python","topics":["note-taking"],"aiDims":["工具调用","AI应用","效率工具"],"zone":"AI","funScore":0,"domainTags":["MCP","笔记应用","Joplin"],"domainKey":"MCP","pushedAt":"2026-09-12T08:34:20Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/alondmnt/joplin-mcp","ts":"2026-09-12T08:34:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/ML-For-Beginners","owner":"microsoft","name":"ML-For-Beginners","desc":"12 weeks, 26 lessons, 52 quizzes, classic Machine Learning for all","summaryCn":"微软出品的机器学习入门课，12周从零基础到实战全能掌握","reasonCn":"微软官方推出的机器学习免费课程，12周26课时52道测验，系统讲解经典ML知识。配套Jupyter Notebook实践代码，涵盖监督学习、无监督学习、强化学习等核心内容。零基础友好，中英双语支持，是全球最受欢迎的ML入门资源之一，学完能独立完成ML项目。","stars":90847,"starGrowth":42,"createdAt":"2021-03-03T01:34:05Z","silentRounds":0,"language":"Jupyter Notebook","topics":["ml"],"aiDims":["教程","学习资源","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["机器学习","教程","课程","Jupyter"],"domainKey":"机器学习","pushedAt":"2026-09-15T05:29:39Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/microsoft/ML-For-Beginners","ts":"2026-09-15T05:29:39Z","category":"learning","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.12738694728499567,"copyOk":false},{"repo":"alivecontext/alive","owner":"alivecontext","name":"alive","desc":"Personal Context Manager for Claude Code. Your life in walnuts.","summaryCn":"不仅是Claude的代码助手，更是你随叫随到的私人数字大脑与记忆管家。","reasonCn":"这是一个为Claude Code量身定制的个人上下文管理器，它能将你庞大的碎片化生活数据结构化为AI可读取的知识库。通过这种方式，AI不仅能理解你的代码逻辑，更能深刻洞察你的工作流习惯，实现真正意义上的个性化辅助。它让你的知识体系不仅存得下，还能让模型精准调用，堪称个人数字化的核心基石。","stars":129,"starGrowth":0,"createdAt":"2026-02-19T07:10:37Z","silentRounds":18,"language":"Python","topics":["knowledge-management"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["上下文管理","AI助手","Claude"],"domainKey":"上下文管理","pushedAt":"2026-09-11T11:39:50Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/alivecontext/alive","ts":"2026-09-11T11:39:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"amazon-science/mm-cot","owner":"amazon-science","name":"mm-cot","desc":"","summaryCn":"项目核心在于采用了独特的解耦训练架构，这一架构将模型分为“理性生成”与“答","reasonCn":"你可以通过该仓库下载项目所需的数据集以及多种预训练的视觉特征映射文件，如 ResNet、CLIP 或 ViT 的提取结果。整个流程的设计非常符合学术研究的标准，模型架构的统一性和训练阶段的差异化，为开发者提供了一个极佳的基准框架。","stars":3983,"starGrowth":0,"createdAt":"2023-02-02T06:31:32Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["模型与训练","多模态Agent"],"zone":"AI","funScore":0.2,"domainTags":["多模态推理","思维链","计算机视觉"],"domainKey":"多模态推理","pushedAt":"2026-09-18T20:05:54.376Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/amazon-science/mm-cot","ts":"2026-09-18T20:05:54.376Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"IvoryHeart/herdr-world","owner":"IvoryHeart","name":"herdr-world","desc":"Herdr World — a multi-surface web experience for Herdr","summaryCn":"想在网页上体验像素风格的沉浸式世界，却不想下载游戏客户端？","reasonCn":"想在网页上体验像素风格的沉浸式世界，却不想下载游戏客户端？Herdr‑World 用纯前端技术把多层次的像素画面和交互逻辑直接渲染在浏览器里，用户只要打开链接就能像在小游戏机上一样随意切换不同表面、探索隐藏彩蛋，随手点击即触发动画和音效，轻松获得玩耍的乐趣。","stars":16,"starGrowth":0,"createdAt":"2026-08-25T22:30:05Z","silentRounds":8,"language":"TypeScript","topics":["pixel-art"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["像素艺术","网页游戏","前端交互","Canvas渲染","游戏开发"],"domainKey":"像素艺术","pushedAt":"2026-09-19T10:22:47Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/IvoryHeart/herdr-world","ts":"2026-09-19T10:22:47Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Coding-Solo/godot-mcp","owner":"Coding-Solo","name":"godot-mcp","desc":"","summaryCn":"这是一个通过模型上下文协议将 AI 智能体与 Godot 游戏引擎深度连接的服务器","reasonCn":"它的核心亮点在于打通了 AI 与游戏引擎之间的直接反馈回路。传统的 AI 辅助编程往往无法验证代码在引擎里的实际运行效果，而该工具允许程序化地控制项目的启动与停止，甚至还能检索已安装的 Godot 版本以及查找本地的项目列表。","stars":5785,"starGrowth":14,"createdAt":"2025-02-26T21:58:45Z","silentRounds":0,"language":"JavaScript","topics":["ai","godot","mcp"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"工具","funScore":0,"domainTags":["MCP协议","Godot引擎","AI集成","游戏开发"],"domainKey":"MCP协议","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"godot","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/Coding-Solo/godot-mcp","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04926015557476232,"copyOk":false},{"repo":"allenai/OLMo","owner":"allenai","name":"OLMo","desc":"","summaryCn":"艾伦研究所打造的开源大语言模型训练与使用仓库，面向科研。","reasonCn":"OLMo 是艾伦人工智能研究所推出的开源大语言模型项目。它提供了从零训练和微调模型的完整代码和流程，旨在为科研人员提供一个透明、可复现的基础模型平台。虽然该仓库已宣告停更，但项目主体已转移至新的核心仓库继续开发。","stars":6684,"starGrowth":2,"createdAt":"2023-02-20T22:29:43Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","开放模型","预训练"],"domainKey":"大语言模型","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/allenai/OLMo","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AI-Efficiency/Awesome-Model-Quantization","owner":"AI-Efficiency","name":"Awesome-Model-Quantization","desc":"","summaryCn":"整理神经网络量化论文、基准测试与开源实现的持续更新导航合集。","reasonCn":"这个项目并非一款可运行的软件，而是一份精心维护的量化领域文献索引。它系统归纳了从基础的二值网络到复杂的KVCache量化技术，帮助研究员快速定位前沿算法与资源。对于关注模型压缩效率的工程师而言，它减少了在海量文献中盲目检索的时间成本，是进入量化领域的入门地图。","stars":2448,"starGrowth":0,"createdAt":"2018-10-18T05:26:41Z","silentRounds":1,"language":"","topics":[],"aiDims":["模型部署","论文","大语言模型"],"zone":"资源","funScore":0.2,"domainTags":["模型量化","低比特LLM","KV缓存","硬件部署"],"domainKey":"模型量化","pushedAt":"2026-09-21T23:17:36.060Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/AI-Efficiency/Awesome-Model-Quantization","ts":"2026-09-21T23:17:36.060Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deepseek-ai/DeepSeek-V3","owner":"deepseek-ai","name":"DeepSeek-V3","desc":"","summaryCn":"DeepSeek-V3 采用了一种庞大的混合专家模型架构，总参数量高达6710亿，通过","reasonCn":"在底层架构上，该项目引入了多头潜在注意力（MLA）技术，这套架构在之前的版本中已经过充分验证，能有效减少键值缓存的内存占用，让推理过程更加流畅顺滑。此外，开发团队还开创性地取消了辅助损失函数，转而采用全新的负载均衡策略，配合多token预测训练目标，使模型在处理长文本","stars":104481,"starGrowth":3,"createdAt":"2024-12-26T09:52:40Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":1,"domainTags":["大语言模型","混合专家模型","推理优化","开源模型"],"domainKey":"大语言模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/deepseek-ai/DeepSeek-V3","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"agentsmd/agents.md","owner":"agentsmd","name":"agents.md","desc":"","summaryCn":"AGENTS.md通过建立一套统一的规范，解决了AI编程工具在面对不同代码库时“摸","reasonCn":"作为一名开发者，你可以将其理解为专门写给AI看的“工程说明书”。在项目的根目录创建该文件，你可以详细罗列项目的架构概览、代码编写习惯、特定技术的偏好设置以及需要AI遵循的禁忌事项。这种方式的强大之处在于其预测性，无论你更换哪种智能编程代理，只要它们支持该协议，都能瞬","stars":24554,"starGrowth":21,"createdAt":"2025-08-19T17:22:54Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"资源","funScore":0,"domainTags":["AI编程","文档规范","提示词工程","开发指南","协议"],"domainKey":"AI编程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/agentsmd/agents.md","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.0655657733974098,"copyOk":false},{"repo":"anthropics/claude-plugins-official","owner":"anthropics","name":"claude-plugins-official","desc":"Official, Anthropic-managed directory of high quality Claude Code Plugins.","summaryCn":"Anthropic官方插件大本营，Claude Code的必藏神器","reasonCn":"这是Anthropic官方维护的Claude Code插件目录，收录了大量社区开发者贡献的高quality插件。插件覆盖代码生成、测试、API调用、文档处理等多种场景，用户可直接在Claude Code中安装使用。项目为Python语言，stars超过3.3万，社区关注度极高。对于日常使用Claude Code的开发者而言，这是扩展AI能力的实用宝库。","stars":36626,"starGrowth":7,"createdAt":"2025-11-20T18:36:20Z","silentRounds":0,"language":"Python","topics":["claude-code","mcp","skills"],"aiDims":["AI应用","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["Claude Code","AI助手","插件市场","代码生成","开发工具"],"domainKey":"Claude Code","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/anthropics/claude-plugins-official","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":["hot","daily"],"fromOfficial":true,"heatScore":0.025888035821547535,"copyOk":false},{"repo":"developersdigest/llm-answer-engine","owner":"developersdigest","name":"llm-answer-engine","desc":"","summaryCn":"项目通过集成Brave Search与Serper API，实现了能够自动抓取实时信息并反馈","reasonCn":"核心亮点在于其架构的高度集成性，它通过Langchain.JS串联起了Groq、Mistral AI和OpenAI等强大的处理引擎，为开发者展示了如何将多个AI组件组装成一个完整的搜索生态。这种模块化的设计思路，让构建复杂的搜索代理变得清晰易懂，非常适合初学者快速上手实操。","stars":5040,"starGrowth":0,"createdAt":"2024-03-07T21:43:22Z","silentRounds":2,"language":"TypeScript","topics":[],"aiDims":["AI搜索","AI应用","RAG与知识"],"zone":"AI","funScore":0.5,"domainTags":["AI搜索引擎","RAG应用","实时信息检索","生成式AI"],"domainKey":"AI搜索引擎","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/developersdigest/llm-answer-engine","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jnMetaCode/agency-agents-zh","owner":"jnMetaCode","name":"agency-agents-zh","desc":"","summaryCn":"用两百多个岗位角色组成中文智能体团队，覆盖多行业业务。","reasonCn":"它把企业经营、工程、设计、营销、产品、游戏、安全、金融等岗位整理成一批可配置的专家角色，放进智能编程工具里使用。每个角色带有岗位人设、工作流程和可交付成果。中文社区版还补入小红书、抖音、微信、飞书等平台运营，以及跨境电商、政务、医疗合规、畜禽养殖档案等本土场景，适合想借智能助手组建虚拟团队的人。","stars":20870,"starGrowth":23,"createdAt":"2026-03-06T11:11:51Z","silentRounds":0,"language":"Shell","topics":["ai-agents","chinese","claude","llm","prompt-engineering","copilot-agent","cursor-rules","qwen","system-prompt","claude-code","agent-definitions","ai-roles","deepseek","multi-agent","no-code","workflow","agency-orchestrator","hermes-agent","workbuddy"],"aiDims":["AI Agent","提示工程","代码助手"],"zone":"AI","funScore":1,"domainTags":["AI智能体","角色扮演","编程助手","专家系统"],"domainKey":"AI智能体","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"prompt-engineering","source":"github","weight":0.7},{"name":"copilot-agent","source":"github","weight":0.7},{"name":"cursor-rules","source":"github","weight":0.7},{"name":"qwen","source":"github","weight":0.7},{"name":"system-prompt","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"agent-definitions","source":"github","weight":0.7},{"name":"ai-roles","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"multi-agent","source":"github","weight":0.7},{"name":"no-code","source":"github","weight":0.7},{"name":"workflow","source":"github","weight":0.7},{"name":"agency-orchestrator","source":"github","weight":0.7},{"name":"hermes-agent","source":"github","weight":0.7},{"name":"workbuddy","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.722,"source":"search","url":"https://github.com/jnMetaCode/agency-agents-zh","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.08676741734547198,"copyOk":true},{"repo":"llmsresearch/llm-flashcards","owner":"llmsresearch","name":"llm-flashcards","desc":"Visual knowledge bank for understanding large language models, with 180 concept cards from tokenization to deployment.","summaryCn":"很多同学一接触大模型就被满屏幕的 Transformer 公式劝退，llm-flashcards ","reasonCn":"项目的技术特点在于其高度模块化的构建方式，基于 MDX 编写使得内容既有文档的结构性，又具备强大的代码展示能力，甚至可以直接嵌入交互式的演示环节。它并没有像传统教科书那样堆砌长篇大论，而是通过类比、流程图和核心代码段，精准击中每一个概念的灵魂。","stars":135,"starGrowth":1,"createdAt":"2026-05-30T01:40:41Z","silentRounds":17,"language":"MDX","topics":["transformers"],"aiDims":["学习资源","大语言模型","教程"],"zone":"资源","funScore":0,"domainTags":["大模型","知识卡片","Transformer","机器学习","教育"],"domainKey":"大模型","pushedAt":"2026-09-16T02:35:06Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"MDX","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/llmsresearch/llm-flashcards","ts":"2026-09-16T02:35:06Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kennylimz/anime-character-guessr","owner":"kennylimz","name":"anime-character-guessr","desc":"猜二次元角色","summaryCn":"对于每个资深漫迷来说，最自信的时刻莫过于一眼认出屏幕里的人物。","reasonCn":"核心架构基于纯前端技术栈开发，使用了高效的图像处理逻辑来实现角色特征的实时比对。它将成千上万的动漫角色数据库进行分类索引，利用预设的权重机制随机抽取题目。这种轻量级的Web架构保证了即便在手机浏览器上，也能实现秒级的加载与响应，不需要任何后端支撑。","stars":355,"starGrowth":0,"createdAt":"2025-03-26T22:19:34Z","silentRounds":17,"language":"JavaScript","topics":["anime"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["动漫","猜谜游戏","Web游戏"],"domainKey":"动漫","pushedAt":"2026-09-15T21:15:07Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/kennylimz/anime-character-guessr","ts":"2026-09-15T21:15:07Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CompVis/stable-diffusion","owner":"CompVis","name":"stable-diffusion","desc":"","summaryCn":"2022年CVPR会议Oral报告揭示了一个关键突破：通过将扩散模型运算移至低维潜","reasonCn":"该项目的核心专利逻辑在于“解耦条件与生成”。它使用冻结的CLIP ViT-L/14文本编码器提取语义特征，再将其作为条件注入潜空间扩散网络。这种设计避免了端到端训练文本-图像映射的巨大成本，同时借助CLIP在海量图文对上的预训练能力，确保了文本提示与生成内容的高度对齐，让“按文作画","stars":73462,"starGrowth":8,"createdAt":"2022-08-10T14:36:44Z","silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["AI应用","模型与训练","多模态Agent"],"zone":"AI","funScore":1,"domainTags":["图像生成","扩散模型","AI 绘画","生成式 AI"],"domainKey":"图像生成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/CompVis/stable-diffusion","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024326699518118212,"copyOk":false},{"repo":"varunshenoy/GraphGPT","owner":"varunshenoy","name":"GraphGPT","desc":"","summaryCn":"GraphGPT 是一个将非结构化自然语言转化为知识图谱的创意实验项目，作者坦言","reasonCn":"该项目的核心亮点在于其交互式的状态更新机制。与一次性提取不同，GraphGPT 支持连续查询，用户可以在后续的对话中注入新信息，系统会智能判断是更新现有节点的颜色、修改已有的边关系，还是构建全新的子图结构。这种动态演化能力让知识图谱不再是静态的数据死胡同，而成为一个可以","stars":4424,"starGrowth":0,"createdAt":"2023-01-31T18:31:32Z","silentRounds":4,"language":"JavaScript","topics":["gpt-3","knowledge-graph"],"aiDims":["知识图谱","AI应用","可视化"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["自然语言处理","知识图谱生成","可视化工具","LLM应用"],"domainKey":"自然语言处理","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"gpt-3","source":"github","weight":0.7},{"name":"knowledge-graph","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/varunshenoy/GraphGPT","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"instantX-research/InstantID","owner":"instantX-research","name":"InstantID","desc":"","summaryCn":"只需单张人脸照片即可实现免微调、高保真的生成式图像身份保持。","reasonCn":"该方案由小红书等团队研发，是一项零样本身份保持图像生成技术。传统人脸定制通常需要收集多张照片并进行耗时的模型微调，而它仅凭单张参考图片即可在数秒内锁定人物面部特征，将其无缝无损地融入各种风格的图像生成中，大幅提升了个性化人工智能绘画的效率与实用价值。","stars":11994,"starGrowth":2,"createdAt":"2023-12-11T13:16:04Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","AI应用"],"zone":"AI","funScore":1,"domainTags":["图像生成","人脸定制","扩散模型"],"domainKey":"图像生成","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/instantX-research/InstantID","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deepseek-ai/DeepGEMM","owner":"deepseek-ai","name":"DeepGEMM","desc":"","summaryCn":"该项目针对现代大语言模型的计算特点进行了深度定制，将矩阵乘法、混合专家","reasonCn":"其最大的技术亮点在于引入了DeepJIT运行时编译机制。开发者在部署时无需进行漫长的安装预编译，环境配置流程得到了极致简化，这对需要频繁实验不同算子配置的科研人员非常友好。它不仅涵盖了基础的计算操作，还针对当下大模型主流的MoE架构做了针对性的通信重叠优化，确保在多卡并","stars":7858,"starGrowth":4,"createdAt":"2025-02-13T09:09:21Z","silentRounds":0,"language":"Cuda","topics":[],"aiDims":["推理引擎","模型部署"],"zone":"工具","funScore":0,"domainTags":["GPU优化","CUDA","矩阵乘法","MoE架构"],"domainKey":"GPU优化","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Cuda","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/deepseek-ai/DeepGEMM","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"komi-store/komi-store","owner":"komi-store","name":"komi-store","desc":"🩵 A free, open-source app store for developers' releases on GitHub, Codeberg & Forgejo — browse, discover, and install apps with one click. Formerly GitHub Store.","summaryCn":"GitHub上的应用商店，一键安装那些好玩又实用的开源软件","reasonCn":"它的主要功能是“发现”和“安装”。你可以在商店里浏览各种分类的应用，比如效率工具、开发环境、游戏等。找到感兴趣的项目后，只需要点击安装，它就会自动处理下载、解压、配置等繁琐步骤。 它支持Windows、macOS和Linux三大主流桌面系统，极大地降低了使用开源软件的门槛。","stars":18689,"starGrowth":27,"createdAt":"2025-11-21T13:03:23Z","silentRounds":0,"language":"Kotlin","topics":["desktop"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["应用商店","软件分发","跨平台"],"domainKey":"应用商店","pushedAt":"2026-09-04T10:31:18Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/komi-store/komi-store","ts":"2026-09-04T10:31:18Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.10240092032746535,"copyOk":false},{"repo":"evidentlyai/evidently","owner":"evidentlyai","name":"evidently","desc":"Evidently is ​​an open-source ML and LLM observability framework. Evaluate, test, and monitor any AI-powered system or data pipeline. From tabular data to Gen AI. 100+ metrics.","summaryCn":"给大模型和机器学习系统做全面体检的开源监控利器","reasonCn":"技术上，它不单单看准确率，还提供超过100种专业指标，从传统机器学习的数据漂移、特征分布变化，到现代大模型的文本质量、幻觉率、安全性评估应有尽有。它能自动生成漂亮的交互式网页报告，支持与CI/CD流水线无缝集成，每次模型更新都能自动跑一遍测试。","stars":7935,"starGrowth":3,"createdAt":"2020-11-25T15:20:08Z","silentRounds":0,"language":"Jupyter Notebook","topics":["data-science"],"aiDims":["AI基础设施","模型与训练","开发者工具"],"zone":"工具","funScore":0,"domainTags":["机器学习监控","模型评估","可观测性","CI/CD","数据漂移"],"domainKey":"机器学习监控","pushedAt":"2026-09-11T09:03:28Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/evidentlyai/evidently","ts":"2026-09-11T09:03:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Akegarasu/lora-scripts","owner":"Akegarasu","name":"lora-scripts","desc":"","summaryCn":"针对 Stable Diffusion 模型微调过程中环境依赖复杂、参数调整繁琐的痛点，","reasonCn":"该项目的核心亮点在于其对不同网络环境的兼容性优化。考虑到国内用户访问 GitHub 或 PyPI 可能遇到的网络波动，项目专门提供了 `install-cn.ps1` 脚本，确保在国内网络环境下也能顺利拉取依赖包并创建独立的虚拟环境。","stars":6143,"starGrowth":2,"createdAt":"2023-01-22T13:50:45Z","silentRounds":0,"language":"Python","topics":["dreambooth","finetune","lora","stable-diffusion"],"aiDims":["模型与训练","微调","AI应用"],"zone":"工具","funScore":0.2,"domainTags":["Stable Diffusion","LoRA训练","Dreambooth","图形界面","环境部署"],"domainKey":"Stable Diffusion","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"dreambooth","source":"github","weight":0.7},{"name":"finetune","source":"github","weight":0.7},{"name":"lora","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/Akegarasu/lora-scripts","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/echarts","owner":"apache","name":"echarts","desc":"Apache ECharts is a powerful, interactive charting and data visualization library for browser","summaryCn":"谷歌亲儿子级的图表库，让枯燥数据一秒变成炫酷动效大屏","reasonCn":"由百度捐献给Apache的开源可视化引擎，基于Canvas/SVG渲染，性能极高且支持复杂交互。拥有数百种图表类型，从基础折线图到3D地球仪都能画。它提供了极其丰富的配置项，几乎任何你想象的图形效果都能实现，是前端数据展示的事实标准。","stars":67376,"starGrowth":2,"createdAt":"2013-04-03T03:18:59Z","silentRounds":0,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["数据可视化","图表库","Canvas渲染"],"domainKey":"数据可视化","pushedAt":"2026-09-16T09:16:02Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/apache/echarts","ts":"2026-09-16T09:16:02Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dk-liang/Awesome-Visual-Transformer","owner":"dk-liang","name":"Awesome-Visual-Transformer","desc":"","summaryCn":"该项目精选了计算机视觉领域的核心论文、综述与中文解读博客清单。","reasonCn":"该项目是一个专注于计算机视觉中 Transformer 研究的开源资源合集。它全面梳理了从注意力机制基础论文到多模态学习的学术前沿，同时精选了多篇长达数万字的中文深度解读博客与系统性综述。它帮助研究者与开发者快速理清算法脉络，省去在海量文献中手动筛选的时间。","stars":3589,"starGrowth":0,"createdAt":"2020-12-03T05:40:09Z","silentRounds":4,"language":"","topics":[],"aiDims":["学习资源","论文"],"zone":"资源","funScore":0.2,"domainTags":["计算机视觉","Visual Transformer","论文精选","深度学习"],"domainKey":"计算机视觉","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1}],"aiScore":0.65,"source":"search","url":"https://github.com/dk-liang/Awesome-Visual-Transformer","ts":"2026-09-18T22:32:16.581Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bowang-lab/MedSAM","owner":"bowang-lab","name":"MedSAM","desc":"","summaryCn":"将SAM分割模型引入医学影像，能自动分割多种器官与病灶。","reasonCn":"MedSAM把Meta的SAM分割模型迁移到医学影像领域，让AI能自动识别CT、MRI等图像中的器官和病灶结构。项目提供完整模型代码和训练流程，已发布支持3D和视频分割的MedSAM2版本。LiteMedSAM轻量版推理速度快10倍，并集成了3D Slicer插件，方便医生直接在专业工具中使用。","stars":4400,"starGrowth":0,"createdAt":"2023-04-22T19:57:25Z","silentRounds":3,"language":"Jupyter Notebook","topics":[],"aiDims":["AI应用","模型与训练"],"zone":"AI","funScore":0.26,"domainTags":["医学图像分割","SAM","3D分割","视频分割"],"domainKey":"医学图像分割","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/bowang-lab/MedSAM","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"google-research/simclr","owner":"google-research","name":"simclr","desc":"","summaryCn":"SimCLR作为自监督视觉表示学习的里程碑式工作，其核心思想是通过对同一图像","reasonCn":"在资源提供方面，该项目展现了极大的诚意，直接在Google Cloud Storage上托管了多达65个SimCLRv2的预训练模型。这些模型不仅包含基础的无监督预训练权重，还提供了在仅使用1%或10%标注数据情况下微调完成的检查点，完美对应了学术论文中的实验数据。","stars":4502,"starGrowth":0,"createdAt":"2020-03-10T16:53:22Z","silentRounds":4,"language":"Jupyter Notebook","topics":["simclr","contrastive-learning","representation-learning","self-supervised-learning","unsupervised-learning","computer-vision","simclrv2"],"aiDims":["模型与训练","微调"],"zone":"AI","funScore":0.2,"domainTags":["对比学习","视觉表示","预训练模型","深度学习"],"domainKey":"对比学习","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"simclr","source":"github","weight":0.7},{"name":"contrastive-learning","source":"github","weight":0.7},{"name":"representation-learning","source":"github","weight":0.7},{"name":"self-supervised-learning","source":"github","weight":0.7},{"name":"unsupervised-learning","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"simclrv2","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/google-research/simclr","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"trykimu/videoeditor","owner":"trykimu","name":"videoeditor","desc":"Your Creative Copilot for Video Editing","summaryCn":"想让视频剪辑不再是专业人士的专属？videoeditor 让任何人都能用浏览器完成","reasonCn":"想让视频剪辑不再是专业人士的专属？videoeditor 让任何人都能用浏览器完成从素材上传到最终输出的全流程。它把繁琐的剪辑步骤拆成可视化的任务块，支持多轨道、分辨率切换、自动字幕、音频匹配等功能。\\n核心技术上，videoeditor 采用 React + TypeScript 搭建前端，使用 WebGL 进","stars":2228,"starGrowth":0,"createdAt":"2025-05-22T18:49:18Z","silentRounds":4,"language":"TypeScript","topics":["editor","video-editor"],"aiDims":["AI应用","工具调用"],"zone":"创意","funScore":1,"domainTags":["视频编辑","浏览器工具","FFmpeg"],"domainKey":"视频编辑","pushedAt":"2026-09-10T16:04:40Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"video-editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/trykimu/videoeditor","ts":"2026-09-10T16:04:40Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenGVLab/DragGAN","owner":"OpenGVLab","name":"DragGAN","desc":"","summaryCn":"基于GAN的交互式图像编辑工具，通过拖拽图像中的点来改变生成结果。","reasonCn":"这是一个将生成式AI与直观交互结合的有趣项目，允许用户通过直接拖拽图像中的特定像素点来操纵生成的图像内容。基于“拖拽你的GAN”论文实现，它利用生成模型背后的潜空间结构，将用户的手动位移转化为模型推理的条件输入，从而实现对人脸、场景等生成内容的局部或全局控制。项目提供了Colab在线演示，用户无需配置复杂环境即可在浏览器中体验这种魔法般的编辑方式，非常适合探索创意图像生成和AI艺术创作，展示了深度学习在交互式媒体中的潜力。","stars":4944,"starGrowth":1,"createdAt":"2023-05-20T08:53:41Z","silentRounds":4,"language":"Python","topics":["draggan","image-editing","image-generation","gradio-interface","interngpt"],"aiDims":["多模态Agent","AI应用","生成艺术"],"zone":"AI","funScore":0.68,"domainTags":["图像生成","拖拽交互","GAN模型","Colab示例"],"domainKey":"图像生成","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"生成艺术","source":"llm","weight":1},{"name":"draggan","source":"github","weight":0.7},{"name":"image-editing","source":"github","weight":0.7},{"name":"image-generation","source":"github","weight":0.7},{"name":"gradio-interface","source":"github","weight":0.7},{"name":"interngpt","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/OpenGVLab/DragGAN","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"SeldonIO/seldon-core","owner":"SeldonIO","name":"seldon-core","desc":"","summaryCn":"Seldon Core 2 作为 MLOps 和 LLMOps 框架，专门用于在 Kubernetes 上部署、","reasonCn":"该框架支持从单一模型到模块化、数据密集型应用的广泛部署场景。这种灵活性使得 Seldon Core 能够适应不同规模和复杂度的 AI 项目需求。无论是在本地数据中心还是在公有云上，你都可以使用相同的配置和流程来部署模型。","stars":4782,"starGrowth":0,"createdAt":"2017-12-20T14:51:54Z","silentRounds":1,"language":"Go","topics":["kubernetes","machine-learning","deployment","serving","mlops","aiops","machine-learning-operations","production-machine-learning"],"aiDims":["模型部署","AI基础设施","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["Kubernetes","MLOps","LLMOps","模型部署"],"domainKey":"模型部署","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"deployment","source":"github","weight":0.7},{"name":"serving","source":"github","weight":0.7},{"name":"mlops","source":"github","weight":0.7},{"name":"aiops","source":"github","weight":0.7},{"name":"machine-learning-operations","source":"github","weight":0.7},{"name":"production-machine-learning","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/SeldonIO/seldon-core","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fastai/courses","owner":"fastai","name":"courses","desc":"","summaryCn":"该课程的核心魅力在于其强烈的实践导向，鼓励学习者在修改代码和观察结果中","reasonCn":"项目不仅包含教学材料，还延伸到了配套的在线论坛和维基百科，为学习者提供了全方位的答疑支持。作者明确建议将所有调试问题发布到专门的论坛中，这使得相关的知识库能够被不断沉淀与索引。这种协作式的学习方式让许多经典的报错和技术难点都变得有迹可循，有效地避免了重复劳动。","stars":5746,"starGrowth":0,"createdAt":"2016-12-17T23:16:41Z","silentRounds":5,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.5,"domainTags":["深度学习","机器学习","教程","神经网络"],"domainKey":"深度学习","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/fastai/courses","ts":"2026-09-15T22:54:20.912Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kodu-ai/claude-coder","owner":"kodu-ai","name":"claude-coder","desc":"","summaryCn":"在许多开发者眼中，将脑海中的创意转化为可运行的代码往往存在一道隐形壁垒","reasonCn":"与传统补全工具不同，Claude Coder 具备更宏观的任务处理能力。用户只需在编辑器中描述具体需求，例如“构建一个用户仪表盘”或“搭建一个简单的后端服务器”，系统便会自动处理相关逻辑，涵盖从前端界面设计到后端服务器配置等多个领域。","stars":5218,"starGrowth":0,"createdAt":"2024-09-01T13:42:47Z","silentRounds":4,"language":"TypeScript","topics":["chatgpt","claude","coding-agents","llm","openai","vscode","vscode-extension"],"aiDims":["AI Agent","代码助手","VS Code 插件"],"zone":"AI","funScore":0.32,"domainTags":["VS Code 扩展","AI 编程","全栈开发","自动化编码","TypeScript"],"domainKey":"VS Code 扩展","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"VS Code 插件","source":"llm","weight":1},{"name":"chatgpt","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"coding-agents","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"vscode","source":"github","weight":0.7},{"name":"vscode-extension","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/kodu-ai/claude-coder","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AILab-CVC/YOLO-World","owner":"AILab-CVC","name":"YOLO-World","desc":"","summaryCn":"YOLO-World 腾讯 AI Lab 与华中科技大学合作开发，2025年2月刚更新了 V2.1 ","reasonCn":"这项技术基于视觉语言模型（VLM）的优势，将自然语言或视觉实例映射到检测框上。传统 YOLO 库通常只适用于预先定义好的 80 类或 120 类 COCO 数据集，而 YOLO-World 实现了真正的零样本（Zero-shot）检测。这种能力让它在工业自动化质检、安防监控等非标准场景下具有巨大潜力，因为","stars":6572,"starGrowth":0,"createdAt":"2024-01-29T02:04:07Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","多模态Agent","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["开放词汇检测","YOLO系列","实时目标检测","图像提示"],"domainKey":"开放词汇检测","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/AILab-CVC/YOLO-World","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tvytlx/ai-agent-deep-dive","owner":"tvytlx","name":"ai-agent-deep-dive","desc":"","summaryCn":"学习和构建 AI Agent 往往需要面对繁杂的工程细节和晦涩的代码架构，许多开","reasonCn":"除了深度的理论分析，该仓库还包含了一个专为教学设计的最小 Python Agent 项目。该项目使用 Poetry 进行依赖管理，核心代码高度集中，去除了所有不必要的工程复杂度。通过查看源码和运行简单的 CLI 命令行入口，学习者可以直观地观察到 AI Agent 的主循环是如何运转的、Fake 语言","stars":5829,"starGrowth":2,"createdAt":"2026-03-31T11:27:01Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","学习资源"],"zone":"资源","funScore":0,"domainTags":["AI Agent","教程","源码分析"],"domainKey":"AI Agent","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/tvytlx/ai-agent-deep-dive","ts":"2026-09-15T22:54:20.912Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huangjunsen0406/py-xiaozhi","owner":"huangjunsen0406","name":"py-xiaozhi","desc":"Open-source AI assistant ecosystem with MCP integrations, multimodal workflows, IoT support, and cross-platform voice interaction.","summaryCn":"把各种智能硬件和MCP协议无缝结合的开源全栈AI语音助手生态","reasonCn":"技术架构上，它最大亮点是全面拥抱了当下大火的MCP（模型上下文协议），这意味着它可以像拼积木一样轻松接入各种第三方工具和数据源。它内置了完整的跨平台语音交互流水线，从音频采集、唤醒词检测、流式语音识别到大模型推理再到文本转语音，整个链路高度优化。","stars":3479,"starGrowth":0,"createdAt":"2025-02-15T18:40:27Z","silentRounds":4,"language":"Python","topics":["raspberry-pi","esp32","iot"],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["AI语音助手","物联网","MCP协议"],"domainKey":"AI语音助手","pushedAt":"2026-09-11T02:20:50Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"esp32","source":"github","weight":0.7},{"name":"iot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/huangjunsen0406/py-xiaozhi","ts":"2026-09-11T02:20:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JekSun97/gdTree3D","owner":"JekSun97","name":"gdTree3D","desc":"🌳 Plugin for procedural generation of 3D trees of varying complexity for Godot Engine.","summaryCn":"Godot引擎的神级插件，一键生成千奇百怪的3D森林。","reasonCn":"这是一个专为Godot游戏引擎设计的程序化3D树木生成插件。它利用分形算法和参数化建模，支持生成从简单几何体到复杂生物形态的树木。插件完全集成到Godot编辑器中，允许开发者通过调整滑块实时控制树的形态、大小和材质，极大地丰富了游戏世界的自然感。","stars":367,"starGrowth":1,"createdAt":"2024-07-09T17:23:25Z","silentRounds":8,"language":"C++","topics":["godot"],"aiDims":["游戏","创意工具"],"zone":"工具","funScore":0,"domainTags":["Godot","3D树木","程序化生成","插件","游戏开发"],"domainKey":"Godot","pushedAt":"2026-09-13T22:22:30Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/JekSun97/gdTree3D","ts":"2026-09-13T22:22:30Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/power-platform-skills","owner":"microsoft","name":"power-platform-skills","desc":"A plugin marketplace for Claude Code/GitHub Copilot that provides Power Platform development plugins, including reusable skills, agents, and commands for building and deploying solutions.","summaryCn":"给Copilot装上Power Platform技能包，开发效率翻倍","reasonCn":"微软官方推出的Claude Code和GitHub Copilot插件集，提供Power Platform开发所需的可复用技能、智能Agent和命令。覆盖Power Apps、Power Automate、Power BI等场景，让AI代码助手深度理解微软低代码生态，显著降低开发门槛和提升开发效率。","stars":909,"starGrowth":1,"createdAt":"2026-01-21T18:25:14Z","silentRounds":1,"language":"JavaScript","topics":["claude-code-plugin","claude-code-skills","power-apps","power-pages","power-platform","code-apps","github-copilot-plugin"],"aiDims":["代码助手","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI技能包","Power Platform","低代码开发"],"domainKey":"AI技能包","pushedAt":"2026-09-20T04:13:28.323Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"claude-code-plugin","source":"github","weight":0.7},{"name":"claude-code-skills","source":"github","weight":0.7},{"name":"power-apps","source":"github","weight":0.7},{"name":"power-pages","source":"github","weight":0.7},{"name":"power-platform","source":"github","weight":0.7},{"name":"code-apps","source":"github","weight":0.7},{"name":"github-copilot-plugin","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/microsoft/power-platform-skills","ts":"2026-09-20T04:13:28.323Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"facebookresearch/jepa","owner":"facebookresearch","name":"jepa","desc":"","summaryCn":"V-JEPA（Video Joint Embedding Predictive Architecture）是 Meta AI 研究","reasonCn":"该项目的代码库基于 PyTorch 框架构建，提供了完整的训练和评估流程。它特别强调了“联合嵌入预测”的概念，即在嵌入空间中进行预测，而不是直接在像素空间中进行重构。这种方法避免了像素级噪声的干扰，使得学习到的特征更加鲁棒和语义丰富。","stars":4149,"starGrowth":1,"createdAt":"2024-02-12T15:34:31Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","多模态Agent"],"zone":"AI","funScore":0.2,"domainTags":["视觉表征","自监督学习","视频理解"],"domainKey":"视觉表征","pushedAt":"2026-09-18T10:52:55.418Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/facebookresearch/jepa","ts":"2026-09-18T10:52:55.418Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"jax-ml/jax","owner":"jax-ml","name":"jax","desc":"","summaryCn":"JAX 是 Google 团队开源的核心自动微分数组计算库，它的定位极其精准：解决","reasonCn":"从使用体验来看，JAX 最大的杀手锏是完全兼容 NumPy 生态。开发者不需要改变代码习惯，只需在函数前加上特定的装饰器，比如用于编译提速的 jit 编译指令，就能让代码在后台自动生成面向底层硬件的并行计算指令，将其扩展到你手边所有的硬件加速设备上。","stars":36326,"starGrowth":3,"createdAt":"2018-10-25T21:25:02Z","silentRounds":0,"language":"Python","topics":["jax"],"aiDims":["AI基础设施","模型与训练"],"zone":"工具","funScore":0,"domainTags":["自动微分","数组计算","GPU加速","深度学习框架"],"domainKey":"自动微分","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"jax","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/jax-ml/jax","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ClickHouse/ClickHouse","owner":"ClickHouse","name":"ClickHouse","desc":"ClickHouse® is a real-time analytics database management system","summaryCn":"俄罗斯开发的实时分析数据库，查询大数据量速度堪比坐火箭","reasonCn":"ClickHouse是Yandex开源的列式分析型数据库，专为实时数据分析设计，查询性能远超传统数据库。支持SQL查询、向量检索、聚合分析等多种操作，单表可容纳数百亿行数据。核心优势在于向量化执行引擎和列式存储，让复杂分析查询在秒级完成。","stars":50025,"starGrowth":20,"createdAt":"2016-06-02T08:28:18Z","silentRounds":0,"language":"C++","topics":["database","ai"],"aiDims":["向量数据库","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["OLAP数据库","大数据分析","列式存储"],"domainKey":"OLAP数据库","pushedAt":"2026-09-23T21:49:02Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/ClickHouse/ClickHouse","ts":"2026-09-23T21:49:02Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.061199400299850076,"copyOk":true},{"repo":"gpustack/gpustack","owner":"gpustack","name":"gpustack","desc":"A GPU cluster manager for high-performance AI model serving (vLLM, SGLang) and on-demand SSH-accessible GPU instances.","summaryCn":"一键搞定GPU集群，秒开AI模型服务和远程GPU实例","reasonCn":"gpustack 用 Python 实现 GPU 资源统一调度，支持 vLLM、SGLang 等高性能推理框架，提供自动扩容、作业排队、SSH 远程登录等功能，用户可通过简洁的 CLI 或 Web UI 快速部署模型服务，兼顾成本与性能，并支持多租户与自定义计费策略。","stars":5741,"starGrowth":0,"createdAt":"2024-05-11T03:41:58Z","silentRounds":2,"language":"Python","topics":["ascend","cuda","deepseek","distributed-inference","genai","inference","llama","llm","maas","openai","qwen","rocm","vllm","mindie","llm-inference","llm-serving","sglang","high-performance-inference"],"aiDims":["AI基础设施","模型部署","推理引擎"],"zone":"工具","funScore":0,"domainTags":["GPU管理","集群调度","vLLM","SGLang","DevOps"],"domainKey":"GPU管理","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"ascend","source":"github","weight":0.7},{"name":"cuda","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"distributed-inference","source":"github","weight":0.7},{"name":"genai","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"maas","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"qwen","source":"github","weight":0.7},{"name":"rocm","source":"github","weight":0.7},{"name":"vllm","source":"github","weight":0.7},{"name":"mindie","source":"github","weight":0.7},{"name":"llm-inference","source":"github","weight":0.7},{"name":"llm-serving","source":"github","weight":0.7},{"name":"sglang","source":"github","weight":0.7},{"name":"high-performance-inference","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/gpustack/gpustack","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Bing-su/adetailer","owner":"Bing-su","name":"adetailer","desc":"","summaryCn":"该项目解决的是 AI 生成图像中最令人头痛的“细节处理”问题。","reasonCn":"在安装方式上，它充分考虑了 WebUI 用户的习惯，只需将仓库地址粘贴进“Install from URL”面板，等待系统自动拉取并重启 UI，即可在扩展标签栏中看到它的身影。这种简单直接的安装方式，使其成为了社区里几乎“必装”的插件，也是检验 AI 绘图工作流是否完善的关键基准。","stars":4788,"starGrowth":0,"createdAt":"2023-04-26T07:54:51Z","silentRounds":1,"language":"Python","topics":["sd-webui","stable-diffusion-webui","stable-diffusion-webui-plugin"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI修图","自动修补","StableDiffusion插件"],"domainKey":"AI修图","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"sd-webui","source":"github","weight":0.7},{"name":"stable-diffusion-webui","source":"github","weight":0.7},{"name":"stable-diffusion-webui-plugin","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/Bing-su/adetailer","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"influxdata/influxdb","owner":"influxdata","name":"influxdb","desc":"Scalable datastore for metrics, events, and real-time analytics","summaryCn":"业界最流行的时序数据库，专门存储监控数据和实时事件。","reasonCn":"InfluxDB是专为时序数据设计的开源数据库，性能卓越，适合存储传感器数据、应用监控指标和实时事件流。支持百万级写入和复杂查询，配备InfluxQL和Flux查询语言。是Prometheus、Grafana等监控生态的核心组件，云原生时代基础设施必备。","stars":31754,"starGrowth":0,"createdAt":"2013-09-26T14:31:10Z","silentRounds":1,"language":"Rust","topics":["database","monitoring"],"aiDims":["AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["时序数据库","监控","物联网","InfluxQL"],"domainKey":"时序数据库","pushedAt":"2026-09-23T20:46:38Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/influxdata/influxdb","ts":"2026-09-23T20:46:38Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"smicallef/spiderfoot","owner":"smicallef","name":"spiderfoot","desc":"SpiderFoot automates OSINT for threat intelligence and mapping your attack surface.","summaryCn":"全自动搜集全网信息的情报搜集神器，安全人员必备","reasonCn":"这是一个开源的自动化开源网络情报搜集(OSINT)平台，能自动从多种数据源收集信息，构建攻击面图，帮助威胁情报分析师和渗透测试人员发现潜在风险。项目基于Python开发，支持多种数据源模块，无需复杂配置即可快速启动。","stars":22455,"starGrowth":1,"createdAt":"2012-04-28T07:10:13Z","silentRounds":1,"language":"Python","topics":["footprinting","osint","threatintel","python","infosec","intelligence-gathering","osint-reconnaissance","pentesting","threat-intelligence","security-tools","information-gathering","cybersecurity","cti","osint-framework","attacksurface","osint-tool","hacking","information-security","recon"],"aiDims":["安全工具","非AI-实用","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["OSINT","渗透测试","威胁情报"],"domainKey":"OSINT","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"footprinting","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"threatintel","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"infosec","source":"github","weight":0.7},{"name":"intelligence-gathering","source":"github","weight":0.7},{"name":"osint-reconnaissance","source":"github","weight":0.7},{"name":"pentesting","source":"github","weight":0.7},{"name":"threat-intelligence","source":"github","weight":0.7},{"name":"security-tools","source":"github","weight":0.7},{"name":"information-gathering","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"cti","source":"github","weight":0.7},{"name":"osint-framework","source":"github","weight":0.7},{"name":"attacksurface","source":"github","weight":0.7},{"name":"osint-tool","source":"github","weight":0.7},{"name":"hacking","source":"github","weight":0.7},{"name":"information-security","source":"github","weight":0.7},{"name":"recon","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/smicallef/spiderfoot","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"op7418/guizang-ppt-skill","owner":"op7418","name":"guizang-ppt-skill","desc":"AI-agent Skill for generating polished HTML slide decks: editorial magazine and Swiss layouts, image prompts, social covers, and a WebGL/low-power presentation runtime.","summaryCn":"让AI帮你自动生成精美HTML幻灯片的开源神器","reasonCn":"基于AI Agent驱动的幻灯片生成工具，支持编辑杂志风格和瑞士极简排版，内置智能图片提示词生成、社交媒体封面制作能力，并配备WebGL低功耗演示运行时。只需输入主题，即可一键产出可直接在浏览器中运行的交互式HTML演示文稿，无需安装PowerPoint等重型软件，彻底解放设计生产力。","stars":26794,"starGrowth":55,"createdAt":"2026-04-23T13:23:07Z","silentRounds":0,"language":"HTML","topics":["ai-agent"],"aiDims":["AI Agent","创意工具","AI应用"],"zone":"AI","funScore":1,"domainTags":["幻灯片生成","排版设计","WebGL演示","AI辅助"],"domainKey":"幻灯片生成","pushedAt":"2026-08-07T03:58:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/op7418/guizang-ppt-skill","ts":"2026-08-07T03:58:08Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.2053897141151004,"copyOk":false},{"repo":"memvid/memvid","owner":"memvid","name":"memvid","desc":"Memory layer for AI Agents. Replace complex RAG pipelines with a serverless, single-file memory layer. Give your agents instant retrieval and long-term memory.","summaryCn":"给AI Agent装即时记忆，用单机内存替代复杂RAG的Rust方案","reasonCn":"memvid是一个用Rust打造的AI Agent内存层，让Agent具备即时检索和长期记忆能力，无需维护复杂的RAG管道。它支持语义搜索和持久化存储，让AI记住对话历史、用户偏好和关键信息。一行命令即可运行，Python和Rust均可快速集成。适合本地部署AI Agent的团队、注重隐私的企业以及想快速验证想法的个人开发者。","stars":16555,"starGrowth":3,"createdAt":"2025-05-27T16:01:08Z","silentRounds":0,"language":"Rust","topics":["ai","context","embedded","faiss","knowledge-base","knowledge-graph","llm","machine-learning","memory","nlp","offline-first","opencv","python","rag","retrieval-augmented-generation","semantic-search","vector-database","video-processing","memvid","mv2"],"aiDims":["AI Agent","RAG与知识","向量数据库"],"zone":"AI","funScore":1,"domainTags":["向量检索","内存层","Rust","Agent"],"domainKey":"Agent","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"context","source":"github","weight":0.7},{"name":"embedded","source":"github","weight":0.7},{"name":"faiss","source":"github","weight":0.7},{"name":"knowledge-base","source":"github","weight":0.7},{"name":"knowledge-graph","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"memory","source":"github","weight":0.7},{"name":"nlp","source":"github","weight":0.7},{"name":"offline-first","source":"github","weight":0.7},{"name":"opencv","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"retrieval-augmented-generation","source":"github","weight":0.7},{"name":"semantic-search","source":"github","weight":0.7},{"name":"vector-database","source":"github","weight":0.7},{"name":"video-processing","source":"github","weight":0.7},{"name":"memvid","source":"github","weight":0.7},{"name":"mv2","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/memvid/memvid","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"reville/lighttable-digital-darkroom","owner":"reville","name":"lighttable-digital-darkroom","desc":"LightTable: a high performance digital darkroom for RAW photography with physical film simulation","summaryCn":"模拟胶片质感的本地高性能 RAW 暗房工具","reasonCn":"技术核心的亮点在于其“物理胶片模拟引擎”。普通的滤镜只是把颜色乘以一个矩阵，而 LightTable 试图建立胶片的科学模型。它理解银盐颗粒分布、染料层响应曲线等概念，因此在高光溢出、阴影死黑、肤色过渡等细节上，比泛用的调色软件更真实。","stars":21,"starGrowth":1,"createdAt":"2026-09-05T17:34:56Z","silentRounds":4,"language":"Python","topics":["photography"],"aiDims":["非AI-实用","创意工具"],"zone":"工具","funScore":0.5,"domainTags":["RAW摄影","数字暗房","胶片模拟","图像编辑","色彩科学"],"domainKey":"RAW摄影","pushedAt":"2026-09-19T12:11:14Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"photography","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/reville/lighttable-digital-darkroom","ts":"2026-09-19T12:11:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dragonflydb/dragonfly","owner":"dragonflydb","name":"dragonfly","desc":"A modern replacement for Redis and Memcached","summaryCn":"Redis平替，单机性能吊打Redis，程序员必备的高速缓存数据库","reasonCn":"这是Redis和Memcached的现代替代品，纯C++编写，多线程架构让单机性能远超Redis。完全兼容Redis协议，支持字符串、列表、集合等数据结构，还有原生pub/sub、流数据和过期key功能。Star近3.1万，性能测试中单机吞吐量可达Redis的数倍，适合需要高性能缓存和实时数据处理的场景。","stars":31661,"starGrowth":1,"createdAt":"2021-12-11T10:00:42Z","silentRounds":1,"language":"C++","topics":["database"],"aiDims":["AI基础设施","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["高性能缓存","分布式存储","数据库"],"domainKey":"高性能缓存","pushedAt":"2026-09-23T15:38:23Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/dragonflydb/dragonfly","ts":"2026-09-23T15:38:23Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dataease/dataease","owner":"dataease","name":"dataease","desc":"🔥 人人可用的开源 BI 工具，数据可视化神器。An open-source BI tool alternative to Tableau.","summaryCn":"国产开源BI神器，拖拽操作搞定企业级数据大屏","reasonCn":"DataEase是国内最受欢迎的开源BI工具，对标Tableau和PowerBI。它支持多种数据源连接，通过拖拽式操作即可生成复杂的图表和数据大屏。内置丰富的模板，操作门槛极低，让不懂代码的业务人员也能轻松完成数据分析与可视化报告。","stars":24534,"starGrowth":1,"createdAt":"2021-02-02T04:10:21Z","silentRounds":1,"language":"Java","topics":["data-visualization"],"aiDims":["AI应用","可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["BI工具","数据可视化","商业智能"],"domainKey":"BI工具","pushedAt":"2026-09-23T15:02:41Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/dataease/dataease","ts":"2026-09-23T15:02:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MoonshotAI/Kimi-K2","owner":"MoonshotAI","name":"Kimi-K2","desc":"","summaryCn":"Kimi K2在15.5万亿token的数据集上完成预训练，且实现零训练不稳定。","reasonCn":"模型架构方面，Kimi K2采用混合专家（MoE）设计，总参数量达1万亿，但每次推理仅激活320亿参数。这种稀疏架构既保证了模型的庞大知识容量，又控制了计算成本。结合专为智能体场景优化的设计，模型在工具调用、复杂推理和自主问题解决上表现出色，能够理解复杂指令并调用外部工具完","stars":11113,"starGrowth":0,"createdAt":"2025-07-03T12:28:22Z","silentRounds":5,"language":"","topics":[],"aiDims":["大语言模型","模型与训练","AI Agent"],"zone":"AI","funScore":0.7,"domainTags":["大模型","MoE","训练"],"domainKey":"训练","pushedAt":"2026-09-15T16:38:13.001Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1}],"aiScore":0.74,"source":"search","url":"https://github.com/MoonshotAI/Kimi-K2","ts":"2026-09-15T16:38:13.001Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"git-ai-project/git-ai","owner":"git-ai-project","name":"git-ai","desc":"A Git extension for tracking the AI-generated code in your repos","summaryCn":"给Git装上眼睛，追踪AI到底改了哪些代码","reasonCn":"这是一个Rust编写的Git扩展工具，专注于追踪代码库中由AI生成的部分。它通过分析代码的元数据和上下文，智能识别哪些文件是AI辅助修改的，从而帮助开发者更好地理解和管理AI参与的开发流程，保持代码库的透明度和可追溯性。","stars":2771,"starGrowth":0,"createdAt":"2025-07-02T16:09:26Z","silentRounds":2,"language":"Rust","topics":["ai","coding-agents","ai-blame"],"aiDims":["开发工具","非AI-实用","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["Git","AI 代码追踪","Rust 工具","开发者工具"],"domainKey":"Git","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"开发工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"coding-agents","source":"github","weight":0.7},{"name":"ai-blame","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/git-ai-project/git-ai","ts":"2026-09-16T07:27:53.690Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"anthropics/anthropic-sdk-python","owner":"anthropics","name":"anthropic-sdk-python","desc":"","summaryCn":"对于需要集成AI能力的Python开发者而言，anthropic-sdk-python 是连接应用与","reasonCn":"从安装与上手角度来看，该工具要求Python环境版本不低于3.10，确保了现代特性如类型提示和异步支持的稳定性。开发者只需通过包管理器安装依赖，并配置好从Anthropic控制台获取的API密钥，即可开始调用服务。SDK内部自动处理了请求的重试机制和错误码映射，使得代码在不同网络环境下","stars":3913,"starGrowth":2,"createdAt":"2023-01-17T20:57:10Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["Claude API","Python SDK","接口开发"],"domainKey":"Claude API","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/anthropics/anthropic-sdk-python","ts":"2026-09-18T20:41:13.636Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"datawhalechina/Agent-Learning-Hub","owner":"datawhalechina","name":"Agent-Learning-Hub","desc":"","summaryCn":"在 AI Agent 领域概念层出不穷的背景下，初学者往往容易陷入资料碎片化的困","reasonCn":"该指南的一个显著特点是其批判性的视角。在“现在该学什么”部分，它明确建议开发者不要将主要精力投入到那些已经泛化为模板的旧式角色扮演多 Agent 框架中，而是关注更贴近真实生产力的方向，如可靠的工具调用机制和闭环反馈系统。","stars":8134,"starGrowth":1,"createdAt":"2026-05-17T06:41:12Z","silentRounds":2,"language":"HTML","topics":[],"aiDims":["学习资源","AI Agent"],"zone":"资源","funScore":0,"domainTags":["AI Agent","学习路线","机器学习","教程"],"domainKey":"AI Agent","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/datawhalechina/Agent-Learning-Hub","ts":"2026-09-14T23:04:28.868Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PacktPublishing/LLM-Engineers-Handbook","owner":"PacktPublishing","name":"LLM-Engineers-Handbook","desc":"","summaryCn":"随着大语言模型在各行各业的普及，如何将模型从实验室推向生产环境成为了工","reasonCn":"项目内容覆盖了端到端大模型系统的全生命周期。在数据层面，它演示了如何进行高质量的数据收集与合成；在训练层面，提供了完整的 LLM 训练流水线代码；在应用层面，则包含了轻量级 RAG 系统的实现方案。这些内容不仅停留在理论教学，更通过实际可运行的代码加深读者的理解。","stars":5348,"starGrowth":1,"createdAt":"2024-04-09T10:56:03Z","silentRounds":3,"language":"Python","topics":["genai","llm","llmops","mlops","rag","aws","fine-tuning-llm","llm-evaluation","ml-system-design"],"aiDims":["学习资源","教程","AI基础设施"],"zone":"资源","funScore":0.2,"domainTags":["大语言模型工程","RAG系统","模型训练","生产部署"],"domainKey":"大语言模型工程","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"genai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"llmops","source":"github","weight":0.7},{"name":"mlops","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"aws","source":"github","weight":0.7},{"name":"fine-tuning-llm","source":"github","weight":0.7},{"name":"llm-evaluation","source":"github","weight":0.7},{"name":"ml-system-design","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/PacktPublishing/LLM-Engineers-Handbook","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bitsandbytes-foundation/bitsandbytes","owner":"bitsandbytes-foundation","name":"bitsandbytes","desc":"","summaryCn":"该库通过向量级量化实现 LLM.int8() 方法，能够以极低的内存开销完成复杂的","reasonCn":"核心亮点在于它提供的 8-bit 优化器，通过分块量化技术，以极小的内存代价维持了传统 32-bit 的计算精度，这对于进行模型训练和精调的任务至关重要。特别是在 QLoRA 技术支持下，它大幅降低了微调大模型所需的硬件门槛，使得中小型项目组甚至独立开发者都能根据自身数据定制专属的","stars":8499,"starGrowth":1,"createdAt":"2021-06-04T00:10:34Z","silentRounds":2,"language":"Python","topics":["llm","machine-learning","pytorch","qlora","quantization"],"aiDims":["模型部署","微调","模型与训练"],"zone":"AI","funScore":0,"domainTags":["模型量化","深度学习","内存优化","QLoRA"],"domainKey":"模型量化","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"qlora","source":"github","weight":0.7},{"name":"quantization","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"tier","url":"https://github.com/bitsandbytes-foundation/bitsandbytes","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chatanywhere/GPT_API_free","owner":"chatanywhere","name":"GPT_API_free","desc":"免费大模型API，支持免费调用GPT、DeepSeek等主流模型，免费额度10000点，每日刷新！另付费价格最低官方1-2折！","summaryCn":"用极低价格调用GPT和DeepSeek等大模型的神器，每日免费额度超多","reasonCn":"这是一个旨在降低大模型API使用门槛的开源项目，通过聚合多个主流模型服务商的接口，提供统一的调用方式。它内置了GPT、DeepSeek、GLM等主流大模型，支持API Key的自动轮换和负载均衡。项目提供了超过10000点的每日免费额度，且额度每日自动刷新，极大地降低了用户的试用成本。同时，它还支持自定义API Key，允许用户接入其他服务商，甚至提供官方价格的1-2折优惠方案，是目前性价比极高的API代理服务。","stars":43045,"starGrowth":0,"createdAt":"2023-04-23T12:42:08Z","silentRounds":1,"language":"","topics":["api"],"aiDims":["AI应用","AI基础设施","大语言模型"],"zone":"AI","funScore":0,"domainTags":["API聚合","大模型","OpenAI","DeepSeek","中转服务"],"domainKey":"API聚合","pushedAt":"2026-09-07T17:19:11Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7}],"aiScore":0.95,"source":"search","url":"https://github.com/chatanywhere/GPT_API_free","ts":"2026-09-07T17:19:11Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bilibili/ailab","owner":"bilibili","name":"ailab","desc":"","summaryCn":"Real-CUGAN 的独特之处在于其数据集的直接来源。","reasonCn":"从技术架构上看，它采用了级联 U-Net 设计。这种结构允许网络分阶段处理图像，逐步提取特征并恢复细节。相比传统的单阶段大网络，级联结构在保持同等恢复效果的同时，显著降低了计算复杂度，使得在普通 GPU 上实现实时或接近实时的视频处理成为可能。","stars":5873,"starGrowth":0,"createdAt":"2021-05-07T03:03:28Z","silentRounds":6,"language":"Python","topics":[],"aiDims":["AI应用","非AI-好玩"],"zone":"AI","funScore":1,"domainTags":["视频超分","动漫","深度学习"],"domainKey":"视频超分","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/bilibili/ailab","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mlfoundations/open_clip","owner":"mlfoundations","name":"open_clip","desc":"An open source implementation of CLIP.","summaryCn":"你有没有想过，为什么AI能通过一句话就画出一张精美的图？","reasonCn":"它厉害的地方在于，通过海量的互联网图像和文字对进行对比学习，模型成功学会了“猫”这个词对应什么样的图像特征，以及“抽象派画作”代表什么样的艺术风格。相比以前那种僵硬的图像分类，它具备惊人的泛化能力，即使是它从未见过的图片类型，只要描述准确，它也能给出极高的匹配度。","stars":14161,"starGrowth":5,"createdAt":"2021-07-28T23:24:39Z","silentRounds":0,"language":"Python","topics":["deep-learning"],"aiDims":["模型与训练","深度学习","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["CLIP","多模态","视觉模型","开源实现"],"domainKey":"多模态","pushedAt":"2026-09-08T19:29:50Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"深度学习","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/mlfoundations/open_clip","ts":"2026-09-08T19:29:50Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.016059247228303083,"copyOk":false},{"repo":"thebriangao/totem","owner":"thebriangao","name":"totem","desc":"Model Context Protocol server giving Claude (or any MCP client) full read + write access to your Whoop fitness data via the private reverse-engineered iOS API. 55 tools: recovery, sleep, strain, HRV trends, Strength Trainer, journal, Whoop Coach, and smart alarm. TypeScript + zod, auto-refresh Cognito auth.","summaryCn":"很多人都有个烦恼：手环里数据不少，但真正想让它帮忙分析时，又得自己截图","reasonCn":"它的技术亮点是通过模型上下文协议暴露一套工具，让支持该协议的助手能查询和写入相关数据。数据维度包括恢复状态、睡眠质量、身体压力、心率变异性等，覆盖日常健身常见指标。这种设计的好处是，助手不只是读一次数据，还能按你的提问组合多个维度做判断。","stars":119,"starGrowth":0,"createdAt":"2026-05-26T20:23:35Z","silentRounds":20,"language":"TypeScript","topics":["fitness"],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":0.5,"domainTags":["MCP协议","健身数据","Whoop集成","健康追踪"],"domainKey":"MCP协议","pushedAt":"2026-09-16T00:45:26Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/thebriangao/totem","ts":"2026-09-16T00:45:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"QuiteAFancyEmerald/InvisiProxy","owner":"QuiteAFancyEmerald","name":"InvisiProxy","desc":"InvisiProxy LTS is a web proxy service that helps you access websites that may be blocked by your network, government or policy all within your browser with no download or setup. It does this securely and with additional privacy features. Browse Tor/Onion sites in any browser, hide browsing activity and bypass filters. (Star if you fork it!)","summaryCn":"不用下载配置就能直接在浏览器里科学上网的网页版代理服务。","reasonCn":"技术层面上，它巧妙地利用了现代Web技术和代理转发机制。当你在界面里输入网址后，服务会在云端代为请求目标网站，并把页面的HTML、CSS和JavaScript等资源进行安全清洗和重写，再实时渲染到你的当前浏览器中。它不仅帮你绕过了网络封锁，还内置了流量加密和隐私保护机制，确保你的","stars":1576,"starGrowth":1,"createdAt":"2020-09-10T19:18:23Z","silentRounds":1,"language":"JavaScript","topics":["nodejs","javascript","fastify","adblocking","bypass","bypass-recaptchav3","https-proxy","onion-service","proxy","socks5-proxy","unblocker","webproxy","tor","vpn","privacy","docker","unblock","web-proxy","anonymity","holyunblocker"],"aiDims":["非AI-实用","安全工具"],"zone":"创意","funScore":1,"domainTags":["网页代理","浏览器插件","网络访问","前端渲染","隐私保护"],"domainKey":"网页代理","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"nodejs","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"fastify","source":"github","weight":0.7},{"name":"adblocking","source":"github","weight":0.7},{"name":"bypass","source":"github","weight":0.7},{"name":"bypass-recaptchav3","source":"github","weight":0.7},{"name":"https-proxy","source":"github","weight":0.7},{"name":"onion-service","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"socks5-proxy","source":"github","weight":0.7},{"name":"unblocker","source":"github","weight":0.7},{"name":"webproxy","source":"github","weight":0.7},{"name":"tor","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"unblock","source":"github","weight":0.7},{"name":"web-proxy","source":"github","weight":0.7},{"name":"anonymity","source":"github","weight":0.7},{"name":"holyunblocker","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/QuiteAFancyEmerald/InvisiProxy","ts":"2026-09-12T22:00:03.438Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NielsRogge/Transformers-Tutorials","owner":"NielsRogge","name":"Transformers-Tutorials","desc":"","summaryCn":"本仓库主要解决初学者面对复杂的 HuggingFace 生态系统时无从下手的问题，通","reasonCn":"项目最大的亮点在于其代码与官方库的紧密结合，所有演示均由社区维护者精心编写并统一采用 PyTorch 实现。不仅展示了模型本身的推理过程，还涉及了 Tokenizers、Datasets 以及 Accelerate 等核心周边工具的使用方法，确保开发者能够全方位掌握模型训练与部署的完整流程。","stars":11754,"starGrowth":0,"createdAt":"2020-08-31T18:00:58Z","silentRounds":2,"language":"Jupyter Notebook","topics":["transformers"],"aiDims":["教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["Transformers","深度学习教程","HuggingFace","Jupyter"],"domainKey":"Transformers","pushedAt":"2026-09-17T08:10:06Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/NielsRogge/Transformers-Tutorials","ts":"2026-09-17T08:10:06Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HolonProduction/godot_kanban_tasks","owner":"HolonProduction","name":"godot_kanban_tasks","desc":"Another kanban board plugin for the godot engine.","summaryCn":"Godot引擎插件，把任务管理直接做进游戏里，做游戏也能理清思路。","reasonCn":"它的核心优势是“上下文一体化”。你不用切到网页或另一个软件去看任务列表，在Godot里选中一个关卡节点，就能打开看板面板。你可以把“设计关卡”、“编写脚本”、“贴图资源”这些任务拖拽到不同的列（比如“待办”、“进行中”、“已完成”），进度一目了然。","stars":260,"starGrowth":0,"createdAt":"2022-02-20T16:08:42Z","silentRounds":12,"language":"GDScript","topics":["godot"],"aiDims":["游戏","创意工具"],"zone":"工具","funScore":0,"domainTags":["任务管理","Godot引擎","看板工具"],"domainKey":"任务管理","pushedAt":"2026-09-11T19:04:48Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/HolonProduction/godot_kanban_tasks","ts":"2026-09-11T19:04:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kimiyoung/transformer-xl","owner":"kimiyoung","name":"transformer-xl","desc":"","summaryCn":"该模型不仅是简单的序列处理工具，更通过在训练过程中保留前一段的隐藏状态","reasonCn":"在技术实现上，该仓库提供了基于 PyTorch 和 TensorFlow 的完整训练与推理代码，支持单机多卡以及多机 TPU 集群训练模式。这种底层架构的完备性，让它不仅是论文中的理论验证，更是可以直接应用于大规模语言模型预训练的实用引擎。","stars":3715,"starGrowth":0,"createdAt":"2019-01-08T12:20:24Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["深度学习","自然语言处理","循环神经网络"],"domainKey":"深度学习","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/kimiyoung/transformer-xl","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/DeepSeek-R1","owner":"deepseek-ai","name":"DeepSeek-R1","desc":"","summaryCn":"该项目由深度求索团队打造，其核心突破在于探索了不依赖传统监督微调、直接","reasonCn":"然而，早期纯强化学习模型在实际表现中面临着无休止重复、可读性差以及多种语言混杂等现实挑战。为了解决这些痛点并进一步提升模型的综合推理表现，研发团队推出了正式的 DeepSeek-R1 模型，在强化学习正式开启之前巧妙地融入了冷启动数据进行预处理。","stars":91974,"starGrowth":0,"createdAt":"2025-01-20T11:57:28Z","silentRounds":4,"language":"","topics":[],"aiDims":["推理引擎","大语言模型","模型与训练"],"zone":"AI","funScore":1,"domainTags":["大语言模型","强化学习","推理模型","开源模型"],"domainKey":"大语言模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1}],"aiScore":0.788,"source":"search","url":"https://github.com/deepseek-ai/DeepSeek-R1","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Acly/krita-ai-diffusion","owner":"Acly","name":"krita-ai-diffusion","desc":"","summaryCn":"通过在Krita中加入对深度图、线稿及参考图的实时引导支持，该插件打破了传统","reasonCn":"该项目的核心亮点在于其强调的「精准与控制」，用户不仅能通过简单的文字生成全图，还能通过对画面的区域权重和强度的自由调节，实现对画作细节的微观干预。无论你是想用AI为草图快速上色，还是在复杂的插画中修正某一小块光影，它提供的生成工具都能在不破坏整体风格的前提下精准","stars":10626,"starGrowth":1,"createdAt":"2023-09-01T22:52:13Z","silentRounds":1,"language":"Python","topics":["generative-ai","krita-plugin","stable-diffusion"],"aiDims":["AI应用","创意工具"],"zone":"工具","funScore":0.8,"domainTags":["AI绘画","Krita","图像生成"],"domainKey":"AI绘画","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"generative-ai","source":"github","weight":0.7},{"name":"krita-plugin","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/Acly/krita-ai-diffusion","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"UX-Decoder/Segment-Everything-Everywhere-All-At-Once","owner":"UX-Decoder","name":"Segment-Everything-Everywhere-All-At-Once","desc":"","summaryCn":"SEEM不仅是一个图像处理模型，它更像是一个拥有多感官能力的智能助手。","reasonCn":"该项目的核心亮点在于其极强的泛化能力。无论是复杂的工业场景，还是日常生活中的照片，它都能通过多样的提示方式给出令人惊叹的结果。研究团队在论文中展示了其如何处理各类视觉任务，这种能力不仅限于简单的物体识别，还涵盖了更高级的图像语义理解。","stars":4791,"starGrowth":0,"createdAt":"2023-04-12T05:08:43Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["AI应用","多模态Agent","文档理解"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["图像分割","多模态模型","交互式智能"],"domainKey":"图像分割","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/UX-Decoder/Segment-Everything-Everywhere-All-At-Once","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mattpocock/skills","owner":"mattpocock","name":"skills","desc":"Skills for Real Engineers. Straight from my .agents directory.","summaryCn":"AI编程专用技能库，让Claude等AI写出更高质量代码","reasonCn":"这是由知名开发者Matt Pocock整理的一套Prompt技能库，专门用于提升AI编程助手（如Claude Code）的代码生成质量。它不是传统意义上的软件，而是一组经过精心设计的指令模板，能指导AI避免常见错误，生成更符合工程标准的代码，是提升AI coding效率的实用利器。","stars":267830,"starGrowth":778,"createdAt":"2026-02-03T11:15:53Z","silentRounds":0,"language":"Shell","topics":[],"aiDims":["提示工程","代码助手","AI应用"],"zone":"AI","funScore":0,"domainTags":["Prompt工程","AI编程","Claude Code"],"domainKey":"Prompt工程","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/mattpocock/skills","ts":"2026-09-07T02:44:20.540Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.370457379681141,"copyOk":false},{"repo":"Picovoice/picollm","owner":"Picovoice","name":"picollm","desc":"On-device LLM Inference Powered by X-Bit Quantization","summaryCn":"用极度量化技术在普通笔记本上跑大模型，不用云端不花电费","reasonCn":"该项目采用X-bit超低位宽量化技术，将大语言模型的权重和激活值压缩到极致，支持在CPU或低功耗设备上本地推理。相比llama.cpp等方案，它在极低比特下保持更好的推理质量。核心创新在于残差缩放和动态量化范围算法，适合资源受限的边缘计算场景。","stars":318,"starGrowth":0,"createdAt":"2024-04-09T21:23:22Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["推理引擎","模型部署","大语言模型"],"zone":"AI","funScore":1,"domainTags":["LLM推理","模型量化","边缘计算","本地AI"],"domainKey":"LLM推理","pushedAt":"2026-09-11T22:27:56Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Picovoice/picollm","ts":"2026-09-11T22:27:56Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"entireio/cli","owner":"entireio","name":"cli","desc":"📜 Entire CLI hooks into your Git workflow to capture AI agent sessions as you work. Sessions are indexed alongside commits, creating a searchable record of how code was written in your repo.","summaryCn":"把AI写代码过程录下来，像Git提交记录一样存起来。","reasonCn":"这是一个嵌入Git工作流的CLI工具，能自动捕获AI Agent的会话。它将交互过程与代码提交关联索引，形成可搜索的记录。开发者能追踪AI是如何一步步修改代码的，甚至回溯对话上下文。这解决了AI辅助开发黑盒化的问题，让代码审查和团队协作更有据可依。","stars":5112,"starGrowth":4,"createdAt":"2026-01-02T17:13:58Z","silentRounds":0,"language":"Go","topics":["agents","ai","claude","developer","developer-platform","gemini"],"aiDims":["AI Agent","代码助手","AI基础设施"],"zone":"工具","funScore":0.6,"domainTags":["Git钩子","AI会话追踪","代码审查","CLI工具"],"domainKey":"Git钩子","pushedAt":"2026-09-07T18:28:48.958Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"developer","source":"github","weight":0.7},{"name":"developer-platform","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/entireio/cli","ts":"2026-09-07T18:28:48.958Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"waditu/czsc","owner":"waditu","name":"czsc","desc":"缠中说禅技术分析工具；缠论；股票；期货；Quant；量化交易","summaryCn":"缠论是A股市场里一套非常独特且复杂的技术分析理论，对于很多交易者来说，掌","reasonCn":"这个工具最核心的功能是帮你识别走势类型、笔、线段、中枢等缠论核心概念。你只需要提供股票代码和日期范围，它就能自动计算出所有的技术指标，并生成可视化的图表。这对于那些想要深入研究缠论，或者想利用缠论逻辑进行量化交易的投资者来说，非常有帮助。","stars":6256,"starGrowth":1,"createdAt":"2019-06-11T07:38:57Z","silentRounds":1,"language":"Rust","topics":["tushare","czsc","quant"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["股票","期货","量化交易","技术分析","缠论"],"domainKey":"股票","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"tushare","source":"github","weight":0.7},{"name":"czsc","source":"github","weight":0.7},{"name":"quant","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/waditu/czsc","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"open-multi-agent/open-multi-agent","owner":"open-multi-agent","name":"open-multi-agent","desc":"","summaryCn":"自托管 AI Agent 运行时，支持私域模型接入与行为可审计。","reasonCn":"该工具构建了严格的审计机制。每一个涉及重大后果的 Agent 行为在被执行前，都必须经过一个持久化且防篡改的审批流程。系统会为每一次运行生成经过字节级校验的记录，这意味着所有的决策路径都可以被离线验证，为企业级应用提供了符合审计标准的追溯能力。","stars":6951,"starGrowth":2,"createdAt":"2026-03-31T19:07:34Z","silentRounds":0,"language":"TypeScript","topics":["agent-framework","ai-agents","claude","llm","multi-agent","openai","typescript","anthropic","ollama","deepseek","gemini","local-llm","mcp","agentic-ai","claude-code","ai-governance","observability","self-hosted","agent-orchestration"],"aiDims":["AI Agent","模型部署","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI Agent","自托管","审计"],"domainKey":"AI Agent","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agent-framework","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"multi-agent","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"local-llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"ai-governance","source":"github","weight":0.7},{"name":"observability","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"agent-orchestration","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/open-multi-agent/open-multi-agent","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"winfunc/deepreasoning","owner":"winfunc","name":"deepreasoning","desc":"","summaryCn":"该平台的核心逻辑是将 DeepSeek R1 的深度思考与推理优势，与 Anthropic Cl","reasonCn":"在使用方式上，项目提供了标准化的统一 API，这意味着你现有的应用无需重构代码即可接入这套增强的推理能力。它同时配套了简洁直观的聊天界面，方便你直接进行测试或辅助日常工作。该项目对于那些希望在开发过程中获得模型逻辑思考辅助，又同时需要高质量工程代码输出的用户尤为适","stars":5364,"starGrowth":1,"createdAt":"2025-01-26T16:25:41Z","silentRounds":3,"language":"Rust","topics":["ai","anthropic","anthropic-claude","api","chain-of-thought","claude","deepseek","deepseek-r1","llm","rust"],"aiDims":["推理引擎","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["API","DeepSeek","Claude"],"domainKey":"API","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"anthropic-claude","source":"github","weight":0.7},{"name":"api","source":"github","weight":0.7},{"name":"chain-of-thought","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"deepseek-r1","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/winfunc/deepreasoning","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KoljaB/RealtimeSTT","owner":"KoljaB","name":"RealtimeSTT","desc":"","summaryCn":"RealtimeSTT 是一款专为实时语音转文字设计的 Python 库。","reasonCn":"该项目的核心亮点在于其高效性和易用性。它采用 `faster_whisper` 引擎，这是一个基于 OpenAI Whisper 模型的优化实现，能够提供快速且准确的转录服务。对于需要 GPU 加速的用户，项目推荐使用 CUDA 设置，这可以进一步提升处理速度。","stars":10143,"starGrowth":2,"createdAt":"2023-08-29T17:58:28Z","silentRounds":0,"language":"Python","topics":["python","realtime","speech-to-text"],"aiDims":["推理引擎"],"zone":"AI","funScore":0,"domainTags":["语音识别","实时转录","Whisper模型","自然语言处理"],"domainKey":"语音识别","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"python","source":"github","weight":0.7},{"name":"realtime","source":"github","weight":0.7},{"name":"speech-to-text","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/KoljaB/RealtimeSTT","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/DeepSeek-VL","owner":"deepseek-ai","name":"DeepSeek-VL","desc":"","summaryCn":"开源视觉语言模型，能处理逻辑图、网页、公式和科学文献等多模态理解任务","reasonCn":"模型的核心亮点在于其跨模态理解能力。它不仅能识别图像中的文字和图表，还能理解复杂的逻辑关系、数学公式和科学文献的结构。这种能力让它区别于普通的图像描述模型，更接近人类对图文信息的综合理解方式。对于需要处理混合内容的场景，比如读取技术文档、解析实验数据或分析工程","stars":4181,"starGrowth":0,"createdAt":"2024-03-07T08:32:57Z","silentRounds":4,"language":"Python","topics":["vision-language-model","vision-language-pretraining","foundation-models"],"aiDims":["多模态Agent","模型与训练","大语言模型"],"zone":"AI","funScore":0.5,"domainTags":["视觉语言模型","多模态理解","公式识别","逻辑图解析","科学文献"],"domainKey":"视觉语言模型","pushedAt":"2026-09-18T00:01:25.926Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"vision-language-model","source":"github","weight":0.7},{"name":"vision-language-pretraining","source":"github","weight":0.7},{"name":"foundation-models","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/deepseek-ai/DeepSeek-VL","ts":"2026-09-18T00:01:25.926Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Kaliiiiiiiiii-Vinyzu/patchright","owner":"Kaliiiiiiiiii-Vinyzu","name":"patchright","desc":"Undetected version of the Playwright testing and automation library.","summaryCn":"无痕Playwright库，绕过安全检测，保持自动化脚本隐蔽运行。","reasonCn":"patchright是Playwright的改写版，利用动态加载和混淆技术规避EDR、AV等安全产品的检测，实现浏览器自动化时保持进程隐身，兼容原生API，支持Chromium、Firefox、WebKit，可直接替换Playwright在CI、爬虫和渗透测试中的使用。","stars":4664,"starGrowth":3,"createdAt":"2023-11-07T19:50:56Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["开发者工具","安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["浏览器自动化","Playwright","反检测"],"domainKey":"浏览器自动化","pushedAt":"2026-09-13T12:47:17Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/Kaliiiiiiiiii-Vinyzu/patchright","ts":"2026-09-13T12:47:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lkwq007/stablediffusion-infinity","owner":"lkwq007","name":"stablediffusion-infinity","desc":"","summaryCn":"在无限画布上不断向外扩展生成图像，AI帮你画出看不见的边界。","reasonCn":"这是一个基于Stable Diffusion的图像外扩工具，让你在无限大的画布上自由挥洒。你可以把一张局部图片不断向外延展，AI会自动补全画面周围的内容，形成浑然一体的全景图。它已经打包成网页应用，浏览器打开就能用。适合喜欢AI绘画、做概念设计、或者单纯想玩玩看AI想象力的朋友。","stars":3890,"starGrowth":0,"createdAt":"2022-09-02T05:07:50Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["AI应用"],"zone":"创意","funScore":0.68,"domainTags":["图像生成","无限画布","图像外扩"],"domainKey":"图像生成","pushedAt":"2026-09-18T20:41:13.636Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/lkwq007/stablediffusion-infinity","ts":"2026-09-18T20:41:13.636Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"RollingGo-AI/RollingGo-hotel-MCP-CN","owner":"RollingGo-AI","name":"RollingGo-hotel-MCP-CN","desc":"RollingGo官方酒店MCP，14年旅游科技数据栈，200万酒店实时库存直连，11万直签，渠道专属价，兼容40+主流大模型代理。企业和个人开发者均可免费申请，无调用上限。可按地点、日期、星级、人数、标签等条件搜索酒店，返回候选酒店及展示价。可查询指定酒店的实时房型列表、价格计划和取消政策。","summaryCn":"计划一次旅行，最头疼的往往是找合适的酒店并且确保价格实时更新。","reasonCn":"计划一次旅行，最头疼的往往是找合适的酒店并且确保价格实时更新。RollingGo‑hotel‑MCP‑CN把这件事交给了AI，只要提供地点、入住日期、星级、人数等参数，它就能在几毫秒内返回包含房型、实时库存、最新价格和取消政策的完整列表，让用户无需手动比价。","stars":31,"starGrowth":1,"createdAt":"2026-08-05T07:16:48Z","silentRounds":3,"language":"Python","topics":["travel"],"aiDims":["AI应用","工具调用","AI Agent"],"zone":"AI","funScore":0,"domainTags":["MCP","酒店搜索","实时库存","大模型代理"],"domainKey":"MCP","pushedAt":"2026-09-08T03:34:41Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"travel","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.93,"source":"search","url":"https://github.com/RollingGo-AI/RollingGo-hotel-MCP-CN","ts":"2026-09-08T03:34:41Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"agentscope-ai/agentscope-java","owner":"agentscope-ai","name":"agentscope-java","desc":"","summaryCn":"内置了31种 typed events 的事件系统，不仅能实现智能体状态的实时前端渲染","reasonCn":"其核心亮点在于对安全性的极致把控，通过内置的 Tool-call gating 系统，可以实现对智能体行为的实时拦截，无论是要求人工审批还是直接禁止，都能保障大模型在执行关键操作时的可控性。这种设计让企业在面对 AI 可能产生的幻觉或风险时，拥有了强有力的防御手段。","stars":5744,"starGrowth":0,"createdAt":"2025-09-23T04:37:43Z","silentRounds":2,"language":"Java","topics":["agent","agentic","agentic-ai","ai","llm","adk","agent-framework","agentscope","harness"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.2,"domainTags":["企业级Agent","分布式系统","沙箱环境"],"domainKey":"企业级Agent","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"agentic","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"adk","source":"github","weight":0.7},{"name":"agent-framework","source":"github","weight":0.7},{"name":"agentscope","source":"github","weight":0.7},{"name":"harness","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/agentscope-ai/agentscope-java","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lm-sys/FastChat","owner":"lm-sys","name":"FastChat","desc":"","summaryCn":"作为 Chatbot Arena 的核心驱动力，FastChat 已经成功支撑了超过一千万次以","reasonCn":"该项目最硬核的价值在于其独特的评估体系，通过收集超过一百五十万条人类侧边对比投票，它建立了一套透明且实时的模型排名机制。这种机制有效解决了模型评估中常见的自吹自擂问题，让排名不再依赖单一的测试集，而是基于真实的交互反馈。","stars":39549,"starGrowth":2,"createdAt":"2023-03-19T00:18:02Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["大语言模型","模型部署","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["大语言模型","聊天机器人","模型评测","分布式服务"],"domainKey":"大语言模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/lm-sys/FastChat","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"skylersaleh/SkyEmu","owner":"skylersaleh","name":"SkyEmu","desc":"Game Boy Advance, Game Boy, Game Boy Color, and DS Emulator","summaryCn":"掌机模拟器SkyEmu，在PC上重温经典游戏。","reasonCn":"它的亮点在于多系统支持和硬件层面的优化。作为DS模拟器，它需要同时模拟ARM7和ARM9双核CPU，以及触摸屏和双屏显示，SkyEmu在这方面表现出了不错的兼容性，能流畅运行《宝可梦》、《马里奥赛车》等大作。作为GB/GBC/GBA模拟器，它则继承了模拟器的经典优势，画面渲染清晰，操作手感","stars":1268,"starGrowth":0,"createdAt":"2021-05-21T00:49:33Z","silentRounds":1,"language":"C","topics":["emulator"],"aiDims":["非AI-好玩","游戏","创意工具"],"zone":"工具","funScore":0.6,"domainTags":["GBA模拟器","DS模拟器","复古游戏","Game Boy"],"domainKey":"GBA模拟器","pushedAt":"2026-09-11T08:18:33Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/skylersaleh/SkyEmu","ts":"2026-09-11T08:18:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"viarotel-org/escrcpy","owner":"viarotel-org","name":"escrcpy","desc":"📱 Display and control your Android device graphically with scrcpy.","summaryCn":"不用拿手机就能在电脑屏幕上直接玩转安卓设备的投屏神仙工具","reasonCn":"技术层面上，它是基于大名鼎鼎的开源投屏项目 scrcpy 打造的图形化界面版本。原版的 scrcpy 虽然强悍，但需要通过命令行敲指令来启动和调节参数，对普通用户不够友好。这个项目用 Electron 和 Vue 等现代前端技术，给它套上了一个极其美观且直观的桌面控制面板。","stars":11878,"starGrowth":2,"createdAt":"2023-09-15T02:42:13Z","silentRounds":0,"language":"JavaScript","topics":["scrcpy","linux","macos","recording","windows","genymobile","gnirehtet","android","scrcpy-gui","screenshots","hacktoberfest","scrcpy-engine","gui","screensharing","mirroring","hacktoberfest2026","mapping","hacktoberfest2027"],"aiDims":["效率工具","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["安卓投屏","设备控制","远程桌面"],"domainKey":"安卓投屏","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"scrcpy","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"recording","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"genymobile","source":"github","weight":0.7},{"name":"gnirehtet","source":"github","weight":0.7},{"name":"android","source":"github","weight":0.7},{"name":"scrcpy-gui","source":"github","weight":0.7},{"name":"screenshots","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"scrcpy-engine","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"screensharing","source":"github","weight":0.7},{"name":"mirroring","source":"github","weight":0.7},{"name":"hacktoberfest2026","source":"github","weight":0.7},{"name":"mapping","source":"github","weight":0.7},{"name":"hacktoberfest2027","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/viarotel-org/escrcpy","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"stacklok/toolhive","owner":"stacklok","name":"toolhive","desc":"ToolHive is an enterprise-grade platform for running and managing Model Context Protocol (MCP) servers.","summaryCn":"给你的 AI 助手打造一个企业级能力插件管理中心","reasonCn":"专注于 Model Context Protocol 协议的平台，支持大规模管理和运行 MCP 服务器。通过 Go 语言实现高性能调度，提供统一的接口让大模型能够安全、稳定地调用各种外部工具，解决了 Agent 在企业环境下部署工具链碎片化的问题。","stars":2206,"starGrowth":2,"createdAt":"2025-03-12T14:49:15Z","silentRounds":0,"language":"Go","topics":["ai","aicodeassistant","mcp","ai-security","mcp-security","mcp-servers","mcp-tools","security","golang","kubernetes","model-context-protocol"],"aiDims":["工具调用","AI Agent","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["MCP协议","AI助手","插件管理","企业级","工具集成"],"domainKey":"MCP协议","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"aicodeassistant","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"ai-security","source":"github","weight":0.7},{"name":"mcp-security","source":"github","weight":0.7},{"name":"mcp-servers","source":"github","weight":0.7},{"name":"mcp-tools","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"model-context-protocol","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/stacklok/toolhive","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"agent-infra/sandbox","owner":"agent-infra","name":"sandbox","desc":"","summaryCn":"很多开发者在构建能够自主操作电脑或执行复杂任务的 AI Agent 时，都会遇到","reasonCn":"在核心亮点方面，该环境集成了网页浏览器、命令行终端、文件管理、VSCode、Jupyter 笔记本以及模型上下文协议等多种实用工具。这意味着开发者无需再费时费力地手动拼凑各种依赖组件，只需通过简单的容器部署命令，就能在极短时间内获得一个麻雀虽小五脏俱全的操作空间，满足智能体","stars":5990,"starGrowth":0,"createdAt":"2025-08-06T14:51:05Z","silentRounds":1,"language":"Python","topics":["agent","all-in-one","sandbox","browser","filesystem","mcp","shell"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI智能体","沙箱","容器"],"domainKey":"AI智能体","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"all-in-one","source":"github","weight":0.7},{"name":"sandbox","source":"github","weight":0.7},{"name":"browser","source":"github","weight":0.7},{"name":"filesystem","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"shell","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/agent-infra/sandbox","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0x4m4/hexstrike-ai","owner":"0x4m4","name":"hexstrike-ai","desc":"HexStrike AI MCP Agents is an advanced MCP server that lets AI agents (Claude, GPT, Copilot, etc.) autonomously run 150+ cybersecurity tools for automated pentesting, vulnerability discovery, bug bounty automation, and security research. Seamlessly bridge LLMs with real-world offensive security capabilities.","summaryCn":"让AI特工自动接管150+黑客工具，秒变全自动渗透测试机器人","reasonCn":"这是一个MCP服务器项目，将Claude、GPT等大模型与150多个网络安全工具连接起来。模型不再只是聊天，而是能直接调用系统命令执行漏洞扫描、端口探测、漏洞验证等操作。用户只需输入自然语言指令，AI就能像专业红队一样执行完整的渗透测试流程，大幅提升安全测试效率。","stars":12048,"starGrowth":0,"createdAt":"2025-07-10T17:53:18Z","silentRounds":1,"language":"Python","topics":["0x4m4","ai","ai-agents","ai-cybersecurity","ai-hacking","ai-penetration-testing","ai-security-tool","artificial-intelligence","ctf-tools","generative-ai","kali-linux","kali-tools","llm","llm-integration","mcp","mcp-server","mcp-tools","pentesting","pentesting-tools","hexstrike"],"aiDims":["AI Agent","工具调用","安全工具"],"zone":"AI","funScore":0,"domainTags":["渗透测试","MCP服务器","安全扫描","漏洞发现","自动化测试"],"domainKey":"渗透测试","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"0x4m4","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-cybersecurity","source":"github","weight":0.7},{"name":"ai-hacking","source":"github","weight":0.7},{"name":"ai-penetration-testing","source":"github","weight":0.7},{"name":"ai-security-tool","source":"github","weight":0.7},{"name":"artificial-intelligence","source":"github","weight":0.7},{"name":"ctf-tools","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"kali-linux","source":"github","weight":0.7},{"name":"kali-tools","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"llm-integration","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"mcp-tools","source":"github","weight":0.7},{"name":"pentesting","source":"github","weight":0.7},{"name":"pentesting-tools","source":"github","weight":0.7},{"name":"hexstrike","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/0x4m4/hexstrike-ai","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"meilisearch/meilisearch-dotnet","owner":"meilisearch","name":"meilisearch-dotnet","desc":".NET client for Meilisearch","summaryCn":"第一段：想在自己的网站或内部系统里装个高性能、支持模糊和向量检索的搜索","reasonCn":"第一段：想在自己的网站或内部系统里装个高性能、支持模糊和向量检索的搜索功能，却不想自己写一堆 HTTP 请求和 JSON 解析代码？Meilisearch‑dotnet 正是为了解决这个痛点而生。它把 Meilisearch 的全部能力包装成 .NET 友好的类库，让即使不熟悉搜索引擎底层实现的开发者，也能几","stars":345,"starGrowth":0,"createdAt":"2020-05-24T08:59:06Z","silentRounds":15,"language":"C#","topics":["vector-db"],"aiDims":["向量数据库","AI搜索","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Meilisearch","搜索SDK","C#","向量搜索","全文检索"],"domainKey":"Meilisearch","pushedAt":"2026-09-16T07:31:19Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/meilisearch/meilisearch-dotnet","ts":"2026-09-16T07:31:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/solr","owner":"apache","name":"solr","desc":"Apache Solr open-source search software","summaryCn":"给海量数据装上的超级搜索引擎，让信息检索快得离谱","reasonCn":"Apache Solr 是企业级文档搜索的基石，基于 Lucene 打造，支持全文检索、高亮、分词及分布式部署。它不仅是传统搜索的最佳选择，更是构建 RAG 检索增强生成系统的核心底层，能高效处理海量非结构化数据，为 AI 应用提供精准的知识召回能力。","stars":1673,"starGrowth":0,"createdAt":"2021-02-23T00:12:42Z","silentRounds":27,"language":"Java","topics":["backend"],"aiDims":["AI搜索","RAG与知识","效率工具"],"zone":"工具","funScore":0,"domainTags":["全文检索","Lucene","搜索引擎","后端服务"],"domainKey":"全文检索","pushedAt":"2026-09-23T17:01:42Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/apache/solr","ts":"2026-09-23T17:01:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dmMaze/BallonsTranslator","owner":"dmMaze","name":"BallonsTranslator","desc":"","summaryCn":"把外文漫画变成可继续编辑的中文页面，兼顾气泡排版与术语一致。","reasonCn":"它是给漫画翻译者准备的辅助程序，能把页面里的外文先检测、识别并抹掉，再调用机器翻译或大模型译文回填到原位置，同时尽量沿用颜色、轮廓和角度。除了自动翻译，还带掩膜修复、富文本改写、查找替换和术语表翻译，便于在机翻后继续校对，也适合处理日漫和美漫页面。","stars":5140,"starGrowth":0,"createdAt":"2022-04-09T13:31:52Z","silentRounds":1,"language":"Python","topics":["comics","computer-aided-translation","deep-learning","manga","ocr","scene-text-detection","inpainting","pytorch","computer-vision","auto-translation","chinese-translation","qt","qt6","pyqt","pyqt6"],"aiDims":["AI应用","文档理解","大语言模型"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["漫画翻译","文本检测","机翻","富文本编辑","术语表翻译"],"domainKey":"漫画翻译","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"comics","source":"github","weight":0.7},{"name":"computer-aided-translation","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"manga","source":"github","weight":0.7},{"name":"ocr","source":"github","weight":0.7},{"name":"scene-text-detection","source":"github","weight":0.7},{"name":"inpainting","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"auto-translation","source":"github","weight":0.7},{"name":"chinese-translation","source":"github","weight":0.7},{"name":"qt","source":"github","weight":0.7},{"name":"qt6","source":"github","weight":0.7},{"name":"pyqt","source":"github","weight":0.7},{"name":"pyqt6","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/dmMaze/BallonsTranslator","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"srbhr/Resume-Matcher","owner":"srbhr","name":"Resume-Matcher","desc":"The #1 AI Harness for Building Resumes, PDFs, Cover Letters & more, locally with 100+ LLMs support.","summaryCn":"本地跑 AI 帮你写简历求职信，支持上百种大模型","reasonCn":"Resume-Matcher 是一个 AI 驱动的简历生成工具，可以在本地运行，支持 100 多种大语言模型。它能帮你生成简历、求职信和 PDF 文档，而且所有数据都在本地处理，保护隐私的同时还能根据职位描述定制内容，让求职效率翻倍。","stars":28506,"starGrowth":3,"createdAt":"2020-04-08T03:26:49Z","silentRounds":0,"language":"TypeScript","topics":["ml"],"aiDims":["AI应用","AI写作","工具调用"],"zone":"工具","funScore":0,"domainTags":["简历生成","本地LLM","求职辅助"],"domainKey":"简历生成","pushedAt":"2026-09-10T17:46:22Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/srbhr/Resume-Matcher","ts":"2026-09-10T17:46:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"artidoro/qlora","owner":"artidoro","name":"qlora","desc":"","summaryCn":"通过高效量化微调技术让单卡运行 65B 大模型微调成为可能。","reasonCn":"QLoRA是由华盛顿大学团队开发的显存高效微调方法，它借助bitsandbytes量化技术并深度融入Hugging Face生态。项目能将650亿参数模型的微调显存压缩至单张48GB显卡上，同时保持16位精度的全量微调效能，极大降低了开发门槛。","stars":11021,"starGrowth":1,"createdAt":"2023-05-11T09:30:23Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["微调","模型与训练","大语言模型"],"zone":"工具","funScore":0,"domainTags":["模型微调","量化","显存优化","大语言模型"],"domainKey":"模型微调","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/artidoro/qlora","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"neka-nat/freecad-mcp","owner":"neka-nat","name":"freecad-mcp","desc":"FreeCAD MCP(Model Context Protocol) server","summaryCn":"让 AI 变身资深机械工程师帮你画 3D 图纸的黑科技","reasonCn":"该项目基于 Anthropic 推出的 MCP 协议，为 FreeCAD 提供了标准化的工具接口。它允许 LLM 直接通过 Python 脚本操作 FreeCAD 的 API，实现从需求描述到自动生成三维几何模型、参数化修改以及导出文件的全流程自动化闭环。","stars":2448,"starGrowth":1,"createdAt":"2023-11-25T08:46:42Z","silentRounds":1,"language":"Python","topics":["mcp","claude","freecad"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":1,"domainTags":["MCP协议","FreeCAD","3D建模","AI辅助设计"],"domainKey":"MCP协议","pushedAt":"2026-09-08T17:39:23.082Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"mcp","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"freecad","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/neka-nat/freecad-mcp","ts":"2026-09-08T17:39:23.082Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JimmyLv/BibiGPT-v1","owner":"JimmyLv","name":"BibiGPT-v1","desc":"","summaryCn":"在信息爆炸的今天，观看动辄数小时的学习视频或讲座往往成本极高，BibiGPT通","reasonCn":"项目的一大核心亮点是其兼容性的广泛度，不仅处理大众熟知的视频媒体，连推特上的多媒体内容、在线会议记录甚至本地的音频文件都在其支持范围内。这意味着无论知识存储在何种介质中，它都能为你提供统一的总结体验，并支持通过浏览器扩展程序实时调用，让学习变得触手可及。","stars":6219,"starGrowth":0,"createdAt":"2023-02-26T16:05:11Z","silentRounds":1,"language":"TypeScript","topics":["bilibili","gpt","gpt-3","chatgpt","openai","youtube"],"aiDims":["AI应用","文档理解","非AI-好玩"],"zone":"AI","funScore":0.5,"domainTags":["视频总结","在线学习","AI对话"],"domainKey":"视频总结","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"bilibili","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"gpt-3","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"youtube","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/JimmyLv/BibiGPT-v1","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GML-MMGroup/Blabber","owner":"GML-MMGroup","name":"Blabber","desc":"Blabber — 首个 Agentic 播客视频生成平台。一句提示词，生成一整期动画播客节目：AI 编写对白 · 双主持配音 · 逐帧口型同步 · 自动运镜剪辑，从想法到成片一站直达。Blabber — The first agentic podcast video generation platform. One prompt, a full animated podcast show: AI-written dialogue · dual-host voices · frame-accurate lip sync · automated camera direction & editing, from idea to final cut.","summaryCn":"想象一下，你脑子里突然蹦出一个关于科技前沿的有趣话题，却苦于不会动画制","reasonCn":"核心架构基于Agent流水线，它将播客任务拆解为脚本生成、语音合成、动画驱动和视听渲染四个环节。系统利用大模型生成生动对白，再对接语音合成技术生成富有情感的双主持人声，最后通过口型驱动技术将音频与动画形象实时绑定。","stars":18,"starGrowth":0,"createdAt":"2026-07-21T12:20:24Z","silentRounds":17,"language":"Python","topics":["podcast"],"aiDims":["多模态Agent","创意工具","AI应用"],"zone":"AI","funScore":0.8,"domainTags":["AI播客生成","语音合成","口型同步","Agent流水线"],"domainKey":"AI播客生成","pushedAt":"2026-09-08T03:39:36Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/GML-MMGroup/Blabber","ts":"2026-09-08T03:39:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"abetlen/llama-cpp-python","owner":"abetlen","name":"llama-cpp-python","desc":"","summaryCn":"llama.cpp 是本地运行大语言模型的高性能引擎，而 abetlen/llama-cpp-pytho","reasonCn":"在功能层面，它不仅仅是简单的 C API 封装。项目提供了高层的 Python API，支持文本补全、函数调用以及视觉 API 等多种场景。特别值得注意的是，它内置了一个 OpenAI 兼容的 Web 服务器，这意味着你可以用它快速搭建一个本地的 AI 服务，任何支持 OpenAI API 格式的前端或客户端都","stars":10630,"starGrowth":0,"createdAt":"2023-03-23T09:30:33Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","推理引擎","模型部署"],"zone":"AI","funScore":0.5,"domainTags":["大语言模型","本地推理","Python封装","llama.cpp"],"domainKey":"大语言模型","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/abetlen/llama-cpp-python","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cilium/cilium","owner":"cilium","name":"cilium","desc":"eBPF-based Networking, Security, and Observability","summaryCn":"基于eBPF技术的K8s网络与安全工具，让集群通信更安全可观测","reasonCn":"Cilium是基于eBPF技术的开源项目，提供Kubernetes集群的网络策略、微隔离和安全可观测能力。用Go语言开发，star近两万五千。虽然主要面向基础设施领域，但对部署AI服务的K8s集群有重要意义，是云原生安全的重要组件。","stars":25459,"starGrowth":4,"createdAt":"2015-12-16T12:33:31Z","silentRounds":0,"language":"Go","topics":["containers","bpf","security","kubernetes","kubernetes-networking","cni","kernel","loadbalancing","monitoring","troubleshooting","xdp","ebpf","k8s","observability","networking","cncf"],"aiDims":["安全工具","开发者工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["eBPF","Kubernetes","网络策略","集群安全"],"domainKey":"eBPF","pushedAt":"2026-09-23T21:47:12Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"containers","source":"github","weight":0.7},{"name":"bpf","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"kubernetes-networking","source":"github","weight":0.7},{"name":"cni","source":"github","weight":0.7},{"name":"kernel","source":"github","weight":0.7},{"name":"loadbalancing","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"troubleshooting","source":"github","weight":0.7},{"name":"xdp","source":"github","weight":0.7},{"name":"ebpf","source":"github","weight":0.7},{"name":"k8s","source":"github","weight":0.7},{"name":"observability","source":"github","weight":0.7},{"name":"networking","source":"github","weight":0.7},{"name":"cncf","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.42,"source":"trending","url":"https://github.com/cilium/cilium","ts":"2026-09-23T21:47:12Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"baichuan-inc/Baichuan-7B","owner":"baichuan-inc","name":"Baichuan-7B","desc":"","summaryCn":"百川智能开源的双语大模型，以70亿参数在1.2万亿tokens上训练。","reasonCn":"这是百川智能开源的可商用大语言模型，基于Transformer结构，参数规模70亿，在约1.2万亿tokens上完成训练，支持中英双语，上下文窗口达4096。在C-Eval、MMLU等主流评测中均取得同尺寸最佳效果，并开放商用，适合学术研究和企业级应用部署。","stars":5649,"starGrowth":0,"createdAt":"2023-06-14T10:57:51Z","silentRounds":7,"language":"Python","topics":["artificial-intelligence","ceval","large-language-models","natural-language-processing","mmlu","chatgpt","gpt-4","huggingface","llama","chinese"],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["大语言模型","中文预训练","双语模型","开源模型"],"domainKey":"大语言模型","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"artificial-intelligence","source":"github","weight":0.7},{"name":"ceval","source":"github","weight":0.7},{"name":"large-language-models","source":"github","weight":0.7},{"name":"natural-language-processing","source":"github","weight":0.7},{"name":"mmlu","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"gpt-4","source":"github","weight":0.7},{"name":"huggingface","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/baichuan-inc/Baichuan-7B","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"joonspk-research/generative_agents","owner":"joonspk-research","name":"generative_agents","desc":"","summaryCn":"构建能自主生活的虚拟小镇，让 AI 代理像真人般互动。","reasonCn":"这是一个基于大语言模型的生成式智能体开源研究项目。它通过创造拥有记忆、反思和规划能力的虚拟居民，在沙盒世界中模拟出逼真的日常社交与人类行为，为你打开通往未来 AI 社会的趣味大门，非常适合对人工智能模拟和互动叙事感兴趣的开发者与爱好者探索。","stars":22132,"starGrowth":2,"createdAt":"2023-07-23T08:26:49Z","silentRounds":0,"language":"","topics":[],"aiDims":["AI Agent","大语言模型","非AI-好玩"],"zone":"AI","funScore":0,"domainTags":["AI","Agent","大语言模型","多智能体","沙盒模拟"],"domainKey":"Agent","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1}],"aiScore":0.77,"source":"tier","url":"https://github.com/joonspk-research/generative_agents","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVIDIA/Isaac-GR00T","owner":"NVIDIA","name":"Isaac-GR00T","desc":"","summaryCn":"NVIDIA Isaac GR00T 这一项目最显著的标志是其正式进入通用可用（GA）阶段。","reasonCn":"该项目的核心亮点在于其对“物理 AI”的深入布局。不同于仅处理文本或图像的传统大模型，GR00T 专门针对机器人交互设计，提供了专门的数据集支持。通过集成推理、微调及评估的完整工具链，它允许用户针对特定机器人本体进行精细化调整。","stars":8116,"starGrowth":0,"createdAt":"2025-03-11T18:34:24Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","模型与训练"],"zone":"AI","funScore":0,"domainTags":["具身智能","机器人控制","视觉语言模型","模型微调"],"domainKey":"具身智能","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/NVIDIA/Isaac-GR00T","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sysadminsmedia/homebox","owner":"sysadminsmedia","name":"homebox","desc":"A continuation of HomeBox the inventory and organization system built for the Home User","summaryCn":"这个工具帮你管理家里所有东西。以前找东西得翻箱倒柜，现在在手机上扫个码","reasonCn":"技术上是Web应用，支持扫码录入、图片上传、标签分类、位置记录。你可以按房间、类别、品牌搜索，设置提醒（比如保修到期）。有Web界面，也有移动端优化，手机扫码录入很方便。数据存在本地，隐私性好。支持导出为JSON或CSV，方便备份。","stars":7364,"starGrowth":0,"createdAt":"2024-06-17T18:07:09Z","silentRounds":1,"language":"Go","topics":["inventory","selfhosted","hacktoberfest","homebox"],"aiDims":["非AI-实用","效率工具","创意工具"],"zone":"工具","funScore":0.5,"domainTags":["库存管理","家庭管理","物品追踪","自托管"],"domainKey":"库存管理","pushedAt":"2026-09-16T07:27:53.690Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"inventory","source":"github","weight":0.7},{"name":"selfhosted","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"homebox","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/sysadminsmedia/homebox","ts":"2026-09-16T07:27:53.690Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chenhg5/cc-connect","owner":"chenhg5","name":"cc-connect","desc":"Bridge local AI coding agents (Claude Code, Cursor, Gemini CLI, Codex) to messaging platforms (Feishu/Lark, DingTalk, Slack, Telegram, Discord, LINE, WeChat Work). Chat with your AI dev assistant from anywhere — no public IP required for most platforms.","summaryCn":"把本地代码助手连进企业聊天，让AI随时帮你写代码。","reasonCn":"cc-connect 用 Go 实现的桥接服务，可把本地部署的 AI 编码助手（Claude Code、Cursor、Gemini CLI、Codex）映射到飞书、钉钉、Slack、Telegram、Discord、LINE、企业微信等主流聊天平台，实现聊天窗口直接生成、调试、测试代码。项目提供统一的插件接口，支持多平台身份认证与消息推送，兼容不同 AI 返回格式，保证上下文连贯，极大提升远程协作效率。","stars":15626,"starGrowth":1,"createdAt":"2026-02-28T02:39:39Z","silentRounds":1,"language":"Go","topics":[],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["AI编码助手","消息桥接","多平台集成"],"domainKey":"AI编码助手","pushedAt":"2026-09-09T20:09:10.798Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/chenhg5/cc-connect","ts":"2026-09-09T20:09:10.798Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"santiandrade/Unity-CityGenerator","owner":"santiandrade","name":"Unity-CityGenerator","desc":"An editor tool for Unity that procedurally generates vibrant cities ready to use in any project.","summaryCn":"几秒钟就能在Unity里自动生成繁华大都市的开发神器","reasonCn":"专为Unity引擎打造的编辑器扩展工具，通过程序化生成技术快速创造高度自定义的城市街区，内置丰富的道路、建筑组合方案，大幅降低开放世界游戏开发中的场景搭建成本，让开发者轻松构建出充满生活气息与视觉细节的虚拟都市。","stars":15,"starGrowth":0,"createdAt":"2026-08-17T13:41:35Z","silentRounds":13,"language":"C#","topics":["procedural-generation"],"aiDims":["游戏","创意工具","可视化"],"zone":"工具","funScore":0,"domainTags":["程序化生成","Unity","游戏开发","城市生成"],"domainKey":"程序化生成","pushedAt":"2026-09-15T14:43:00Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/santiandrade/Unity-CityGenerator","ts":"2026-09-15T14:43:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/gemma.cpp","owner":"google","name":"gemma.cpp","desc":"","summaryCn":"gemma.cpp 定位为填补现代LLM推理生态中特定空白的轻量级独立引擎。","reasonCn":"这款工具的核心价值在于其“实验友好”的特性。它不仅仅是一个运行模型的程序，更是一个允许开发者深入到底层机制进行研究的沙盒。通过剥离掉非必要的生产级特性，它让研究人员能够更直接地观察和优化神经网络在C++环境下的执行过程。","stars":7041,"starGrowth":0,"createdAt":"2024-02-13T04:36:03Z","silentRounds":1,"language":"C++","topics":[],"aiDims":["推理引擎","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["Gemma模型","C++推理","轻量化运行"],"domainKey":"Gemma模型","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/google/gemma.cpp","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"InternLM/InternLM","owner":"InternLM","name":"InternLM","desc":"","summaryCn":"InternLM3-8B-Instruct 的诞生背景是对训练效率的极致追求。","reasonCn":"在实际应用中，InternLM3 适合用于通用对话、逻辑推理和知识问答场景。由于参数量控制在 8B 级别，它可以在单张消费级 GPU（如 24GB 显存的 RTX 3090/4090）或普通云服务器上高效运行。这降低了企业和个人开发者部署私有化大模型门槛，无需昂贵的集群即可体验先进的大语言模型能力","stars":7279,"starGrowth":0,"createdAt":"2023-07-06T04:52:06Z","silentRounds":6,"language":"Python","topics":["chatbot","gpt","large-language-model","long-context","rlhf","fine-tuning-llm","llm","chinese","flash-attention","pretrained-models"],"aiDims":["大语言模型","模型与训练","模型部署"],"zone":"AI","funScore":0.2,"domainTags":["指令模型","7B参数","高效训练","开源LLM"],"domainKey":"指令模型","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"large-language-model","source":"github","weight":0.7},{"name":"long-context","source":"github","weight":0.7},{"name":"rlhf","source":"github","weight":0.7},{"name":"fine-tuning-llm","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"flash-attention","source":"github","weight":0.7},{"name":"pretrained-models","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/InternLM/InternLM","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vllm-project/agentic-api","owner":"vllm-project","name":"agentic-api","desc":"Stateful API logic for agentic applications using vLLM","summaryCn":"给AI大脑配个有记忆的智能管家，让Agent干活更靠谱","reasonCn":"它给vLLM推理引擎加上了状态管理能力，解决了传统Agent调用工具时必须每次从头开始、容易遗忘上下文的问题。通过Rust编写的高性能API，实现了Session级别的会话记忆和上下文流转，让Agent能像人一样记住之前的对话并据此调用工具，大幅提升了复杂任务的执行成功率。","stars":294,"starGrowth":0,"createdAt":"2026-03-23T12:48:34Z","silentRounds":3,"language":"Rust","topics":[],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["Agent框架","vLLM","状态管理","RAG基础设施","推理引擎"],"domainKey":"推理引擎","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/vllm-project/agentic-api","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rwv/chinese-dos-games","owner":"rwv","name":"chinese-dos-games","desc":"","summaryCn":"对于喜爱复古电子游戏的玩家来说，寻找并配置早期的运行环境往往是一件极其","reasonCn":"项目核心亮点在于其庞大的游戏阵容与极低的上手门槛。目前该合集已经收录了近两千款经过整理的中文DOS游戏，涵盖了早期电脑游戏发展的黄金时期。它依托了浏览器端的DOSBox模拟技术，不仅还原度极高，而且在操作响应和运行流畅度上都有着不错的表现，让老游戏在现代设备上焕发了新的","stars":10319,"starGrowth":0,"createdAt":"2018-08-30T14:12:12Z","silentRounds":4,"language":"Python","topics":["dos","game"],"aiDims":["非AI-好玩","游戏","创意工具"],"zone":"创意","funScore":1,"domainTags":["DOS游戏","复古游戏","浏览器模拟器","游戏存档"],"domainKey":"DOS游戏","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"dos","source":"github","weight":0.7},{"name":"game","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/rwv/chinese-dos-games","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jlcodes99/cockpit-tools","owner":"jlcodes99","name":"cockpit-tools","desc":"🚀 通用 AI IDE 账号管理工具：支持 Antigravity / Codex / GitHub Copilot / Windsurf / Kiro / Cursor / Gemini-cli / CodeBuddy，多账号切换、配额监控、自动唤醒与多开实例管理。 🚀 Universal AI IDE account manager for Antigravity / Codex / GitHub Copilot / Windsurf / Kiro / Cursor / Gemini-cli / CodeBuddy, with multi-account switching, quota monitoring, wake-up automation, and multi-insta","summaryCn":"管理几十个 AI 开发者账号的神器，一键切换配额不手软。","reasonCn":"在技术实现上，它用 Rust 编写，保证了极高的运行效率和稳定性。它支持多种主流 AI IDE 的协议，能够模拟真实的客户端行为，实现无缝切换。它内置了智能配额监控引擎，能实时追踪每个账号的剩余 Token 和额度，一旦接近上限自动切换。","stars":18217,"starGrowth":4,"createdAt":"2026-01-16T07:09:32Z","silentRounds":0,"language":"Rust","topics":["account-manager","ai","antigravity","codex","github-copilot","windsurf","kiro","cursor","codebuddy"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"AI","funScore":0.6,"domainTags":["AI IDE","账号管理","配额监控"],"domainKey":"AI IDE","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"account-manager","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"github-copilot","source":"github","weight":0.7},{"name":"windsurf","source":"github","weight":0.7},{"name":"kiro","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"codebuddy","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/jlcodes99/cockpit-tools","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"QwenLM/Qwen3-TTS","owner":"QwenLM","name":"Qwen3-TTS","desc":"","summaryCn":"该项目基于 Qwen3-TTS-Tokenizer-12Hz 架构推出了 0.6B 和 1.7B 两种参数规","reasonCn":"核心亮点在于其极其丰富的语音功能矩阵。无论是需要高度个性化的声音克隆、充满想象力的声音设计，还是媲美真人发音的超高质量语音生成，它都能轻松应对。同时，项目提供了非常友好的生态支持，不仅在 Hugging Face 和 ModelScope 上架了在线体验 Demo，还配备了详细的博客说明与 ","stars":13501,"starGrowth":0,"createdAt":"2026-01-21T06:41:32Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","AI应用","模型与训练"],"zone":"AI","funScore":0.6,"domainTags":["语音合成","声音克隆","AI模型","文本转语音"],"domainKey":"语音合成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/QwenLM/Qwen3-TTS","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"oasis-surveys/oasis-platform","owner":"oasis-surveys","name":"oasis-platform","desc":"Self-hosted open-source platform / harness for AI-powered survey interviews. Voice-to-voice and text chat agents with semi-structured interview guides, multi-provider LLM support (OpenAI, Gemini, Scaleway), RAG knowledge base, and Twilio telephony. Built for researchers. Docker, one command, full control.","summaryCn":"Oasis 平台的出现是为了让企业和研究者不再依赖第三方调研服务，直接在自己","reasonCn":"Oasis 平台的出现是为了让企业和研究者不再依赖第三方调研服务，直接在自己的服务器上跑起 AI 访谈机器人。传统的问卷调研往往耗时长、回收率低，而本项目通过语音交互和大语言模型的理解能力，能够在几分钟内完成一次结构化访谈，并自动生成可直接使用的报告。","stars":26,"starGrowth":0,"createdAt":"2026-03-14T08:15:08Z","silentRounds":11,"language":"Python","topics":["interview"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI 访谈","语音交互","市场调研","LLM Agent","自动报告"],"domainKey":"AI 访谈","pushedAt":"2026-09-11T00:05:18Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"interview","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.86,"source":"search","url":"https://github.com/oasis-surveys/oasis-platform","ts":"2026-09-11T00:05:18Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wavetermdev/waveterm","owner":"wavetermdev","name":"waveterm","desc":"An open-source, AI-integrated, cross-platform terminal for seamless workflows","summaryCn":"把AI装进终端里，让命令行从此会思考能对话","reasonCn":"开源的AI集成终端，用Go和Rust开发，底层基于Cairo和HarfBuzz渲染，原生支持AI对话、智能命令建议、输出自动解析解释。内置插件系统可自定义功能，SSE流式输出让AI回复实时呈现。让命令行从冷冰冰的工具变成能理解你意图的智能助手。","stars":22343,"starGrowth":0,"createdAt":"2022-06-08T00:26:00Z","silentRounds":1,"language":"Go","topics":["command-line","developer-tools"],"aiDims":["AI应用","开发者工具","工具调用"],"zone":"工具","funScore":0.8,"domainTags":["终端","AI工具","开发者工具"],"domainKey":"终端","pushedAt":"2026-09-10T05:30:34Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/wavetermdev/waveterm","ts":"2026-09-10T05:30:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepspeedai/DeepSpeedExamples","owner":"deepspeedai","name":"DeepSpeedExamples","desc":"","summaryCn":"该资源库包含了深度学习模型从训练微调到推理加速的完整实践范例，涵盖了 D","reasonCn":"在推理领域，该项目重点介绍了 DeepSpeed-MII 与 FastGen 技术，这些工具专为低延迟、高吞吐的生成式 AI 任务设计。通过其提供的指南，开发者可以快速实现复杂模型的落地部署。此外，针对模型计算量过大的问题，库中还包含了一系列压缩实践，帮助用户在保证精度的前提下提升模型的","stars":6848,"starGrowth":1,"createdAt":"2020-01-29T21:27:22Z","silentRounds":6,"language":"Python","topics":[],"aiDims":["模型部署","模型与训练"],"zone":"AI","funScore":0.2,"domainTags":["模型训练","推理加速","模型压缩"],"domainKey":"模型训练","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/deepspeedai/DeepSpeedExamples","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"theagentrouter/agent-router","owner":"theagentrouter","name":"agent-router","desc":"Manages Unified Access to Generative AI Services built on Envoy Gateway","summaryCn":"一个智能路由器，帮你统一管理各种AI模型和服务","reasonCn":"它的核心是构建了一个基于Envoy的高性能代理网关。Envoy是业界公认的高性能网络代理，Agent Router 利用了它的强大能力来实现智能路由。它可以根据请求的意图、内容或者负载情况，自动把请求转发到最适合的AI服务上。","stars":2130,"starGrowth":0,"createdAt":"2024-10-21T02:59:59Z","silentRounds":1,"language":"Go","topics":["ai","ai-gateway","api-gateway","kubernetes","llm","inference","envoy","envoyproxy"],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI 网关","服务路由","代理","Envoy"],"domainKey":"AI 网关","pushedAt":"2026-09-10T17:23:15.587Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"api-gateway","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"envoy","source":"github","weight":0.7},{"name":"envoyproxy","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/theagentrouter/agent-router","ts":"2026-09-10T17:23:15.587Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chiphuyen/aie-book","owner":"chiphuyen","name":"aie-book","desc":"","summaryCn":"这是一本系统介绍基础模型落地与开发的开源 AI 工程全书与资源指南。","reasonCn":"除了理论知识，仓库还整理了大量实用的提示词示例与真实案例研究，甚至包含针对模型错位等进阶主题的探讨。对于希望深入了解如何把大模型应用到实际产品开发中的技术人员来说，这是一个极具参考价值的知识宝库。仓库未来几周还将持续更新更多资源，涵盖从架构设计到有趣小工具的多","stars":17524,"starGrowth":5,"createdAt":"2024-12-03T19:43:36Z","silentRounds":0,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["AI工程","大模型","提示词","教程"],"domainKey":"AI工程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/chiphuyen/aie-book","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":["daily"],"fromOfficial":false,"heatScore":0.015855968956859163,"copyOk":true},{"repo":"garrytan/gstack","owner":"garrytan","name":"gstack","desc":"Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA","summaryCn":"一套AI角色团队帮你打工，CEO设计师工程师全到位","reasonCn":"这个项目把Garry Tan（YC CEO）实际使用的Claude Code配置开源了，包含23个精心设计的角色，每个角色对应一个专业职位，比如CEO做决策、Designer出界面、Eng Manager管进度、QA负责测试。它们通过工具调用互相协作，形成完整开发流程。","stars":133936,"starGrowth":6,"createdAt":"2026-03-11T21:22:45Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":1,"domainTags":["Claude Code","AI角色","开发流程","工具链","配置"],"domainKey":"Claude Code","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/garrytan/gstack","ts":"2026-09-07T22:38:09.240Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02176127105483216,"copyOk":false},{"repo":"cbh123/narrator","owner":"cbh123","name":"narrator","desc":"","summaryCn":"\"David Attenborough narrates your life\"（大卫·爱登堡解说你的人生）是一","reasonCn":"项目的技术实现涉及多个AI服务的协作。它首先需要接入OpenAI或类似的大语言模型，用于根据输入的生活场景文本生成符合纪录片风格的解说词。随后，系统调用Replicate平台运行的模型进行中间处理或增强。最后，最关键的步骤是接入ElevenLabs的语音合成API，使用预先配置好的、模仿大","stars":4427,"starGrowth":0,"createdAt":"2023-11-14T20:10:10Z","silentRounds":6,"language":"Python","topics":[],"aiDims":["AI应用","创意工具"],"zone":"创意","funScore":0.68,"domainTags":["语音合成","自然历史解说","风格模仿","ElevenLabs","趣味应用"],"domainKey":"语音合成","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/cbh123/narrator","ts":"2026-09-16T22:54:48.334Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lemony-ai/cascadeflow","owner":"lemony-ai","name":"cascadeflow","desc":"Cascading runtime for AI agents. Optimize cost, latency, quality, and policy decisions inside the agent loop.","summaryCn":"在智能体运行循环中自动优化成本、延迟与质量的级联运行时框架。","reasonCn":"核心技术亮点在于它的多层级联调度算法。它能在Agent处理每一个步骤时，实时评估当前任务的复杂度，如果是个简单查询就立刻扔给轻量快速的小模型，如果遇到硬骨头再无缝切换到顶配大模型。同时它内置了策略控制和质量守护机制，确保在追求低成本和低延迟的同时，不会翻车降低输出质","stars":3944,"starGrowth":0,"createdAt":"2025-10-24T11:08:44Z","silentRounds":18,"language":"Python","topics":["api"],"aiDims":["AI Agent","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","级联调度","运行时","LLM","成本优化"],"domainKey":"AI Agent","pushedAt":"2026-09-08T21:50:34Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/lemony-ai/cascadeflow","ts":"2026-09-08T21:50:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"safing/portmaster","owner":"safing","name":"portmaster","desc":"🏔 Love Freedom - ❌ Block Mass Surveillance","summaryCn":"开源网络防火墙，像玩游戏一样阻断巨头的隐私监控，守护你的网络自由。","reasonCn":"这是一款基于Rust开发的开源网络防火墙和监控工具，旨在帮助用户检测并阻断恶意连接，特别是那些未经授权的数据传输。它运行在本地，通过监控网络流量来识别可疑行为，并允许用户通过直观的界面管理哪些应用可以访问互联网。","stars":13759,"starGrowth":1,"createdAt":"2018-11-13T15:31:45Z","silentRounds":1,"language":"Go","topics":["privacy"],"aiDims":["安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网络防火墙","深度包检测","隐私保护","流量监控","物联网安全"],"domainKey":"网络防火墙","pushedAt":"2026-09-10T08:44:59Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"privacy","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/safing/portmaster","ts":"2026-09-10T08:44:59Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"askrella/whatsapp-chatgpt","owner":"askrella","name":"whatsapp-chatgpt","desc":"","summaryCn":"项目通过 Puppeteer 驱动真实的网页端 WhatsApp 实例，确保了机器人能够以合","reasonCn":"作为一名重度依赖聊天软件的用户，你最直观的感受是它可以处理语音消息并给出精准回复，甚至能利用 DALL-E 2 直接在对话框中生成图片。这种将语音转录、大语言模型理解和图像生成三种能力打包在一起的操作，极大地拓宽了即时通讯的使用边界，让你可以无需频繁切换软件即可完成资料","stars":3785,"starGrowth":0,"createdAt":"2022-12-08T19:44:20Z","silentRounds":4,"language":"TypeScript","topics":["artificial-intelligence","chatbot","chatgpt","openai","whatsapp","bot","whatsapp-web","dalle"],"aiDims":["AI应用"],"zone":"AI","funScore":0.5,"domainTags":["即时通讯","语音识别","AI助手"],"domainKey":"语音识别","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"artificial-intelligence","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"whatsapp","source":"github","weight":0.7},{"name":"bot","source":"github","weight":0.7},{"name":"whatsapp-web","source":"github","weight":0.7},{"name":"dalle","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/askrella/whatsapp-chatgpt","ts":"2026-09-18T22:32:16.581Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ashawkey/stable-dreamfusion","owner":"ashawkey","name":"stable-dreamfusion","desc":"","summaryCn":"本项目是 Dreamfusion 在 PyTorch 环境下的实现，它打破了传统 3D 建模依赖","reasonCn":"项目的一大亮点在于对生成效果的持续优化，比如引入了 Perp-Neg 算法，专门针对三维生成中常见的“多头怪物”现象进行了修正，使得生成的物体几何形状更加合理、纹理更加连贯。对于想要将 AI 生成内容用于游戏开发、动画制作或仅仅是个人艺术创作的人来说，这种工具能大幅降低 3D 内","stars":8860,"starGrowth":0,"createdAt":"2022-10-06T06:18:39Z","silentRounds":4,"language":"Python","topics":["text-to-3d","gui","nerf","stable-diffusion","dreamfusion","image-to-3d"],"aiDims":["创意工具","多模态Agent"],"zone":"创意","funScore":1,"domainTags":["3D 生成","NeRF","文本生成","可视化艺术"],"domainKey":"3D 生成","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"text-to-3d","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"nerf","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"dreamfusion","source":"github","weight":0.7},{"name":"image-to-3d","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/ashawkey/stable-dreamfusion","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lllyasviel/ControlNet","owner":"lllyasviel","name":"ControlNet","desc":"","summaryCn":"“零卷积”权重的创新设计让 ControlNet 在进行微调时能够保持原始扩散模型的","reasonCn":"该项目的核心精髓在于“锁定副本”与“可训练副本”的联动机制。当你在进行创作时，锁定副本负责维持模型原有的艺术审美与逻辑，而可训练副本则负责学习你输入的特定图像条件。这种分工协作确保了在添加控制力度的同时，模型不会像传统微调那样因过度训练而变得“痴呆”，保证了产出作品","stars":34121,"starGrowth":1,"createdAt":"2023-02-01T02:43:30Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","模型部署"],"zone":"AI","funScore":1,"domainTags":["ControlNet","扩散模型","图像生成","Stable Diffusion"],"domainKey":"图像生成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/lllyasviel/ControlNet","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EthanAlgoX/AIStock","owner":"EthanAlgoX","name":"AIStock","desc":"A multi-agent AI trading system using LLMs to optimize strategies and adapt to market conditions in real-time.","summaryCn":"用多智能体系统炒股，AI 像基金经理一样实时优化交易策略。","reasonCn":"最有趣的是它的“实时适应”能力。传统的量化策略往往是死板的规则，而 AIStock 利用 LLM 的推理能力，可以根据当天的突发新闻、市场情绪或技术指标变化，实时修改交易逻辑。比如，当市场波动剧烈时，AI 会自动调整仓位风险；当某个板块出现异动时，它会快速生成新的交易计划。","stars":332,"starGrowth":0,"createdAt":"2025-12-20T10:04:55Z","silentRounds":3,"language":"Python","topics":["trading"],"aiDims":["AI Agent","多模态Agent","AI应用"],"zone":"AI","funScore":1,"domainTags":["量化交易","多智能体","LLM","金融科技"],"domainKey":"量化交易","pushedAt":"2026-09-10T17:39:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/EthanAlgoX/AIStock","ts":"2026-09-10T17:39:22Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google-deepmind/deepmind-research","owner":"google-deepmind","name":"deepmind-research","desc":"","summaryCn":"该项目主要致力于打破前沿人工智能研究的黑盒，通过开源DeepMind在各大顶级","reasonCn":"项目最具吸引力的亮点在于它不仅提供理论上的算法实现，更直接开放了多款用于科学实验的高级环境。用户可以直接将代码接入DeepMind实验室或是星际争霸II等高难度仿真平台中，体验智能体在复杂动态环境下的决策与训练过程。","stars":15202,"starGrowth":0,"createdAt":"2019-01-15T09:54:13Z","silentRounds":5,"language":"Jupyter Notebook","topics":[],"aiDims":["AI基础设施","模型与训练","学习资源"],"zone":"工具","funScore":0,"domainTags":["深度学习","强化学习","AI研究","算法实现","仿真环境"],"domainKey":"深度学习","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/google-deepmind/deepmind-research","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wechaty/wechaty","owner":"wechaty","name":"wechaty","desc":"Conversational RPA SDK for Chatbot Makers. Join our Discord: https://discord.gg/7q8NBZbQzt","summaryCn":"聊天机器人开发的瑞士军刀，一套代码打通微信WhatsApp等多平台。","reasonCn":"作为一个对话式 RPA（机器人流程自动化）SDK，Wechaty 的真正威力在于将人类常用的即时通讯渠道转化为可编程的智能入口。无论是 automate 日常消息处理，还是构建复杂的交互逻辑，它都屏蔽了平台差异。开发者可以通过 TypeScript 友好的接口，轻松读取消息、发送回复或操作联系人，","stars":23312,"starGrowth":8,"createdAt":"2016-05-01T14:36:45Z","silentRounds":0,"language":"TypeScript","topics":["wechaty","chatbot","wechat","whatsapp","conversational-ai","robotics","qq","framework","bot"],"aiDims":["AI Agent","工具调用"],"zone":"工具","funScore":0,"domainTags":["聊天机器人","RPA","SDK","跨平台开发"],"domainKey":"聊天机器人","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"wechaty","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"wechat","source":"github","weight":0.7},{"name":"whatsapp","source":"github","weight":0.7},{"name":"conversational-ai","source":"github","weight":0.7},{"name":"robotics","source":"github","weight":0.7},{"name":"qq","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"bot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"trending","url":"https://github.com/wechaty/wechaty","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02502951269732327,"copyOk":false},{"repo":"Imbad0202/academic-research-skills","owner":"Imbad0202","name":"academic-research-skills","desc":"Academic Research Skills for Claude Code: research → write → review → revise → finalize","summaryCn":"把Claude变成你的全能学术科研助手，从选题调研到论文终稿一气呵成","reasonCn":"这是一个专为学术研究打造的AI智能体技能包，无缝嵌入Claude Code生态中。它能够自动化处理文献检索、大纲规划、论文初稿撰写、同行评审模拟以及多轮修改润色，极大解放了科研人员的双手，大幅提升高质量学术论文的产出效率。","stars":49163,"starGrowth":2,"createdAt":"2026-02-26T06:31:01Z","silentRounds":0,"language":"Python","topics":["academic-writing","ai-research","claude","claude-code","literature-review","peer-review","prompt-engineering","academic-pipeline"],"aiDims":["AI Agent","AI写作","代码助手"],"zone":"AI","funScore":0,"domainTags":["AI","科研","论文写作","技能包","Claude"],"domainKey":"AI","pushedAt":"2026-09-06T11:32:12.876Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"academic-writing","source":"github","weight":0.7},{"name":"ai-research","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"literature-review","source":"github","weight":0.7},{"name":"peer-review","source":"github","weight":0.7},{"name":"prompt-engineering","source":"github","weight":0.7},{"name":"academic-pipeline","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/Imbad0202/academic-research-skills","ts":"2026-09-06T11:32:12.876Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"anthropics/claude-code-action","owner":"anthropics","name":"claude-code-action","desc":"","summaryCn":"Claude Code Action 的核心在于能够根据 GitHub PR 或 Issue 的实际上下文，","reasonCn":"在身份认证的灵活性上，该项目表现得非常专业，它支持从 Anthropic 直连 API 到亚马逊 Bedrock、Google Vertex AI 以及微软 Foundry 的广泛后端平台。这种设计意味着开发团队可以根据自身成本控制、算力需求或已有的云资源配置，自由选择最优的大模型供应商，而不必被单一厂商的技","stars":8927,"starGrowth":4,"createdAt":"2025-05-19T15:26:50Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI 辅助编程","GitHub Action","自动化工具","代码生成"],"domainKey":"AI 辅助编程","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/anthropics/claude-code-action","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"MustBeSimo/web-design-studio","owner":"MustBeSimo","name":"web-design-studio","desc":"Web Design Studio: free agent skill for distinctive websites, interactive 3D and motion. 28 live examples, art direction and accessible fallbacks. Formerly Cinematic Scroll.","summaryCn":"免费 web 设计工作室，提供交互式 3D 与动态效果模板","reasonCn":"想快速制作炫酷网页却不想写代码？web-design-studio 就是为此而生的在线工具。它把复杂的 3D 场景、粒子特效与动画拆解为可视化组件，用户只需拖拽即可得到完整页面。\\n核心技术使用 Three.js 进行 3D 渲染，GSAP 处理动画，PostCSS 与 Tailwind 负责样式与响应式。","stars":36,"starGrowth":0,"createdAt":"2026-05-30T07:01:42Z","silentRounds":12,"language":"JavaScript","topics":["creative-coding"],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":0.7,"domainTags":["Web 设计","3D 渲染","交互模板","可视化编辑","前端开发"],"domainKey":"Web 设计","pushedAt":"2026-09-16T13:13:10Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/MustBeSimo/web-design-studio","ts":"2026-09-16T13:13:10Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nashsu/llm_wiki","owner":"nashsu","name":"llm_wiki","desc":"LLM Wiki is a cross-platform desktop application that turns your documents into an organized, interlinked knowledge base — automatically. Instead of traditional RAG (retrieve-and-answer from scratch every time), the LLM incrementally builds and maintains a persistent wiki from your sources。","summaryCn":"把文件自动转成知识库，让你随时轻松检索、互链、组织。","reasonCn":"LLM Wiki 利用大模型提取文档内容，自动生成语义索引，构建可交互的知识图谱，支持多语言与自定义插件。\n用户可通过桌面客户端，快速搜索、编辑并互链文档，支持多平台同步与离线使用，可离线工作，无需网络。\n项目基于TypeScript+Electron，开箱即用，适合科研、技术文档管理和知识库构建，兼容本地数据库，易于集成。","stars":19869,"starGrowth":0,"createdAt":"2026-04-08T11:24:01Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["AI搜索","文档理解","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["知识管理","文档组织","向量检索","离线应用"],"domainKey":"知识管理","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/nashsu/llm_wiki","ts":"2026-09-15T01:07:58.514Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KlingAIResearch/LivePortrait","owner":"KlingAIResearch","name":"LivePortrait","desc":"","summaryCn":"快手开源的人像视频动画工具，支持人和动物，被主流平台广泛采用。","reasonCn":"快手技术研究院开源的 LivePortrait 是一款高效的人像动画解决方案，不仅能驱动人类面部，还能适配猫狗等动物。该项目已被快手、抖音、剪映、视频号等多个主流视频平台广泛采用，为创作者提供了从源图像到驱动视频的快速生成能力，让普通人也能轻松制作生动有趣的面部动画视频内容，大幅降低了视频创作的门槛。","stars":19099,"starGrowth":4,"createdAt":"2024-07-03T16:15:03Z","silentRounds":0,"language":"Python","topics":["image-animation","video-editing","video-generation","face-animation"],"aiDims":["多模态Agent","AI应用","创意工具"],"zone":"AI","funScore":1,"domainTags":["人像动画","视频生成","AI图像驱动","多模态"],"domainKey":"视频生成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"image-animation","source":"github","weight":0.7},{"name":"video-editing","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"face-animation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/KlingAIResearch/LivePortrait","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"MoonshotAI/kimi-code","owner":"MoonshotAI","name":"kimi-code","desc":"","summaryCn":"该项目具备自主读取并修改代码文件的能力，它根据用户反馈自动判断下一步执","reasonCn":"项目的使用过程高度依赖于终端环境，它不仅是一个简单的接口封装，更内置了针对复杂工程任务的决策逻辑。安装时，该项目官方提供了脚本安装方式，无需预装 Node.js 即可运行。对于 macOS 或 Linux 用户，通过官方脚本即可快速部署；而 Windows 用户则需注意，由于其依赖内置的 Git","stars":7612,"starGrowth":0,"createdAt":"2026-05-22T08:02:03Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","代码生成","Agent","命令行工具"],"domainKey":"Agent","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/MoonshotAI/kimi-code","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apple-aiml-research/ml-fastvlm","owner":"apple-aiml-research","name":"ml-fastvlm","desc":"","summaryCn":"项目通过设计创新的混合架构，专门解决了多模态大模型在处理高清晰度图片时","reasonCn":"为了展示其在端侧的优越性能，研究团队还发布了一个 iOS 的演示应用。通过该应用，开发者和用户可以直接在真实的 iPhone 设备上感受到该模型带来的极致流畅感。这种从算法架构到移动端优化的全链条思路，充分证明了该模型不仅适合在高性能计算平台上运行，在资源受限的移动环境下同","stars":7418,"starGrowth":0,"createdAt":"2025-05-01T22:56:43Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["推理引擎","模型部署"],"zone":"AI","funScore":0.2,"domainTags":["多模态模型","视觉编码器","模型压缩"],"domainKey":"多模态模型","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/apple-aiml-research/ml-fastvlm","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HKUSTDial/Supervisor-Skills","owner":"HKUSTDial","name":"Supervisor-Skills","desc":"将博导十年科研经验炼化为可直接调用的 AI 技能。从 Idea 构思到论文投稿，你的 AI 科研副导师。","summaryCn":"博导经验浓缩的 AI 科研副导师，帮你一键完成科研全流程","reasonCn":"HKUSTDial/Supervisor‑Skills 把博导的科研套路变成可调用的 AI 工具，涵盖论文写作、实验设计、数据分析等全链路，帮助科研人员省时省力。通过预训练模型与插件式 API，用户只需输入问题即可得到实验设计方案、论文框架，甚至自动生成 LaTeX 代码。该系统已在 5 位博士生和 10 位科研团队中验证，平均节省 30% 的实验和写作时间。","stars":7427,"starGrowth":0,"createdAt":"2026-04-19T03:33:02Z","silentRounds":2,"language":"Python","topics":["awesome"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["科研助手","论文写作","AI技能","实验设计"],"domainKey":"科研助手","pushedAt":"2026-09-05T10:53:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/HKUSTDial/Supervisor-Skills","ts":"2026-09-05T10:53:22Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liyupi/yu-ai-code-mother","owner":"liyupi","name":"yu-ai-code-mother","desc":"编程导航 AI + 微服务全栈新项目，基于 Spring Boot 3 + LangChain4j 的大厂 AI 应用生成平台，支持多种模式的智能代码生成、可视化编辑预览、一键部署分享、应用管理、AI 智能路由等。覆盖 AI 智能体开发、LangGraph4j 工作流、Tool Calling、SSE、Vue 3 全栈开发、Spring Cloud 微服务架构、MySQL / Redis / COS / Caffeine 存储、Nginx、多种设计模式、ARMS + Prometheus + Grafana 监控、Cursor Vibe Coding 等核心技术。 用一套教程掌握大厂级 AI 智能体开发+前后端核心技术，成为 AI 时代企业的香饽饽，给你的简历增加竞争力","summaryCn":"跟AI说句话就能自动生成全栈应用的大厂级平台","reasonCn":"基于Spring Boot 3加LangChain4j的AI全栈应用生成平台，支持自然语言描述生成代码、可视化编辑预览、一键部署分享。内置LangGraph工作流、Tool Calling、SSE流式输出，覆盖AI智能体开发和微服务架构，让普通人也能快速搭建AI应用。","stars":1958,"starGrowth":0,"createdAt":"2025-07-24T04:14:10Z","silentRounds":4,"language":"Java","topics":["backend","frontend"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"工具","funScore":0,"domainTags":["LangChain","代码生成","Spring Boot","全栈开发","可视化"],"domainKey":"LangChain","pushedAt":"2026-09-12T09:21:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/liyupi/yu-ai-code-mother","ts":"2026-09-12T09:21:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google/model_search","owner":"google","name":"model_search","desc":"","summaryCn":"Model Search 解决的是深度学习研究中“模型结构设计”这一高耗时、高依赖人工","reasonCn":"其核心亮点在于灵活性和对多种自动化策略的支持。用户可以根据具体分类问题，自定义神经网络的搜索空间，指定允许出现的层类型组合。这种“开箱即用”与“高度可定制”的结合，使得它既适合快速基线测试，也适合深度定制化研究。","stars":3234,"starGrowth":0,"createdAt":"2021-01-19T18:26:34Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["模型与训练","大语言模型"],"zone":"AI","funScore":0.2,"domainTags":["AutoML","架构搜索","模型蒸馏"],"domainKey":"AutoML","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/google/model_search","ts":"2026-09-19T22:19:04.205Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"karpathy/build-nanogpt","owner":"karpathy","name":"build-nanogpt","desc":"","summaryCn":"build nanogpt 仓库的独特之处在于其核心载体是“Git 提交历史”本身，而非传","reasonCn":"为了强化理解，项目配套了 YouTube 视频讲座，视频中作者逐一对应 Git Commit 进行讲解。这种“代码+视频”的双轨模式，解决了纯文字源码晦涩难懂的问题，也避免了纯视频难以跟随敲代码的痛点。对于学习者来说，这意味着可以一边看视频听作者解释设计动机，一边在本地环境暂停视频、","stars":5519,"starGrowth":0,"createdAt":"2024-06-09T05:16:13Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["学习资源","教程","大语言模型"],"zone":"资源","funScore":0.2,"domainTags":["GPT复现","深度学习","Git历史","配套视频"],"domainKey":"GPT复现","pushedAt":"2026-09-16T18:07:39.903Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/karpathy/build-nanogpt","ts":"2026-09-16T18:07:39.903Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tursodatabase/turso","owner":"tursodatabase","name":"turso","desc":"A SQL database in Rust: SQLite-compatible, now also speaking Postgres (experimental). The LLVM of databases.","summaryCn":"SQLite出身的边缘数据库新星，AI应用存储层的新选择","reasonCn":"基于SQLite开发的分布式数据库，采用边缘计算架构将数据部署到全球多个地理位置，现在也实验性支持PostgreSQL协议。性能强劲且成本极低，特别适合AI应用的向量存储、配置管理和边缘计算场景。GitHub获得23000多颗星，是开源数据库领域最受关注的创新项目之一。","stars":24361,"starGrowth":13,"createdAt":"2023-08-26T09:21:36Z","language":"Rust","topics":["database"],"aiDims":["AI基础设施","向量数据库"],"zone":"工具","funScore":0,"domainTags":["分布式数据库","SQLite","边缘计算","PostgreSQL","Rust"],"domainKey":"分布式数据库","pushedAt":"2026-09-23T21:53:43.262Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.4,"source":"trending","url":"https://github.com/tursodatabase/turso","ts":"2026-09-23T21:53:43.262Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04060091950248348,"copyOk":false},{"repo":"bytebot-ai/bytebot","owner":"bytebot-ai","name":"bytebot","desc":"","summaryCn":"开源AI桌面代理，可自主操控你的电脑帮你完成各类任务。","reasonCn":"Bytebot是一款开源AI桌面代理工具，核心是为AI分配独立操作桌面，让它能像人类一样自主操控电脑完成任务，无需用户手动拆分步骤指引，真正把AI从对话式助手变成能自主干实事的数字员工，大幅降低重复性电脑操作的时间成本。","stars":11087,"starGrowth":1,"createdAt":"2025-02-03T18:18:50Z","silentRounds":7,"language":"TypeScript","topics":["ai-agents","anthropic","computer-use","docker","llm","agentic-ai","agents","ai-tools","cua","gemini","openai","bytebot","automation","desktop-automation","agent","ai","desktop","mcp","computer-use-agent"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI桌面代理","桌面自动化","操作系统","多语言"],"domainKey":"AI桌面代理","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"computer-use","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"ai-tools","source":"github","weight":0.7},{"name":"cua","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"bytebot","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"desktop-automation","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"desktop","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"computer-use-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/bytebot-ai/bytebot","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"LSSTDESC/imSim","owner":"LSSTDESC","name":"imSim","desc":"GalSim based Rubin Observatory image simulation package","summaryCn":"在鲁宾天文台（即将开启的LSST项目）开始真正观测星空之前，科学家们需要大","reasonCn":"它的核心能力在于能够精确模拟天文成像过程中的各种物理现象，比如星星的光点是怎么散开的（PSF），背景里那些模糊的光斑是星系还是光污染。通过调整参数，它可以模拟出从白天到黑夜、从城市到深空的各种观测环境。这对天文学家来说非常重要，相当于在电脑上做了一次“预演”，确保数","stars":42,"starGrowth":0,"createdAt":"2016-10-15T00:53:14Z","silentRounds":13,"language":"Python","topics":["astronomy"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["天文模拟","图像处理","望远镜数据"],"domainKey":"天文模拟","pushedAt":"2026-09-08T21:29:37Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"astronomy","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/LSSTDESC/imSim","ts":"2026-09-08T21:29:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"matt-w-horn/morningprint","owner":"matt-w-horn","name":"morningprint","desc":"One original print, every morning. A receipt printer wakes before you do and prints a one-of-a-kind piece of character art themed to the date, the weather, and the mood of the news.","summaryCn":"每天清晨自动打印一份结合天气与新闻心情的独家复古字符画报。","reasonCn":"技术上，它采用TypeScript开发，巧妙地将外部天气API、新闻情绪分析服务与小型热敏收据打印机的硬件控制结合起来。通过精心设计的算法把各种实时数据转化为精美的ASCII字符艺术，并自动排版输出到物理打印设备上，展现了极高的极客动手能力和审美。","stars":15,"starGrowth":0,"createdAt":"2026-07-03T16:42:25Z","silentRounds":17,"language":"TypeScript","topics":["ascii-art","generative-art"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["热敏打印机","ASCII艺术","极客硬件","数据可视化","生活仪式感"],"domainKey":"热敏打印机","pushedAt":"2026-09-11T15:43:48Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ascii-art","source":"github","weight":0.7},{"name":"generative-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/matt-w-horn/morningprint","ts":"2026-09-11T15:43:48Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JoshuaC215/agent-service-toolkit","owner":"JoshuaC215","name":"agent-service-toolkit","desc":"","summaryCn":"该项目通过整合LangGraph、FastAPI与Streamlit，构建了一套完整的AI智能体运","reasonCn":"项目的一大核心亮点在于对Pydantic的深度集成，系统内所有的数据结构与运行参数均通过Pydantic进行规范化管理，这不仅保证了数据在各个组件之间传输时的严谨性与健壮性，还极大地降低了调试过程中因数据格式错误引发的异常。","stars":4499,"starGrowth":0,"createdAt":"2024-08-04T23:13:44Z","silentRounds":2,"language":"Python","topics":["agents","langgraph","streamlit"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["LangGraph","FastAPI","Streamlit","AI Agent构建","数据结构定义"],"domainKey":"LangGraph","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"langgraph","source":"github","weight":0.7},{"name":"streamlit","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/JoshuaC215/agent-service-toolkit","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tobspr-games/shapez.io","owner":"tobspr-games","name":"shapez.io","desc":"shapez is an open source base building game on Steam inspired by factorio!","summaryCn":"Steam上开源的类异星工厂基地建设游戏，好玩又硬核。","reasonCn":"这款开源游戏灵感源自异星工厂，基于JavaScript构建。玩家通过无限地图管理自动化生产线，将不同形状的方块加工成更高级的货物，并配置物流网络实现自动化运输。游戏核心是逻辑规划与资源调度，支持Steam平台游玩，拥有超过6900星。","stars":6976,"starGrowth":0,"createdAt":"2020-04-13T15:26:48Z","silentRounds":1,"language":"JavaScript","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["模拟经营","自动化流水线","Steam游戏"],"domainKey":"模拟经营","pushedAt":"2026-09-07T13:22:33Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/tobspr-games/shapez.io","ts":"2026-09-07T13:22:33Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DasterProkio/awesome-ai-companion","owner":"DasterProkio","name":"awesome-ai-companion","desc":"Open-source AI companions, desktop pets, long-term memory & proactive chat. 人机恋开源项目大全。让你的家机能够脱离官端自主存在，拥有记忆和主动性。","summaryCn":"让桌面小宠物拥有记忆和主动聊天，像真人伴侣一样陪你玩儿。","reasonCn":"它提供可本地运行的 AI 桌面宠物，支持长期记忆、情感模型和主动发起对话，采用跨平台前端与轻量后端，用户可自行编辑角色设定，实现个人化陪伴与任务提醒等功能。此外项目提供 RESTful 接口和插件机制，社区贡献多种角色模板，已获 600+ 星，活跃度高，适合想玩 AI 伴侣或做交互实验的开发者。","stars":725,"starGrowth":1,"createdAt":"2026-07-03T04:24:24Z","silentRounds":2,"language":"HTML","topics":["chatbot"],"aiDims":["AI Agent","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["桌面宠物","AI伴侣","本地AI","情感模型","长期记忆"],"domainKey":"桌面宠物","pushedAt":"2026-09-09T07:24:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/DasterProkio/awesome-ai-companion","ts":"2026-09-09T07:24:11Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"hua1995116/awesome-ai-painting","owner":"hua1995116","name":"awesome-ai-painting","desc":"","summaryCn":"一份持续更新的AI绘画与AIGC资料宝典，带新手从入门到进阶。","reasonCn":"这是AI绘画爱好者秋风的私人收藏夹，他从2022年开始整理，收录了各类AI绘画教程、工具和模型资源，比如ChatTTS、Flux.1、stable-cascade等。项目以清单和导航形式呈现，方便查阅，适合想入门AI绘画或追踪最新AI创作工具的人。作者还在同步分享自己打造100个AI产品的进展。","stars":11787,"starGrowth":1,"createdAt":"2022-10-08T07:03:22Z","silentRounds":7,"language":"","topics":["disco-diffusion","stable-diffusion","ai-painting","dd5","stable-diffusion-diffusers","stable-diffusion-embedding","stable-diffusion-tutorial","stable-diffusion-v1-5","stable-diffusion-webui"],"aiDims":["学习资源","教程","AI应用"],"zone":"资源","funScore":0,"domainTags":["AI绘画","AIGC教程","资源导航","模型工具清单"],"domainKey":"AI绘画","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"disco-diffusion","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"ai-painting","source":"github","weight":0.7},{"name":"dd5","source":"github","weight":0.7},{"name":"stable-diffusion-diffusers","source":"github","weight":0.7},{"name":"stable-diffusion-embedding","source":"github","weight":0.7},{"name":"stable-diffusion-tutorial","source":"github","weight":0.7},{"name":"stable-diffusion-v1-5","source":"github","weight":0.7},{"name":"stable-diffusion-webui","source":"github","weight":0.7}],"aiScore":0.74,"source":"search","url":"https://github.com/hua1995116/awesome-ai-painting","ts":"2026-09-14T23:04:28.868Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fastapi/full-stack-fastapi-template","owner":"fastapi","name":"full-stack-fastapi-template","desc":"Full-stack web application template with FastAPI, React, SQLModel, PostgreSQL, Vite, Tailwind CSS, shadcn/ui, FastAPI Cloud, and Docker Compose.","summaryCn":"全栈开发者的起飞跑道，复制粘贴就能拥有专业网站原型。","reasonCn":"这套模板把FastAPI后端、React前端、PostgreSQL数据库和Docker容器化打包好了。你不用再纠结技术选型，直接克隆下来就能跑。它自带完整的用户认证系统、自动HTTPS生成，还有GitHub Actions持续集成。对于想快速验证AI应用想法或搭建内部工具的开发者来说，它是最高效的起步方式，节省你几周的环境配置时间。","stars":45681,"starGrowth":0,"createdAt":"2019-02-23T15:08:34Z","silentRounds":1,"language":"TypeScript","topics":["docker"],"aiDims":["开发者工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["全栈开发","Docker","FastAPI","React模板"],"domainKey":"全栈开发","pushedAt":"2026-09-18T17:40:40Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/fastapi/full-stack-fastapi-template","ts":"2026-09-18T17:40:40Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/BitNet","owner":"microsoft","name":"BitNet","desc":"","summaryCn":"微软推出的 bitnet.cpp 在近期更新中发布了 VibeASR.cpp 实时多语种语音识别","reasonCn":"项目核心亮点在于其针对常规 x86 架构 CPU 深度优化的转换内核，以及在极低内存占用下依然能保持极高竞争力的嵌入质量。对于希望在没有高端显卡支援的普通电脑或服务器上部署实时语音转写、向量检索和语言模型的开发者而言，它提供了一种兼顾速度与精度的全新底层技术方案，有效缓","stars":40336,"starGrowth":3,"createdAt":"2024-08-05T09:07:38Z","silentRounds":0,"language":"C++","topics":[],"aiDims":["推理引擎","模型部署","大语言模型"],"zone":"AI","funScore":0.6,"domainTags":["模型压缩","低比特推理","边缘计算","语音识别"],"domainKey":"语音识别","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/microsoft/BitNet","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"d2l-ai/d2l-zh","owner":"d2l-ai","name":"d2l-zh","desc":"","summaryCn":"一套将数学原理与可运行代码紧密结合的开源深度学习中文教材。","reasonCn":"本书通过将理论剖析、数学推导与可运行工程代码紧密交织，为读者提供兼具技术深度与动手实践能力的深度学习学习路径。它不仅直观地把数学公式转化为代码，还支持读者在线修改并实时观察结果，同时配备技术论坛方便社区答疑与经验交换，是进入人工智能领域的优质全景指南。","stars":81020,"starGrowth":1,"createdAt":"2017-08-23T04:40:24Z","silentRounds":1,"language":"Python","topics":["deep-learning","book","notebook","natural-language-processing","computer-vision","machine-learning","python","chinese"],"aiDims":["教程","模型与训练","学习资源"],"zone":"资源","funScore":0.5,"domainTags":["深度学习","教程","Jupyter Notebook","中文教材"],"domainKey":"深度学习","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"book","source":"github","weight":0.7},{"name":"notebook","source":"github","weight":0.7},{"name":"natural-language-processing","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/d2l-ai/d2l-zh","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"FB208/OpenBidKit_Yibiao","owner":"FB208","name":"OpenBidKit_Yibiao","desc":"开箱即用的AI标书编写工具，标书AI生成工具，投标工具箱、知识库、标书查重、废标项检查，完全开源免费，欢迎使用","summaryCn":"标书神器！AI一键生成查重，投标再也不怕写废标。","reasonCn":"基于大模型构建的开源标书全流程助手，集成了AI自动撰写、查重检测及废标风险排查等核心功能。它支持多行业标书模板，通过RAG技术调用企业知识库，辅助用户快速生成合规且高质量的投标文件，解决传统人工撰写效率低、易漏项的问题。","stars":3025,"starGrowth":0,"createdAt":"2025-08-04T03:22:17Z","silentRounds":1,"language":"JavaScript","topics":["services","ai","document-ai","electron","llm","office-automation","proposal-generator","react","tender","typescript","ai-writing","bid","bidding","proposal-writing","rag","tender-document","service"],"aiDims":["AI写作","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["标书生成","RAG","文档处理","投标工具"],"domainKey":"RAG","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"services","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"document-ai","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"office-automation","source":"github","weight":0.7},{"name":"proposal-generator","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"tender","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"ai-writing","source":"github","weight":0.7},{"name":"bid","source":"github","weight":0.7},{"name":"bidding","source":"github","weight":0.7},{"name":"proposal-writing","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"tender-document","source":"github","weight":0.7},{"name":"service","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/FB208/OpenBidKit_Yibiao","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"duckdb/duckdb","owner":"duckdb","name":"duckdb","desc":"DuckDB is an analytical in-process SQL database management system","summaryCn":"数据库界的瑞士军刀，能在笔记本上把万亿行数据秒级查完","reasonCn":"DuckDB 专为分析型查询（OLAP）设计，采用列式存储与向量化执行引擎，支持在进程内直接运行 SQL，无需部署独立数据库服务。它支持 Parquet、CSV、JSON 等多种数据源，API 友好，常被用于数据分析、机器学习特征工程及 AI 管道中的数据预处理。","stars":41647,"starGrowth":1,"createdAt":"2018-06-26T15:04:45Z","silentRounds":1,"language":"C++","topics":["database"],"aiDims":["非AI-实用","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["列式数据库","分析型查询","SQL","数据科学"],"domainKey":"列式数据库","pushedAt":"2026-09-23T19:09:31Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/duckdb/duckdb","ts":"2026-09-23T19:09:31Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Data-Centric-AI-Community/fg-data-profiling","owner":"Data-Centric-AI-Community","name":"fg-data-profiling","desc":"1 Line of code data quality profiling & exploratory data analysis for Pandas and Spark DataFrames. ","summaryCn":"写数据分析代码前先体检，一行代码自动生成全套特征报告的神器","reasonCn":"专为Pandas和Spark数据框打造的自动化数据质量检测与探索性分析工具。通过精简接口快速输出包含缺失值、数据分布及异常检测在内的全面统计报告，大幅减少数据科学家的前期摸底时间。项目在开源社区热度较高，非常适合处理日常数据预处理与模型特征工程。","stars":13708,"starGrowth":1,"createdAt":"2016-01-09T23:47:55Z","silentRounds":1,"language":"Python","topics":["ml"],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"资源","funScore":0,"domainTags":["数据分析","数据质量","特征工程","Pandas","数据挖掘"],"domainKey":"数据分析","pushedAt":"2026-09-11T19:49:45Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/Data-Centric-AI-Community/fg-data-profiling","ts":"2026-09-11T19:49:45Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVlabs/SPADE","owner":"NVlabs","name":"SPADE","desc":"","summaryCn":"在处理复杂的图像合成任务时，传统的批归一化操作往往会无意中抹去输入的语","reasonCn":"作为 CVPR 2019 的 Oral 论文成果，该开源项目由英伟达研究团队打造，并提供了学术研究使用的基础代码。虽然原仓库建议访问更新的 Imaginaire 库获取更高性能的版本，但其核心思想依然是深度学习图像生成的重要里程碑。","stars":7717,"starGrowth":1,"createdAt":"2019-03-14T03:08:27Z","silentRounds":8,"language":"Python","topics":[],"aiDims":["模型部署","多模态Agent","创意工具"],"zone":"AI","funScore":0,"domainTags":["图像生成","GAN","语义分割","计算机视觉"],"domainKey":"图像生成","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/NVlabs/SPADE","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mit-han-lab/streaming-llm","owner":"mit-han-lab","name":"streaming-llm","desc":"","summaryCn":"MIT Han Lab发布的Streaming-LLM项目，直击当前大语言模型（LLM）推理中的一","reasonCn":"这一技术的实际落地效果相当震撼。项目新闻中特别提到，Streaming-LLM已经被集成到NVIDIA的TensorRT-LLM推理引擎中，这意味着英伟达官方认可了这一技术路线在生产环境中的价值。此外，HuggingFace Transformers也在相关PR中集成了该技术，使得普通开发者可以轻易在本地或云端复现这","stars":7269,"starGrowth":2,"createdAt":"2023-09-29T17:45:40Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["推理引擎","大语言模型","模型部署"],"zone":"AI","funScore":0,"domainTags":["流式推理","长文本","KV Cache","注意力机制"],"domainKey":"流式推理","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/mit-han-lab/streaming-llm","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mrdbourke/tensorflow-deep-learning","owner":"mrdbourke","name":"tensorflow-deep-learning","desc":"","summaryCn":"对于想要踏入人工智能领域但不知从何开始的自学者来说，这个项目提供了一条","reasonCn":"在内容组织上，项目通过循序渐进的 Jupyter Notebook 文件引导用户动手实践，从基础概念一直过渡到复杂的模型训练。这种理论与代码紧密结合的方式，极大地降低了初学者的上手门槛。课程材料不仅适合具备基础编程能力的开发者，也适合希望将深度学习应用到实际问题中的数据爱好者。","stars":5942,"starGrowth":0,"createdAt":"2020-11-23T23:11:42Z","silentRounds":7,"language":"Jupyter Notebook","topics":["deep-learning","deep-neural-networks","tensorflow","tensorflow2","tensorflow-tutorials","tensorflow-course","curriculum"],"aiDims":["模型与训练","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["深度学习","TensorFlow","Keras","教程"],"domainKey":"深度学习","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"deep-neural-networks","source":"github","weight":0.7},{"name":"tensorflow","source":"github","weight":0.7},{"name":"tensorflow2","source":"github","weight":0.7},{"name":"tensorflow-tutorials","source":"github","weight":0.7},{"name":"tensorflow-course","source":"github","weight":0.7},{"name":"curriculum","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/mrdbourke/tensorflow-deep-learning","ts":"2026-09-15T22:54:20.912Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Javis603/token-monitor","owner":"Javis603","name":"token-monitor","desc":"Local-first desktop widget for tracking token usage, costs, and limits across 35+ AI coding tools—including Claude Code, Codex, Cursor, OpenCode, and OpenClaw—with multi-device sync.","summaryCn":"写代码时常常因为调用大模型而不知 token 消耗，导致账单爆表。","reasonCn":"写代码时常常因为调用大模型而不知 token 消耗，导致账单爆表。token‑monitor 把这个盲区搬到了桌面，提供一个随时可见的窗口，实时展示各大 AI 编码工具的 token 用量、费用累计和剩余额度，帮助你在写代码的同时做好成本控制。","stars":2291,"starGrowth":0,"createdAt":"2026-05-19T09:33:54Z","silentRounds":1,"language":"JavaScript","topics":["ai-tools","claude-code","codex","llm","local-first","self-hosted","token-tracker","token-usage","cursor","openclaw","opencode","ai","linux","macos","windows","hermes-agent","antigravity","deepseek","deepseek-harness","dsh"],"aiDims":["工具调用","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Token监控","AI编码工具","成本追踪"],"domainKey":"Token监控","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai-tools","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"local-first","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"token-tracker","source":"github","weight":0.7},{"name":"token-usage","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"hermes-agent","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"deepseek-harness","source":"github","weight":0.7},{"name":"dsh","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.81,"source":"trending","url":"https://github.com/Javis603/token-monitor","ts":"2026-09-11T22:17:29.879Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"XTLS/Xray-core","owner":"XTLS","name":"Xray-core","desc":"Xray, Penetrates Everything. Also the best v2ray-core. Where the magic happens. An open platform for various uses.","summaryCn":"网络穿透神器，把你的电脑变成一个隐形节点。","reasonCn":"Xray 是一个强大的网络代理工具，支持多种协议（如 VMess、VLESS、Trojan）和加密方式。它采用模块化设计，具备强大的抗封锁能力和灵活的配置选项，被广泛应用于科学上网、内网穿透和网络安全测试等领域，是开发者和网络爱好者的必备工具。","stars":41731,"starGrowth":6,"createdAt":"2020-11-09T11:23:10Z","silentRounds":0,"language":"Go","topics":["xray","vless","tunnel","anticensorship","vmess","network","shadowsocks","trojan","proxy","xtls","socks5","reality","xudp","wireguard","dns","utls","vpn","tls","vision","xhttp"],"aiDims":["非AI-实用","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网络代理","穿透","加密协议","网络工具","科学上网"],"domainKey":"网络代理","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"xray","source":"github","weight":0.7},{"name":"vless","source":"github","weight":0.7},{"name":"tunnel","source":"github","weight":0.7},{"name":"anticensorship","source":"github","weight":0.7},{"name":"vmess","source":"github","weight":0.7},{"name":"network","source":"github","weight":0.7},{"name":"shadowsocks","source":"github","weight":0.7},{"name":"trojan","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"xtls","source":"github","weight":0.7},{"name":"socks5","source":"github","weight":0.7},{"name":"reality","source":"github","weight":0.7},{"name":"xudp","source":"github","weight":0.7},{"name":"wireguard","source":"github","weight":0.7},{"name":"dns","source":"github","weight":0.7},{"name":"utls","source":"github","weight":0.7},{"name":"vpn","source":"github","weight":0.7},{"name":"tls","source":"github","weight":0.7},{"name":"vision","source":"github","weight":0.7},{"name":"xhttp","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/XTLS/Xray-core","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.018431334020272697,"copyOk":false},{"repo":"alembics/disco-diffusion","owner":"alembics","name":"disco-diffusion","desc":"","summaryCn":"把Python文件一键转成可玩笔记的AI艺术生成器。","reasonCn":"Disco Diffusion 是一个让人通过代码生成惊艳AI绘画与动画的开源项目。它解决了传统绘图软件门槛高、无法直接用自然语言生成艺术的问题，核心在于通过一系列精心编排的AI模型与技巧，将文字描述转化为极具视觉冲击力的画面。开发者无需复杂配置，只需通过转换工具将Python脚本转为交互式笔记本，即可在浏览器中一步步操作，体验从文本到像素的奇妙旅程。","stars":7394,"starGrowth":0,"createdAt":"2022-02-24T23:52:28Z","silentRounds":7,"language":"Jupyter Notebook","topics":[],"aiDims":["创意工具","生成艺术"],"zone":"创意","funScore":1,"domainTags":["AI绘画","生成艺术","扩散模型","创意编程"],"domainKey":"AI绘画","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"生成艺术","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/alembics/disco-diffusion","ts":"2026-09-15T21:03:53.738Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Physical-Intelligence/openpi","owner":"Physical-Intelligence","name":"openpi","desc":"","summaryCn":"Physical Intelligence 团队推出的 openpi 仓库，核心交付物是三类用于机器","reasonCn":"该仓库最实用的特性是提供了经过超过一万小时真实机器人数据预训练的基座模型检查点。这意味着开发者无需从零开始训练，可以直接加载这些权重获取具备基础操作能力的机器人，或者利用提供的示例代码在自己的数据集上进行微调，从而快速适配人形机械臂或移动底盘等具体硬件设备。","stars":13955,"starGrowth":1,"createdAt":"2024-10-21T15:23:28Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","多模态Agent","微调"],"zone":"AI","funScore":0.6,"domainTags":["机器人控制","视觉语言动作","具身智能","强化学习"],"domainKey":"机器人控制","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/Physical-Intelligence/openpi","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bytedance/UI-TARS-desktop","owner":"bytedance","name":"UI-TARS-desktop","desc":"","summaryCn":"该系统通过识别计算机图形界面实现自动化操作，能像真人一样通过视觉交互直","reasonCn":"核心亮点在于它不仅是本地运行的图形界面代理，还支持连接远程计算机，这让它在管理复杂云端环境或协作办公时显得格外高效。通过结合多模态大模型的视觉感知能力，它能处理不同窗口和网页布局的变化，在面对非标准化的 UI 界面时依然保持鲁棒性，填补了传统自动化工具在处理动态图","stars":39089,"starGrowth":1,"createdAt":"2025-01-19T09:04:43Z","silentRounds":1,"language":"TypeScript","topics":["agent","vlm","vision","computer-use","mcp","mcp-server","gui-operator","browser-use","gui-agent","multimodal","tars","ui-tars","agent-tars","cowork"],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.8,"domainTags":["UI自动化","多模态大模型","桌面代理","视觉交互"],"domainKey":"UI自动化","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"vlm","source":"github","weight":0.7},{"name":"vision","source":"github","weight":0.7},{"name":"computer-use","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"gui-operator","source":"github","weight":0.7},{"name":"browser-use","source":"github","weight":0.7},{"name":"gui-agent","source":"github","weight":0.7},{"name":"multimodal","source":"github","weight":0.7},{"name":"tars","source":"github","weight":0.7},{"name":"ui-tars","source":"github","weight":0.7},{"name":"agent-tars","source":"github","weight":0.7},{"name":"cowork","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/bytedance/UI-TARS-desktop","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langchain-ai/local-deep-researcher","owner":"langchain-ai","name":"local-deep-researcher","desc":"","summaryCn":"基于本地大模型的深度联网搜索助手，能够自主迭代检索与反思并生成报告。","reasonCn":"这个项目将 OpenAI 的深度研究功能完全本地化。你只需输入一个研究主题，它就会调用 Ollama 或 LMStudio 托管的本地模型生成搜索词、抓取网页信息、总结结果并检查知识盲区，不断循环迭代直至补全信息，最后输出带完整引用来源的 Markdown 研究报告。","stars":9357,"starGrowth":0,"createdAt":"2024-12-04T23:57:20Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","AI搜索","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["RAG","本地搜索","Agent","研究助手"],"domainKey":"RAG","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.752,"source":"search","url":"https://github.com/langchain-ai/local-deep-researcher","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"microsoft/SkillOpt","owner":"microsoft","name":"SkillOpt","desc":"","summaryCn":"微软研究院推出的SkillOpt项目，让智能体技能的优化过程获得了前所未有的透","reasonCn":"该框架最核心的机制在于完全解耦了模型权重与技能逻辑。在训练过程中，系统会利用验证门控机制对每一批次的技能调整效果进行评估，只有当新策略在验证集上表现优越时才会被保留。这意味着开发者可以在不重新训练大语言模型、不消耗昂贵GPU算力的情况下，仅仅通过优化提示工程和执行","stars":17366,"starGrowth":3,"createdAt":"2026-05-08T06:41:01Z","silentRounds":0,"language":"Python","topics":["agent-skills","self-evolving-agents"],"aiDims":["AI Agent","提示工程","模型与训练"],"zone":"AI","funScore":0,"domainTags":["智能体","强化学习","提示工程"],"domainKey":"提示工程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"self-evolving-agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/microsoft/SkillOpt","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"kyegomez/zeta","owner":"kyegomez","name":"zeta","desc":"Build high-performance AI models with modular building blocks","summaryCn":"在深度学习模型研发的过程中，研究人员常常需要在不同的 Transformer 结构之","reasonCn":"技术实现上，zeta 完全基于 PyTorch，遵循 nn.Module 的标准接口，每个模块都支持前向、反向以及梯度裁剪等完整训练流程。它提供了标准的 MultiHeadAttention、LinearAttention、SparseAttention 等实现，并允许用户自定义稀疏模式或混合注意力。","stars":603,"starGrowth":0,"createdAt":"2023-07-09T23:39:33Z","silentRounds":5,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","AI基础设施","AI应用"],"zone":"AI","funScore":0,"domainTags":["深度学习","Transformer架构","PyTorch库","模型训练"],"domainKey":"模型训练","pushedAt":"2026-09-07T16:14:36Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/kyegomez/zeta","ts":"2026-09-07T16:14:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenRA/OpenRA","owner":"OpenRA","name":"OpenRA","desc":"Open Source real-time strategy game engine for early Westwood games such as Command & Conquer: Red Alert written in C# using SDL and OpenGL. Runs on Windows, Linux, *BSD and Mac OS X.","summaryCn":"经典红色警戒开源重制版，跨平台免费畅玩，老玩家必入","reasonCn":"OpenRA是完全开源的实时战略游戏引擎，专门复刻Westwood经典作品。它用C#重写核心逻辑，结合SDL和OpenGL渲染，支持Windows、Linux、macOS和BSD。项目不仅保留原版玩法，还内置模组工具和地图编辑器，方便玩家自定义内容。","stars":17419,"starGrowth":1,"createdAt":"2010-10-04T07:38:05Z","silentRounds":2,"language":"C#","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["RTS游戏","开源引擎","经典重制"],"domainKey":"RTS游戏","pushedAt":"2026-09-14T20:23:46Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/OpenRA/OpenRA","ts":"2026-09-14T20:23:46Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"wtfutil/wtf","owner":"wtfutil","name":"wtf","desc":"The personal information dashboard for your terminal","summaryCn":"把天气日历股票全塞进终端，一眼掌握人生大事","reasonCn":"基于Go开发的终端仪表板，内置20余种数据源插件，涵盖天气、日历、股票、待办事项、服务器监控等。采用模块化架构，各插件可自由排列组合成不同面板，支持WebSocket实时推送。无需图形界面，纯命令行运行，还能嵌套进tmux，对开发者友好，能同步看CI/CD状态、GitHub通知、云主机负载等。","stars":17100,"starGrowth":1,"createdAt":"2018-03-29T02:42:12Z","silentRounds":1,"language":"Go","topics":["devops"],"aiDims":["效率工具","可视化","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["终端仪表板","数据监控","DevOps","生产力"],"domainKey":"终端仪表板","pushedAt":"2026-09-17T21:56:33Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/wtfutil/wtf","ts":"2026-09-17T21:56:33Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"WhatDreamsCost/WhatDreamsCost-ComfyUI","owner":"WhatDreamsCost","name":"WhatDreamsCost-ComfyUI","desc":"LTX Director and a variety of other custom ComfyUI nodes and workflows","summaryCn":"ComfyUI的自定义节点库，轻松搭建酷炫AI图生图工作流","reasonCn":"WhatDreamsCost-ComfyUI 为 Stable Diffusion 提供丰富的自定义节点和预设工作流，涵盖图像放大、噪声消除、提示词增强等功能，用户可通过可视化编辑器拖拽组合，实现复杂的图生图、风格迁移和多模态生成，支持 GPU 加速和批量处理，极大降低 AI 艺术创作的技术门槛。","stars":2051,"starGrowth":0,"createdAt":"2026-03-20T07:26:23Z","silentRounds":2,"language":"JavaScript","topics":[],"aiDims":["AI应用","创意工具"],"zone":"AI","funScore":0.7,"domainTags":["Stable Diffusion","ComfyUI","AI绘画","节点编辑","图像生成"],"domainKey":"图像生成","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.86,"source":"trending","url":"https://github.com/WhatDreamsCost/WhatDreamsCost-ComfyUI","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zilliztech/attu","owner":"zilliztech","name":"attu","desc":"The Best GUI for Milvus","summaryCn":"如果你觉得在命令行里敲Milvus命令很枯燥，那这个项目就是你的救星。","reasonCn":"它的核心优势在于把专业的向量检索概念“翻译”成了人类能看懂的操作。你可以通过点击按钮来创建集合、插入数据、配置索引，而不是手写代码。它内置了一个查询构建器，支持布尔逻辑组合和范围查询，还能实时看到查询结果的分布图，这对调参和调试特别有用。","stars":3171,"starGrowth":0,"createdAt":"2021-12-09T09:15:33Z","silentRounds":1,"language":"Shell","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","可视化"],"zone":"工具","funScore":0,"domainTags":["Milvus","向量数据库","GUI","RAG"],"domainKey":"Milvus","pushedAt":"2026-09-08T01:57:55Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/zilliztech/attu","ts":"2026-09-08T01:57:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MicrosoftDocs/mcp","owner":"MicrosoftDocs","name":"mcp","desc":"Official Microsoft Learn MCP Server and CLI tool – powering LLMs and AI agents with real-time, trusted Microsoft docs & code samples.","summaryCn":"有时候问AI技术问题，它总喜欢一本正经地胡说八道。","reasonCn":"它的核心技术基础是MCP协议（Model Context Protocol）。它提供了一套通用的通讯方式，让AI大模型能够像调用API一样，实时查询特定的知识库或代码仓库。相较于传统的RAG（检索增强生成），这种方式架构更轻、响应更直接，且微软文档的结构化输出使得模型在解析复杂指令时具备极高的","stars":1900,"starGrowth":1,"createdAt":"2025-06-09T03:31:49Z","silentRounds":2,"language":"TypeScript","topics":["documentation"],"aiDims":["工具调用","AI Agent","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["MCP协议","知识库","文档检索"],"domainKey":"MCP协议","pushedAt":"2026-09-10T06:31:29Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"documentation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/MicrosoftDocs/mcp","ts":"2026-09-10T06:31:29Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bohyy/academic-ai-prompt","owner":"bohyy","name":"academic-ai-prompt","desc":"一套为研究生和学术研究者设计的完整AI Prompt库  📖 包含内容： ✨ 40+ 精心设计的AI Prompt ✨ 论文选题系统方法（生成、评估、论证） ✨ 论文查找快速方案（8个不同方案） ✨ 文献综述框架和工具 ✨ Excel自动评估表格 ✨ 3个完整的论证模板  🚀 核心优势： ⚡ 节省时间 50-70%（选题3-5天而不是2-3周） 🎯 科学方法（基于系统的5维度评估体系） 💡 即插即用（所有Prompt直接复制可用） 📚 全流程覆盖（从选题到出版的完整方案）  🎓 适用人群： 👨‍🎓 硕士研究生 | 博士研究生 | 本科毕业设计 | 学术研究者 | 内容创作者  ","summaryCn":"让研究生告别选题焦虑，把写论文的时间省下一半的神级Prompt库。","reasonCn":"这套库专门为学术场景打造，提供了40多个经过验证的AI提示词，覆盖论文全流程。它不仅包含选题系统的生成与评估方法，还提供了8种文献查找方案和文献综述框架。通过Excel自动评估表格和3个完整论证模板，能帮助研究者科学地利用AI辅助科研，大幅缩短从选题到成文的时间。","stars":1951,"starGrowth":1,"createdAt":"2026-02-12T06:45:41Z","silentRounds":1,"language":"","topics":["productivity"],"aiDims":["提示工程","AI写作","论文"],"zone":"资源","funScore":0,"domainTags":["学术Prompt","论文写作","文献综述","科研效率","模板库"],"domainKey":"学术Prompt","pushedAt":"2026-09-02T08:28:40Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7}],"aiScore":0.95,"source":"search","url":"https://github.com/bohyy/academic-ai-prompt","ts":"2026-09-02T08:28:40Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"babaohuang/GeminiProChat","owner":"babaohuang","name":"GeminiProChat","desc":"","summaryCn":"GeminiProChat的定位非常纯粹——一个针对Gemini Pro模型的极简Web UI。","reasonCn":"项目的核心价值在于部署的便捷性。它兼容多种主流云平台：最推荐的Vercel部署，只需点击项目仓库中的Deploy按钮，按提示填入Gemini API密钥即可完成搭建，整个过程不超过两分钟。此外还支持Railway、Zeabur以及Docker方式部署。","stars":4893,"starGrowth":0,"createdAt":"2023-12-14T06:17:33Z","silentRounds":3,"language":"TypeScript","topics":["astro","gemini","gemini-api","gemini-client","gemini-pro","gemini-server","google","google-api"],"aiDims":["AI应用"],"zone":"AI","funScore":0.44,"domainTags":["聊天机器人","AI对话","Web应用"],"domainKey":"聊天机器人","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"astro","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"gemini-api","source":"github","weight":0.7},{"name":"gemini-client","source":"github","weight":0.7},{"name":"gemini-pro","source":"github","weight":0.7},{"name":"gemini-server","source":"github","weight":0.7},{"name":"google","source":"github","weight":0.7},{"name":"google-api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/babaohuang/GeminiProChat","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apple/tensorflow_macos","owner":"apple","name":"tensorflow_macos","desc":"","summaryCn":"该工程通过将底层的 ML Compute 框架与 TensorFlow 进行深度耦合，使得用户","reasonCn":"对于深度学习从业者而言，该项目的主要价值在于其提供了一套本地化的训练环境，减少了模型在云端与本地同步的时间损耗。项目支持 macOS 11.0 及以上版本，要求配套使用 Xcode 的命令行工具，用户可以通过下载官方 Release 中提供的安装包来完成配置。","stars":3645,"starGrowth":0,"createdAt":"2020-11-05T18:47:47Z","silentRounds":4,"language":"Shell","topics":[],"aiDims":["模型部署","模型与训练"],"zone":"工具","funScore":0.2,"domainTags":["TensorFlow","macOS","硬件加速","深度学习开发"],"domainKey":"TensorFlow","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/apple/tensorflow_macos","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NoFxAiOS/nofx","owner":"NoFxAiOS","name":"nofx","desc":"Your AI trading terminal assistant for US stocks, commodities, forex, and crypto.","summaryCn":"美股交易必备的AI终端助手，能帮你自动盯盘决策","reasonCn":"这是一个专为美股、大宗商品、外汇和加密货币设计的AI交易终端助手，核心能力是利用大语言模型实时分析市场数据并辅助决策。它支持多种交易接口，能自动抓取行情、分析图表、监控持仓，并通过自然语言交互给出买卖建议。","stars":12945,"starGrowth":3,"createdAt":"2025-10-28T07:17:53Z","silentRounds":0,"language":"Go","topics":["trading"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI交易","量化投资","多市场数据"],"domainKey":"AI交易","pushedAt":"2026-09-05T02:55:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/NoFxAiOS/nofx","ts":"2026-09-05T02:55:46Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mvanhorn/cli-printing-press","owner":"mvanhorn","name":"cli-printing-press","desc":"Every API has a secret identity. This finds it, absorbs every feature from every competing tool, then builds the GOAT CLI — designed for AI agents first, with SQLite sync, offline search, and compound insight commands.","summaryCn":"把所有AI工具的优点集一身的全能CLI工具，专治各种工具不兼容","reasonCn":"这是一个基于Go语言构建的命令行工具，旨在为AI智能体提供统一的API交互接口。它通过SQLite实现数据持久化，支持离线搜索和复合查询，解决了不同AI工具间接口不统一、数据难以共享的痛点。项目设计初衷就是为了让AI代理能更高效地调用各种工具，具备极强的可扩展性和兼容性。","stars":4738,"starGrowth":3,"createdAt":"2026-03-23T16:02:09Z","silentRounds":0,"language":"Go","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI智能体","命令行工具","API聚合","SQLite","离线搜索"],"domainKey":"AI智能体","pushedAt":"2026-09-09T13:10:46Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/mvanhorn/cli-printing-press","ts":"2026-09-09T13:10:46Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"najmuzzaman-mohammad/gawkbot","owner":"najmuzzaman-mohammad","name":"gawkbot","desc":"open source grok bot. gawk bots automate your menial work via AI models and build you microapps to manage the outcome, so that you have a false sense of control.","summaryCn":"用AI把琐碎任务变微应用，帮你省事又省心的自动化小管家","reasonCn":"gawkbot 基于 Go，实现多模型 API 调用，解析自然语言指令自动生成微服务代码或脚本，支持 Slack、Telegram 等交互。内置工作流编排，将任务串联并提供可视化面板监控。插件机制可扩展模型或接入企业内部工具，实现任务接收、处理到结果输出的链路自动化。","stars":1456,"starGrowth":4,"createdAt":"2026-03-25T13:27:01Z","silentRounds":0,"language":"Go","topics":["agents","autonomous-agents","claude-code","knowledge-graph","grokbot","ai-agents","ai-teammates","codex","grok-bot","grok-bot-alternative","local-llm","microapps","open-source","self-hosted","workflow-automation"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["AI自动化","自然语言编程","聊天机器人"],"domainKey":"AI自动化","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"autonomous-agents","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"knowledge-graph","source":"github","weight":0.7},{"name":"grokbot","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-teammates","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"grok-bot","source":"github","weight":0.7},{"name":"grok-bot-alternative","source":"github","weight":0.7},{"name":"local-llm","source":"github","weight":0.7},{"name":"microapps","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"workflow-automation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/najmuzzaman-mohammad/gawkbot","ts":"2026-09-12T22:00:03.438Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dmlc/gluon-cv","owner":"dmlc","name":"gluon-cv","desc":"","summaryCn":"Gluon-CV 的核心价值在于它不仅仅是一个模型库，更是一个完整的视觉算法工程","reasonCn":"在技术选型上，该项目展示了较强的兼容性。它同时支持 PyTorch 和 MXNet 两大主流深度学习框架，这让不同技术背景的团队都能在自己的生态系统中使用这些模型。这对于处于框架迁移期或需要维护双栈代码的工程项目来说，是一个显著的实用优势，避免了重写底层算子或适配不同框架API的","stars":5915,"starGrowth":0,"createdAt":"2018-02-26T01:33:21Z","silentRounds":7,"language":"Python","topics":["deep-learning","computer-vision","neural-network","gluon","mxnet","machine-learning","image-classification","object-detection","semantic-segmentation","gan","person-reid","action-recognition","pose-estimation"],"aiDims":["模型与训练","多模态Agent"],"zone":"工具","funScore":0,"domainTags":["计算机视觉","模型库","图像处理","PyTorch"],"domainKey":"计算机视觉","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"neural-network","source":"github","weight":0.7},{"name":"gluon","source":"github","weight":0.7},{"name":"mxnet","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"image-classification","source":"github","weight":0.7},{"name":"object-detection","source":"github","weight":0.7},{"name":"semantic-segmentation","source":"github","weight":0.7},{"name":"gan","source":"github","weight":0.7},{"name":"person-reid","source":"github","weight":0.7},{"name":"action-recognition","source":"github","weight":0.7},{"name":"pose-estimation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/dmlc/gluon-cv","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"multica-ai/andrej-karpathy-skills","owner":"multica-ai","name":"andrej-karpathy-skills","desc":"A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.","summaryCn":"Karpathy亲传AI编程咒语，修正大模型代码幻觉","reasonCn":"这个项目提炼了OpenAI联合创始人Andrej Karpathy关于LLM编码陷阱的观察，浓缩为一个CLAUDE.md配置文件。它能让Claude Code等AI编程助手避免常见错误，显著提升代码正确率。适合追求极致代码质量的开发者，是把大师经验转化为生产力的绝佳工具。","stars":214660,"starGrowth":33,"createdAt":"2026-01-27T03:53:13Z","silentRounds":0,"language":"","topics":[],"aiDims":["提示工程","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["CLAUDE.md","提示模板","编码规范","幻觉修正","AI编程"],"domainKey":"CLAUDE.md","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1}],"aiScore":0.7,"source":"trending","url":"https://github.com/multica-ai/andrej-karpathy-skills","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.11935343333643902,"copyOk":false},{"repo":"chatchat-space/Langchain-Chatchat","owner":"chatchat-space","name":"Langchain-Chatchat","desc":"","summaryCn":"受 GanymedeNil 项目与 ChatGLM-6B 相关 Pull Request 的启发，开发者们打造","reasonCn":"在技术架构上，项目深度整合了 Langchain 思想，支持通过 Xinference、Ollama 等主流模型推理框架无缝接入 GLM-4-Chat、Qwen2-Instruct、Llama3 等各类前沿大模型。它不仅兼容市面上主流的 Embedding 模型与向量数据库，还支持 OpenAI GPT API 的灵活调用，为开发者和企业构建定制","stars":38663,"starGrowth":1,"createdAt":"2023-03-31T12:12:45Z","silentRounds":2,"language":"Python","topics":["chatglm","langchain","llm","knowledge-base","llama","chatbot","chatgpt","embedding","faiss","fastchat","gpt","milvus","streamlit","chatchat","langchain-chatglm","qwen","rag","retrieval-augmented-generation","xinference","ollama"],"aiDims":["AI基础设施","模型部署","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["RAG","知识库","大语言模型","本地部署"],"domainKey":"RAG","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"chatglm","source":"github","weight":0.7},{"name":"langchain","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"knowledge-base","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"embedding","source":"github","weight":0.7},{"name":"faiss","source":"github","weight":0.7},{"name":"fastchat","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"milvus","source":"github","weight":0.7},{"name":"streamlit","source":"github","weight":0.7},{"name":"chatchat","source":"github","weight":0.7},{"name":"langchain-chatglm","source":"github","weight":0.7},{"name":"qwen","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"retrieval-augmented-generation","source":"github","weight":0.7},{"name":"xinference","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/chatchat-space/Langchain-Chatchat","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Splode/pomotroid","owner":"Splode","name":"pomotroid","desc":":tomato: Simple and visually-pleasing Pomodoro timer","summaryCn":"Rust写的简洁番茄钟，界面美观、计时精准、支持多任务切换。","reasonCn":"很多人用手机或者网页番茄钟，却总被广告、弹窗打扰，pomotroid提供了一款干净、离线、资源占用极低的本地计时工具，让你专注于手头任务而不被外界干扰。它的核心理念是把番茄工作法的时间块和休息提醒做成一个可视化的仪表盘，打开后即可看到当前进度条和待办任务列表。","stars":5513,"starGrowth":0,"createdAt":"2018-01-26T04:47:51Z","silentRounds":2,"language":"Rust","topics":["productivity"],"aiDims":["效率工具","非AI-好玩","非AI-实用"],"zone":"工具","funScore":1,"domainTags":["番茄钟","时间管理","生产力","Rust"],"domainKey":"番茄钟","pushedAt":"2026-09-07T23:38:41Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Splode/pomotroid","ts":"2026-09-07T23:38:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"facebookresearch/sam2","owner":"facebookresearch","name":"sam2","desc":"","summaryCn":"Meta 通过将图像视为单帧视频的巧妙逻辑，构建了一个通用的视觉分割底层架构","reasonCn":"这套系统内置了精密的 Transformer 架构，专门为实时视频处理优化，能够精准捕捉到高速移动物体的边缘变化。对于从事视频剪辑、特效合成或是自动驾驶数据标注的专业人员来说，它不仅节省了海量手动逐帧修图的时间，还能在处理复杂遮挡场景时展现出惊人的稳定性，是提升视觉工作流效","stars":19907,"starGrowth":0,"createdAt":"2024-07-29T21:48:23Z","silentRounds":3,"language":"Jupyter Notebook","topics":[],"aiDims":["模型与训练","文档理解","模型部署"],"zone":"AI","funScore":0,"domainTags":["视频分割","实时视频处理","目标跟踪","视觉模型","计算机视觉"],"domainKey":"视频分割","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/facebookresearch/sam2","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"567-labs/instructor","owner":"567-labs","name":"instructor","desc":"","summaryCn":"在调用大语言模型处理业务时，模型经常会返回格式混乱的文本，导致程序解析","reasonCn":"除了专注于快速稳定地提取结构化数据之外，该项目还拥有极低的上手门槛和清晰的架构设计。在实际使用时，开发者只需要按照规范定义好 Pydantic 模型，就能直接下发指令并拿到符合预期的数据对象。对于需要构建更复杂智能体应用的用户，官方生态还提供了延伸方案，在保留原有特性的","stars":13935,"starGrowth":0,"createdAt":"2023-06-14T10:42:23Z","silentRounds":2,"language":"Python","topics":["openai","python","pydantic-v2","openai-functions","validation","openai-function-calli"],"aiDims":["大语言模型","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["LLM","数据提取","Python 库","API 集成"],"domainKey":"LLM","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"openai","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"pydantic-v2","source":"github","weight":0.7},{"name":"openai-functions","source":"github","weight":0.7},{"name":"validation","source":"github","weight":0.7},{"name":"openai-function-calli","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/567-labs/instructor","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cvg/LightGlue","owner":"cvg","name":"LightGlue","desc":"","summaryCn":"LightGlue 的核心技术在于它引入了基于网络宽度与深度的自适应剪枝，这使得","reasonCn":"作为一款专注于推理层面的高性能工具，它主要通过输入图像的关键点坐标与描述符，直接输出对应点的索引关系。这意味着开发者无需从零训练模型，通过简单的接口调用即可将此能力集成到现有的视觉流水线中。无论是全景拼接、SLAM 定位还是三维重建任务，LightGlue 都能在保证结果质量","stars":4791,"starGrowth":1,"createdAt":"2023-06-25T09:39:51Z","silentRounds":1,"language":"Python","topics":["deep-learning","image-matching","pose-estimation","transformers"],"aiDims":["推理引擎","模型部署"],"zone":"AI","funScore":0.2,"domainTags":["图像匹配","特征点检测","计算机视觉"],"domainKey":"图像匹配","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"image-matching","source":"github","weight":0.7},{"name":"pose-estimation","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/cvg/LightGlue","ts":"2026-09-16T22:54:48.334Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"RickStrahl/MarkdownMonster","owner":"RickStrahl","name":"MarkdownMonster","desc":"An extensible Markdown Editor, Viewer and Weblog Publisher for Windows","summaryCn":"Windows上最好用的Markdown编辑器兼发布器","reasonCn":"写技术文档或者写博客的时候，Markdown是最流行的格式，因为它简单直观。但很多编辑器要么预览不实时，要么功能太简陋。这个软件就像一个全栈工具，左边写代码，右边实时显示渲染结果，所见即所得。它最厉害的地方在于“发布”功能，写完文档可以直接一键发布到WordPress、Blogger或","stars":1694,"starGrowth":0,"createdAt":"2016-05-05T22:22:59Z","silentRounds":14,"language":"HTML","topics":["documentation","writing"],"aiDims":["开发者工具","文档理解"],"zone":"工具","funScore":0,"domainTags":["Markdown编辑器","博客发布","Windows工具","静态站点"],"domainKey":"Markdown编辑器","pushedAt":"2026-09-15T19:13:03Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"documentation","source":"github","weight":0.7},{"name":"writing","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/RickStrahl/MarkdownMonster","ts":"2026-09-15T19:13:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Lordog/dive-into-llms","owner":"Lordog","name":"dive-into-llms","desc":"","summaryCn":"源自上海交大课程的公益大模型编程实践，含昇腾国产适配与前沿安全主题。","reasonCn":"该项目源于上海交大李瑞东团队拓展的自然语言处理与安全课程，旨在帮助读者通过代码实践快速掌握大模型开发。它不仅覆盖了基础理论，还深入探讨了隐写术、数学推理和大模型对齐等进阶话题。结合华为昇腾社区提供的国产化开发指南，学习者能在不同硬件环境下进行全流程实操，是连接学术研究与工业落地的优质桥梁。","stars":55073,"starGrowth":1,"createdAt":"2024-04-08T13:11:40Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["大语言模型","微调","教程"],"zone":"资源","funScore":0,"domainTags":["大模型教程","自然语言处理","AI安全","昇腾适配"],"domainKey":"大模型教程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/Lordog/dive-into-llms","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"stitionai/devika","owner":"stitionai","name":"devika","desc":"","summaryCn":"能将人类高阶指令拆解为具体步骤、自动调研并编写代码的实验性自主智能体。","reasonCn":"Devika 是一款处于早期实验阶段的 AI 软件工程师智能体。它能够理解用户的高层需求，自主将复杂目标拆解为可执行步骤。在项目过程中，它会主动检索相关信息，并直接编写代码以实现既定目标。尽管当前版本存在较多未完善的功能，但它展示了从指令到代码的端到端自动化潜力，适合对自主 AI 编程感兴趣的技术爱好者探索。","stars":19564,"starGrowth":1,"createdAt":"2024-03-21T03:03:21Z","silentRounds":10,"language":"Python","topics":[],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":0.7,"domainTags":["AI智能体","自动编程","软件工程","自主Agent","Python"],"domainKey":"AI智能体","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/stitionai/devika","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ShishirPatil/gorilla","owner":"ShishirPatil","name":"gorilla","desc":"","summaryCn":"Gorilla推出的BFCL V4 Agentic基准测试，直接定义了智能体在复杂生产环境下","reasonCn":"在与LMSYS合作的Agent Arena中，你可以直观地进行多维度实测，横向比对不同模型在复杂任务流下的交互表现。这种实战型Arena不仅包含传统的金融数据处理，还涵盖了具备记忆管理与格式敏感性的高级评估任务。对于那些致力于打造高效能、低幻觉率AI智能体的工程师来说，这里是观察各种","stars":13038,"starGrowth":0,"createdAt":"2023-05-19T00:46:45Z","silentRounds":3,"language":"Python","topics":["api","llm","api-documentation","chatgpt","gpt-4-api","claude-api","openai-api","openai-functions"],"aiDims":["工具调用","AI Agent","大语言模型"],"zone":"AI","funScore":0,"domainTags":["AI基准测试","Agent评估","LLM评测","模型评估","Benchmark"],"domainKey":"AI基准测试","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"api-documentation","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"gpt-4-api","source":"github","weight":0.7},{"name":"claude-api","source":"github","weight":0.7},{"name":"openai-api","source":"github","weight":0.7},{"name":"openai-functions","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/ShishirPatil/gorilla","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NateBJones-Projects/OB1","owner":"NateBJones-Projects","name":"OB1","desc":"Open Brain — The infrastructure layer for your thinking. One database, one AI gateway, one chat channel — any AI plugs in. No middleware, no SaaS.","summaryCn":"像搭积木一样把AI能力塞进你的项目里，一劳永逸","reasonCn":"OB1 是一个统一的 AI 基础设施层，旨在解决 AI 开发中的碎片化问题。它提供单一的数据库、网关和聊天通道，支持任何 AI 模型无缝接入，无需中间件或 SaaS 服务。项目基于 TypeScript 构建，专注于简化 AI 应用的集成与部署流程，适合希望构建可扩展、无依赖 AI 系统的开发者。","stars":4634,"starGrowth":0,"createdAt":"2026-03-11T17:59:59Z","silentRounds":1,"language":"TypeScript","topics":["ai-agents","mcp","memory-layer","memory-layer-for-ai","open-brain","open-source","personal-knowledge","personal-knowledge-management","supabase"],"aiDims":["AI基础设施","AI应用","工具调用"],"zone":"工具","funScore":0,"domainTags":["AI基础设施","模型网关","开发者工具"],"domainKey":"AI基础设施","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"memory-layer","source":"github","weight":0.7},{"name":"memory-layer-for-ai","source":"github","weight":0.7},{"name":"open-brain","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"personal-knowledge","source":"github","weight":0.7},{"name":"personal-knowledge-management","source":"github","weight":0.7},{"name":"supabase","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/NateBJones-Projects/OB1","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TabbyML/tabby","owner":"TabbyML","name":"tabby","desc":"","summaryCn":"在当前众多依赖云端的 AI 编程工具中，代码安全和隐私常常成为企业和个人的","reasonCn":"作为一款强大的私有化 AI 编码助手，它不仅支持常见的代码自动补全，还在不断演进中加入了丰富的工程协作能力。例如，它能够通过连接任务直接实现 GitHub 问题的自动处理，并从侧边栏直接创建包含测试和检查结果的拉取请求；同时还支持将 GitLab 的合并请求纳入上下文索引，让 AI ","stars":33888,"starGrowth":0,"createdAt":"2023-03-16T09:18:01Z","silentRounds":2,"language":"Rust","topics":["ai","codegen","coding-assistant","coding-language","developer-experience","developer-tools","gen-ai","llms","ide"],"aiDims":["代码助手","AI基础设施","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI编程助手","代码补全","自托管","私有化部署","Rust"],"domainKey":"AI编程助手","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"codegen","source":"github","weight":0.7},{"name":"coding-assistant","source":"github","weight":0.7},{"name":"coding-language","source":"github","weight":0.7},{"name":"developer-experience","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"gen-ai","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"ide","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/TabbyML/tabby","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"thunlp/PLMpapers","owner":"thunlp","name":"PLMpapers","desc":"","summaryCn":"由清华 THUNLP 实验室推出的这个仓库，是了解预训练语言模型（PLM）技术分支","reasonCn":"除了理论梳理，该项目还兼具了“开源实践”的价值。仓库内专门开辟了“Open PLMs”章节，列出了清华团队近年来持续训练并对外发布的开源模型，如 CPM-2 及更早版本的 CPM-1。这种“论文清单一站式汇总”与“高质量模型即时下载”的结合，使得该项目从单纯的知识库升级为了一个可工作的基础","stars":3360,"starGrowth":0,"createdAt":"2019-09-23T15:34:06Z","silentRounds":4,"language":"","topics":[],"aiDims":["大语言模型","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["预训练语言模型","NLP论文检索","模型演进图谱","大模型开源"],"domainKey":"预训练语言模型","pushedAt":"2026-09-19T18:31:48.187Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.5599999999999999,"source":"tier","url":"https://github.com/thunlp/PLMpapers","ts":"2026-09-19T18:31:48.187Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"modelscope/facechain","owner":"modelscope","name":"facechain","desc":"","summaryCn":"FaceChain 由阿里通义团队开发，旨在解决传统 AI 绘画中人物肖像风格不一致","reasonCn":"除了速度，FaceChain 在风格一致性上也做了深度优化。它提供了针对新风格的训练脚本，能够自动生成 LoRA（Low-Rank Adaptation）模型，并配套相应的风格提示词。这意味着用户不仅可以生成标准肖像，还可以通过简单的操作训练出自己专属的画师风格或摄影风格，且这些 LoRA 模型能与","stars":9509,"starGrowth":1,"createdAt":"2023-08-10T10:46:54Z","silentRounds":4,"language":"Jupyter Notebook","topics":[],"aiDims":["AI应用","模型与训练"],"zone":"创意","funScore":0.5,"domainTags":["肖像生成","图像风格化","AI 绘画","生成艺术"],"domainKey":"肖像生成","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/modelscope/facechain","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"IDEA-Research/GroundingDINO","owner":"IDEA-Research","name":"GroundingDINO","desc":"","summaryCn":"结合文本提示与图像识别，实现开放世界任意物体的精准检测定位。","reasonCn":"该项目通过将视觉检测架构 DINO 与文本跨模态预训练结合，实现了无需预设类别的开放世界目标检测。用户只需在输入框填入自然语言描述，模型就能迅速在图像中定位并框出对应目标。此外该模型已全面集成至 Hugging Face，并能联合 SAM 2 实现复杂的视频物体追踪。","stars":10608,"starGrowth":0,"createdAt":"2023-03-09T06:14:41Z","silentRounds":3,"language":"Python","topics":["object-detection","open-world","open-world-detection","vision-language","vision-language-transformer"],"aiDims":["模型与训练"],"zone":"AI","funScore":0,"domainTags":["目标检测","开放词汇","多模态","计算机视觉"],"domainKey":"多模态","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"object-detection","source":"github","weight":0.7},{"name":"open-world","source":"github","weight":0.7},{"name":"open-world-detection","source":"github","weight":0.7},{"name":"vision-language","source":"github","weight":0.7},{"name":"vision-language-transformer","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/IDEA-Research/GroundingDINO","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"agentcodee/cursor-free-everyday","owner":"agentcodee","name":"cursor-free-everyday","desc":"","summaryCn":"一款为AI编程助手提供续杯服务的实用工具。","reasonCn":"项目的一个核心亮点是其广泛的兼容性。根据文档介绍，它“同时支持 Cursor、windsurf、kiro、codex无限额度使用”。这意味着开发者不需要为不同的AI工具分别寻找解决方案，这款工具可以一站式满足多种主流AI编程助手的需求。","stars":6099,"starGrowth":0,"createdAt":"2025-05-14T07:24:05Z","silentRounds":6,"language":"Rust","topics":[],"aiDims":["AI应用","代码助手","效率工具"],"zone":"工具","funScore":0.26,"domainTags":["AI编程助手","软件激活","跨平台应用","Rust"],"domainKey":"AI编程助手","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/agentcodee/cursor-free-everyday","ts":"2026-09-15T22:54:20.912Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"brycedrennan/imaginAIry","owner":"brycedrennan","name":"imaginAIry","desc":"","summaryCn":"该项目集成了 spandrel 库，专门用于处理图像超分辨率增强，这意味着你可以","reasonCn":"它的核心优势在于对多平台的支持，无论是 Linux 还是 macOS（尤其是 M1 芯片）用户，都能享受到“开箱即用”的体验。对于追求创作效率的用户来说，它不再要求你具备深厚的深度学习背景，只要系统里安装好了 Python 环境，简单的命令即可触发复杂的模型运算，非常适合想在本地折腾 AI","stars":8187,"starGrowth":0,"createdAt":"2022-09-12T04:23:10Z","silentRounds":5,"language":"Python","topics":[],"aiDims":["模型部署","AI应用","模型与训练"],"zone":"创意","funScore":0,"domainTags":["图像增强","超分辨率","图像处理","AI 绘图"],"domainKey":"图像增强","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/brycedrennan/imaginAIry","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nwjs/nw.js","owner":"nwjs","name":"nw.js","desc":"Call all Node.js modules directly from DOM/WebWorker and enable a new way of writing applications with all Web technologies.","summaryCn":"用写网页的本事造桌面软件的神器，让前端代码直接调用系统底层能力。","reasonCn":"NW.js 让开发者能用 HTML、CSS 和 JavaScript 构建跨平台桌面应用，并能直接调用 Node.js 的强大能力。对于 AI 项目来说，它是构建本地 AI 助手、数据可视化工具或桌面端 Agent 界面的绝佳载体，让 Web 技术无缝对接原生系统。","stars":41164,"starGrowth":0,"createdAt":"2012-01-04T06:21:10Z","silentRounds":11,"language":"JavaScript","topics":["desktop"],"aiDims":["开发者工具","AI应用","效率工具"],"zone":"工具","funScore":0,"domainTags":["桌面应用","前端开发","Node.js","跨平台"],"domainKey":"桌面应用","pushedAt":"2026-09-23T15:56:46Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/nwjs/nw.js","ts":"2026-09-23T15:56:46Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/open-r1","owner":"huggingface","name":"open-r1","desc":"","summaryCn":"在当前大模型技术狂飙突进的背景下，顶级推理模型往往掌握在少数机构手中，","reasonCn":"项目核心亮点在于完整覆盖了模型训练的关键环节，特别是提供了基于群组相对策略优化的训练脚本，以及监督微调的具体实现方案。除此之外，它还支持从蒸馏小模型或大模型本身出发来生成高质量的合成数据，解决了训练高质量推理模型时最核心的数据瓶颈问题，让整个生态闭环变得更加清","stars":26477,"starGrowth":0,"createdAt":"2025-01-24T15:44:11Z","silentRounds":3,"language":"Python","topics":[],"aiDims":["大语言模型","微调","模型与训练"],"zone":"AI","funScore":0,"domainTags":["大模型","推理","训练"],"domainKey":"训练","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/huggingface/open-r1","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"meta-llama/codellama","owner":"meta-llama","name":"codellama","desc":"","summaryCn":"Code Llama系列模型在训练时采用了长达16k token的序列，这使得它在处理超长","reasonCn":"针对不同规模的计算资源需求，该模型家族提供了7B、13B和34B三种参数规模的变体，灵活满足从个人开发者到企业级的部署环境。其中，特别推出的Python专用版本经过了针对性的深度强化训练，能够更好地理解Python的语法规范和库函数调用习惯。","stars":16250,"starGrowth":0,"createdAt":"2023-08-24T14:25:15Z","silentRounds":9,"language":"Python","topics":[],"aiDims":["大语言模型","代码助手","模型与训练"],"zone":"AI","funScore":0.5,"domainTags":["代码生成","AI编程助手","大语言模型"],"domainKey":"大语言模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/meta-llama/codellama","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"2noise/ChatTTS","owner":"2noise","name":"ChatTTS","desc":"","summaryCn":"ChatTTS 是一款专门针对大语言模型助手等日常对话场景深度优化的生成式语音","reasonCn":"作为开源社区备受关注的项目，它不仅提供了基础的算法基础设施，还搭配了丰富的社区衍生产品生态。从功能特性来看，它目前已经原生支持了中文和英文两种主流语言，满足不同开发者的本地化部署需求。无论是在线客服、语音小说朗读还是智能陪伴应用，它都能派上用场。","stars":39859,"starGrowth":0,"createdAt":"2024-05-27T18:26:49Z","silentRounds":1,"language":"Python","topics":["agent","text-to-speech","chat","chatgpt","chattts","chinese","chinese-language","english","english-language","gpt","llm","llm-agent","natural-language-inference","python","torch","torchaudio","tts"],"aiDims":["AI应用","大语言模型","推理引擎"],"zone":"AI","funScore":0.5,"domainTags":["语音合成","对话生成","生成式 AI","文本转语音"],"domainKey":"语音合成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"text-to-speech","source":"github","weight":0.7},{"name":"chat","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"chattts","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"chinese-language","source":"github","weight":0.7},{"name":"english","source":"github","weight":0.7},{"name":"english-language","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"llm-agent","source":"github","weight":0.7},{"name":"natural-language-inference","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"torch","source":"github","weight":0.7},{"name":"torchaudio","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/2noise/ChatTTS","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MochiDiffusion/MochiDiffusion","owner":"MochiDiffusion","name":"MochiDiffusion","desc":"","summaryCn":"MochiDiffusion解决了在Mac设备上高效、原生运行扩散模型（如Stable Diffus","reasonCn":"在功能层面，MochiDiffusion提供了完整的端到端生图体验。它不仅支持基于文本提示词生成图像，还内建了图库用于保存和同步作品，并支持Image2Image和ControlNet等高级功能，允许用户基于现有图像进行重绘或控制生成结构的细致程度。","stars":7956,"starGrowth":0,"createdAt":"2022-12-15T16:30:56Z","silentRounds":3,"language":"Swift","topics":["apple","apple-silicon","macos","stable-diffusion","coreml","ane","neural-engine","swift","swiftui"],"aiDims":["AI应用","模型部署"],"zone":"创意","funScore":1,"domainTags":["AI绘画","Stable Diffusion","Mac应用","图像生成"],"domainKey":"AI绘画","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"apple","source":"github","weight":0.7},{"name":"apple-silicon","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"coreml","source":"github","weight":0.7},{"name":"ane","source":"github","weight":0.7},{"name":"neural-engine","source":"github","weight":0.7},{"name":"swift","source":"github","weight":0.7},{"name":"swiftui","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/MochiDiffusion/MochiDiffusion","ts":"2026-09-15T21:03:53.738Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bytedance/UI-TARS","owner":"bytedance","name":"UI-TARS","desc":"","summaryCn":"能像人一样操作电脑界面并执行复杂任务的智能体。","reasonCn":"UI-TARS 是一款基于大模型的智能体框架，旨在解决复杂 GUI 操作难题。它不仅能通过视觉理解界面元素，还能像人一样点击、输入，甚至调用外部工具。项目最新版本大幅增强了游戏、代码编写和工具调用能力，通过绝对坐标定位技术精准控制界面。","stars":11511,"starGrowth":0,"createdAt":"2025-01-19T09:04:17Z","silentRounds":2,"language":"Python","topics":["research"],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.5599999999999999,"domainTags":["GUI自动化","网页自动化","模型推理","工具调用","大模型"],"domainKey":"GUI自动化","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"research","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/bytedance/UI-TARS","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"claude-code-best/claude-code","owner":"claude-code-best","name":"claude-code","desc":"","summaryCn":"Claude Code Best 是一个旨在完整复原 Anthropic 官方 Claude Code 功能的开","reasonCn":"在部署便捷性方面，该项目提供了基于 NPM 的安装方式，用户无需克隆庞大的源码仓库，只需通过 npm i -g claude-code-best@latest 即可全局安装最新版本。这种方式显著简化了从安装到启动的流程，即使出现兼容性问题，也有明确的清理和重装指引，确保了主流开发环境下的稳定运行。","stars":22701,"starGrowth":0,"createdAt":"2026-03-31T13:33:21Z","silentRounds":2,"language":"TypeScript","topics":[],"aiDims":["代码助手","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","代码生成","开源替代"],"domainKey":"AI编程助手","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/claude-code-best/claude-code","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chatfire-AI/huobao-drama","owner":"chatfire-AI","name":"huobao-drama","desc":"","summaryCn":"这是一个专为短剧创作者打造的全栈 TypeScript 平台，旨在将繁琐的短剧制作","reasonCn":"该项目的核心亮点在于其强大的多模态 AI 整合能力，真正实现了“一个密钥解锁一切”的便捷体验。用户只需在部署后填入相应的 API 密钥，便能一键写入系统推荐的三项配置，省去了复杂繁琐的环境调试过程。同时，官方贴心地提供了覆盖 macOS 和 Windows 系统的桌面端应用程序，让不习惯","stars":15438,"starGrowth":0,"createdAt":"2026-01-05T11:38:30Z","silentRounds":2,"language":"Vue","topics":[],"aiDims":["AI应用","多模态Agent","创意工具"],"zone":"AI","funScore":0,"domainTags":["AI视频","短剧制作","全栈平台"],"domainKey":"AI视频","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/chatfire-AI/huobao-drama","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"metainternal/llama-cookbook","owner":"metainternal","name":"llama-cookbook","desc":"","summaryCn":"Meta官方Llama模型开发指南，涵盖推理微调与创意应用全流程","reasonCn":"这个项目覆盖的领域非常广泛，从最简单的API入门到复杂的端到端应用都有详细的recipes。其中将Llama API与WhatsApp集成的方案特别实用，让开发者可以轻松把AI能力接入日常通讯场景。此外还有分析研究论文、创建角色心智图等有趣的创意应用，展示了大语言模型在不同行业的潜力。","stars":18559,"starGrowth":0,"createdAt":"2023-07-17T07:33:48Z","silentRounds":11,"language":"Jupyter Notebook","topics":["ai","finetuning","langchain","llama","llama2","llm","machine-learning","python","pytorch","vllm"],"aiDims":["大语言模型","微调","提示工程"],"zone":"资源","funScore":0,"domainTags":["Llama","大模型","微调","教程","API集成"],"domainKey":"Llama","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"finetuning","source":"github","weight":0.7},{"name":"langchain","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"llama2","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"vllm","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/metainternal/llama-cookbook","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"infinispan/infinispan","owner":"infinispan","name":"infinispan","desc":"Infinispan is an open source data grid platform and highly scalable NoSQL cloud data store.","summaryCn":"Java数据网格神器的新玩法，缓存+向量检索一体搞定","reasonCn":"Infinispan是红帽旗下高性能数据网格平台，原生支持Java内存网格、分布式缓存和事务。新增向量数据库能力后，可直接支撑语义搜索和RAG架构，提供REST API和gRPC接口，多语言客户端覆盖Java、Python等，适合需要高并发AI后端的企业级项目。","stars":1352,"starGrowth":0,"createdAt":"2010-11-04T12:33:19Z","silentRounds":28,"language":"Java","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","开发者工具"],"zone":"AI","funScore":0,"domainTags":["向量数据库","分布式缓存","企业级中间件"],"domainKey":"向量数据库","pushedAt":"2026-09-23T21:10:49Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/infinispan/infinispan","ts":"2026-09-23T21:10:49Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"BuilderIO/gpt-crawler","owner":"BuilderIO","name":"gpt-crawler","desc":"","summaryCn":"把任意网站变成专属知识库，一键生成 OpenAI 可用的数据文件。","reasonCn":"这是一个专为 AI 定制知识库打造的开源爬虫工具，能够自动抓取指定网站的文档内容，将其转化为结构化的知识文件。用户只需提供目标网址，它就能像蜘蛛一样扫描网页，提取关键信息并生成供 GPT 模型使用的训练素材。这解决了传统 AI 缺乏特定领域知识、回答不够精准的问题，让模型能基于自有数据进行对话，特别适合构建垂直领域的智能助手或客服系统。","stars":22402,"starGrowth":0,"createdAt":"2023-11-14T00:31:11Z","silentRounds":4,"language":"TypeScript","topics":["ai"],"aiDims":["RAG与知识","AI应用","文档理解"],"zone":"AI","funScore":0,"domainTags":["网络爬虫","知识库构建","数据抓取"],"domainKey":"网络爬虫","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/BuilderIO/gpt-crawler","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nxrix/pixel-8","owner":"nxrix","name":"pixel-8","desc":"Web based fantasy console and JavaScript library","summaryCn":"想象一下，你不需要安装任何复杂的游戏引擎，只要打开网页就能直接开始写代","reasonCn":"从技术架构来看，它完全基于现代 JavaScript 构建，利用 HTML5 Canvas 提供了极其轻量且流畅的像素级渲染能力。它内置了一整套完整的虚拟主机架构，包括经典的像素画板、声音合成器以及简化的输入处理系统，让开发者能够用极少的代码实现复杂的复古视听效果，在性能和包体积上都做","stars":17,"starGrowth":0,"createdAt":"2023-01-06T16:39:44Z","silentRounds":12,"language":"JavaScript","topics":["pixel-art"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["复古游戏","像素艺术","JavaScript","浏览器模拟器"],"domainKey":"复古游戏","pushedAt":"2026-09-10T14:32:36Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/nxrix/pixel-8","ts":"2026-09-10T14:32:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"datawhalechina/easy-rl","owner":"datawhalechina","name":"easy-rl","desc":"","summaryCn":"对于想要踏入人工智能前沿领域的自学者来说，理解复杂的智能决策算法一直是","reasonCn":"该书的核心亮点在于其强大的名师阵容与名校硕博团队的倾力打磨。教程不仅吸收了幽默生动的经典视频讲义，还专门补充了系统性的纲要与从零实践的丰富案例，从基础概念一直铺垫到进阶算法。书中生动的马里奥打游戏等经典比喻，极大地降低了读者的理解门槛，让原本枯燥的理论学习变成","stars":14694,"starGrowth":0,"createdAt":"2020-07-03T02:08:38Z","silentRounds":1,"language":"Jupyter Notebook","topics":["deep-reinforcement-learning","reinforcement-learning","dqn","ppo","a3c","q-learning","sarsa","imitation-learning","policy-gradient","ddpg","easy-rl","double-dqn","dueling-dqn","td3"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["强化学习","机器学习","教程"],"domainKey":"强化学习","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"deep-reinforcement-learning","source":"github","weight":0.7},{"name":"reinforcement-learning","source":"github","weight":0.7},{"name":"dqn","source":"github","weight":0.7},{"name":"ppo","source":"github","weight":0.7},{"name":"a3c","source":"github","weight":0.7},{"name":"q-learning","source":"github","weight":0.7},{"name":"sarsa","source":"github","weight":0.7},{"name":"imitation-learning","source":"github","weight":0.7},{"name":"policy-gradient","source":"github","weight":0.7},{"name":"ddpg","source":"github","weight":0.7},{"name":"easy-rl","source":"github","weight":0.7},{"name":"double-dqn","source":"github","weight":0.7},{"name":"dueling-dqn","source":"github","weight":0.7},{"name":"td3","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/datawhalechina/easy-rl","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GetBindu/Bindu","owner":"GetBindu","name":"Bindu","desc":"Bindu: The identity, communication, and payments layer for AI agents.","summaryCn":"给AI代理发身份证开银行账户，让它们能转账能通信","reasonCn":"Bindu是为AI代理设计的身份通信支付基础设施。每个AI代理可以拥有独特的身份标识，通过它进行安全的身份验证、消息传递和支付交易。支持多种支付渠道，让AI代理能够像人类一样在互联网上自主行动，是构建多代理系统和AI经济生态的重要基础设施。","stars":9820,"starGrowth":1,"createdAt":"2025-03-16T22:37:03Z","silentRounds":5,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"工具","funScore":0.6,"domainTags":["AI代理","数字身份","支付基础设施"],"domainKey":"AI代理","pushedAt":"2026-09-06T14:24:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/GetBindu/Bindu","ts":"2026-09-06T14:24:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AntonOsika/gpt-engineer","owner":"AntonOsika","name":"gpt-engineer","desc":"","summaryCn":"基于大语言模型的代码生成平台，支持用自然语言编写并自动执行软件。","reasonCn":"gpt-engineer 是一款经典的开源 AI 代码生成实验平台。用户只需用自然语言描述所需的软件功能，AI 就能自动生成整套工程代码并尝试运行，随后还能根据用户的改进指令不断迭代项目。它不仅适合快速构建原型，也为探索编程 Agent 提供了可扩展基座。","stars":55089,"starGrowth":0,"createdAt":"2023-04-29T12:52:15Z","silentRounds":10,"language":"Python","topics":["ai","coding-assistant","gpt-4","openai","python","autonomous-agent","codebase-generation","code-generation","codegen","gpt-engineer"],"aiDims":["AI Agent","代码助手","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["代码生成","AI编程","自然语言","Agent"],"domainKey":"Agent","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"coding-assistant","source":"github","weight":0.7},{"name":"gpt-4","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"autonomous-agent","source":"github","weight":0.7},{"name":"codebase-generation","source":"github","weight":0.7},{"name":"code-generation","source":"github","weight":0.7},{"name":"codegen","source":"github","weight":0.7},{"name":"gpt-engineer","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/AntonOsika/gpt-engineer","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"karpathy/micrograd","owner":"karpathy","name":"micrograd","desc":"","summaryCn":"一个极简自动求导引擎，用百行代码教你反向传播原理","reasonCn":"这是特斯拉AI总监Karpathy的经典教学项目。核心是一个迷你自动微分引擎，实现了反向传播算法，只有约100行代码，上面还搭了个50行的神经网络库，API长得像PyTorch。它处理的是标量运算，把每个神经元拆成一个个小的加法和乘法，足以完成二分类任务，适合教学演示。","stars":17629,"starGrowth":0,"createdAt":"2020-04-13T04:31:18Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["模型与训练"],"zone":"资源","funScore":0.53,"domainTags":["自动微分","反向传播","神经网络","深度学习"],"domainKey":"自动微分","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/karpathy/micrograd","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AIGC-Audio/AudioGPT","owner":"AIGC-Audio","name":"AudioGPT","desc":"","summaryCn":"统一整合语音、歌唱、音效与数字人生成的综合多模态 AI 框架。","reasonCn":"AudioGPT 是一个将大语言模型与多种基础音频模型无缝连接的多模态系统。它打破了传统音频 AI 功能单一的限制，不仅能处理人声语音合成与识别，还能进行歌声合成、生成复杂的环境音效，甚至直接配合图像生成说话的数字人视频。通过对话指令即可驱动多类模型协作。","stars":10171,"starGrowth":0,"createdAt":"2023-03-16T07:12:18Z","silentRounds":8,"language":"Python","topics":["audio","gpt","music","sound","speech","talking-head"],"aiDims":["多模态Agent","AI Agent","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["多模态","音频生成","语音合成","数字人"],"domainKey":"多模态","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"audio","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"music","source":"github","weight":0.7},{"name":"sound","source":"github","weight":0.7},{"name":"speech","source":"github","weight":0.7},{"name":"talking-head","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/AIGC-Audio/AudioGPT","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hazel-js/hazeljs","owner":"hazel-js","name":"hazeljs","desc":"HazelJS - Agent OS for TypeScript backends.","summaryCn":"简单来说，这个项目就像是给程序员准备的“智能体操作系统”。","reasonCn":"技术亮点在于它深度融合了TypeScript的类型系统，提供了极佳的代码提示和类型安全保障。它内置了强大的状态管理和记忆回话机制，能够让AI Agent在执行多步复杂任务时不会迷失方向。同时它对工具调用的封装极其优雅，支持异步并发流式输出，性能和扩展性表现都相当优异。","stars":48,"starGrowth":0,"createdAt":"2025-12-08T20:30:48Z","silentRounds":19,"language":"TypeScript","topics":["web-framework"],"aiDims":["AI Agent","开发者工具","AI应用"],"zone":"AI","funScore":0,"domainTags":["智能体框架","TypeScript后端","Agent运行时","状态管理","异步流式输出"],"domainKey":"智能体框架","pushedAt":"2026-09-06T20:25:24Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/hazel-js/hazeljs","ts":"2026-09-06T20:25:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FoundationAgents/MetaGPT","owner":"FoundationAgents","name":"MetaGPT","desc":"","summaryCn":"让多个AI角色组建虚拟团队，协作完成复杂软件项目开发。","reasonCn":"MetaGPT是一个多智能体框架，核心思路给不同GPT分配产品经理、架构师、工程师等角色，组成虚拟软件公司协作完成复杂任务。它解决的问题是让AI不只做单点问答，而是能分工协作完成系统性工程。你给它一个自然语言需求，它就能产出需求文档、设计文档、任务拆解甚至代码，适合想探索AI自动化开发流程的开发者和团队。","stars":70560,"starGrowth":5,"createdAt":"2023-06-30T09:04:55Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0.32,"domainTags":["多智能体","软件开发","自然语言编程","智能体工作流","LLM应用"],"domainKey":"多智能体","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/FoundationAgents/MetaGPT","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015212585034013605,"copyOk":true},{"repo":"karpathy/nn-zero-to-hero","owner":"karpathy","name":"nn-zero-to-hero","desc":"","summaryCn":"这门课程的第一讲就带着学员从零开始编写名为micrograd的微型自动微分库，完","reasonCn":"和市面上多数直接调用PyTorch、TensorFlow等框架教学的入门课程不同，Karpathy选择用最基础的Python代码手动实现自动微分和神经网络层，把抽象的张量计算、链式法则等概念转化为可运行的代码片段。课程配套的Jupyter笔记本直接对应YouTube视频的每一节内容，学员可以一边看视频讲解","stars":24501,"starGrowth":1,"createdAt":"2022-09-08T21:23:25Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["模型与训练","教程"],"zone":"资源","funScore":0.5599999999999999,"domainTags":["神经网络","反向传播","深度学习","micrograd"],"domainKey":"神经网络","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/karpathy/nn-zero-to-hero","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"impredicative/podgenai","owner":"impredicative","name":"podgenai","desc":"OpenAI GPT based informational audiobook/podcast mp3 generator","summaryCn":"输入任意主题就能自动生成专属播客音频文件的GPT生成器","reasonCn":"在技术实现上，它巧妙地串联了文本生成与音频合成两大AI能力。前端通过Python脚本接收用户的输入需求，利用OpenAI的强大理解与创作能力，将枯燥的纯文本转化为带有主持人和嘉宾互动感的播客剧本。随后，它会调用高质量的TTS（语音合成）接口，将文稿转换成音色自然、富有情感的语音","stars":157,"starGrowth":0,"createdAt":"2024-03-29T23:42:20Z","silentRounds":22,"language":"Python","topics":["podcast"],"aiDims":["AI应用","AI写作"],"zone":"创意","funScore":0.8,"domainTags":["Podcast","OpenAI","Audio Generation","Python","GPT"],"domainKey":"Podcast","pushedAt":"2026-09-13T19:05:22Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/impredicative/podgenai","ts":"2026-09-13T19:05:22Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ATH-MaaS/Pixelle-Video","owner":"ATH-MaaS","name":"Pixelle-Video","desc":"","summaryCn":"输入主题即可全自动完成文案、语音和剪辑，让视频创作变成一句话的事。","reasonCn":"它是一个能够将文字创意直接转化为完整视频的全自动AI短视频引擎。用户无需具备任何专业的视频剪辑经验，只要输入一个简单的主题，系统便会自动撰写文案、合成语音、添加背景音乐并一键合成视频，大幅降低了短视频的创作门槛。","stars":28343,"starGrowth":0,"createdAt":"2025-11-07T06:34:27Z","silentRounds":2,"language":"Python","topics":["aigc","comfyui","image-generation","tts","video-generation"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI视频生成","短视频","TTS语音合成","数字人"],"domainKey":"AI视频生成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"aigc","source":"github","weight":0.7},{"name":"comfyui","source":"github","weight":0.7},{"name":"image-generation","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"search","url":"https://github.com/ATH-MaaS/Pixelle-Video","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ScrapeGraphAI/Scrapegraph-ai","owner":"ScrapeGraphAI","name":"Scrapegraph-ai","desc":"Python scraper based on AI","summaryCn":"AI驱动的网页爬虫，自动提取网页信息超方便","reasonCn":"技术上，它结合了LLM的理解能力和RAG（检索增强生成）架构。支持多种LLM后端，包括OpenAI、Anthropic、本地部署的模型等。内置了多种爬虫模式：Graph CRM模式从页面提取结构化数据，Full Site模式可递归爬取多个页面。","stars":31207,"starGrowth":2,"createdAt":"2024-01-27T16:54:38Z","silentRounds":0,"language":"Python","topics":["rag","llm-model","llm"],"aiDims":["AI Agent","RAG与知识","AI应用"],"zone":"AI","funScore":0,"domainTags":["网页爬虫","LLM","RAG"],"domainKey":"RAG","pushedAt":"2026-09-07T13:49:52Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm-model","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/ScrapeGraphAI/Scrapegraph-ai","ts":"2026-09-07T13:49:52Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"alondmnt/joplin-plugin-jarvis","owner":"alondmnt","name":"joplin-plugin-jarvis","desc":"Joplin (note-taking) assistant running a very intelligent system (GPT, Claude, Gemini, Ollama, Hugging Face)","summaryCn":"给知名开源笔记软件装个超级大脑，让你的知识库瞬间拥有AI助手。","reasonCn":"技术上，它作为一个高内聚的TypeScript插件，深度嵌入到了Joplin客户端内部。它支持灵活接入各种主流AI后端，既可以通过API连接云端的GPT、Claude、Gemini，也支持通过Ollama等工具连接本地运行的开源大模型，兼顾了云端的大容量智能与本地的数据隐私安全。","stars":359,"starGrowth":0,"createdAt":"2022-11-20T19:36:25Z","silentRounds":11,"language":"TypeScript","topics":["note-taking"],"aiDims":["AI应用","AI写作","文档理解"],"zone":"AI","funScore":0.6,"domainTags":["Joplin","AI 助手","知识管理","插件"],"domainKey":"Joplin","pushedAt":"2026-09-09T04:28:11Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/alondmnt/joplin-plugin-jarvis","ts":"2026-09-09T04:28:11Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sdaqo/anipy-cli","owner":"sdaqo","name":"anipy-cli","desc":"Little tool in python to watch and download anime from the terminal  (the better way to watch anime). Also applicable as an API ","summaryCn":"Little tool in python to watch and download anime from the terminal.","reasonCn":"这个项目是一个Python工具，用于在终端中观看和下载动漫，非常适合喜欢动漫文化的用户。","stars":521,"starGrowth":0,"createdAt":"2021-12-15T19:43:54Z","silentRounds":1,"language":"Python","topics":["anime"],"aiDims":["动漫"],"zone":"工具","funScore":0,"domainTags":["终端工具","动漫","视频下载"],"domainKey":"终端工具","pushedAt":"2026-09-13T18:15:34Z","tags":[{"name":"动漫","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/sdaqo/anipy-cli","ts":"2026-09-13T18:15:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"leookun/cursor-byok","owner":"leookun","name":"cursor-byok","desc":"cursor-byok is a local implementation of Cursor's backend","summaryCn":"把Cursor的AI后端搬回家，本地也能用Cursor","reasonCn":"这是一个本地实现的Cursor后端项目，允许用户在本地环境中运行Cursor的AI编程功能，而不需要依赖云端服务。通过自部署Cursor的核心引擎，它为开发者提供了更私密、更可控的AI编程体验，特别适合对数据安全和隐私有高要求的场景。","stars":3044,"starGrowth":2,"createdAt":"2026-04-12T09:50:23Z","silentRounds":0,"language":"Rust","topics":["cursor","cursor-cli"],"aiDims":["开发工具","非AI-实用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI编程","自托管","代码补全","隐私保护"],"domainKey":"AI编程","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"开发工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"cursor","source":"github","weight":0.7},{"name":"cursor-cli","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/leookun/cursor-byok","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/autogen","owner":"microsoft","name":"autogen","desc":"","summaryCn":"微软开源的多智能体应用创建框架迎来了重大节点，官方正式宣告该项目已经全","reasonCn":"随着技术的迭代，作为其企业级成熟接班人的微软代理框架如今已正式发布，具备稳定的应用程序接口和长期的官方支持承诺。对于那些希望继续升级现有系统的开发者来说，官方编写的迁移文档提供了清晰的指引，帮助大家将基于旧版框架构建的智能体顺利迁移至新环境。","stars":61115,"starGrowth":4,"createdAt":"2023-08-18T11:43:45Z","silentRounds":0,"language":"Python","topics":["chatgpt","llm-agent","llm-framework","agentic","agentic-agi","agents","ai","autogen","framework","autogen-ecosystem"],"aiDims":["AI Agent","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["多智能体","智能体编排","大模型应用开发"],"domainKey":"多智能体","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"chatgpt","source":"github","weight":0.7},{"name":"llm-agent","source":"github","weight":0.7},{"name":"llm-framework","source":"github","weight":0.7},{"name":"agentic","source":"github","weight":0.7},{"name":"agentic-agi","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"autogen","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"autogen-ecosystem","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/microsoft/autogen","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"mongodb/mongo","owner":"mongodb","name":"mongo","desc":"The MongoDB Database","summaryCn":"全球最火的NoSQL数据库，存数据跑应用都靠它","reasonCn":"由MongoDB Inc开发的开源文档数据库，用C++编写，采用BSON格式存储数据，支持水平扩展和多种索引类型。近年强化向量搜索能力，可直接存储和查询向量嵌入，成为RAG和AI应用常用基础设施。文档型数据模型让开发上手极快，无需预先设计表结构。","stars":28580,"starGrowth":6,"createdAt":"2009-01-15T16:15:18Z","silentRounds":0,"language":"C++","topics":["database"],"aiDims":["AI基础设施","非AI-实用","向量数据库"],"zone":"工具","funScore":0,"domainTags":["NoSQL数据库","文档存储","向量搜索","RAG"],"domainKey":"NoSQL数据库","pushedAt":"2026-09-23T20:42:06Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/mongodb/mongo","ts":"2026-09-23T20:42:06Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018629811056682995,"copyOk":false},{"repo":"jianchang512/ChatTTS-ui","owner":"jianchang512","name":"ChatTTS-ui","desc":"","summaryCn":"本地运行的网页版文字转语音工具，支持中英混排且无需联网即可离线使用","reasonCn":"ChatTTS-ui是一个基于ChatTTS模型的本地网页界面和API服务，让用户方便地将文字合成为自然语音。它特别擅长处理中英文、数字混杂的复杂文本，能保持语气的连贯性。项目提供了Windows预打包版，用户只需解压双击即可运行，极大地降低了使用门槛。同时，系统内置了从Hugging Face或ModelScope自动下载模型的机制，确保在无外网环境下也能顺利获取资源，非常适合需要本地化语音合成的开发者或个人用户。","stars":7659,"starGrowth":0,"createdAt":"2024-05-30T03:11:05Z","silentRounds":4,"language":"Python","topics":["tts","chattts"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.6,"domainTags":["语音合成","TTS","ChatTTS","本地运行"],"domainKey":"语音合成","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"tts","source":"github","weight":0.7},{"name":"chattts","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/jianchang512/ChatTTS-ui","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVlabs/alpamayo","owner":"NVlabs","name":"alpamayo","desc":"NVIDIA Alpamayo 1 Nano is an open 10B reasoning VLA model for autonomous vehicles that pairs driving trajectories with Chain-of-Causation reasoning.","summaryCn":"给自动驾驶汽车装上思考大脑，能边开车边分析因果。","reasonCn":"在技术上，Alpamayo 结合了视觉感知和语言理解。它通过多模态大模型架构，将摄像头捕捉的视觉信息转化为语义描述，再结合推理引擎进行决策。其“Chain-of-Causation”（因果链）机制是核心亮点，它不是简单地识别物体，而是构建事件之间的逻辑链条，从而做出更安全、更符合人类直觉的","stars":2025,"starGrowth":0,"createdAt":"2025-11-19T06:07:06Z","silentRounds":2,"language":"Python","topics":["robotics"],"aiDims":["多模态Agent","机器人","推理引擎"],"zone":"AI","funScore":0,"domainTags":["自动驾驶","视觉语言模型","具身智能"],"domainKey":"自动驾驶","pushedAt":"2026-09-09T23:38:54Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"机器人","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/NVlabs/alpamayo","ts":"2026-09-09T23:38:54Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"BloopAI/bloop","owner":"BloopAI","name":"bloop","desc":"","summaryCn":"开发者在购买大型开源库时，往往面临一个尴尬的困境：文档滞后且晦涩，而代","reasonCn":"在实际应用场景中，Bloop 展现出了极强的务实性。比如当你接手一个旧的遗留系统，想要理解某个业务逻辑块时，你可以直接问它“这个模块是如何处理用户权限的”，它会用大白话解释流程。更重要的是，它支持直接生成代码。","stars":9491,"starGrowth":0,"createdAt":"2022-12-10T14:04:30Z","silentRounds":8,"language":"Rust","topics":[],"aiDims":["代码助手","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["RAG","代码理解","AI编程助手","本地部署"],"domainKey":"RAG","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/BloopAI/bloop","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sameemul-haque/GitArt","owner":"sameemul-haque","name":"GitArt","desc":"📊 GitHub contribution graph customizer • JavaScript, Git, Express.js, HTML","summaryCn":"想象一下，你每天辛勤写代码在GitHub留下的那一串绿格子，其实是可以被“涂鸦","reasonCn":"技术核心在于Git底层对提交时间戳的灵活处理。它通过Express构建起一个后端服务，利用Node.js脚本自动化生成带有特定时间戳的伪提交，从而在GitHub的贡献面板上“染色”。与普通的脚本不同，它实现了图形与时间的坐标映射，不仅能自定义图案内容，还提供了简洁的像素编辑器界面，解决","stars":16,"starGrowth":0,"createdAt":"2024-04-06T16:59:40Z","silentRounds":15,"language":"HTML","topics":["pixel-art"],"aiDims":["非AI-好玩","可视化","开发者工具"],"zone":"创意","funScore":1,"domainTags":["像素艺术","Git操作","GitHub","自动化","创意编程"],"domainKey":"像素艺术","pushedAt":"2026-09-13T04:46:30Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/sameemul-haque/GitArt","ts":"2026-09-13T04:46:30Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ageron/handson-ml3","owner":"ageron","name":"handson-ml3","desc":"","summaryCn":"该仓库核心内容是O'Reilly出版的《Hands-on Machine Learning with Scikit-","reasonCn":"这个项目最大的亮点在于其“零摩擦”的在线运行支持。根据仓库说明，用户无需在本地安装任何Python环境或依赖库，即可通过提供的在线链接（如Colab）直接打开并运行所有笔记。这种设计彻底降低了机器学习入门的环境配置门槛，让学习者可以专注于算法逻辑本身，而不是在折腾虚拟环境上","stars":14196,"starGrowth":0,"createdAt":"2022-02-19T09:43:22Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程","模型与训练"],"zone":"资源","funScore":0,"domainTags":["机器学习","教程","Jupyter","Scikit-learn"],"domainKey":"机器学习","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/ageron/handson-ml3","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"karpathy/minGPT","owner":"karpathy","name":"minGPT","desc":"","summaryCn":"Kaggle明星作者Andrej Karpathy推出的minGPT，将GPT模型的核心逻辑压缩至三","reasonCn":"在实现细节上，minGPT忠实还原了GPT「序列指数输入、下一位置概率分布输出」的基本范式。开发者可以直接通过阅读mingpt/model.py文件，理解全注意力机制如何实现上下文建模。这种代码级透明度使得它成为溯源大模型原理的最佳入口，比黑盒API或庞大框架更具教育价值。","stars":24911,"starGrowth":0,"createdAt":"2020-08-17T07:08:48Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","学习资源"],"zone":"AI","funScore":0,"domainTags":["GPT","Transformer","深度学习","教学"],"domainKey":"GPT","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/karpathy/minGPT","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenCoworkAI/open-codesign","owner":"OpenCoworkAI","name":"open-codesign","desc":"","summaryCn":"OpenCoDesign 在 AI 设计工具领域走出了一条独特的路，其最核心的特性是强调","reasonCn":"在实际功能上，该项目瞄准了从设计到代码的自动化闭环。除了常规的提示词生成 UI 界面外，它最新的更新分支提供了一个令人兴奋的功能：能够将图像直接分解为组件化的 UI 资产包（ui_kits），并可直接移交给编码智能体使用。","stars":7961,"starGrowth":2,"createdAt":"2026-04-18T03:14:33Z","silentRounds":0,"language":"TypeScript","topics":["ai-design","anthropic","byok","claude-design","desktop-app","electron","local-first","multi-model","ollama","openai","claude","claude-design-alternative","claude-code","deepseek","design-to-code","figma-alternative","gemini","openrouter","typescript","ui-generator"],"aiDims":["AI应用","代码助手","创意工具"],"zone":"AI","funScore":0.6,"domainTags":["AI设计","UI生成","本地运行","前端开发"],"domainKey":"AI设计","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"ai-design","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"byok","source":"github","weight":0.7},{"name":"claude-design","source":"github","weight":0.7},{"name":"desktop-app","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"local-first","source":"github","weight":0.7},{"name":"multi-model","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-design-alternative","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"design-to-code","source":"github","weight":0.7},{"name":"figma-alternative","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"openrouter","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"ui-generator","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/OpenCoworkAI/open-codesign","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ggerganov/kbd-audio","owner":"ggerganov","name":"kbd-audio","desc":"","summaryCn":"通过麦克风采集环境音，利用声学特征反推用户敲击的键盘按键序列。","reasonCn":"这是一个基于信号处理的奇思妙想项目。它不依赖视觉或硬件改装，直接分析电脑麦克风录制的键盘敲击声，推测你按了哪些键。除了最有趣的 Keytap 工具，还有无需训练数据的 Keytap2，展示了一种独特的盲打监测思路，非常适合技术爱好者玩味。","stars":9028,"starGrowth":0,"createdAt":"2018-08-27T17:19:02Z","silentRounds":7,"language":"C++","topics":["acoustic","microphone-audio-capture","eavesdrop"],"aiDims":["非AI-好玩","多模态Agent"],"zone":"工具","funScore":1,"domainTags":["声学信号处理","键盘监听","语音分析","麦克风"],"domainKey":"声学信号处理","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"acoustic","source":"github","weight":0.7},{"name":"microphone-audio-capture","source":"github","weight":0.7},{"name":"eavesdrop","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/ggerganov/kbd-audio","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"Solr159/JavBoss","owner":"Solr159","name":"JavBoss","desc":"开箱即用的本地 JAV/视频 刮削、管理、播放软件，支持命令行一键安装和 Docker 部署。只需简单添加目录，即可打造你的私人 JAV/视频 媒体库，带给你顶级的浏览体验，懒人必备。| Your local JAV/video manager.","summaryCn":"本地视频刮削管理神器，一键打造你的私人影视库","reasonCn":"JavBoss是一款用Go语言开发的本地媒体管理软件，支持视频刮削、元数据下载和Docker一键部署。它能够自动整理本地视频文件，生成精美的海报墙和详细信息，支持命令行和Docker两种安装方式，极大提升了本地影音库的管理效率。","stars":660,"starGrowth":0,"createdAt":"2026-03-21T08:09:36Z","silentRounds":1,"language":"Go","topics":["jav","porn","javbus","javlibrary","scraper","javdb","golang"],"aiDims":["非AI-实用"],"zone":"创意","funScore":1,"domainTags":["媒体管理","视频刮削","本地播放","Docker","元数据"],"domainKey":"媒体管理","pushedAt":"2026-09-15T01:07:58.514Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"jav","source":"github","weight":0.7},{"name":"porn","source":"github","weight":0.7},{"name":"javbus","source":"github","weight":0.7},{"name":"javlibrary","source":"github","weight":0.7},{"name":"scraper","source":"github","weight":0.7},{"name":"javdb","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/Solr159/JavBoss","ts":"2026-09-15T01:07:58.514Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"amElnagdy/delegate-skills","owner":"amElnagdy","name":"delegate-skills","desc":"Delegate a coding task to a separate coding agent CLI, review the diff, land the commit yourself — one per implementer.","summaryCn":"让AI自动写代码，生成差异后你自行审阅并提交，省时又可靠","reasonCn":"delegate-skills 提供一个 CLI，能把功能需求交给内部代码生成代理（基于大语言模型），自动产出实现文件并生成Git差异。用户只需审阅差异，确认后工具完成提交，整个过程让 AI 负责繁琐编码，同时保留人工把关，适合多人协作时每人负责一个实现任务，提升效率并降低出错风险。","stars":2156,"starGrowth":13,"createdAt":"2026-06-14T10:35:48Z","silentRounds":0,"language":"JavaScript","topics":["agent-skills","claude-code","coding-agent","openai-codex","ai-coding-agent","antigravity","claude-code-skills","codex","cursor","grok","kimi","mistral","opencode","qoder","skills-sh"],"aiDims":["AI Agent","代码助手","效率工具"],"zone":"AI","funScore":0.6,"domainTags":["AI编程","代码生成","代码审查","Git集成","开发者工具"],"domainKey":"AI编程","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"coding-agent","source":"github","weight":0.7},{"name":"openai-codex","source":"github","weight":0.7},{"name":"ai-coding-agent","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"claude-code-skills","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"grok","source":"github","weight":0.7},{"name":"kimi","source":"github","weight":0.7},{"name":"mistral","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"qoder","source":"github","weight":0.7},{"name":"skills-sh","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.81,"source":"trending","url":"https://github.com/amElnagdy/delegate-skills","ts":"2026-09-07T22:38:09.240Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.06850686456400742,"copyOk":false},{"repo":"Facico/Chinese-Vicuna","owner":"Facico","name":"Chinese-Vicuna","desc":"","summaryCn":"“can be trained on a single Nvidia RTX-2080TI”是该项目最吸引人的技术亮","reasonCn":"项目区分了基础指令模型和多轮对话模型的训练需求。虽然 2080Ti 足以应对前者，但若希望模型具备较长的上下文记忆能力以支持多轮流畅对话，则需要升级到 RTX-3090 级别显卡，且上下文长度限制在 2048 字左右。这种清晰的硬件参数说明，帮助用户根据自己的显卡配置合理规划训练目标","stars":4108,"starGrowth":0,"createdAt":"2023-03-23T01:54:50Z","silentRounds":6,"language":"C","topics":[],"aiDims":["大语言模型","微调"],"zone":"AI","funScore":0.2,"domainTags":["中文模型","LoRA 微调","本地部署"],"domainKey":"中文模型","pushedAt":"2026-09-18T10:52:55.418Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/Facico/Chinese-Vicuna","ts":"2026-09-18T10:52:55.418Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"revaldianggara/PEFT-SingLoRA","owner":"revaldianggara","name":"PEFT-SingLoRA","desc":"PEFT-compatible SingLoRA: single-matrix low-rank adaptation with half the adapter parameters of standard LoRA, drop-in for transformers workflows.","summaryCn":"把LoRA参数减半的轻量微调插件，直接塞进Transformers。","reasonCn":"SingLoRA通过单矩阵低秩适配，将LoRA参数量减至原来一半，兼容PEFT接口，可直接在HuggingFaceTransformers训练流水线中替换使用，保持模型性能的同时显著降低显存占用，并支持梯度累积和混合精度训练，适用于大模型微调。","stars":76,"starGrowth":0,"createdAt":"2022-05-13T06:39:39Z","silentRounds":12,"language":"Python","topics":["transformers"],"aiDims":["微调","模型与训练","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["LoRA","模型微调","PEFT","显存优化","Transformer"],"domainKey":"LoRA","pushedAt":"2026-09-10T07:56:03Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/revaldianggara/PEFT-SingLoRA","ts":"2026-09-10T07:56:03Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"linkedin/Liger-Kernel","owner":"linkedin","name":"Liger-Kernel","desc":"","summaryCn":"LinkedIn发布的Liger Kernel项目，核心目标是通过高效的Triton内核来加速大","reasonCn":"该项目之所以高效，是因为它深入底层，使用Triton这种专为GPU设计的编程语言重写关键算子。在0.3.0版本之前，团队就已经开始与TorchTune协作，发布了关于如何通过torch.compile和Liger Kernel联合优化性能的联合博文，展示了“峰值性能、最小化显存”的技术路线。","stars":6629,"starGrowth":2,"createdAt":"2024-08-06T17:47:52Z","silentRounds":0,"language":"Python","topics":["llm-training","triton","finetuning","gemma2","llama","llama3","llms","mistral","phi3","triton-kernels","hacktoberfest"],"aiDims":["模型与训练","大语言模型","开发者工具"],"zone":"AI","funScore":0,"domainTags":["大模型训练","GPU加速","Triton内核","深度学习"],"domainKey":"大模型训练","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"llm-training","source":"github","weight":0.7},{"name":"triton","source":"github","weight":0.7},{"name":"finetuning","source":"github","weight":0.7},{"name":"gemma2","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"llama3","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"mistral","source":"github","weight":0.7},{"name":"phi3","source":"github","weight":0.7},{"name":"triton-kernels","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/linkedin/Liger-Kernel","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/text-generation-inference","owner":"huggingface","name":"text-generation-inference","desc":"","summaryCn":"该项目已进入维护模式，仅接受轻量更新。它的核心是一个使用 Rust 编写的高","reasonCn":"Text Generation Inference 在 Hugging Face 生产环境中已被广泛使用，支持 Hugging Chat 对话机器人、Inference API 以及各种推理端点。通过它，开发者能够将大语言模型如 GPT、Llama 等部署在自己的基础设施上，而不必担心底层细节。","stars":10885,"starGrowth":0,"createdAt":"2022-10-08T10:26:28Z","silentRounds":8,"language":"Python","topics":["bloom","nlp","pytorch","inference","gpt","deep-learning","transformer","falcon","starcoder"],"aiDims":["推理引擎","大语言模型","模型部署"],"zone":"AI","funScore":0,"domainTags":["大模型推理","模型部署","文本生成"],"domainKey":"模型部署","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"bloom","source":"github","weight":0.7},{"name":"nlp","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"transformer","source":"github","weight":0.7},{"name":"falcon","source":"github","weight":0.7},{"name":"starcoder","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.728,"source":"search","url":"https://github.com/huggingface/text-generation-inference","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"rustic-ai/uni-db","owner":"rustic-ai","name":"uni-db","desc":"Uni is a modern, embedded database that combines property graph (OpenCypher), vector search, and columnar storage (Lance) into a single, cohesive engine. It is designed for applications requiring local, fast, and multimodal data access, backed by object storage (S3/GCS) durability.","summaryCn":"将图数据库、向量搜索和列式存储融为一体的嵌入式 AI 数据库","reasonCn":"Uni 是一款现代嵌入式数据库，同时支持属性图查询 OpenCypher、向量搜索和列式存储 Lance 格式。它专为本地 AI 应用设计，将知识图谱和向量检索能力集成在单一引擎中，避免在不同数据库间切换。Rust 编写保证性能，适合需要本地运行的 RAG 系统和知识管理应用。","stars":54,"starGrowth":0,"createdAt":"2026-01-19T16:55:51Z","silentRounds":18,"language":"Rust","topics":["knowledge-management"],"aiDims":["向量数据库","知识图谱","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["RAG","嵌入式数据库","向量检索","图数据库"],"domainKey":"RAG","pushedAt":"2026-09-11T00:37:43Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/rustic-ai/uni-db","ts":"2026-09-11T00:37:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Aider-AI/aider","owner":"Aider-AI","name":"aider","desc":"","summaryCn":"通过为整个庞大代码库绘制一张清晰的全局结构地图，它彻底解决了大模型在面","reasonCn":"在实际编码过程中，它最令人省心的特性是深度集成了 Git 版本控制系统。当你确认 AI 的修改方案后，它会自动提炼出合情合理的提交说明并完成 commit，你只需要像往常一样通过熟悉的 Git 命令去对比差异、审查变更或者随时撤销不满意的内容，整个过程既流畅又安全。","stars":49118,"starGrowth":4,"createdAt":"2023-05-09T18:57:49Z","silentRounds":0,"language":"Python","topics":["chatgpt","cli","command-line","gpt-4","openai","gpt-3","gpt-35-turbo","claude-3","gpt-4o","anthropic","gemini","llama","sonnet"],"aiDims":["AI Agent","代码助手","大语言模型"],"zone":"工具","funScore":0,"domainTags":["代码编辑","Git集成","AI编程","命令行工具","开发辅助"],"domainKey":"代码编辑","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"chatgpt","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"command-line","source":"github","weight":0.7},{"name":"gpt-4","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"gpt-3","source":"github","weight":0.7},{"name":"gpt-35-turbo","source":"github","weight":0.7},{"name":"claude-3","source":"github","weight":0.7},{"name":"gpt-4o","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"sonnet","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.734,"source":"search","url":"https://github.com/Aider-AI/aider","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Tencent/teamai-cli","owner":"Tencent","name":"teamai-cli","desc":"Make Every Team AI Native","summaryCn":"让团队像原生AI应用一样工作，AI集成进你的开发流程","reasonCn":"它的架构设计非常灵活。它允许你配置不同的AI服务，不管是OpenAI、DeepSeek还是其他兼容接口，都能统一管理。它还支持上下文管理，可以把项目代码、文档、历史对话记录喂给AI，让AI生成的回答更准确、更贴合项目实际。","stars":4905,"starGrowth":0,"createdAt":"2026-04-27T09:42:49Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["AI应用","效率工具","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI集成","CLI工具","开发流程","团队协作","Prompt工程"],"domainKey":"AI集成","pushedAt":"2026-09-08T17:39:23.082Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/Tencent/teamai-cli","ts":"2026-09-08T17:39:23.082Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"koharu-rs/koharu","owner":"koharu-rs","name":"koharu","desc":"AI-powered manga translator, written in Rust.","summaryCn":"用AI一键翻译日漫，让追番不再受语言限制","reasonCn":"基于Rust打造的高性能漫画翻译工具，利用AI视觉模型自动识别对话框文字区域，结合OCR与机器翻译技术将日文精准转为中文。支持多种翻译后端接口，可本地部署保护隐私。自动排版适配，一键完成整本漫画翻译，无需手动调试，对漫画爱好者和汉化组都极为实用。","stars":5624,"starGrowth":5,"createdAt":"2025-04-09T10:52:34Z","silentRounds":0,"language":"Rust","topics":["deep-learning","manga","rust","japanese","computer-vision","gpu","gui","translation","text-rendering","webgpu"],"aiDims":["AI应用","非AI-好玩","多模态Agent"],"zone":"AI","funScore":0.9,"domainTags":["漫画翻译","OCR","机器翻译","图像处理","隐私保护"],"domainKey":"漫画翻译","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"manga","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"japanese","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"gpu","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"translation","source":"github","weight":0.7},{"name":"text-rendering","source":"github","weight":0.7},{"name":"webgpu","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/koharu-rs/koharu","ts":"2026-09-07T22:38:09.240Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.017667140825035562,"copyOk":false},{"repo":"suyuan32/simple-admin-core","owner":"suyuan32","name":"simple-admin-core","desc":"Simple Admin是一个基于Go Zero开发面向小型到大型项目的分布式微服务后端管理系统脚手架，提供丰富的后台管理功能，支持k8s快速部署，助力快速开发高并发微服务集群，适合学习和商用。Simple Admin is a powerful microservice framework for large management system. It is based on go-zero and supports several advanced features. It can help you to develop a microservice back-end management system in a short time. ","summaryCn":"用Go语言一键生成微服务后台管理系统，高并发项目必备神器","reasonCn":"基于 go-zero 框架构建的微服务管理后台脚手架，内置 RBAC 权限、操作日志、代码生成等核心功能。支持一键部署至 Kubernetes 集群，解决了从零搭建后端系统的重复劳动问题。代码结构规范，适合作为 Go 语言后端开发的入门模板或企业级项目起点，帮你节省大量基础架构搭建时间。","stars":2060,"starGrowth":0,"createdAt":"2021-11-01T12:14:11Z","silentRounds":13,"language":"Go","topics":["backend"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["微服务架构","Go语言","Kubernetes部署","RBAC权限","代码生成"],"domainKey":"微服务架构","pushedAt":"2026-09-17T21:57:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/suyuan32/simple-admin-core","ts":"2026-09-17T21:57:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"liyupi/niulai-game","owner":"liyupi","name":"niulai-game","desc":"🐂 牛来格斗 NIU LAI FIGHT — 3D 双人格斗游戏 / A 3D two-player fighting game with Three.js: 7 fighters, 3 stages, combo & super moves, local versus + LAN multiplayer. 零资源全程序化生成，致敬拳皇与电影《牛来》","summaryCn":"拳皇同款3D格斗，牛来电影致敬，零资源程序生成","reasonCn":"基于Three.js开发的3D双人格斗游戏，拥有7位可选角色、3个特色场景，支持连招、必杀技及局域网对战。游戏特色在于完全程序化生成资源，无需下载任何素材包，致敬经典格斗游戏与电影《牛来》，支持本地与联网双模式。","stars":14,"starGrowth":0,"createdAt":"2026-08-28T03:44:35Z","silentRounds":18,"language":"JavaScript","topics":["procedural-generation"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["格斗游戏","Three.js","程序化生成"],"domainKey":"格斗游戏","pushedAt":"2026-08-28T03:44:55Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/liyupi/niulai-game","ts":"2026-08-28T03:44:55Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"autowarefoundation/vision_pilot","owner":"autowarefoundation","name":"vision_pilot","desc":"Free and fully open-source L2 ADAS stack powered by End-to-End AI technology","summaryCn":"开源的端到端AI自动驾驶系统，让你的智能车也能拥有老司机般的视觉脑。","reasonCn":"该项目是由知名自动驾驶基金会推出的全开源L2级高级辅助驾驶技术栈。它完全基于前沿的端到端人工智能算法构建，打破了传统自动驾驶方案模块过多的痛点，通过强大的深度学习模型直接将传感器数据转化为精准的车辆控制指令，极具前瞻性。","stars":901,"starGrowth":0,"createdAt":"2024-09-24T17:49:00Z","silentRounds":2,"language":"C++","topics":["robotics"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["自动驾驶","端到端AI","机器人"],"domainKey":"自动驾驶","pushedAt":"2026-09-11T16:43:21Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/autowarefoundation/vision_pilot","ts":"2026-09-11T16:43:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mindfold-ai/Trellis","owner":"mindfold-ai","name":"Trellis","desc":"","summaryCn":"Trellis 核心逻辑在于将项目记忆持久化到代码仓库中，确保任何编码代理都能","reasonCn":"上手使用该项目非常直观，它将繁琐的上下文同步过程简化为简单的交互步骤。用户只需要根据项目指引，在终端进行简单的初始化配置，即可让 AI 介入开发流程。在实际使用中，你只需用自然语言描述需求，系统会引导你与 AI 一步步拆解任务，这种互动方式有效地将开发者的思路与 AI 的","stars":14775,"starGrowth":3,"createdAt":"2026-01-26T11:49:10Z","silentRounds":0,"language":"TypeScript","topics":["agentic-coding","ai-workflow","codex","harness","claudecode"],"aiDims":["AI Agent","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["项目记忆","AI编程","代码仓库","上下文管理"],"domainKey":"项目记忆","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"agentic-coding","source":"github","weight":0.7},{"name":"ai-workflow","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"harness","source":"github","weight":0.7},{"name":"claudecode","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/mindfold-ai/Trellis","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"linearmouse/linearmouse","owner":"linearmouse","name":"linearmouse","desc":"The mouse and trackpad utility for Mac.","summaryCn":"很多换到Mac的老用户都有一个痛点：怎么买了个几百块的鼠标，在苹果电脑上用","reasonCn":"这款软件的核心逻辑在于其极其轻量化的驱动拦截方案。它不需要频繁读取硬件数据，而是在系统层级对鼠标的输入信号进行修正，保证了极低的CPU占用率。它不仅能让鼠标移动变得线性、可控，还支持为不同设备设置不同的滚轮翻滚逻辑，比如你想要触控板自然滚动，但鼠标滚轮必须是传统逻","stars":6881,"starGrowth":0,"createdAt":"2021-06-10T10:09:21Z","silentRounds":1,"language":"Swift","topics":["productivity"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Mac","鼠标优化","输入设备"],"domainKey":"Mac","pushedAt":"2026-09-14T16:29:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/linearmouse/linearmouse","ts":"2026-09-14T16:29:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EthanAlgoX/LLM-TradeBot","owner":"EthanAlgoX","name":"LLM-TradeBot","desc":"A multi-agent AI trading system using LLMs to optimize strategies and adapt to market conditions in real-time.","summaryCn":"让多个AI角色组成炒股委员会，全自动帮你盯着市场实时分析并调整策略","reasonCn":"在底层架构上，项目运用了现代的多智能体协同框架，每个AI代理都拥有特定的提示词和工具调用权限。通过大模型强大的推理能力，系统能够理解复杂的市场情绪并动态调整交易策略，而不是死守传统的机械化阈值。这种设计让它比传统的量化网格交易更加灵活，能够适应突发利好或黑天鹅事","stars":332,"starGrowth":0,"createdAt":"2025-12-20T10:04:55Z","silentRounds":2,"language":"Python","topics":["trading"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["LLM 应用","量化交易","多智能体","金融科技","策略优化"],"domainKey":"LLM 应用","pushedAt":"2026-09-06T18:54:04Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/EthanAlgoX/LLM-TradeBot","ts":"2026-09-06T18:54:04Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"henrypp/simplewall","owner":"henrypp","name":"simplewall","desc":"","summaryCn":"基于底层平台构建的轻量级 Windows 网络活动配置工具。","reasonCn":"这是一个专为高级 Windows 用户打造的轻量级网络过滤配置工具。它不依附于系统自带的防火墙控制界面，而是直接基于微软的 Windows Filtering Platform 内部接口与服务运行。通过它，用户可以精细地掌控电脑上每一个应用程序的网络访问权限，有效拦截未经允许的联网请求。程序本身极为小巧且兼容主流的微软操作系统，是管理设备网络安全与隐私保护的实用利器。","stars":9039,"starGrowth":3,"createdAt":"2016-08-18T17:10:56Z","silentRounds":0,"language":"C","topics":["wfp","network","firewall","windows","windows-firewall","foss","simplewall","arm64","win64"],"aiDims":["安全工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网络安全","防火墙","Windows 工具","网络管理"],"domainKey":"网络安全","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"wfp","source":"github","weight":0.7},{"name":"network","source":"github","weight":0.7},{"name":"firewall","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"windows-firewall","source":"github","weight":0.7},{"name":"foss","source":"github","weight":0.7},{"name":"simplewall","source":"github","weight":0.7},{"name":"arm64","source":"github","weight":0.7},{"name":"win64","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/henrypp/simplewall","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"datahub-project/datahub","owner":"datahub-project","name":"datahub","desc":"The Context Platform for your Data and AI Stack","summaryCn":"企业数据资产管家，让数据像水一样透明好找。","reasonCn":"DataHub的核心能力在于它的元数据摄取和图谱构建。它支持通过REST API、数据库触发器等多种方式自动抓取数据结构。更重要的是，它能理解数据之间的血缘关系，当你修改了一个表字段时，它能告诉你这会影响哪些BI大屏和机器学习模型。","stars":12751,"starGrowth":1,"createdAt":"2015-11-18T05:47:40Z","silentRounds":1,"language":"Python","topics":["metadata","datahub","data-catalog","data-discovery","data-governance","agent-platform","context-management","data-observability"],"aiDims":["AI基础设施","文档理解","知识图谱"],"zone":"工具","funScore":0,"domainTags":["元数据管理","数据血缘","数据资产"],"domainKey":"元数据管理","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"metadata","source":"github","weight":0.7},{"name":"datahub","source":"github","weight":0.7},{"name":"data-catalog","source":"github","weight":0.7},{"name":"data-discovery","source":"github","weight":0.7},{"name":"data-governance","source":"github","weight":0.7},{"name":"agent-platform","source":"github","weight":0.7},{"name":"context-management","source":"github","weight":0.7},{"name":"data-observability","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/datahub-project/datahub","ts":"2026-09-11T22:17:29.879Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apple-aiml-research/ml-ferret","owner":"apple-aiml-research","name":"ml-ferret","desc":"","summaryCn":"苹果人工智能研究团队推出的这个开源项目，致力于解决计算机视觉领域中精细","reasonCn":"在核心亮点方面，它改变了以往多模态模型只能处理全局图像或固定矩形框的局限，允许用户以任意形式在图像上划定范围并获取针对性的响应结果。这种突破性的全颗粒度指代能力，使得它在智能终端界面分析、复杂图像问答以及细粒度视觉推理等场景中展现出巨大潜力。","stars":8662,"starGrowth":0,"createdAt":"2023-10-06T21:39:03Z","silentRounds":9,"language":"Python","topics":[],"aiDims":["多模态Agent","大语言模型","文档理解"],"zone":"AI","funScore":0,"domainTags":["多模态视觉","精细图像理解","UI界面分析","细粒度视觉推理","计算机视觉"],"domainKey":"多模态视觉","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/apple-aiml-research/ml-ferret","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/DeepSeek-OCR","owner":"deepseek-ai","name":"DeepSeek-OCR","desc":"","summaryCn":"2025年10月20日，DeepSeek 团队正式发布了 DeepSeek-OCR，这是一款专门用于","reasonCn":"在技术落地方面，该模型已正式获得上游 vLLM 框架的原生支持。这意味着开发者无需进行复杂的自定义适配，即可利用 vLLM 强大的推理加速能力来运行 DeepSeek-OCR。这一特性显著降低了部署门槛，使得高性能的视觉-文本压缩推理能够在生产环境中快速应用，同时也体现了该项目在开源社","stars":23905,"starGrowth":0,"createdAt":"2025-10-17T06:14:27Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["文档理解","大语言模型","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["光学字符识别","多模态","视觉语言模型","vLLM","深度学习"],"domainKey":"多模态","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"tier","url":"https://github.com/deepseek-ai/DeepSeek-OCR","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"facebookresearch/seamless_communication","owner":"facebookresearch","name":"seamless_communication","desc":"","summaryCn":"Seamless 多语言模型家族能边听边译并带走说话人的语气。","reasonCn":"它面向跨语言沟通中的机械感，用一组模型同时处理翻译、语音识别和语气节奏。你输入不同语言，它能尽量保留说话人的语气、停顿和声音风格，并支持接近实时的同传与流式识别。相比只追求字面准确，它更在意听上去自然，也适合研究者快速验证多语种语音交互产品。","stars":11863,"starGrowth":0,"createdAt":"2023-08-01T20:36:21Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["AI应用","模型与训练","模型部署"],"zone":"AI","funScore":0,"domainTags":["多语言翻译","语音识别","实时同传","语音交互","跨语言沟通"],"domainKey":"语音识别","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.716,"source":"tier","url":"https://github.com/facebookresearch/seamless_communication","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"lllyasviel/stable-diffusion-webui-forge","owner":"lllyasviel","name":"stable-diffusion-webui-forge","desc":"","summaryCn":"Forge的诞生源于一个痛点：Stable Diffusion WebUI虽然功能强大，但显存占用","reasonCn":"项目名字致敬Minecraft Forge——那个让《我的世界》玩法无限扩展的模组加载器。开发者的野心很明确：不仅要做一个优化版，更要成为SD WebUI生态的底层平台，让后续各种实验性功能、第三方插件都能在Forge上快速落地。","stars":13029,"starGrowth":1,"createdAt":"2024-01-14T11:39:30Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI应用","推理引擎","模型部署"],"zone":"AI","funScore":0.8,"domainTags":["Stable Diffusion","AI绘画","显存优化","WebUI"],"domainKey":"Stable Diffusion","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/lllyasviel/stable-diffusion-webui-forge","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"akdeb/ElatoAI","owner":"akdeb","name":"ElatoAI","desc":"Realtime Voice AI with 100+ Models on Arduino ESP32 with Secure Websockets and Edge Functions for AI Companions, and Devices","summaryCn":"让你的 ESP32 开发板秒变能听会说的 AI 智能陪伴机器人","reasonCn":"该项目利用 ESP32 硬件配合安全 WebSocket 和边缘函数，实现了低延迟的实时语音交互。它支持集成超过一百种 AI 模型，将语音采集、流式传输与云端推理无缝结合，让简单的微控制器也能拥有强大的 AI 语音对话和设备控制能力。","stars":2017,"starGrowth":0,"createdAt":"2025-04-08T13:03:05Z","silentRounds":3,"language":"TypeScript","topics":["arduino"],"aiDims":["AI应用","多模态Agent","非AI-好玩"],"zone":"AI","funScore":1,"domainTags":["ESP32","语音交互","边缘AI","WebSocket","智能硬件"],"domainKey":"ESP32","pushedAt":"2026-09-02T08:38:22Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/akdeb/ElatoAI","ts":"2026-09-02T08:38:22Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"didilili/ai-agents-from-zero","owner":"didilili","name":"ai-agents-from-zero","desc":" 🚀 2026 最系统的 AI Agent 速成指南｜智能体实战教程 · 完整学习路径  + 实战项目 + 面试题库 · 对标大模型应用开发工程师岗位 · 覆盖LangChain / LangGraph / Coze / Dify / MCP / skills / LLM / RAG / 提示词 · 企业级部署与微调 · 从0到企业级落地 + 从学习到上线项目 + 面试准备一体化 ","summaryCn":"手把手教你从零搭建企业级 AI 智能体，配实战项目和面试题。","reasonCn":"本仓库提供系统化学习路径，涵盖 LangChain、LangGraph、Coze、Dify 等主流框架，深入讲解 RAG、提示词、技能库构建，并配套企业级部署、模型微调、面试题库，帮助从理论到落地全链路提升。还提供代码示例、最佳实践及常见坑点，支持本地和云端多种部署，兼容 OpenAI、Claude、Gemini 等模型。","stars":4838,"starGrowth":1,"createdAt":"2026-01-29T02:49:11Z","silentRounds":2,"language":"Python","topics":["rag","ai-agent"],"aiDims":["AI Agent","教程","RAG与知识"],"zone":"资源","funScore":0.5,"domainTags":["AI Agent","学习路径","教程","面试准备"],"domainKey":"AI Agent","pushedAt":"2026-09-10T01:47:11Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/didilili/ai-agents-from-zero","ts":"2026-09-10T01:47:11Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jina-ai/clip-as-service","owner":"jina-ai","name":"clip-as-service","desc":"","summaryCn":"基于CLIP模型的图文嵌入微服务，支持高吞吐与水平扩展。","reasonCn":"除了速度，该服务在设计上极具弹性。它支持在单一GPU实例上水平扩展多个CLIP模型自动负载均衡，用户无需编写复杂的调度逻辑，即可实现服务的平滑扩容。同时，非阻塞的双工流式处理机制针对大数据传输和长周期任务进行了专门优化，确保在处理大图或长文档时不会占用主线程，从而保障","stars":12831,"starGrowth":0,"createdAt":"2018-11-12T10:48:50Z","silentRounds":11,"language":"Python","topics":["bert","sentence-encoding","deep-learning","clip-model","clip-as-service","bert-as-service","cross-modal-retrieval","multi-modality","neural-search","openai","pytorch","onnx","cross-modality","image2vec","sentence2vec"],"aiDims":["模型部署","AI基础设施","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["CLIP模型","多模态检索","微服务"],"domainKey":"CLIP模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"bert","source":"github","weight":0.7},{"name":"sentence-encoding","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"clip-model","source":"github","weight":0.7},{"name":"clip-as-service","source":"github","weight":0.7},{"name":"bert-as-service","source":"github","weight":0.7},{"name":"cross-modal-retrieval","source":"github","weight":0.7},{"name":"multi-modality","source":"github","weight":0.7},{"name":"neural-search","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"onnx","source":"github","weight":0.7},{"name":"cross-modality","source":"github","weight":0.7},{"name":"image2vec","source":"github","weight":0.7},{"name":"sentence2vec","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/jina-ai/clip-as-service","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tensorflow/tfjs","owner":"tensorflow","name":"tfjs","desc":"","summaryCn":"TensorFlow.js 是一个开源的硬件加速 JavaScript 库，它彻底改变了机器学习","reasonCn":"这个工具的核心亮点在于强大的模型迁移与端侧重训能力。你不需要重写复杂的后端服务，只需使用官方提供的模型转换器，就能把现存的 TensorFlow 模型直接无缝加载到浏览器里跑起来。更厉害的是，它支持收集浏览器连接的各种传感器数据或本地客户端数据，直接在用户的设备端对已有模","stars":19142,"starGrowth":0,"createdAt":"2018-03-05T05:41:02Z","silentRounds":9,"language":"TypeScript","topics":["deep-learning","typescript","webgl","javascript","machine-learning","neural-network","deep-neural-network","gpu-acceleration","wasm","web-assembly"],"aiDims":["AI基础设施","模型部署","开发者工具"],"zone":"AI","funScore":0,"domainTags":["JavaScript","机器学习","浏览器AI","模型部署"],"domainKey":"模型部署","pushedAt":"2026-09-14T22:13:28.570Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"neural-network","source":"github","weight":0.7},{"name":"deep-neural-network","source":"github","weight":0.7},{"name":"gpu-acceleration","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"web-assembly","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/tensorflow/tfjs","ts":"2026-09-14T22:13:28.570Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"project-zot/zot","owner":"project-zot","name":"zot","desc":"zot - A scale-out production-ready vendor-neutral OCI-native container image/artifact registry (purely based on OCI Distribution Specification)","summaryCn":"一个轻量级且高性能的容器镜像仓库，帮你安全存储和分发模型文件。","reasonCn":"Zot 是一款基于 OCI 标准的容器镜像仓库，完全用 Go 语言编写，旨在提供高性能和易用性。它支持分布式部署，非常适合企业级环境，能够无缝集成到现有的 Kubernetes 或 Docker 生态中。作为云原生时代的“数字货架”，Zot 让 AI 模型分发变得像管理 Docker 镜像一样简单。","stars":2794,"starGrowth":1,"createdAt":"2019-06-21T21:40:58Z","silentRounds":1,"language":"Go","topics":["zot","opencontainers","oci","distribution-spec","kubernetes","oci-distribution","helm"],"aiDims":["AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["容器镜像仓库","OCI","Kubernetes","容器编排","私有云"],"domainKey":"容器镜像仓库","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"zot","source":"github","weight":0.7},{"name":"opencontainers","source":"github","weight":0.7},{"name":"oci","source":"github","weight":0.7},{"name":"distribution-spec","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"oci-distribution","source":"github","weight":0.7},{"name":"helm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/project-zot/zot","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langchain-ai/chat-langchain","owner":"langchain-ai","name":"chat-langchain","desc":"","summaryCn":"项目核心采用了托管深度代理（Managed Deep Agents）技术，确保用户在获取解","reasonCn":"作为一套完整的解决方案，它不仅包含了后端的逻辑处理，还贴心地提供了Next.js前端界面。这意味着开发者无需从零开始编写交互逻辑，只需直接部署其前端模块，即可迅速获得一个具有现代UI风格的文档对话框，大幅降低了从原型到上线的产品化周期。","stars":6461,"starGrowth":0,"createdAt":"2023-01-16T20:43:13Z","silentRounds":3,"language":"TypeScript","topics":[],"aiDims":["AI Agent","RAG与知识","文档理解"],"zone":"AI","funScore":0.2,"domainTags":["文档助手","智能代理","知识检索"],"domainKey":"文档助手","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/langchain-ai/chat-langchain","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"tomicz/unity-procedurally-generated-grid","owner":"tomicz","name":"unity-procedurally-generated-grid","desc":"📐 Procedurally generated grid for Unity rendered as a single vertex-colored mesh. One draw call at any size, per-node occupancy and color, built-in A* pathfinding, saveable layouts, custom inspector. 2D & 3D, UPM package, Unity 2021.3+, MIT.","summaryCn":"程序化生成网格，像搭积木一样玩Unity","reasonCn":"它的技术亮点非常硬核。它使用“单顶点颜色网格”技术，不管网格有多大，Unity引擎只需要一次 `DrawCall` 就能画完。这意味着你可以轻松创建数千个格子的地图，而不会卡顿。同时，它内置了A*寻路算法，你不用自己写算法，直接调用API就能得到最短路径。","stars":18,"starGrowth":0,"createdAt":"2021-03-24T17:14:47Z","silentRounds":11,"language":"C#","topics":["procedural-generation"],"aiDims":["非AI-好玩","游戏开发","可视化"],"zone":"工具","funScore":0,"domainTags":["Unity","程序化生成","游戏开发"],"domainKey":"Unity","pushedAt":"2026-09-12T21:25:48Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏开发","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/tomicz/unity-procedurally-generated-grid","ts":"2026-09-12T21:25:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"acheong08/ChatGPT","owner":"acheong08","name":"ChatGPT","desc":"","summaryCn":"让开发者能直接在代码里调用ChatGPT，而不必依赖官方提供的付费API，这是该","reasonCn":"在使用限制方面，项目明确设定了代理服务器每 10 秒仅允许 5 次请求，而从 OpenAI 端来看单个账号每小时上限为 50 次请求。这些硬性限制提醒用户，这并非用于高并发生产环境，而是适合个人开发调试、学习笔记制作或小范围的内部分享场景。","stars":27891,"starGrowth":0,"createdAt":"2022-12-03T14:35:24Z","silentRounds":1,"language":"Python","topics":["chatgpt","cli","library","pypi-package","gptchat","revchatgpt","gpt-35-turbo"],"aiDims":["AI应用","大语言模型"],"zone":"AI","funScore":0,"domainTags":["ChatGPT","API代理","非官方客户端"],"domainKey":"ChatGPT","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"chatgpt","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"library","source":"github","weight":0.7},{"name":"pypi-package","source":"github","weight":0.7},{"name":"gptchat","source":"github","weight":0.7},{"name":"revchatgpt","source":"github","weight":0.7},{"name":"gpt-35-turbo","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/acheong08/ChatGPT","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jd-opensource/joyagent-jdgenie","owner":"jd-opensource","name":"joyagent-jdgenie","desc":"","summaryCn":"在 GAIA 榜单 Validation 集上，JoyAgent‑JDGenie 的准确率达到了 75.15%，","reasonCn":"JoyAgent‑JDGenie 采用轻量化的多智能体架构，核心是可插拔子智能体与工具。用户只需挂载自定义的子智能体或第三方工具，即可扩展功能。例如，内置的 JoyDataAgent 负责数据治理与智能问数，支持 DGP 协议、数据治理与诊断分析。","stars":11923,"starGrowth":0,"createdAt":"2025-07-16T02:59:53Z","silentRounds":1,"language":"Java","topics":[],"aiDims":["AI Agent","多模态Agent","RAG与知识"],"zone":"AI","funScore":0.8,"domainTags":["多模态知识管理","RAG","多智能体","数据治理","Java"],"domainKey":"RAG","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/jd-opensource/joyagent-jdgenie","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cumulo-autumn/StreamDiffusion","owner":"cumulo-autumn","name":"StreamDiffusion","desc":"","summaryCn":"传统扩散模型虽然画质出色，但生成一张图往往需要十几秒甚至更久，这种延迟","reasonCn":"其核心创新在于“流式批处理”策略。系统将原本一次性完成的大步长去噪过程，拆解为多个可重叠执行的小批次流水线，隐藏了计算等待时间。这种工程化的优化让 GPU 利用率维持在高位，使得原本离线的生成任务具备了视频帧率般的实时性，用户输入提示词或拖动参数时，画面能即时反馈变化","stars":10814,"starGrowth":0,"createdAt":"2023-11-28T13:40:30Z","silentRounds":7,"language":"Python","topics":[],"aiDims":["推理引擎","模型部署","AI应用"],"zone":"AI","funScore":1,"domainTags":["实时图像生成","扩散模型","AI绘画","流式处理"],"domainKey":"实时图像生成","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/cumulo-autumn/StreamDiffusion","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/DeepSeek-Coder","owner":"deepseek-ai","name":"DeepSeek-Coder","desc":"","summaryCn":"DeepSeek-Coder 每一个版本均从零开始使用了 2 万亿个 token 进行高强度预训","reasonCn":"针对项目级代码补全的需求，该模型通过采用 16K 的上下文窗口和专门设计的填空任务（Infilling）进行优化。这种机制让模型不仅仅是简单地预测下一行代码，而是能够根据周围现有的代码结构，智能地填充缺失部分，这对于处理具有相互依赖关系的复杂项目代码至关重要，极大减少了开发","stars":24284,"starGrowth":0,"createdAt":"2023-10-20T06:38:01Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["代码助手","大语言模型","模型与训练"],"zone":"AI","funScore":0,"domainTags":["代码生成","代码补全","大语言模型","编程模型","DeepSeek"],"domainKey":"大语言模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/deepseek-ai/DeepSeek-Coder","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"catdad/canvas-confetti","owner":"catdad","name":"canvas-confetti","desc":"","summaryCn":"只需将这个库通过CDN直接引入到HTML文档中，原本枯燥的页面交互瞬间就能变成","reasonCn":"该项目的核心亮点在于极致的易用性与精良的动画逻辑。无论是粒子的大小、数量、颜色还是飘落的速度，都能通过简单的参数进行调整，使其能完美适配不同风格的网站设计。这不仅仅是一个简单的绘图工具，它更像是一个为网页增加“情绪价值”的开关，让枯燥的代码响应变得富有情感交流的","stars":12760,"starGrowth":1,"createdAt":"2018-01-26T03:38:52Z","silentRounds":2,"language":"JavaScript","topics":["canvas","confetti","jsdelivr","animation","burst","fireworks","snow","particles","canvas-confetti","hacktoberfest"],"aiDims":["非AI-好玩","可视化"],"zone":"创意","funScore":1,"domainTags":["Canvas动画","网页特效","庆祝动画","交互反馈","轻量级"],"domainKey":"Canvas动画","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"canvas","source":"github","weight":0.7},{"name":"confetti","source":"github","weight":0.7},{"name":"jsdelivr","source":"github","weight":0.7},{"name":"animation","source":"github","weight":0.7},{"name":"burst","source":"github","weight":0.7},{"name":"fireworks","source":"github","weight":0.7},{"name":"snow","source":"github","weight":0.7},{"name":"particles","source":"github","weight":0.7},{"name":"canvas-confetti","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/catdad/canvas-confetti","ts":"2026-09-13T22:08:09.499Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liyupi/ai-guide","owner":"liyupi","name":"ai-guide","desc":"程序员鱼皮的 AI 资源大全 + Vibe Coding 零基础教程，分享 OpenClaw 保姆级教程、大模型玩法（DeepSeek / GPT / Gemini / Claude / GLM）、最新 AI 资讯、Prompt 提示词大全、AI 知识百科（Agent Skills / RAG / MCP / A2A）、AI 编程教程（Harness Engineering）、AI 工具用法（Cursor / Claude Code / TRAE / Codex / Copilot）、AI 开发框架教程（Spring AI / LangChain）、AI 产品变现指南，帮你快速掌握 AI 技术，走在时代前沿。本项目为开源文档 aiguide，已升级为鱼皮 AI 导航网站","summaryCn":"程序员鱼皮整理的AI学习资源大全，从入门到进阶一站搞定","reasonCn":"这个项目是程序员鱼皮精心整理的AI学习资源集合，涵盖主流大模型玩法教程、Prompt提示词大全、AI知识百科等内容。收录了OpenClaw保姆级教程、Agent Skills、RAG、MCP等热门技术方向，以及最新AI资讯和编程工具用法，适合想系统学习AI的入门者和进阶开发者。","stars":20337,"starGrowth":1,"createdAt":"2025-02-13T05:47:46Z","silentRounds":1,"language":"JavaScript","topics":["rag","ai-agent"],"aiDims":["学习资源","教程","提示工程"],"zone":"资源","funScore":0,"domainTags":["AI 学习","教程导航","Prompt"],"domainKey":"AI 学习","pushedAt":"2026-09-11T08:10:09Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/liyupi/ai-guide","ts":"2026-09-11T08:10:09Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"awesome-foss/awesome-sysadmin","owner":"awesome-foss","name":"awesome-sysadmin","desc":"A curated list of amazingly awesome open-source sysadmin resources.","summaryCn":"运维人必备的宝藏清单，管理服务器有它就够了","reasonCn":"这是一个精心整理的运维资源合集，涵盖自动化部署、监控、日志分析、容器编排等全栈技术栈。由社区维护，包含大量高质量工具和最佳实践，是系统管理员和DevOps工程师的案头必备指南，帮助快速查找替代方案和提升工作效率。","stars":35235,"starGrowth":2,"createdAt":"2014-12-23T08:41:22Z","silentRounds":0,"language":"","topics":["devops","awesome","self-hosted"],"aiDims":["非AI-实用","开发工具","效率工具"],"zone":"资源","funScore":0,"domainTags":["运维","开源工具","系统管理"],"domainKey":"运维","pushedAt":"2026-09-17T15:16:20Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7}],"aiScore":0.6,"source":"search","url":"https://github.com/awesome-foss/awesome-sysadmin","ts":"2026-09-17T15:16:20Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cursor/cursor","owner":"cursor","name":"cursor","desc":"","summaryCn":"从官方提供的下载入口开始，用户可以直接获取 Cursor 客户端。","reasonCn":"在使用体验上，Cursor 提供了专门的特性页面来展示其功能边界。开发者可以看到它如何利用上下文感知能力来推荐代码或解释现有逻辑。这种设计意味着它不仅仅是一个自动补全工具，而是一个理解项目结构的助手。当用户遇到陌生的代码模块时，可以借助其智能功能快速理清思路，而不必陷","stars":33248,"starGrowth":0,"createdAt":"2023-03-12T00:19:42Z","silentRounds":2,"language":"","topics":[],"aiDims":["代码助手","AI应用","大语言模型"],"zone":"工具","funScore":0,"domainTags":["AI编程","代码编辑器","IDE","开发工具","智能编辑"],"domainKey":"AI编程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1}],"aiScore":0.74,"source":"search","url":"https://github.com/cursor/cursor","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cs-lazy-tools/ChatGPT-On-CS","owner":"cs-lazy-tools","name":"ChatGPT-On-CS","desc":"拼多多、千牛、抖店 AI 客服机器人：自动回复客户咨询、商品答疑、售后申诉处理，支持微信、小红书、京东、抖音、B站、微博等多平台统一接待；可接入 DeepSeek / 通义千问 等大模型，支持自有知识库定制。","summaryCn":"想象一下，你经营着多个电商平台的网店，每天要回复成百上千条重复的顾客咨","reasonCn":"技术上，该项目采用了模块化的中转设计，通过统一的接口层适配了不同平台的私有协议，将复杂的网页端交互转化为标准化的消息流。核心大脑支持接入多种高性能模型，开发者可以通过简单的提示词（Prompt）工程配置来定义AI的人设。","stars":4414,"starGrowth":0,"createdAt":"2024-03-07T09:06:25Z","silentRounds":2,"language":"TypeScript","topics":["automation"],"aiDims":["AI Agent","AI应用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["AI客服","电商自动化","RAG","多平台接入","智能问答"],"domainKey":"RAG","pushedAt":"2026-08-27T05:10:32Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/cs-lazy-tools/ChatGPT-On-CS","ts":"2026-08-27T05:10:32Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepinsight/insightface","owner":"deepinsight","name":"insightface","desc":"","summaryCn":"InsightFace 2.0 版本带来了专为本地视频处理打造的 PrivateFrame 功能，它","reasonCn":"在底层架构和运行支持方面，该项目的 Python 软件包新增了对 raccoon_s 和 raccoon_l 模型包的支持，并且具备自动选择 CoreML、CUDA 或 CPU 硬件提供商的能力，同时还带有可重用的 CoreML 编译缓存。这些底层的智能适配大大减轻了开发者在不同硬件平台上部署和配置环境的负担。","stars":29807,"starGrowth":5,"createdAt":"2017-09-01T00:36:51Z","silentRounds":0,"language":"Python","topics":["face-recognition","face-detection","mxnet","face-alignment","age-estimation","arcface","retinaface","pytorch","oneflow","paddlepaddle"],"aiDims":["模型部署","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["人脸识别","计算机视觉","OpenCV"],"domainKey":"人脸识别","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"face-recognition","source":"github","weight":0.7},{"name":"face-detection","source":"github","weight":0.7},{"name":"mxnet","source":"github","weight":0.7},{"name":"face-alignment","source":"github","weight":0.7},{"name":"age-estimation","source":"github","weight":0.7},{"name":"arcface","source":"github","weight":0.7},{"name":"retinaface","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"oneflow","source":"github","weight":0.7},{"name":"paddlepaddle","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/deepinsight/insightface","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01550323749454826,"copyOk":false},{"repo":"Spandan-Madan/DeepLearningProject","owner":"Spandan-Madan","name":"DeepLearningProject","desc":"","summaryCn":"这个项目最独特的地方在于它极其硬核的学习逻辑。","reasonCn":"在课程的学习路径中，你将从基础的传统机器学习算法起步，理解算法背后的数学逻辑和参数调优细节，随后才会自然过渡到深度学习框架的使用。这种由浅入深的安排，确保了学习者不仅仅是在调用库函数，而是能够真正理解每一个接口背后的数据流向。","stars":4817,"starGrowth":1,"createdAt":"2017-07-12T16:03:24Z","silentRounds":4,"language":"HTML","topics":["machine-learning","deep-learning","neural-networks","tutorial"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.2,"domainTags":["机器学习","神经网络","端到端开发"],"domainKey":"机器学习","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"machine-learning","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"neural-networks","source":"github","weight":0.7},{"name":"tutorial","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/Spandan-Madan/DeepLearningProject","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"google-gemini/gemini-fullstack-langgraph-quickstart","owner":"google-gemini","name":"gemini-fullstack-langgraph-quickstart","desc":"","summaryCn":"该系统通过动态生成搜索项并反复反思研究结果来弥补知识漏洞，这意味着你可","reasonCn":"项目核心亮点在于其迭代式的工作逻辑，后端智能体会在初步搜索后主动判断信息完整度，一旦发现逻辑链条缺失，便会自动发起新一轮查询。这种自我纠错的设计极大提升了回答的可靠性，特别是在需要跨领域整合知识或进行深度课题调研的场景下，它就像一个不知疲倦的助理，始终在为你寻","stars":18339,"starGrowth":0,"createdAt":"2025-05-22T11:59:02Z","silentRounds":1,"language":"Jupyter Notebook","topics":["gemini","gemini-api"],"aiDims":["AI Agent","AI搜索","工具调用"],"zone":"AI","funScore":0,"domainTags":["LangGraph","智能体","工作流编排","搜索助手","科研工具"],"domainKey":"LangGraph","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"gemini","source":"github","weight":0.7},{"name":"gemini-api","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/google-gemini/gemini-fullstack-langgraph-quickstart","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"yemount/pose-animator","owner":"yemount","name":"pose-animator","desc":"","summaryCn":"这个项目的核心在于将计算机图形学中的骨骼动画概念，优雅地应用到了 SVG 矢","reasonCn":"其核心亮点在于完全在浏览器内运行，无需后端渲染，极大地降低了用户体验门槛。用户不仅可以看到模型如何平滑地跟随动作，还可以通过修改 SVG 路径来调整角色的外观细节，从而创造出属于自己的虚拟 Avatar。对于那些喜欢二次元绘画或是想要在直播中尝试低成本虚拟偶像的创作者来说","stars":8850,"starGrowth":0,"createdAt":"2020-03-30T09:43:38Z","silentRounds":9,"language":"JavaScript","topics":[],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["骨骼动画","SVG","计算机视觉","浏览器应用"],"domainKey":"骨骼动画","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/yemount/pose-animator","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EutropicAI/Final2x","owner":"EutropicAI","name":"Final2x","desc":"","summaryCn":"跨平台的图像高清化工具，利用AI算法提升模糊图片分辨率并适配最新显卡","reasonCn":"Final2x是一款基于AI的跨平台图像超分辨率工具，旨在解决低分辨率图片模糊不清的问题。它通过深度学习模型对图像进行细节重建，使其看起来更加清晰锐利。最新版本v4.0.0引入了cccv后端，不仅提升了性能，还允许用户加载和使用自定义模型，极大地扩展了功能可能性。项目对硬件兼容性做得很好，明确支持Nvidia 50系列等新显卡，确保用户能充分利用硬件加速获得更快的处理速度，是图像处理爱好者和设计师的得力助手。","stars":7332,"starGrowth":0,"createdAt":"2023-06-19T02:31:56Z","silentRounds":6,"language":"TypeScript","topics":["cross-platform","electron","typescript","vue3","i18n","computer-vision","super-resolution","image-processing","pytorch","low-level-vision"],"aiDims":["AI应用","模型部署"],"zone":"AI","funScore":0.7,"domainTags":["图像超分辨率","AI图像增强","老照片修复"],"domainKey":"图像超分辨率","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"cross-platform","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"vue3","source":"github","weight":0.7},{"name":"i18n","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"super-resolution","source":"github","weight":0.7},{"name":"image-processing","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"low-level-vision","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/EutropicAI/Final2x","ts":"2026-09-15T21:03:53.738Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"browser-act/skills","owner":"browser-act","name":"skills","desc":"Browser automation CLI built for AI agents. Break through anti-bot walls, hand off to humans across platforms when stuck. Parallel multi-task execution, independent multi-session operation, isolated multi-account browsing.","summaryCn":"让AI代理学会操作浏览器，攻克反爬虫并接管网页任务的神器","reasonCn":"这是一个专为AI智能体设计的浏览器自动化CLI工具，核心功能是让AI通过工具调用模拟人类操作浏览器。它不仅能突破网站的反爬虫机制，还能在遇到困难时无缝切换人工接管，支持多任务并行和独立会话。该工具底层利用浏览器原生平滑交互，解决了传统爬虫容易被识别的痛点，是目前构建复杂AI Agent应用的重要基础设施。","stars":5981,"starGrowth":2,"createdAt":"2026-02-06T10:35:35Z","silentRounds":0,"language":"Python","topics":["automation"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["浏览器自动化","爬虫","AI Agent","反爬虫","多任务"],"domainKey":"AI Agent","pushedAt":"2026-08-24T14:56:52Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/browser-act/skills","ts":"2026-08-24T14:56:52Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"karpathy/llm.c","owner":"karpathy","name":"llm.c","desc":"","summaryCn":"这个项目由知名 AI 研究者 karpathy 发起，目标是用最纯粹的 C 和 CUDA 语言","reasonCn":"从具体实现来看，项目包含两个核心版本。CUDA 版本专注于性能，实测比 PyTorch Nightly 快约 7%，适合想要极致效率的场景。CPU 版本则约 1000 行代码，以 fp32 精度运行，代码干净易读，更适合学习和调试。两个版本都以 GPT-2 和 GPT-3 迷你系列为目标，并配有 PyTorch 参考实现 t","stars":31046,"starGrowth":0,"createdAt":"2024-04-08T16:58:11Z","silentRounds":2,"language":"Cuda","topics":[],"aiDims":["模型与训练","大语言模型","推理引擎"],"zone":"AI","funScore":0.6,"domainTags":["GPT","C语言","CUDA","模型实现"],"domainKey":"GPT","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Cuda","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/karpathy/llm.c","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AiHubCN/Awesome-Chinese-LLM","owner":"AiHubCN","name":"Awesome-Chinese-LLM","desc":"","summaryCn":"这是一个集中收录上百个中文大模型、数据集与微调教程的资源导航库。","reasonCn":"随着开源大语言模型的快速发展，开发者急需一个全面且系统的中文资源导航。该项目收集并梳理了上百个中文大模型相关的开源模型、垂直领域微调案例、预训练及SFT数据集、训练部署框架与评测教程。无论是想快速寻找基座模型还是学习提示工程，都能在这里找到指引。","stars":22772,"starGrowth":0,"createdAt":"2023-05-22T12:27:03Z","silentRounds":1,"language":"","topics":["llm","nlp","chatglm","chinese","llama","awesome-lists"],"aiDims":["学习资源","教程","大语言模型"],"zone":"资源","funScore":0,"domainTags":["中文大模型","资源导航","数据集","微调教程","模型评测"],"domainKey":"中文大模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"nlp","source":"github","weight":0.7},{"name":"chatglm","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"awesome-lists","source":"github","weight":0.7}],"aiScore":0.728,"source":"search","url":"https://github.com/AiHubCN/Awesome-Chinese-LLM","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mayswind/AriaNg-Native","owner":"mayswind","name":"AriaNg-Native","desc":"A better aria2 desktop frontend than AriaNg, with all features of AriaNg and providing more features for desktop usage.","summaryCn":"如果你是个经常需要从各种渠道下载大文件的资深用户，肯定被浏览器的下载管","reasonCn":"技术上，它是基于 Electron 开发的跨平台应用，这意味着它把网页端的强大 UI 和后台 aria2 进程紧密绑定在了一起。它最厉害的地方在于“原生感”，相比于网页端，它能直接调用系统的托盘图标、进行后台静默运行，并处理复杂的网络协议逻辑。","stars":2719,"starGrowth":0,"createdAt":"2017-03-26T06:49:18Z","silentRounds":12,"language":"JavaScript","topics":["desktop"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["下载管理","Aria2","Electron","文件传输"],"domainKey":"下载管理","pushedAt":"2026-09-12T06:57:45Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/mayswind/AriaNg-Native","ts":"2026-09-12T06:57:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opengeos/geoai","owner":"opengeos","name":"geoai","desc":"GeoAI: Artificial Intelligence for Geospatial Data","summaryCn":"用 AI 处理卫星地图和地理数据，生成地理信息。","reasonCn":"在技术实现上，GeoAI 涵盖了从图像分割到目标检测的多种 AI 任务。它支持处理遥感影像（RSI）、高分辨率卫星图甚至街景数据。项目提供了丰富的预训练模型和工具链，专门针对地理数据的非欧几里得特性（如网格数据）进行了优化，解决了传统图像处理算法在地理场景下的局限性。","stars":3391,"starGrowth":2,"createdAt":"2023-08-11T20:05:04Z","silentRounds":0,"language":"Python","topics":["data-science"],"aiDims":["AI应用","数据科学","可视化"],"zone":"AI","funScore":0.6,"domainTags":["地理空间 AI","遥感影像","目标检测","GIS","深度学习"],"domainKey":"地理空间 AI","pushedAt":"2026-09-07T03:05:02Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"数据科学","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.87,"source":"search","url":"https://github.com/opengeos/geoai","ts":"2026-09-07T03:05:02Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langchain-ai/rag-from-scratch","owner":"langchain-ai","name":"rag-from-scratch","desc":"","summaryCn":"从零构建RAG系统的教程合集，配套视频课程讲解检索增强生成原理。","reasonCn":"这是一套系统讲解RAG技术的Jupyter Notebook教程，配合视频课程从零开始构建知识库。解决LLM知识固定问题，通过检索外部文档扩展模型能力。适合想学习RAG实现原理的开发者，内容扎实、结构清晰，是学习RAG的优质资源。","stars":9356,"starGrowth":0,"createdAt":"2024-01-31T01:23:48Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["RAG与知识","大语言模型"],"zone":"资源","funScore":0.26,"domainTags":["检索增强生成","大语言模型","教程"],"domainKey":"检索增强生成","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/langchain-ai/rag-from-scratch","ts":"2026-09-14T23:04:28.868Z","category":"learning","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"llm-jp/awesome-japanese-llm","owner":"llm-jp","name":"awesome-japanese-llm","desc":"日本語LLMまとめ - Overview of Japanese LLMs","summaryCn":"日系大模型资源库，一站式掌握日本AI技术进展","reasonCn":"这不是一个具体能用的AI工具，而是日本开源大模型项目的权威目录。收录了Japanese LLM和JapaneseLLM等核心项目，涵盖预训练、微调和应用。对想研究日语NLP或了解日本AI生态的人来说，这是最省时间的导航图，能帮你快速定位所需技术栈，避免在海量信息中迷路，是进入日本AI领域的必备钥匙。","stars":1432,"starGrowth":0,"createdAt":"2023-07-09T04:36:38Z","silentRounds":12,"language":"TypeScript","topics":["llm-model"],"aiDims":["大语言模型","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["日语LLM","大模型","资源导航","Japanese NLP"],"domainKey":"日语LLM","pushedAt":"2026-09-23T01:58:22Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.42,"source":"search","url":"https://github.com/llm-jp/awesome-japanese-llm","ts":"2026-09-23T01:58:22Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"humanlayer/skills","owner":"humanlayer","name":"skills","desc":"","summaryCn":"给AI装技能的万能插件库，让它干活更厉害","reasonCn":"基于TypeScript构建的AI技能框架，让大语言模型能够调用预定义工具来完成各类任务。支持自定义技能开发，可对接文件系统、数据库、API等多种外部资源，帮助开发者快速为Agent赋予专业能力，大幅扩展LLM的实际应用边界。","stars":4434,"starGrowth":6,"createdAt":"2026-03-18T03:23:23Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI技能","Agent框架","开发工具","TypeScript"],"domainKey":"AI技能","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/humanlayer/skills","ts":"2026-09-07T02:44:20.540Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.026471447902571037,"copyOk":true},{"repo":"MoonshotAI/Kimi-K3","owner":"MoonshotAI","name":"Kimi-K3","desc":"Open Frontier Intelligence","summaryCn":"月之暗面发布的Kimi K3大模型，长文本理解和智能体能力炸裂","reasonCn":"这是Moonshot AI（月之暗面）发布的Kimi K3开源模型，定位为基础模型（Foundation Model）。它继承了Kimi系列在超长上下文处理上的优势，同时在推理能力和多语言支持上进行了显著增强。对于AI研究者和应用开发者来说，这是一个可以直接微调或部署的高性能基座模型。","stars":8844,"starGrowth":1,"createdAt":"2026-07-27T08:01:37Z","silentRounds":1,"language":"","topics":[],"aiDims":["大语言模型","推理引擎"],"zone":"AI","funScore":0.6,"domainTags":["大语言模型","长文本","Transformer"],"domainKey":"大语言模型","pushedAt":"2026-08-06T10:17:28Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1}],"aiScore":0.95,"source":"bigbro","url":"https://github.com/MoonshotAI/Kimi-K3","ts":"2026-08-06T10:17:28Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AGI-is-going-to-arrive/ahadiff","owner":"AGI-is-going-to-arrive","name":"ahadiff","desc":"Learn from your vibe coding instead of just clicking Accept. AhaDiff turns each AI diff into a code-verified lesson, quiz, and review. · 别再无脑 vibe coding 啦！让你从 vibe coding 中学到真东西。每次改动都变成能查证的课程、测验和复习。","summaryCn":"把AI改代码的瞬间变成带测验的刷题课，拒绝无脑接受","reasonCn":"这是一个专为开发者设计的AI代码审查学习插件。它利用AI Agent技术分析代码变更，自动生成可验证的测验题和详细复习笔记。当开发者使用Copilot等工具生成代码时，插件会实时介入，将每一次AI的Diff（差异）转化为知识点考核，帮助用户从被动接受转变为主动学习，确保真正理解每一行代码的来龙去脉。","stars":232,"starGrowth":0,"createdAt":"2026-04-19T13:53:59Z","silentRounds":14,"language":"Python","topics":["learning"],"aiDims":["AI应用","代码助手","学习资源"],"zone":"AI","funScore":0.6,"domainTags":["代码审查","AI 学习","测验系统","代码教育"],"domainKey":"代码审查","pushedAt":"2026-09-08T23:04:09Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/AGI-is-going-to-arrive/ahadiff","ts":"2026-09-08T23:04:09Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"datawhalechina/llm-universe","owner":"datawhalechina","name":"llm-universe","desc":"本项目是一个面向小白开发者的大模型应用开发教程，在线阅读地址：https://datawhalechina.github.io/llm-universe/","summaryCn":"手把手教你把大模型变成超实用AI工具的保姆级教程","reasonCn":"这是一套专为编程新手设计的大模型应用开发课程，覆盖从零基础到实战的全过程。它使用Jupyter Notebook交互式教学，通过RAG技术实现知识库问答和智能问答机器人。内容涵盖Prompt工程、LangChain框架、向量数据库搭建及实战案例，帮助小白开发者快速上手构建大模型应用。","stars":14025,"starGrowth":0,"createdAt":"2023-10-29T16:01:22Z","silentRounds":1,"language":"Jupyter Notebook","topics":["rag"],"aiDims":["教程","RAG与知识","AI应用"],"zone":"资源","funScore":0,"domainTags":["大模型教程","RAG","LangChain","向量数据库","Prompt工程"],"domainKey":"大模型教程","pushedAt":"2026-08-27T03:08:26Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/datawhalechina/llm-universe","ts":"2026-08-27T03:08:26Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jianchang512/clone-voice","owner":"jianchang512","name":"clone-voice","desc":"","summaryCn":"一键式多语言声音克隆工具，支持16种语言，可离线运行，无需GPU。","reasonCn":"clone‑voice 是基于 Coqui AI xtts_v2 的声音克隆工具，支持文字到语音和语音到语音的转换。项目不需 GPU，提供预编译 exe，用户可直接双击运行。它支持 16 种语言，并允许用户用麦克风录制 5~20 秒的音频以获得最佳合成效果。适合内容创作者、配音工作者或爱好者快速生成多种音色的音频。","stars":8974,"starGrowth":0,"createdAt":"2023-11-19T16:29:06Z","silentRounds":5,"language":"Python","topics":["clonevoice","tts","voice-assistant","speech-analysis","sts"],"aiDims":["AI应用","模型部署"],"zone":"AI","funScore":0.6,"domainTags":["语音克隆","TTS","多语言语音合成","离线AI"],"domainKey":"语音克隆","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"clonevoice","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"voice-assistant","source":"github","weight":0.7},{"name":"speech-analysis","source":"github","weight":0.7},{"name":"sts","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/jianchang512/clone-voice","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ts1/BLEUnlock","owner":"ts1","name":"BLEUnlock","desc":"","summaryCn":"BLEUnlock 是一款运行在 Mac 菜单栏的小型实用工具，它利用蓝牙低能耗（BLE","reasonCn":"这个工具的安全机制基于静态 MAC 地址的周期性信号传输，意味着它不依赖云端同步，所有数据都在本地处理，隐私性极强。除了基础的锁屏解锁功能，它还提供了丰富的自动化钩子。例如，当检测到设备远离自动锁屏时，你可以选择暂停正在播放的音乐或视频，防止敏感内容被旁人看到；当设","stars":4442,"starGrowth":0,"createdAt":"2019-04-28T08:19:20Z","silentRounds":1,"language":"Swift","topics":["swift","macos","bluetooth-low-energy","ble","corebluetooth"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.2,"domainTags":["Mac自动化","蓝牙解锁","菜单栏工具","隐私安全"],"domainKey":"Mac自动化","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"swift","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"bluetooth-low-energy","source":"github","weight":0.7},{"name":"ble","source":"github","weight":0.7},{"name":"corebluetooth","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/ts1/BLEUnlock","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"datawhalechina/happy-llm","owner":"datawhalechina","name":"happy-llm","desc":"📚 从零开始构建大模型","summaryCn":"从零手搓大模型的中文入门教程，小白也能学会","reasonCn":"Happy-LLM是Datawhale开源的从零构建大语言模型系统教程，涵盖Transformer、预训练、微调、RAG等完整技术栈。用Jupyter Notebook形式讲解，配合代码实践，适合想深入理解LLM原理的开发者。star数32000+，是国内最热门AI学习项目之一。","stars":33993,"starGrowth":4,"createdAt":"2024-05-28T03:22:50Z","silentRounds":0,"language":"Jupyter Notebook","topics":["rag","llm"],"aiDims":["大语言模型","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["大模型教程","Transformer","RAG","深度学习"],"domainKey":"大模型教程","pushedAt":"2026-08-08T17:18:48Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/datawhalechina/happy-llm","ts":"2026-08-08T17:18:48Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Nutlope/logocreator","owner":"Nutlope","name":"logocreator","desc":"A free + OSS logo generator powered by Flux on Together AI","summaryCn":"免费输入一句描述就能生成品牌标志的开源神器","reasonCn":"它用现代前端框架搭起生成界面，把品牌名、行业、风格和配色交给后台提示词拼接，再调用云端图像大模型出图。项目封装了模板选择、随机种子、尺寸比例、风格约束和批量生成逻辑，还能保存预览结果与多种尺寸，方便快速对比多个标志方案，适合小团队低成本做视觉探索。","stars":8771,"starGrowth":1,"createdAt":"2024-11-06T03:55:48Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["AI应用","创意工具"],"zone":"AI","funScore":1,"domainTags":["图像生成","品牌设计","提示词工程","文生图"],"domainKey":"图像生成","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/Nutlope/logocreator","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wong2/chatgpt-google-extension","owner":"wong2","name":"chatgpt-google-extension","desc":"","summaryCn":"在信息检索时，我们常常需要从海量网页中手动提炼核心观点，而这款开源浏览","reasonCn":"该项目的核心亮点在于其出色的生态兼容性与沉浸式体验。它不仅原生支持官方 OpenAI API 和 ChatGPT Plus，还自带 Markdown 格式渲染、代码高亮显示以及护眼的暗黑模式。无论是开发者查阅技术文档，还是普通人检索日常问题，都能获得极佳的阅读质感，并且它还允许用户随时提供反馈来","stars":13046,"starGrowth":0,"createdAt":"2022-12-04T13:33:22Z","silentRounds":11,"language":"TypeScript","topics":["browser-extension","chatgpt","chrome-extension","firefox-addon"],"aiDims":["AI应用","AI搜索"],"zone":"AI","funScore":0.6,"domainTags":["搜索引擎","浏览器插件","文本摘要","OpenAI"],"domainKey":"搜索引擎","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"browser-extension","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"chrome-extension","source":"github","weight":0.7},{"name":"firefox-addon","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"tier","url":"https://github.com/wong2/chatgpt-google-extension","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vercel-labs/agent-skills","owner":"vercel-labs","name":"agent-skills","desc":"Vercel's official collection of agent skills","summaryCn":"给 AI 赋予动手能力，让它像专业人士一样搞定复杂任务","reasonCn":"这是 Vercel 官方提供的一套 AI Agent 技能库。它定义了一套标准化的接口，让 AI 模型（如 OpenAI 的 GPT-4）能够通过“工具调用”来执行具体的操作，如发送邮件、查询数据库、执行 shell 命令等。这套技能库简化了 Agent 开发流程，开发者只需组合不同的技能，就能快速构建出能自动处理任务的智能体。","stars":31471,"starGrowth":11,"createdAt":"2025-12-08T19:10:06Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","工具调用","技能编排","自动化"],"domainKey":"AI Agent","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/vercel-labs/agent-skills","ts":"2026-09-06T13:43:54.327Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04085830129325411,"copyOk":true},{"repo":"run-llama/liteparse","owner":"run-llama","name":"liteparse","desc":"A fast, helpful, and open-source document parser","summaryCn":"文档解析界的瑞士军刀，各种格式都能给你剥干净","reasonCn":"基于Rust开发的高性能文档解析器，支持PDF、Word、PPT、Excel等20多种格式。内置智能分页分段逻辑，能自动识别表格、列表、代码块等特殊结构。专为RAG系统设计，输出结构化文本，兼容LangChain等主流框架，解析速度比同类工具快数倍。","stars":12397,"starGrowth":38,"createdAt":"2026-02-09T22:16:30Z","silentRounds":0,"language":"Rust","topics":["document-ocr","document-processing","ocr","ocr-recognition","pdf","pdf-parser","text-extraction"],"aiDims":["文档理解","RAG与知识","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["文档解析","RAG","Rust","OCR"],"domainKey":"文档解析","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"document-ocr","source":"github","weight":0.7},{"name":"document-processing","source":"github","weight":0.7},{"name":"ocr","source":"github","weight":0.7},{"name":"ocr-recognition","source":"github","weight":0.7},{"name":"pdf","source":"github","weight":0.7},{"name":"pdf-parser","source":"github","weight":0.7},{"name":"text-extraction","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/run-llama/liteparse","ts":"2026-08-13T01:37:33.040Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.14783492780511412,"copyOk":false},{"repo":"juicedata/juicefs","owner":"juicedata","name":"juicefs","desc":"JuiceFS is a distributed POSIX file system built on top of Redis and S3.","summaryCn":"把云端文件挂载到本地，像操作硬盘一样存取 S3 数据的神器","reasonCn":"JuiceFS 是一个基于 Redis 和 S3 构建的高性能分布式 POSIX 文件系统。它将海量对象存储封装成标准的文件系统接口，让应用像读写本地文件一样操作云端数据，无需修改任何代码。支持高并发读写，具备数据强一致性和容灾能力。","stars":14454,"starGrowth":1,"createdAt":"2021-01-08T09:39:46Z","silentRounds":1,"language":"Go","topics":["filesystem","cloud-native","golang","redis","distributed-systems","storage","object-storage","posix","hdfs","s3","bigdata","go"],"aiDims":["非AI-实用","开发工具"],"zone":"工具","funScore":0,"domainTags":["分布式文件系统","对象存储","POSIX","Redis","S3"],"domainKey":"分布式文件系统","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"filesystem","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"redis","source":"github","weight":0.7},{"name":"distributed-systems","source":"github","weight":0.7},{"name":"storage","source":"github","weight":0.7},{"name":"object-storage","source":"github","weight":0.7},{"name":"posix","source":"github","weight":0.7},{"name":"hdfs","source":"github","weight":0.7},{"name":"s3","source":"github","weight":0.7},{"name":"bigdata","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/juicedata/juicefs","ts":"2026-08-15T00:57:24.351Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"akitaonrails/ai-usagebar","owner":"akitaonrails","name":"ai-usagebar","desc":"Rust-based waybar widget to monitor status of Claude, GPT, GLM, OpenRouter plans/credits - inspired by claudebar/codexbar","summaryCn":"给Waybar加个实时监控AI额度的小插件","reasonCn":"这个项目的核心亮点在于它的轻量和高效。它完全用Rust编写，这意味着它不仅运行速度极快，而且对系统资源的占用非常低。它不需要后台常驻进程，而是通过监听Waybar的信号来更新数据，非常省电。更重要的是，它支持多种主流的AI服务，包括Claude、OpenAI、GLM等，甚至还能显示OpenR","stars":541,"starGrowth":1,"createdAt":"2026-05-24T00:29:51Z","silentRounds":1,"language":"Rust","topics":[],"aiDims":["工具调用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Waybar","AI额度监控","Linux桌面"],"domainKey":"Waybar","pushedAt":"2026-09-11T17:24:39.834Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/akitaonrails/ai-usagebar","ts":"2026-09-11T17:24:39.834Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"oramasearch/orama","owner":"oramasearch","name":"orama","desc":"🌌  A complete search engine and RAG pipeline in your browser, server or edge network with support for full-text, vector, and hybrid search in less than 2kb.","summaryCn":"比弹珠还小的搜索引擎，浏览器里就能跑RAG","reasonCn":"超轻量级搜索引擎，体积不到2KB却支持全文搜索、向量搜索和混合搜索，可在浏览器、服务器或边缘网络运行。将向量搜索和全文搜索完美融合于一个库中，客户端直接运行搜索无需后端，大幅降低延迟且保护用户隐私。特别适合需要在前端实现智能搜索和RAG功能的场景，让开发者无需引入重型数据库就能实现AI搜索能力。","stars":10559,"starGrowth":0,"createdAt":"2022-05-10T12:01:50Z","silentRounds":1,"language":"TypeScript","topics":["vector-db"],"aiDims":["RAG与知识","向量数据库","AI搜索"],"zone":"工具","funScore":0.8,"domainTags":["搜索引擎","向量检索","RAG","TypeScript","混合搜索"],"domainKey":"搜索引擎","pushedAt":"2026-09-11T05:09:46Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/oramasearch/orama","ts":"2026-09-11T05:09:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"magenta/magenta","owner":"magenta","name":"magenta","desc":"","summaryCn":"这个项目原本致力于探索机器学习在艺术和音乐创作中的角色，其核心目标并非","reasonCn":"Magenta 的一个独特定位是强调AI工具与人类创作的共生关系。项目自述中明确提到，其开发的智能工具和界面旨在让艺术家和音乐家延伸（而不是替换）他们的创作流程。这意味着AI在这里扮演的是辅助者角色，帮助创作者突破传统技法限制或激发灵感，例如生成意想不到的旋律片段或视觉元","stars":19796,"starGrowth":0,"createdAt":"2016-05-05T20:10:40Z","silentRounds":9,"language":"Python","topics":[],"aiDims":["模型与训练","AI应用","非AI-好玩"],"zone":"创意","funScore":0.7,"domainTags":["音乐生成","艺术创作","机器学习","AI艺术"],"domainKey":"音乐生成","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/magenta/magenta","ts":"2026-09-13T22:08:09.499Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mvanhorn/last30days-skill","owner":"mvanhorn","name":"last30days-skill","desc":"AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary","summaryCn":"一键把全网热门信息揉碎了喂给大模型，省去自己刷几十个网页的功夫","reasonCn":"这是一个AI智能体技能，专门帮用户从Reddit、X、YouTube、HN等平台抓取最新信息，经过分析后生成一份有依据的总结报告。它能自动访问不同网站并提取关键内容，把零散的帖子、视频和新闻变成结构化的综述，解决信息过载问题。","stars":62668,"starGrowth":3,"createdAt":"2026-01-23T20:37:37Z","silentRounds":0,"language":"Python","topics":["ai-prompts","claude","claude-code","reddit","twitter","hackernews","polymarket","research","youtube","ai-skill","bluesky","clawhub","deep-research","instagram","openclaw","recency","social-media","tiktok","trends","web-search"],"aiDims":["AI Agent","AI搜索","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","信息聚合","技能包","数据抓取","研究工具"],"domainKey":"AI Agent","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-prompts","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"reddit","source":"github","weight":0.7},{"name":"twitter","source":"github","weight":0.7},{"name":"hackernews","source":"github","weight":0.7},{"name":"polymarket","source":"github","weight":0.7},{"name":"research","source":"github","weight":0.7},{"name":"youtube","source":"github","weight":0.7},{"name":"ai-skill","source":"github","weight":0.7},{"name":"bluesky","source":"github","weight":0.7},{"name":"clawhub","source":"github","weight":0.7},{"name":"deep-research","source":"github","weight":0.7},{"name":"instagram","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"recency","source":"github","weight":0.7},{"name":"social-media","source":"github","weight":0.7},{"name":"tiktok","source":"github","weight":0.7},{"name":"trends","source":"github","weight":0.7},{"name":"web-search","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/mvanhorn/last30days-skill","ts":"2026-08-19T01:08:47.324Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"genieincodebottle/generative-ai","owner":"genieincodebottle","name":"generative-ai","desc":"Comprehensive resources on Generative AI, including a detailed roadmap, projects, use cases, interview preparation, and coding preparation.","summaryCn":"从入门到面试的生成式AI全栈学习路线图，干货满满","reasonCn":"内容覆盖了从基础概念到高级应用的完整链条，包括经典机器学习、深度学习、大语言模型原理、提示工程、RAG检索增强、多模态AI等核心主题。每个模块都配有Jupyter Notebook实战代码，你可以边看边跑。还整理了行业应用案例和面试高频题目，相当于一本可执行的AI教科书。","stars":2640,"starGrowth":0,"createdAt":"2024-01-09T11:10:20Z","silentRounds":4,"language":"Jupyter Notebook","topics":["llm-model"],"aiDims":["学习资源","教程","大语言模型"],"zone":"资源","funScore":0,"domainTags":["生成式AI","学习路线图","大语言模型"],"domainKey":"生成式AI","pushedAt":"2026-09-13T17:46:14Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/genieincodebottle/generative-ai","ts":"2026-09-13T17:46:14Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bmild/nerf","owner":"bmild","name":"nerf","desc":"","summaryCn":"用神经网络编码3D场景，从几张图片就能渲染出任意角度的逼真新画面","reasonCn":"NeRF是2020年ECCV最佳论文荣誉提名作品，核心创新是用神经网络学习连续辐射场，将3D场景编码进一个小型网络，通过体渲染从任意视角生成逼真图像。作者来自UC Berkeley、Google Research和UC San Diego，提供了TensorFlow实现和示例数据，初学者可直接在本地观察Fern场景的训练过程。","stars":10938,"starGrowth":0,"createdAt":"2020-03-19T01:24:18Z","silentRounds":1,"language":"Jupyter Notebook","topics":["nerf","neural-radiance-fields"],"aiDims":["模型与训练","多模态Agent"],"zone":"AI","funScore":1,"domainTags":["神经辐射场","3D重建","计算机视觉","体渲染"],"domainKey":"神经辐射场","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"nerf","source":"github","weight":0.7},{"name":"neural-radiance-fields","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/bmild/nerf","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"iamtouchskyer/memex","owner":"iamtouchskyer","name":"memex","desc":"Zettelkasten-based persistent memory for AI coding agents. Works with Claude Code, Cursor, VS Code Copilot, Codex, Windsurf & any MCP client. No vector DB — just markdown + git sync.","summaryCn":"给AI写代码助手装个永久记忆，不用记事本记笔记了","reasonCn":"这是一个专为AI编程助手设计的本地化知识库系统，通过Zettelkasten卡片笔记法将项目上下文、代码片段和配置信息持久化存储。它不依赖外部向量数据库，仅用Git进行版本同步，完美解决了Claude Code、Cursor等工具在长时间工作中遗忘上下文的痛点。","stars":142,"starGrowth":0,"createdAt":"2026-03-20T03:06:21Z","silentRounds":16,"language":"TypeScript","topics":["knowledge-management"],"aiDims":["AI Agent","知识图谱","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["Zettelkasten","AI Coding Assistant","MCP Client"],"domainKey":"Zettelkasten","pushedAt":"2026-09-08T16:20:21Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/iamtouchskyer/memex","ts":"2026-09-08T16:20:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"karpathy/nanoGPT","owner":"karpathy","name":"nanoGPT","desc":"","summaryCn":"为读懂和跑通中型语言模型训练流程准备的极简仓库。","reasonCn":"它把训练和微调中型语言模型这件事压得很轻，核心训练脚本与模型定义都只有约三百行，适合想亲手跑通模型结构的人。项目能复现较小规模的语言模型，也可选择加载已有权重继续训练，帮助读者理解从数据到训练循环的完整过程；但它已经较旧，更适合作为代码骨架参考，而不是直接用于生产。","stars":63322,"starGrowth":5,"createdAt":"2022-12-28T00:51:12Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","大语言模型","微调"],"zone":"AI","funScore":0.2,"domainTags":["GPT-2","OpenWebText","训练脚本","模型定义","权重加载"],"domainKey":"GPT-2","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.704,"source":"search","url":"https://github.com/karpathy/nanoGPT","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015236884495120178,"copyOk":true},{"repo":"meta-llama/llama","owner":"meta-llama","name":"llama","desc":"","summaryCn":"Meta 公司在发布 Llama 3.1 系列模型时，对原有的 GitHub 仓库结构进行了重","reasonCn":"核心变化之一是分拆出了 PurpleLlama 这一专门组件。它并非模型本身，而是 Llama Stack 中负责安全的关键部分，专注于推理阶段的安全风险检测和缓解措施。对于企业级应用而言，模型能力再强，若缺乏对生成内容的安全把控也是隐患。","stars":59620,"starGrowth":0,"createdAt":"2023-02-14T09:29:12Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","模型部署"],"zone":"AI","funScore":0,"domainTags":["大语言模型","LLM","模型权重","Meta","开源AI"],"domainKey":"大语言模型","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/meta-llama/llama","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"go-ego/gse","owner":"go-ego","name":"gse","desc":"Go efficient multilingual NLP and text segmentation; support English, Chinese, Japanese and others.","summaryCn":"Go语言写的中文分词利器，多语言NLP处理一步到位","reasonCn":"基于Go语言开发的多语言自然语言处理工具，支持中文、英文、日文等多种语言的分词和文本处理。核心采用HMM隐马尔可夫模型和Viterbi算法，分词准确率高。内置词典和自定义词典扩展功能，适合做中文文本分析、关键词提取、搜索引擎等NLP场景，填补了Go生态高质量中文分词的空白。","stars":2845,"starGrowth":0,"createdAt":"2017-06-23T15:42:35Z","silentRounds":13,"language":"Go","topics":["english"],"aiDims":["AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["分词","NLP","Go语言","中文处理"],"domainKey":"分词","pushedAt":"2026-09-12T13:41:56Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"english","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/go-ego/gse","ts":"2026-09-12T13:41:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nanobrowser/nanobrowser","owner":"nanobrowser","name":"nanobrowser","desc":"Open-Source Chrome extension for AI-powered web automation. Run multi-agent workflows using your own LLM API key. Alternative to OpenAI Operator.","summaryCn":"浏览器里的AI特工，一键自动搞定网页操作","reasonCn":"这是一个基于TypeScript开发的浏览器插件，允许用户通过LLM API密钥运行多智能体工作流。它作为OpenAI Operator的开源替代方案，支持复杂的自动化任务。通过智能体协作，用户可以定义任务链，让AI自主探索网页、提取数据和执行操作，无需手动编写代码。","stars":13828,"starGrowth":5,"createdAt":"2024-12-31T08:11:27Z","silentRounds":0,"language":"TypeScript","topics":["automation"],"aiDims":["AI Agent","工具调用","自动化"],"zone":"AI","funScore":0,"domainTags":["AI Agent","浏览器自动化","RPA","插件","多智能体"],"domainKey":"AI Agent","pushedAt":"2026-08-18T00:42:48Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"自动化","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/nanobrowser/nanobrowser","ts":"2026-08-18T00:42:48Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.016084755568411917,"copyOk":false},{"repo":"agentscope-ai/agentscope-studio","owner":"agentscope-ai","name":"agentscope-studio","desc":"A development-oriented visualization toolkit","summaryCn":"给AI Agent装上可视化仪表盘，调试开发更直观。","reasonCn":"它不仅仅是一个简单的调试器，更是一个完整的 Agent 开发工作台。最核心的功能是“可视化编排”，你可以通过图形化的界面，像搭积木一样拖拽不同的组件（如 LLM 调用、代码执行、条件分支）来构建 Agent 的逻辑流，而不是在代码里一层层嵌套 if-else。","stars":652,"starGrowth":1,"createdAt":"2025-07-01T08:33:44Z","silentRounds":10,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","工具调用","可视化"],"zone":"AI","funScore":0.7,"domainTags":["Agent 开发","可视化","调试工具","AI 应用"],"domainKey":"Agent 开发","pushedAt":"2026-09-02T06:48:23Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/agentscope-ai/agentscope-studio","ts":"2026-09-02T06:48:23Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"josStorer/RWKV-Runner","owner":"josStorer","name":"RWKV-Runner","desc":"A RWKV management and startup tool, full automation, only 8MB. And provides an interface compatible with the OpenAI API. RWKV is a large language model that is fully open source and available for commercial use.","summaryCn":"仅8MB大小却能全自动管理和运行国产开源RWKV大模型的轻量神器","reasonCn":"技术上，RWKV本身就是一种颠覆传统Transformer架构的创新大模型，它既有RNN的超低显存消耗和线性推理速度，又具备Transformer的强大并行训练能力。而这个Runner工具则是连接普通用户和RWKV模型的最佳桥梁。它内置了高效的推理后端，优化了CPU和GPU的协同运算。","stars":6484,"starGrowth":1,"createdAt":"2023-05-05T04:21:56Z","silentRounds":3,"language":"TypeScript","topics":["api"],"aiDims":["大语言模型","模型部署","推理引擎"],"zone":"AI","funScore":0.8,"domainTags":["RWKV","本地推理","大模型","API"],"domainKey":"RWKV","pushedAt":"2026-09-04T13:37:10Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/josStorer/RWKV-Runner","ts":"2026-09-04T13:37:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jacobgil/pytorch-grad-cam","owner":"jacobgil","name":"pytorch-grad-cam","desc":"Advanced AI Explainability for computer vision.  Support for CNNs, Vision Transformers, Classification, Object detection, Segmentation, Image similarity and more.","summaryCn":"PyTorch Grad-CAM利用深度学习技术，通过生成与输入图像相匹配的图像来解释","reasonCn":"PyTorch Grad-CAM利用深度学习技术，通过生成与输入图像相匹配的图像来解释模型的输出。它支持多模态输入（如文本、图片等），并能根据输入内容自动调整生成图像的风格。这使得研究者能够更直观地理解模型的决策过程，从而提高研究的效率和质量。","stars":12980,"starGrowth":0,"createdAt":"2017-05-31T19:55:15Z","silentRounds":5,"language":"Python","topics":["ml"],"aiDims":["AI应用","机器学习"],"zone":"AI","funScore":0,"domainTags":["可解释性","计算机视觉","PyTorch"],"domainKey":"可解释性","pushedAt":"2026-08-13T07:15:26Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"机器学习","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":1,"source":"search","url":"https://github.com/jacobgil/pytorch-grad-cam","ts":"2026-08-13T07:15:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chenyuntc/pytorch-book","owner":"chenyuntc","name":"pytorch-book","desc":"","summaryCn":"深入讲解PyTorch从底层张量到实战应用的操作指南，附可执行代码。","reasonCn":"这是配套《深度学习框架PyTorch》书籍的代码仓库，采用Jupyter Notebook形式，涵盖从Tensor底层原理到实战应用的三大模块。它帮助读者理解自动微分机制，并通过不到五十行代码实现ResNet网络，是系统掌握PyTorch框架的高效学习资源。","stars":12850,"starGrowth":1,"createdAt":"2017-05-24T07:49:38Z","silentRounds":1,"language":"Jupyter Notebook","topics":["pytorch","pytorch-tutorials","pytorch-tutorials-cn","deep-learning","neural-style","charrnn","gan","caption","neuraltalk","image-classification","visdom","tensorboard","nn","tensor","autograd","jupyter-notebook"],"aiDims":["学习资源","模型与训练","教程"],"zone":"资源","funScore":0,"domainTags":["PyTorch","深度学习","教程","Jupyter"],"domainKey":"PyTorch","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"pytorch","source":"github","weight":0.7},{"name":"pytorch-tutorials","source":"github","weight":0.7},{"name":"pytorch-tutorials-cn","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"neural-style","source":"github","weight":0.7},{"name":"charrnn","source":"github","weight":0.7},{"name":"gan","source":"github","weight":0.7},{"name":"caption","source":"github","weight":0.7},{"name":"neuraltalk","source":"github","weight":0.7},{"name":"image-classification","source":"github","weight":0.7},{"name":"visdom","source":"github","weight":0.7},{"name":"tensorboard","source":"github","weight":0.7},{"name":"nn","source":"github","weight":0.7},{"name":"tensor","source":"github","weight":0.7},{"name":"autograd","source":"github","weight":0.7},{"name":"jupyter-notebook","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/chenyuntc/pytorch-book","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jackwener/OpenCLI","owner":"jackwener","name":"OpenCLI","desc":"Make Any Website into CLI & Use your logged-in browser by AI agent. ","summaryCn":"这个工具打破了网页操作的交互壁垒，让你能用命令行掌控任何网站。","reasonCn":"技术实现上，项目通过解析网站DOM结构自动生成命令接口。核心创新在于AI Agent模块，它结合大语言模型理解用户意图，转化为浏览器操作序列。Playwright引擎确保状态保持，登录后操作不会踢出。 安装：`npm install -g open-cli`\n使用：`open-cli https://app.example.com --ai`  \n","stars":29573,"starGrowth":1,"createdAt":"2026-03-14T22:10:23Z","silentRounds":1,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","浏览器自动化","CLI工具"],"domainKey":"AI Agent","pushedAt":"2026-08-30T17:36:26Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/jackwener/OpenCLI","ts":"2026-08-30T17:36:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"modelcontextprotocol/servers","owner":"modelcontextprotocol","name":"servers","desc":"Model Context Protocol Servers","summaryCn":"让AI助手一键对接各种外部工具和数据源的万能桥梁","reasonCn":"这是Anthropic推出的Model Context Protocol官方服务器集合，为AI应用提供统一的外部数据接入能力。支持文件系统、数据库、GitHub、Slack、Google Workspace等数十种常见工具连接。采用客户端-服务器架构，AI助手通过配置即可调用各种MCP服务器，无需为每个工具单独开发适配层。安装简单，用NPM即可快速部署，适合各种需要AI接入外部数据的场景。","stars":90553,"starGrowth":5,"createdAt":"2024-11-19T01:10:17Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI基础设施","工具调用","AI Agent"],"zone":"AI","funScore":0,"domainTags":["AI Agent","MCP","协议","工具集成","生态"],"domainKey":"AI Agent","pushedAt":"2026-08-18T16:04:56Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"bigbro","url":"https://github.com/modelcontextprotocol/servers","ts":"2026-08-18T16:04:56Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015165648846531864,"copyOk":false},{"repo":"pyg-team/pytorch_geometric","owner":"pyg-team","name":"pytorch_geometric","desc":"Graph Neural Network Library for PyTorch","summaryCn":"把复杂图神经网络变简单的神器，装在PyTorch里随调随用。","reasonCn":"从技术架构上看，PyG 是 PyTorch 的扩展库，利用了 PyTorch 的动态计算图优势。它特别擅长处理超大规模图数据，引入了采样算法，能在不把整张图加载到内存的情况下训练模型。比如在分子动力学模拟中，它能高效处理原子间的连接关系；在推荐系统中，它能捕捉用户与物品之间的复杂交","stars":24098,"starGrowth":0,"createdAt":"2017-10-06T16:03:03Z","silentRounds":3,"language":"Python","topics":["deep-learning"],"aiDims":["深度学习","知识图谱","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["图神经网络","PyTorch","图学习","深度学习"],"domainKey":"图神经网络","pushedAt":"2026-09-01T23:22:42Z","tags":[{"name":"深度学习","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/pyg-team/pytorch_geometric","ts":"2026-09-01T23:22:42Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mage-ai/mage-ai","owner":"mage-ai","name":"mage-ai","desc":"🧙 Build, run, and manage data pipelines for integrating and transforming data.","summaryCn":"把枯燥的数据处理变成搭积木的现代化开源工具箱","reasonCn":"技术上它采用了非常清爽的前后端分离架构，核心后端由Python驱动。它最大的亮点是把每个处理步骤都变成了一个个可复用的积木块，支持Python、SQL和R语言混搭。和老牌的Airflow相比，它不需要你写一堆复杂的配置文件，所有的操作都在一个极其丝滑的网页端完成，还自带实时日志和交互","stars":8826,"starGrowth":0,"createdAt":"2022-05-16T22:11:39Z","silentRounds":11,"language":"Python","topics":["data-science"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["数据工程","数据管道","Airflow","可视化","Python"],"domainKey":"数据工程","pushedAt":"2026-09-11T19:20:35Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/mage-ai/mage-ai","ts":"2026-09-11T19:20:35Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"iOfficeAI/AionUi","owner":"iOfficeAI","name":"AionUi","desc":"Open-source 24/7 Cowork app for OpenClaw, Hermes, Claude Code, Codex, OpenCode and 20+ more CLI Agent | Customize your assistants | Team them up｜Star if you like it!","summaryCn":"24小时在线的AI助手协作台，多Agent组队干活","reasonCn":"这是一个开源的24/7全天候协作应用，支持接入OpenClaw、Hermes、Claude Code、Codex等20多种CLI Agent。你可以像组团队一样，让不同的AI助手协同工作，并通过可视化的界面管理它们，是AI Agent聚合与协同的创意实践。","stars":33043,"starGrowth":4,"createdAt":"2025-08-07T10:29:51Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","ai","llm","chatbot"],"aiDims":["AI Agent","AI应用","可视化"],"zone":"AI","funScore":1,"domainTags":["AI Agent","多Agent协作","CLI Agent","任务编排","自托管AI"],"domainKey":"AI Agent","pushedAt":"2026-09-09T07:57:01Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/iOfficeAI/AionUi","ts":"2026-09-09T07:57:01Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mayocream/koharu","owner":"mayocream","name":"koharu","desc":"ML-powered manga translator, written in Rust.","summaryCn":"一键把日漫翻译成中文，像魔法一样自动识别角色说话。","reasonCn":"这是一个基于机器学习的日漫翻译工具，用 Rust 编写。它能精准识别漫画中的日语文本和说话人角色，自动翻译成中文并排版。相比传统 OCR，它能理解上下文语境，把生硬的机器翻译变成通顺的中文对话，完美适配漫画分镜。","stars":5624,"starGrowth":5,"createdAt":"2025-04-09T10:52:34Z","silentRounds":0,"language":"Rust","topics":["deep-learning","manga","rust","japanese","computer-vision","gpu","gui","translation","text-rendering","webgpu"],"aiDims":["多模态Agent","文档理解","非AI-实用"],"zone":"AI","funScore":0.7,"domainTags":["漫画翻译","OCR","日译中","机器学习","Rust"],"domainKey":"漫画翻译","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"manga","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"japanese","source":"github","weight":0.7},{"name":"computer-vision","source":"github","weight":0.7},{"name":"gpu","source":"github","weight":0.7},{"name":"gui","source":"github","weight":0.7},{"name":"translation","source":"github","weight":0.7},{"name":"text-rendering","source":"github","weight":0.7},{"name":"webgpu","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/mayocream/koharu","ts":"2026-08-17T00:55:40.962Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.017667140825035562,"copyOk":false},{"repo":"DLR-RM/BlenderProc","owner":"DLR-RM","name":"BlenderProc","desc":"","summaryCn":"BlenderProc解决的是计算机视觉领域中高质量光影一致仿真实时采集难的问题。","reasonCn":"该流水线的核心亮点在于物理真实性与自动化程度的平衡。它支持对物体姿态进行采样，并内置了物理引擎和碰撞检查，确保生成的场景中物体不会穿透或悬浮，从而保证了物理规律的合理性。同时，它支持物理基于渲染（PBR）材质，能模拟真实世界的光反射和折射特性，这对于训练对纹理和光","stars":3725,"starGrowth":1,"createdAt":"2019-10-10T11:29:14Z","silentRounds":1,"language":"Python","topics":["blender-pipeline","segmentation","depth-images","camera-positions","suncg-scene","camera-sampling","blender-installation","synthetic","blender","rendering","pose-estimation","synthetic-data","python","3d-graphics","computer-graphics","3d-reconstruction","3d-engines","3d-front-dataset"],"aiDims":["AI基础设施","模型与训练","可视化"],"zone":"工具","funScore":0.2,"domainTags":["3D渲染","计算机视觉","程序化建模","仿真数据"],"domainKey":"3D渲染","pushedAt":"2026-09-18T22:32:16.581Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"blender-pipeline","source":"github","weight":0.7},{"name":"segmentation","source":"github","weight":0.7},{"name":"depth-images","source":"github","weight":0.7},{"name":"camera-positions","source":"github","weight":0.7},{"name":"suncg-scene","source":"github","weight":0.7},{"name":"camera-sampling","source":"github","weight":0.7},{"name":"blender-installation","source":"github","weight":0.7},{"name":"synthetic","source":"github","weight":0.7},{"name":"blender","source":"github","weight":0.7},{"name":"rendering","source":"github","weight":0.7},{"name":"pose-estimation","source":"github","weight":0.7},{"name":"synthetic-data","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"3d-graphics","source":"github","weight":0.7},{"name":"computer-graphics","source":"github","weight":0.7},{"name":"3d-reconstruction","source":"github","weight":0.7},{"name":"3d-engines","source":"github","weight":0.7},{"name":"3d-front-dataset","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/DLR-RM/BlenderProc","ts":"2026-09-18T22:32:16.581Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"palxiao/poster-design","owner":"palxiao","name":"poster-design","desc":"迅排设计 - 图片编辑器、AI 在线海报设计，仿稿定设计，适用于多种场景：海报生成、电商产品图、文章长图、视频/公众号封面等。A beautiful online image designer, suitable for various scenarios like generate posters, making design easier!","summaryCn":"在线AI海报神器，点几下生成电商、公众号炫酷图，省时省力又好玩。","reasonCn":"迅排设计基于 Vue 前端和后端 AI 模型，提供海报、电商主图、文章长图等多场景模板，支持文字自动生成、图片智能裁剪与配色，一键即可输出高质量视觉素材，并支持批量生成、自定义品牌配色，可通过API与内部系统对接，实现自动化设计流。","stars":4872,"starGrowth":1,"createdAt":"2022-02-01T04:46:12Z","silentRounds":1,"language":"Vue","topics":["design"],"aiDims":["AI应用","创意工具","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["海报设计","AI生成","电商图","图像编辑","自动化设计"],"domainKey":"海报设计","pushedAt":"2026-09-10T16:28:13Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/palxiao/poster-design","ts":"2026-09-10T16:28:13Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TechyCSR/OpenCluely","owner":"TechyCSR","name":"OpenCluely","desc":"OpenCluely is a free, open source Cluely (alternative), built for technical interviews like DSA, OAs, and CP. It offers an invisible overlay, real-time AI help, Smart Image Processing for question capture, and multi-language support : 100% customizable and private.","summaryCn":"面试刷题神器，实时AI提示，画面不见却帮你答题。","reasonCn":"OpenCluely 通过隐形网页覆盖捕获题目截图，实时调用大模型解析并给出分步解答和代码框架；支持智能图片文字识别，兼容多平台，支持 JavaScript、Python 等语言生成，可离线使用本地模型，帮助求职者在练习或现场面试中快速获得 AI 辅助。","stars":995,"starGrowth":2,"createdAt":"2025-08-05T08:20:08Z","silentRounds":0,"language":"JavaScript","topics":["ai-assistant","cluely","dsa","interview-helper","open-source","trending","open-source-cluely","cluely-alternative","gemini","interview-preparation","whisper"],"aiDims":["AI应用","代码助手","AI写作"],"zone":"AI","funScore":0.6,"domainTags":["AI面试","浏览器插件","OCR","解题"],"domainKey":"AI面试","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"ai-assistant","source":"github","weight":0.7},{"name":"cluely","source":"github","weight":0.7},{"name":"dsa","source":"github","weight":0.7},{"name":"interview-helper","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"trending","source":"github","weight":0.7},{"name":"open-source-cluely","source":"github","weight":0.7},{"name":"cluely-alternative","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"interview-preparation","source":"github","weight":0.7},{"name":"whisper","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/TechyCSR/OpenCluely","ts":"2026-09-05T02:38:58.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Open330/kiwimu","owner":"Open330","name":"kiwimu","desc":"Turn textbooks, PDFs, and web content into your own interlinked learning wiki powered by LLM","summaryCn":"一款能将教科书与文档自动转化为带知识库和问答功能的维基系统。","reasonCn":"面对厚重的教科书或 PDF 文献，学习效率往往很低。Kiwi Mu 利用大模型技术，将复杂的文档内容自动处理成结构化的维基百科页面，并内置了 RAG（检索增强生成）问答功能。它不仅能整理文字，还能自动从页面中提取图片并生成描述，甚至能追踪每一句 AI 回答的原始出处，是深度学习者进行知识梳理和自动化知识管理的强大辅助工具。","stars":30,"starGrowth":0,"createdAt":"2026-03-17T10:03:19Z","silentRounds":11,"language":"TypeScript","topics":["textbook"],"aiDims":["AI搜索","文档理解","RAG与知识"],"zone":"AI","funScore":0.6,"domainTags":["RAG","知识库","文档解析","学习工具"],"domainKey":"RAG","pushedAt":"2026-09-14T00:18:41Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"textbook","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/Open330/kiwimu","ts":"2026-09-14T00:18:41Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"garrytan/gbrain","owner":"garrytan","name":"gbrain","desc":"Garry's Opinionated OpenClaw/Hermes Agent Brain","summaryCn":"给AI大脑装上外骨骼，让它干活比人类还溜。","reasonCn":"这是一个高度模块化的智能体框架，专为Claude/OpenAI等大模型构建。它通过定义工具调用接口，让AI不仅能聊天，还能读写文件、控制服务器、调用API，像真正的助手一样执行复杂任务。代码基于TypeScript，适合构建自主运行的AI应用。","stars":30239,"starGrowth":5,"createdAt":"2026-04-05T14:40:56Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","推理引擎"],"zone":"AI","funScore":0,"domainTags":["Agent框架","工具调用","自动化","TypeScript"],"domainKey":"Agent框架","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/garrytan/gbrain","ts":"2026-08-12T01:22:42.342Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01859525777968848,"copyOk":false},{"repo":"thoughtbot/shoulda-matchers","owner":"thoughtbot","name":"shoulda-matchers","desc":"Simple one-liner tests for common Rails functionality","summaryCn":"一款简洁的测试工具，用于Rails应用中常见的功能测试。","reasonCn":"它通过提供一种简单的方式来编写和运行针对Rails应用的单元测试，帮助开发者快速验证代码的正确性。","stars":3581,"starGrowth":1,"createdAt":"2010-12-15T22:41:52Z","silentRounds":19,"language":"Ruby","topics":["testing"],"aiDims":["测试","自动化测试"],"zone":"工具","funScore":0,"domainTags":["Ruby","Rails","测试框架","RSpec","单元测试"],"domainKey":"Ruby","pushedAt":"2026-08-10T13:14:13Z","tags":[{"name":"测试","source":"llm","weight":1},{"name":"自动化测试","source":"llm","weight":1},{"name":"testing","source":"github","weight":0.7},{"name":"Ruby","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/thoughtbot/shoulda-matchers","ts":"2026-08-10T13:14:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nowork-studio/notfair-plugin","owner":"nowork-studio","name":"notfair-plugin","desc":"Open-source SEO, GEO, and marketing skills for AI agents.","summaryCn":"给你的AI助手装上营销SEO技能，一键提升内容排名","reasonCn":"为AI Agent提供SEO、GEO和营销专业技能的开源插件，让大模型直接具备搜索引擎优化和内容营销能力。支持自动化内容生成、关键词分析、地理位置优化等实用功能，可无缝集成到主流AI Agent框架中。TypeScript开发，3600+star验证社区认可。","stars":3846,"starGrowth":2,"createdAt":"2026-03-27T06:39:58Z","silentRounds":0,"language":"TypeScript","topics":["claude-code-plugin","claude-skills","geo","googleads","mcp","metaads","seo","googleadsmcp","agents","gemini-cli-extension"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["SEO","AI Agent","技能包","营销","Agent插件"],"domainKey":"AI Agent","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"claude-code-plugin","source":"github","weight":0.7},{"name":"claude-skills","source":"github","weight":0.7},{"name":"geo","source":"github","weight":0.7},{"name":"googleads","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"metaads","source":"github","weight":0.7},{"name":"seo","source":"github","weight":0.7},{"name":"googleadsmcp","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"gemini-cli-extension","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/nowork-studio/notfair-plugin","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lllyasviel/IC-Light","owner":"lllyasviel","name":"IC-Light","desc":"","summaryCn":"IC-Light是一个专注于图像照明操纵的AI项目，它的核心能力是“非真实地”改变","reasonCn":"项目目前发布了两种类型的模型：一种是文本条件重光照模型，允许用户通过自然语言描述光线变化（如“晴天正午”、“夜晚路灯”）；另一种是背景条件模型，即将前景图像放入特定背景中，AI会根据背景的光源自动调整前景的阴影和高光。","stars":8530,"starGrowth":1,"createdAt":"2024-05-07T13:26:05Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["多模态Agent","AI应用","图像生成"],"zone":"AI","funScore":0.8,"domainTags":["图像光照","AI绘画","光影合成","图像编辑"],"domainKey":"图像光照","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"图像生成","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/lllyasviel/IC-Light","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"The-Swarm-Corporation/AutoHedge","owner":"The-Swarm-Corporation","name":"AutoHedge","desc":"Build your autonomous hedge fund in minutes. AutoHedge harnesses the power of swarm intelligence and AI agents to automate market analysis, risk management, and trade execution.","summaryCn":"用 AI 群体智能帮你秒建自动化对冲基金，省时省力省风险。","reasonCn":"AutoHedge 利用 Python 编写的多智能体框架，结合大语言模型与强化学习，实现市场数据抓取、风险评估、策略生成、交易执行全流程自动化。系统内置群体协作机制，AI 代理相互竞争共享信息，以 swarm intelligence 提升策略鲁棒性，支持自定义因子、回测与实时监控，帮助用户在数分钟内搭建可交易的对冲基金。","stars":6177,"starGrowth":3,"createdAt":"2024-12-10T20:34:46Z","silentRounds":0,"language":"Python","topics":["ai","blackrock","finance","goldmansachs","hedgefund","jpmorgan","ml","quant","quantitative-finance","quantitative-trading","trading"],"aiDims":["AI Agent","AI应用","非AI-实用"],"zone":"AI","funScore":0.6,"domainTags":["量化交易","金融科技","多智能体","对冲基金"],"domainKey":"量化交易","pushedAt":"2026-09-06T14:58:06.418Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"blackrock","source":"github","weight":0.7},{"name":"finance","source":"github","weight":0.7},{"name":"goldmansachs","source":"github","weight":0.7},{"name":"hedgefund","source":"github","weight":0.7},{"name":"jpmorgan","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"quant","source":"github","weight":0.7},{"name":"quantitative-finance","source":"github","weight":0.7},{"name":"quantitative-trading","source":"github","weight":0.7},{"name":"trading","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/The-Swarm-Corporation/AutoHedge","ts":"2026-09-06T14:58:06.418Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"R-s0n/ars0n-framework-v2","owner":"R-s0n","name":"ars0n-framework-v2","desc":"AI Native Bug Bounty Hunting Framework Designed to Help Beginners Compete w/ the Pros","summaryCn":"帮新手用智能体打安全赏金，边学边抓漏洞的开源工具箱","reasonCn":"这个 Go 项目把漏洞赏金流程拆成目标范围管理、信息收集、指纹识别、漏洞验证和报告生成等阶段，再接入大模型做任务规划、证据整理与工具调用。它像给新人配了安全实验室，能自动串起代理、浏览器、目录扫描和笔记模块，把零散命令变成可复用流水线。能根据页面内容判断入口、参数和潜在风险点，减少新手盲目试错。","stars":821,"starGrowth":0,"createdAt":"2024-12-30T01:11:42Z","silentRounds":1,"language":"Go","topics":[],"aiDims":["AI Agent","安全工具","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["漏洞赏金","安全测试","AI辅助","自动化工具","渗透测试"],"domainKey":"漏洞赏金","pushedAt":"2026-09-10T17:23:15.587Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/R-s0n/ars0n-framework-v2","ts":"2026-09-10T17:23:15.587Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Mikubill/sd-webui-controlnet","owner":"Mikubill","name":"sd-webui-controlnet","desc":"","summaryCn":"ControlNet让AI绘画可按边缘、深度等条件精确控制生成结果。","reasonCn":"这个项目是Stable Diffusion WebUI的ControlNet扩展插件，让AI图像生成不再盲抽，而是能通过边缘图、深度图、人体姿态等条件精确控制生成结果。比如给AI一张线稿，它就能按线稿结构生成完整图像；输入一张深度图，就能控制画面的远近层次。目前持续跟进Depth Anything V2等新技术。","stars":17840,"starGrowth":0,"createdAt":"2023-02-12T16:26:27Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["创意工具","提示工程"],"zone":"工具","funScore":0,"domainTags":["AI绘画","Stable Diffusion","图像生成"],"domainKey":"AI绘画","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/Mikubill/sd-webui-controlnet","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"google-deepmind/alphagenome","owner":"google-deepmind","name":"alphagenome","desc":"This API provides programmatic access to the AlphaGenome model developed by Google DeepMind.","summaryCn":"谷歌DeepMind开源的基因组模型API，让生物学家轻松调用","reasonCn":"这是一个专为基因组学设计的生成式模型，通过API提供程序化访问。它利用Transformer架构处理DNA序列，能生成与自然语言处理模型类似的文本。该模型在基准测试中表现优异，为生物信息学领域的AI应用提供了强大工具。项目托管在GitHub上，便于科研人员复现和二次开发。","stars":2154,"starGrowth":0,"createdAt":"2024-10-16T09:33:11Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["模型与训练","文档理解","知识图谱"],"zone":"AI","funScore":0,"domainTags":["基因组学","生成式模型","DNA 序列","生物信息学","API"],"domainKey":"基因组学","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/google-deepmind/alphagenome","ts":"2026-09-10T22:16:01.544Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"The-Vibe-Company/quivr","owner":"The-Vibe-Company","name":"quivr","desc":"Opiniated RAG for integrating GenAI in your apps 🧠   Focus on your product rather than the RAG. Easy integration in existing products with customisation!  Any LLM: GPT4, Groq, Llama. Any Vectorstore: PGVector, Faiss. Any Files. Anyway you want.","summaryCn":"把你的文档变成随时问的智能助理，装进任何应用里","reasonCn":"这是一个专为产品经理设计的RAG框架，旨在降低AI集成门槛。它支持任意大模型和向量库，通过简单的API即可将企业知识库无缝嵌入现有产品，无需重构代码。用户可以自定义Prompt模板，实现精准的知识检索与生成，极大提升开发效率。","stars":39550,"starGrowth":0,"createdAt":"2023-05-12T23:53:30Z","silentRounds":1,"language":"Python","topics":["rag","ai","llm","chatbot"],"aiDims":["RAG与知识","AI应用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["RAG","向量数据库","大模型","知识库","API集成"],"domainKey":"RAG","pushedAt":"2026-08-31T18:02:45Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/The-Vibe-Company/quivr","ts":"2026-08-31T18:02:45Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kurianbenoy/Indic-Subtitler","owner":"kurianbenoy","name":"Indic-Subtitler","desc":"Open source subtitling platform 💻 for transcribing and translating videos/audios in Indic languages.","summaryCn":"开源视频字幕工具，支持印地语等印度语言自动转写与翻译","reasonCn":"Indic-Subtitler 结合 OpenAI Whisper 语音识别与 MarianMT 翻译模型，提供 Jupyter Notebook 界面。用户上传视频后，脚本自动生成时间轴、原始字幕并翻译成目标印度语言，支持批量处理。它适合研究者、内容创作者与低资源环境字幕制作，已在 GitHub 公开，社区欢迎贡献。","stars":95,"starGrowth":0,"createdAt":"2024-02-05T07:17:22Z","silentRounds":17,"language":"Jupyter Notebook","topics":["transformers"],"aiDims":["AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["语音识别","字幕生成","视频翻译","Whisper","多语言处理"],"domainKey":"语音识别","pushedAt":"2026-09-07T14:24:00Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.84,"source":"search","url":"https://github.com/kurianbenoy/Indic-Subtitler","ts":"2026-09-07T14:24:00Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SlimeBoyOwO/LingChat","owner":"SlimeBoyOwO","name":"LingChat","desc":"Immersive AI-driven Galgame chat with emotional expressions, desktop pet, scheduling, and interactive story modules. / 一款沉浸式 AI-Galgame 聊天软件，附带桌宠，日程，剧情功能","summaryCn":"把你的桌面变成AI Galgame，不仅好玩还能陪聊","reasonCn":"LingChat 是一款基于 Rust 构建的沉浸式 AI 聊天应用，完美融合了 Galgame（美少女游戏）的视觉体验与 AI 交互。项目集成了情感表情包、桌面宠物、日程管理以及可定制的剧情模块，为用户打造了一个既有 AI 智慧又充满人情味的虚拟伴侣。其桌面宠物功能允许 AI 角色在你的屏幕上活动，增强了陪伴感，同时 Rust 的性能优势确保了在处理复杂剧情和实时对话时的流畅度。","stars":2245,"starGrowth":4,"createdAt":"2025-04-06T12:06:42Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["AI应用","游戏","非AI-好玩"],"zone":"创意","funScore":0,"domainTags":["AI Galgame","桌面宠物","交互式叙事","虚拟伴侣"],"domainKey":"AI Galgame","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/SlimeBoyOwO/LingChat","ts":"2026-08-19T01:08:47.324Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DaizeDong/Stanford-CS231n-2021-and-2022","owner":"DaizeDong","name":"Stanford-CS231n-2021-and-2022","desc":"Notes and slides for Stanford CS231n 2021 & 2022 in English. I merged the contents together to get a better version. Assignments are not included. 斯坦福cs231n的课程笔记(英文版本，不含实验代码)，将2021与2022两年的课程进行了合并，分享以供交流。","summaryCn":"把斯坦福神级课笔记合并成一本AI入门神书","reasonCn":"该仓库合并了斯坦福CS231n 2021与2022两届课程的官方讲义与PPT，旨在提供更完善的视觉学习体验。内容涵盖CNN、RNN、优化器、迁移学习等计算机视觉核心理论，配有详细公式推导与图表解释，适合深度学习初学者系统梳理知识体系。","stars":26,"starGrowth":0,"createdAt":"2022-07-02T11:14:19Z","silentRounds":13,"language":"","topics":["course"],"aiDims":["学习资源","教程","模型与训练"],"zone":"资源","funScore":0,"domainTags":["计算机视觉","深度学习","课程笔记","CNN","AI教程"],"domainKey":"计算机视觉","pushedAt":"2026-08-29T22:38:14Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"course","source":"github","weight":0.7}],"aiScore":0.92,"source":"search","url":"https://github.com/DaizeDong/Stanford-CS231n-2021-and-2022","ts":"2026-08-29T22:38:14Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"llm-workflow-engine/llm-workflow-engine","owner":"llm-workflow-engine","name":"llm-workflow-engine","desc":"Power CLI and Workflow manager for LLMs (core package)","summaryCn":"专为大语言模型打造的灵活高效命令行与工作流管理核心引擎","reasonCn":"技术层面上，它采用了高度模块化的核心包设计，用纯Python语言编写。它不仅支持灵活的链式调用和状态管理，还能无缝对接各种主流大模型API和外部工具函数。通过精心优化的调度逻辑，它能够高效处理工作流中的并发请求、异常捕获和数据传递，确保整个AI流水线稳健运行。","stars":3715,"starGrowth":0,"createdAt":"2022-12-03T14:46:24Z","silentRounds":23,"language":"Python","topics":["chatbot"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["LLM","工作流","Python","命令行工具"],"domainKey":"LLM","pushedAt":"2026-09-05T21:19:52Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.86,"source":"search","url":"https://github.com/llm-workflow-engine/llm-workflow-engine","ts":"2026-09-05T21:19:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"CorentinJ/Real-Time-Voice-Cloning","owner":"CorentinJ","name":"Real-Time-Voice-Cloning","desc":"","summaryCn":"Real-Time-Voice-Cloning 是一个基于 SV2TTS 论文实现的深度学习项目，它的","reasonCn":"项目的核心逻辑分为三个紧密相连的阶段。第一阶段是“听”，它分析输入的几秒音频，提取出代表说话人身份的特征向量，忽略具体的说话内容，只保留音色特质。第二阶段和第三阶段则是“说”，利用第一阶段得到的参考特征，将任意输入的文本转换为波形数据。","stars":60144,"starGrowth":1,"createdAt":"2019-05-26T08:56:15Z","silentRounds":6,"language":"Python","topics":["deep-learning","pytorch","tensorflow","tts","voice-cloning","python"],"aiDims":["AI应用","模型与训练","非AI-好玩"],"zone":"AI","funScore":0.6,"domainTags":["语音克隆","深度学习","语音合成","实时处理"],"domainKey":"语音克隆","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"tensorflow","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"voice-cloning","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/CorentinJ/Real-Time-Voice-Cloning","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"x1xhlol/system-prompts-and-models-of-ai-tools","owner":"x1xhlol","name":"system-prompts-and-models-of-ai-tools","desc":"FULL Augment Code, Claude Code, Cluely, CodeBuddy, Comet, Cursor, Devin AI, Junie, Kiro, Leap.new, Lovable, Manus, NotionAI, Orchids.app, Perplexity, Poke, Qoder, Replit, Same.dev, Trae, Traycer AI, VSCode Agent, Warp.dev, Windsurf, Xcode, Z.ai Code, Dia & v0. (And other Open Sourced) System Prompts, Internal Tools & AI Models","summaryCn":"扒光了各大AI工具的底裤，看看它们都是怎么思考的","reasonCn":"这是一个巨大的GitHub仓库，收录了Cursor、Claude Code、Perplexity等几十款主流AI编程和生产力工具的系统提示词。对于想了解头部产品如何设计AI行为、优化提示词策略的用户来说，这是最直接的偷师素材，能帮你快速理解顶级AI工具的设计逻辑。","stars":143814,"starGrowth":0,"createdAt":"2025-03-05T16:38:29Z","silentRounds":1,"language":"","topics":["ai"],"aiDims":["提示工程","大语言模型","AI应用"],"zone":"资源","funScore":1,"domainTags":["系统提示词","AI工具","提示词工程","Cursor","Claude"],"domainKey":"系统提示词","pushedAt":"2026-08-11T13:01:09Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7}],"aiScore":0.98,"source":"trending","url":"https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools","ts":"2026-08-11T13:01:09Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yezz123/authx","owner":"yezz123","name":"authx","desc":"Ready-to-use and customizable Authentications and Oauth2 management for FastAPI ✨","summaryCn":"FastAPI项目的身份认证神器，开箱即用","reasonCn":"核心技术亮点在于它和FastAPI的深度集成。它提供了装饰器、中间件和路由生成器，可以一行代码添加登录、注册、OAuth2授权等功能。支持JWT和OAuth2两种主流协议，还能自动集成OpenID Connect标准，支持Google、GitHub等第三方登录。","stars":1196,"starGrowth":0,"createdAt":"2021-08-23T16:18:04Z","silentRounds":27,"language":"Python","topics":["backend"],"aiDims":["AI基础设施","安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["身份认证","OAuth2","JWT","FastAPI","Python库"],"domainKey":"身份认证","pushedAt":"2026-09-21T22:21:43Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/yezz123/authx","ts":"2026-09-21T22:21:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"github/copilot-cli","owner":"github","name":"copilot-cli","desc":"","summaryCn":"GitHub Copilot CLI 把 Copilot coding agent 的能力直接搬到了你的终端里。","reasonCn":"在实际使用中，交互方式完全依靠自然语言。你不需要写复杂的 prompt 或记忆各种命令参数，直接用日常语言告诉它你想做什么就行。比如你可以说「帮我写一个处理 JSON 数据的函数」，或者「这段代码为什么会报错」，它就能理解你的意图并给出准确的回应。","stars":11195,"starGrowth":0,"createdAt":"2023-01-06T09:22:30Z","silentRounds":2,"language":"Shell","topics":[],"aiDims":["代码助手"],"zone":"AI","funScore":0.7,"domainTags":["AI编程助手","终端工具","代码生成","CLI"],"domainKey":"AI编程助手","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/github/copilot-cli","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GeeeekExplorer/nano-vllm","owner":"GeeeekExplorer","name":"nano-vllm","desc":"","summaryCn":"nano-vLLM用约1200行Python代码从零构建了一个完整的离线推理引擎，代码量精","reasonCn":"项目内置了完整的优化技术栈，包括前缀缓存、张量并行、Torch编译和CUDA图等。前缀缓存可以复用相同输入前缀的计算结果，大幅减少重复推理的延迟；张量并行让单个GPU能处理更大的模型；Torch编译和CUDA图则进一步优化了推理速度。","stars":15583,"starGrowth":3,"createdAt":"2025-06-09T16:22:14Z","silentRounds":0,"language":"Python","topics":["inference","llm","pytorch","transformer","deep-learning","nlp"],"aiDims":["推理引擎","大语言模型","模型部署"],"zone":"AI","funScore":1,"domainTags":["推理引擎","大模型","CUDA优化","vLLM","教学源码"],"domainKey":"推理引擎","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"inference","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"transformer","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"nlp","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/GeeeekExplorer/nano-vllm","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"threestudio-project/threestudio","owner":"threestudio-project","name":"threestudio","desc":"","summaryCn":"项目通过引入诸如Score Distillation等先进算法，成功将2D生成模型的潜力挖","reasonCn":"除了基础的生成能力，该框架还具备强大的扩展性。通过threestudio-extension仓库，社区贡献者不断加入如MagicClay、3DFuse等新算法，使得生成的物体细节和拓扑结构越来越精准。这种模块化的设计思路，让开发者能够根据具体任务需求，灵活更换底层的蒸馏策略或模型结构，是目前学术","stars":7065,"starGrowth":0,"createdAt":"2023-04-06T08:43:22Z","silentRounds":2,"language":"Jupyter Notebook","topics":[],"aiDims":["模型部署","AI应用","多模态Agent"],"zone":"AI","funScore":1,"domainTags":["3D生成","文本转3D","AI建模","扩散模型"],"domainKey":"3D生成","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/threestudio-project/threestudio","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenGVLab/LLaMA-Adapter","owner":"OpenGVLab","name":"LLaMA-Adapter","desc":"","summaryCn":"LLaMA-Adapter 的诞生源于对大模型微调效率的极致追求。","reasonCn":"从单一文本模态向多模态的扩展是目前该项目的最大亮点。V2 版本不再局限于处理纯文本指令，而是引入了视觉编码器，让模型能够“看懂”图片。这意味着用户可以上传图片并提出相关问题，模型能基于视觉内容和上下文给出回答。","stars":5913,"starGrowth":0,"createdAt":"2023-03-19T15:08:12Z","silentRounds":7,"language":"Python","topics":[],"aiDims":["微调","大语言模型","多模态Agent"],"zone":"AI","funScore":0.6,"domainTags":["模型微调","多模态","参数高效"],"domainKey":"多模态","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/OpenGVLab/LLaMA-Adapter","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"abi/screenshot-to-code","owner":"abi","name":"screenshot-to-code","desc":"","summaryCn":"该工具支持将网站的录屏直接转化为可交互的原型代码，意味着你可以像给手机","reasonCn":"项目内部集成了一套强大的模型阵列，包括 Gemini 3 Flash、GPT-5.5 以及 Claude Opus 4.8 等业界顶尖模型。这些大模型的混合调用，确保了它在处理不同风格界面时的识别精度和代码健壮性，即便是复杂的布局结构也能通过指令调整得到理想的还原效果，是前端工程师和产品经理不可多得","stars":79542,"starGrowth":0,"createdAt":"2023-11-14T17:53:32Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI应用","文档理解","代码助手"],"zone":"AI","funScore":1,"domainTags":["截图转代码","录屏转代码","前端生成","AI编程"],"domainKey":"截图转代码","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.74,"source":"search","url":"https://github.com/abi/screenshot-to-code","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alirezarezvani/claude-skills","owner":"alirezarezvani","name":"claude-skills","desc":"345 Claude Code skills & agent skills & plugins (30+ Agents, 70+ custom commands, 330+ skills, customizable references, scripts)for Claude Code, Codex, Gemini CLI, Cursor, and 8 more coding agents — engineering, marketing, product, compliance, C-level advisory, research, business operations, commercial & finance, and your daily productivity skills.","summaryCn":"给AI编码助手发的技能升级包，30多款Agent秒变编程大神","reasonCn":"这是一个专为Claude Code、Cursor、Codex等编码Agent准备的技能集合仓库，包含330多个可复用skills、70多个自定义命令和30多个预置Agent。每个skill都是一段精心设计的提示词或脚本，能让AI编码助手在代码生成、架构设计、Git操作、测试编写等场景下更智能。支持多平台，安装即用，无需复杂配置，是目前GitHub上最热门的开发效率增强项目之一。","stars":26261,"starGrowth":2,"createdAt":"2025-10-19T04:04:05Z","silentRounds":0,"language":"Python","topics":["developer-tools"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI编程助手","技能包","Agent","Cursor","Claude Code"],"domainKey":"Agent","pushedAt":"2026-08-30T09:46:16Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/alirezarezvani/claude-skills","ts":"2026-08-30T09:46:16Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"titanwings/colleague-skill","owner":"titanwings","name":"colleague-skill","desc":"将冰冷的离别化为温暖的 Skill，欢迎加入数字生命1.0！Transforming cold farewells into warm skills? It's giving rebirth era. Welcome to Digital Life 1.0. 🫶","summaryCn":"把离职同事变成AI助手，离职也能继续打工","reasonCn":"这是一个将离职同事的聊天记录转化为AI Agent的技能项目。它通过自然语言处理技术，提取并重构离职员工的知识、风格和沟通习惯，生成一个可以继续协助工作的数字生命体。这不仅是技术演示，更提供了一套从数据清洗到模型部署的完整方案。","stars":24970,"starGrowth":4,"createdAt":"2026-03-30T05:22:23Z","silentRounds":0,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":1,"domainTags":["AI Agent","知识传承","数字生命"],"domainKey":"AI Agent","pushedAt":"2026-08-18T13:35:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/titanwings/colleague-skill","ts":"2026-08-18T13:35:40Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ahmadrosid/nakama","owner":"ahmadrosid","name":"nakama","desc":"It's like Hermes Agent &amp; OpenClaw but designed to work nicely with teams.","summaryCn":"让团队像打配合一样高效指挥多个AI助手的协作中枢","reasonCn":"这是一款基于TypeScript开发的AI团队协作文具，旨在优化Hermes与OpenClaw等智能体在多人环境下的运作体验。它通过任务分配、状态同步及权限管理模块，解决了单兵作战式AI助手在团队场景中容易出现的上下文断裂与责任模糊问题，助力构建稳定的多智能体工作流，提升复杂业务场景下的自动化处理效率与可维护性。","stars":409,"starGrowth":2,"createdAt":"2026-05-23T06:02:30Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["多智能体","AI协作","团队管理","Agent编排","工作流"],"domainKey":"多智能体","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/ahmadrosid/nakama","ts":"2026-09-10T22:16:01.544Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AsyncFuncAI/deepwiki-open","owner":"AsyncFuncAI","name":"deepwiki-open","desc":"Open Source DeepWiki: AI-Powered Wiki Generator for GitHub/Gitlab/Bitbucket Repositories. Join the discord: https://discord.gg/gMwThUMeme","summaryCn":"很多开源项目最让人抓狂的就是没有文档，或者文档跟代码版本不同步。","reasonCn":"它的核心技术基于RAG（检索增强生成），通过向量化处理代码库中的语义，从而实现对复杂代码库的精准问答。相比传统的自动化文档工具，DeepWiki更智能的地方在于它能理解代码的深层意图，而不仅仅是提取注释。它支持与Git平台联动，只要代码有更新，文档也会自动尝试同步，保持信息","stars":18043,"starGrowth":0,"createdAt":"2025-04-30T06:01:29Z","silentRounds":1,"language":"Python","topics":["self-hosted"],"aiDims":["文档理解","AI应用","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["RAG","代码文档生成","Git集成","交互式问答"],"domainKey":"RAG","pushedAt":"2026-09-03T20:29:26Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/AsyncFuncAI/deepwiki-open","ts":"2026-09-03T20:29:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0toshigami/orbis-pictus","owner":"0toshigami","name":"orbis-pictus","desc":"A tap-to-explore picture book where an AI draws every page in real time — type anything, click anything inside, and it draws a new page about what you clicked. No links, no markup, every pixel made on demand. An open-source homage to flipbook.page.","summaryCn":"点一下就能让AI当场画出一页全新绘本，无厘头互动艺术神器","reasonCn":"基于TypeScript构建的实时AI绘本引擎，用户输入任意主题、点击画布任意位置，AI即时生成对应内容的新页面。没有预设链接、没有固定剧本，每点击一次都是独一无二的创作。技术栈简洁，前端调用AI绘图API，真正实现了人机共创的开放式绘本体验，无需任何技术背景即可上手玩。","stars":27,"starGrowth":0,"createdAt":"2026-08-11T00:16:54Z","silentRounds":17,"language":"TypeScript","topics":["creative-coding"],"aiDims":["非AI-好玩","创意工具","AI应用"],"zone":"创意","funScore":1,"domainTags":["互动绘本","实时绘图","生成艺术","创意编程","WebGL"],"domainKey":"互动绘本","pushedAt":"2026-08-11T16:56:20Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/0toshigami/orbis-pictus","ts":"2026-08-11T16:56:20Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"e2b-dev/awesome-ai-agents","owner":"e2b-dev","name":"awesome-ai-agents","desc":"A list of AI autonomous agents","summaryCn":"现在AI Agent的概念非常火，但市面上项目多如牛毛，让人眼花缭乱。","reasonCn":"它之所以厉害，在于其维护者对“Agent”定义的深刻理解。列表不是简单堆砌，而是分门别类：包括支持自主推理的框架（如AutoGPT）、支持多智能体协作的平台（如CrewAI）、以及针对特定工具链优化的代理方案。通过这些项目，你能清晰地看到AI从单一的聊天机器人，是如何演变成能够自动","stars":30125,"starGrowth":7,"createdAt":"2023-06-19T00:20:06Z","silentRounds":0,"language":"","topics":["ai"],"aiDims":["AI Agent","学习资源","论文"],"zone":"资源","funScore":0,"domainTags":["AI Agent","项目导航","资源列表","精选集"],"domainKey":"AI Agent","pushedAt":"2026-08-21T18:52:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7}],"aiScore":0.92,"source":"search","url":"https://github.com/e2b-dev/awesome-ai-agents","ts":"2026-08-21T18:52:45Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.021697095435684648,"copyOk":false},{"repo":"basecamp/omarchy","owner":"basecamp","name":"omarchy","desc":"Beautiful, Modern &amp; Opinionated Linux","summaryCn":"让Linux装起来像装Windows一样爽，极简又漂亮","reasonCn":"这是BaseCamp团队维护的一款极简Linux发行版，主打“现代且带有观点”的体验。它基于Arch Linux，预装了大量经过精心挑选的桌面环境、终端工具和开发环境，开箱即用。核心特点是对系统进行深度定制，移除了所有臃肿的默认软件，只保留最必要的组件，让系统轻快且美观。","stars":42696,"starGrowth":46,"createdAt":"2025-06-01T07:26:22Z","silentRounds":0,"language":"Shell","topics":[],"aiDims":["非AI-实用","非AI-好玩","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["Linux发行版","Arch Linux","极简桌面"],"domainKey":"Linux发行版","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/basecamp/omarchy","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.14123215289488478,"copyOk":false},{"repo":"markmoxon/elite-source-code-commodore-64","owner":"markmoxon","name":"elite-source-code-commodore-64","desc":"Fully documented source code for the classic game Elite on the Commodore 64","summaryCn":"把上古神作游戏每一行汇编代码都拆解注释的考古宝库","reasonCn":"技术上，它完全是用古老的汇编语言写成的。最让人惊叹的是原作者在只有几KB内存和极弱CPU的电脑上，硬生生通过精妙的数学算法实现了流畅的3D矢量图形渲染。这个开源项目的作者花费了数年时间，把这些天书般的机器代码一行行逆向并加上了万字长文级别的技术解释，堪称计算机考古学的","stars":468,"starGrowth":0,"createdAt":"2024-10-22T14:38:32Z","silentRounds":23,"language":"Assembly","topics":["retrocomputing"],"aiDims":["非AI-好玩","游戏"],"zone":"资源","funScore":1,"domainTags":["复古计算","Commodore 64","汇编语言","游戏源码","3D渲染"],"domainKey":"复古计算","pushedAt":"2026-09-15T08:34:49Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"Assembly","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/markmoxon/elite-source-code-commodore-64","ts":"2026-09-15T08:34:49Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"amanmaqsood/prose-humanizer","owner":"amanmaqsood","name":"prose-humanizer","desc":"A universal AI writing skill that turns generic drafts into specific, truthful, voice-consistent prose.","summaryCn":"专门给AI写作去味的神器，一秒把机器味十足的假大空草稿改得像真人写的。","reasonCn":"该项目提供了一个通用的AI写作技能，能够剥离AI文本中常见的冗长陈腔滥调和机械腔调。它通过精细的提示词策略和风格重构，让生成的文章更具体、真实且保持自然一致的语调，大幅提升中文或英文写作的可读性与人类质感。","stars":20,"starGrowth":0,"createdAt":"2026-08-22T15:14:33Z","silentRounds":14,"language":"JavaScript","topics":["writing"],"aiDims":["AI写作","提示工程"],"zone":"AI","funScore":1,"domainTags":["提示词工程","文本润色","写作助手","AI 技能","风格迁移"],"domainKey":"提示词工程","pushedAt":"2026-09-08T03:42:13Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/amanmaqsood/prose-humanizer","ts":"2026-09-08T03:42:13Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"taosdata/TDengine","owner":"taosdata","name":"TDengine","desc":"High-performance, scalable time-series database designed for Industrial IoT (IIoT) scenarios","summaryCn":"高吞吐量时序数据库，物联网和工业场景的数据存储利器","reasonCn":"专为物联网和工业场景打造的高性能时序数据库，支持分布式部署和 SQL 查询，写入性能远超传统数据库。内置聚合、降采样、数据裁剪等时序特有功能，特别适合处理设备传感器数据、监控日志、实时指标等时间序列数据，是工业级应用的首选数据库之一。","stars":25143,"starGrowth":0,"createdAt":"2019-07-11T08:33:48Z","silentRounds":1,"language":"C","topics":["database","iot","monitoring"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["时序数据库","物联网","工业数据","数据库","高性能"],"domainKey":"时序数据库","pushedAt":"2026-09-17T02:02:07Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"iot","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/taosdata/TDengine","ts":"2026-09-17T02:02:07Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"holaboss-ai/holaOS","owner":"holaboss-ai","name":"holaOS","desc":"Open-source All in One AI agent workspace. Run any agent — Claude Code, Codex — across your tools (100+ integrations + MCP), apps, browser, and files, with shared memory. Built-in models or BYOK.","summaryCn":"把所有AI工具串起来，一键搞定你的工作流","reasonCn":"这是一个类似操作系统的AI Agent平台，它把Claude Code、OpenAI Codex等主流模型通过MCP协议统一管理。用户可以在一个界面里同时控制浏览器、本地文件、100多个第三方应用，所有Agent共享内存上下文，实现跨工具协作。支持自带模型和BYOK两种模式，让个人也能低成本搭建私有AI工作台。","stars":11354,"starGrowth":3,"createdAt":"2026-03-22T07:36:16Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI agent","MCP协议","浏览器自动化","自托管","工作流"],"domainKey":"AI agent","pushedAt":"2026-08-21T13:11:13Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/holaboss-ai/holaOS","ts":"2026-08-21T13:11:13Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liustack/modlens","owner":"liustack","name":"modlens","desc":"The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics). | 全网最强 DeepSeek Harness 外挂视觉插件，为 DeepSeek、GLM 等纯文本模型外挂视觉能力，粘贴图片即得结构化 JSON 证据（OCR、版面、语义）。","summaryCn":"让只会写代码的AI学会“看”图，直接读出图片里的内容","reasonCn":"这是全网首个专为DeepSeek Harness打造的视觉插件，作为文本型编码Agent的视觉桥梁。用户只需粘贴图片，它就能通过OCR、布局分析和语义理解，将图片转化为结构化JSON数据。这使得仅处理文本的AI智能体能够直接“看见”并处理视觉信息，大幅提升了代码助手在处理UI截图、图表或文档时的能力，是AI Agent向多模态进化的关键一步。","stars":4015,"starGrowth":0,"createdAt":"2026-02-22T07:12:00Z","silentRounds":1,"language":"TypeScript","topics":["agent-skills","claude-code","claude-skills","codex","deepseek","dsh-plugin","glm","harness","harness-engineering","hermes-agent","openclaw","pi-agent","text-only-llm","vision-transformer","image-to-text","multimodal","ocr","vision","cordis","dsh"],"aiDims":["多模态Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","视觉插件","OCR","编码助手","插件"],"domainKey":"AI Agent","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"claude-skills","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"glm","source":"github","weight":0.7},{"name":"harness","source":"github","weight":0.7},{"name":"harness-engineering","source":"github","weight":0.7},{"name":"hermes-agent","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"pi-agent","source":"github","weight":0.7},{"name":"text-only-llm","source":"github","weight":0.7},{"name":"vision-transformer","source":"github","weight":0.7},{"name":"image-to-text","source":"github","weight":0.7},{"name":"multimodal","source":"github","weight":0.7},{"name":"ocr","source":"github","weight":0.7},{"name":"vision","source":"github","weight":0.7},{"name":"cordis","source":"github","weight":0.7},{"name":"dsh","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/liustack/modlens","ts":"2026-08-18T00:56:27.274Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tabularis-ai/be_great","owner":"tabularis-ai","name":"be_great","desc":"A novel approach for synthesizing tabular data using pretrained large language models","summaryCn":"使用预训练大型语言模型，创新地合成表格数据，提升信息检索和分析效率","reasonCn":"通过结合深度学习技术和大规模预训练模型，该开源项目旨在提高表格数据的处理速度和准确性。它能够将文本数据转换为表格格式，使得数据分析更加直观和高效。","stars":370,"starGrowth":0,"createdAt":"2022-09-14T14:21:23Z","silentRounds":20,"language":"Python","topics":["transformers"],"aiDims":["AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["表格数据","LLM","数据合成"],"domainKey":"表格数据","pushedAt":"2026-08-16T16:08:24Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/tabularis-ai/be_great","ts":"2026-08-16T16:08:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"thedaviddias/Front-End-Checklist","owner":"thedaviddias","name":"Front-End-Checklist","desc":"🗂 The essential checklist for modern web development, for humans and AI agents","summaryCn":"现代前端开发者的终极清单，AI和人类都能用的宝典","reasonCn":"这是一个面向现代Web开发的综合检查清单，涵盖HTML、CSS、JavaScript、无障碍、SEO、性能优化等全栈要素，同时集成了AI智能体协作的最佳实践。项目采用MDX格式编写，兼顾文档可读性与代码演示能力，被全球数万开发者视为前端工程化的黄金标准。","stars":74233,"starGrowth":4,"createdAt":"2017-10-16T10:12:36Z","silentRounds":0,"language":"MDX","topics":["ai-agent"],"aiDims":["AI基础设施","开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["Web开发","前端工程化","代码规范","MDX","最佳实践"],"domainKey":"Web开发","pushedAt":"2026-08-14T05:59:25Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"MDX","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/thedaviddias/Front-End-Checklist","ts":"2026-08-14T05:59:25Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rlabbe/Kalman-and-Bayesian-Filters-in-Python","owner":"rlabbe","name":"Kalman-and-Bayesian-Filters-in-Python","desc":"","summaryCn":"SpaceX 的工程师们在内部教学中采用这套教程来传授关键的状态估计概念，这足","reasonCn":"项目采用 Jupyter Notebook 格式编写是其最大的亮点，你不需要进行繁琐的本地配置，只需在浏览器中打开，就能直接运行、修改代码并实时观察数据变化。这种“所见即所得”的互动式学习方式，能够帮助你深刻理解滤波过程中的参数设置对结果的影响。","stars":19358,"starGrowth":1,"createdAt":"2014-05-16T19:24:36Z","silentRounds":1,"language":"Jupyter Notebook","topics":[],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0.6,"domainTags":["卡尔曼滤波","贝叶斯估计","状态估计"],"domainKey":"卡尔曼滤波","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/rlabbe/Kalman-and-Bayesian-Filters-in-Python","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"THUDM/slime","owner":"THUDM","name":"slime","desc":"slime is an LLM post-training framework for RL Scaling.","summaryCn":"让大模型像打游戏一样越练越强的RL训练框架","reasonCn":"Slime是THUDM团队开源的LLM后训练框架，专注于强化学习缩放。它通过自研算法大幅提升模型在复杂任务上的表现，尤其擅长让模型学会主动思考和多步推理。该框架支持高效分布式训练，能显著降低计算资源消耗，已在多项基准测试中刷新记录。","stars":8523,"starGrowth":2,"createdAt":"2025-06-18T07:15:54Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","微调","大语言模型"],"zone":"AI","funScore":0,"domainTags":["强化学习","大模型训练","RLHF","深度学习"],"domainKey":"强化学习","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/THUDM/slime","ts":"2026-08-17T00:55:40.962Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"arcboxlabs/arcbox","owner":"arcboxlabs","name":"arcbox","desc":"Run AI agents on real and isolated machines — own kernel, filesystem, and network — with &lt;100ms boot. Local first, OCI compatible, pure Rust.","summaryCn":"让AI智能体在隔离沙盒里跑，独立内核文件系统，启动不到0.1秒","reasonCn":"让AI Agent在完全隔离的环境中运行，每个实例有独立的内核、文件系统和网络。启动时间低于100毫秒，基于Rust实现，兼容OCI标准。通过REST API与Agent交互，适合需要安全隔离和批量运行的场景，比如金融、医疗数据处理。","stars":6600,"starGrowth":19,"createdAt":"2026-01-04T14:46:50Z","silentRounds":0,"language":"Rust","topics":["ai-agents","computer-use","containers","docker","firecracker","microvm","rust","sandbox","virtual-machine","virtualization"],"aiDims":["AI基础设施","安全工具","AI Agent"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","沙盒隔离","Rust","虚拟机","OCI标准"],"domainKey":"AI Agent","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"computer-use","source":"github","weight":0.7},{"name":"containers","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"firecracker","source":"github","weight":0.7},{"name":"microvm","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"sandbox","source":"github","weight":0.7},{"name":"virtual-machine","source":"github","weight":0.7},{"name":"virtualization","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/arcboxlabs/arcbox","ts":"2026-09-07T02:44:20.540Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.07876363636363636,"copyOk":true},{"repo":"JerBouma/FinanceToolkit","owner":"JerBouma","name":"FinanceToolkit","desc":"Transparent and Efficient Financial Analysis","summaryCn":"一套简洁高效的金融分析工具包，让数据分析师不再为繁杂的财务计算发愁。","reasonCn":"这是一个为金融专业人士与数据爱好者设计的财务分析工具集，提供透明且高效的财务比率计算、股票指标分析与数据可视化功能。它内置了多种专业的金融模型与计算公式，支持从多种数据源拉取并一键生成分析报表，极大降低了复杂财务建模的门槛，是财务审计、个股分析与金融量化研究领域中不可多得的高质量开源工具，显著提升了金融数据处理的专业度与响应速度。","stars":5371,"starGrowth":2,"createdAt":"2019-04-08T11:44:25Z","silentRounds":0,"language":"Python","topics":["finance"],"aiDims":["非AI-实用","可视化","效率工具"],"zone":"工具","funScore":0,"domainTags":["金融分析","财务指标","股票数据"],"domainKey":"金融分析","pushedAt":"2026-09-10T16:10:58Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/JerBouma/FinanceToolkit","ts":"2026-09-10T16:10:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Devin-AXIS/deepseek-design","owner":"Devin-AXIS","name":"deepseek-design","desc":"DeepSeek Harness 可编辑设计系统：AI 生成、可视化编辑、模板市场与 PPT｜Native Design & PPT Studio for DeepSeek Harness.","summaryCn":"如果你常被“方案写得很好，但PPT排版怎么看都土”的问题困扰，这个项目就是你","reasonCn":"其核心逻辑在于构建了一个“AI驱动的设计语义”，它通过DeepSeek作为底层大脑，理解你的设计意图，并将其转化为可视化的布局参数。这不仅仅是填空，而是通过Native Design架构，让每一张幻灯片都能拥有模块化的逻辑，这种深度与大型设计工具的结合，使其在处理复杂数据展示和逻辑图示","stars":1474,"starGrowth":3,"createdAt":"2026-08-15T13:08:53Z","silentRounds":0,"language":"JavaScript","topics":["design"],"aiDims":["创意工具","可视化","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI演示文稿生成","设计系统","PPT自动化"],"domainKey":"AI演示文稿生成","pushedAt":"2026-09-04T10:40:28Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/Devin-AXIS/deepseek-design","ts":"2026-09-04T10:40:28Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"inboxpraveen/LLM-Minutes-of-Meeting","owner":"inboxpraveen","name":"LLM-Minutes-of-Meeting","desc":"A tool that transforms audio or video files into text transcripts and generates concise meeting minutes. Stay organized and efficient in your meetings, and get ready for Phase 2 where we'll be open for contributions to enable real-time meeting transcription! ","summaryCn":"一键上传音视频，LLM 自动转写并生成精炼会议纪要，省时又不漏要点","reasonCn":"开会时常常被繁杂的讨论淹没，事后又要花半小时甚至更久手动整理纪要？LLM‑Minutes‑of‑Meeting 通过“一键上传”实现全流程自动化，先用 OpenAI Whisper 将音视频转成文字稿，再交给 GPT‑4/Claude 等大模型进行要点抽取、决策归纳和待办事项生成，最后输出结构化的 Markdown 或 PDF，","stars":175,"starGrowth":0,"createdAt":"2023-10-11T19:22:52Z","silentRounds":16,"language":"Python","topics":["transformers"],"aiDims":["AI应用","文档理解","工具调用"],"zone":"AI","funScore":0,"domainTags":["会议纪要","Whisper","LLM"],"domainKey":"会议纪要","pushedAt":"2026-09-07T09:27:21Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/inboxpraveen/LLM-Minutes-of-Meeting","ts":"2026-09-07T09:27:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CodeSoul-co/Plasmod","owner":"CodeSoul-co","name":"Plasmod","desc":"Plasmod: An Agent-Native Database for Cognitive Objects, Event-Driven Materialization, and Structured Evidence Retrieval","summaryCn":"专为AI Agent设计的认知型数据库，存事件更存证据链","reasonCn":"Plasmod是一个实验性的Agent原生数据库，不只是一般的向量存储，而是专注于认知对象的持久化和事件驱动的材料化。它为AI Agent提供了结构化证据检索能力，解决传统向量库缺乏上下文和因果关联的问题，非常适合需要复杂推理和记忆追踪的Agent系统。","stars":25,"starGrowth":0,"createdAt":"2026-03-16T12:31:13Z","silentRounds":27,"language":"Go","topics":["llm-model"],"aiDims":["AI基础设施","向量数据库","AI Agent"],"zone":"AI","funScore":0,"domainTags":["数据库","事件驱动","RAG","Agent"],"domainKey":"RAG","pushedAt":"2026-09-05T05:00:37Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/CodeSoul-co/Plasmod","ts":"2026-09-05T05:00:37Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"larashero3-dotcom/lieflat-charts","owner":"larashero3-dotcom","name":"lieflat-charts","desc":"Data visualization Skill for AI Agents, turning data into polished, interactive HTML charts. 面向 AI Agents 的数据可视化 Skill，将数据快速生成精致、可交互的 HTML 图表。","summaryCn":"让AI像人类一样生成精美图表，再也不用担心审美问题了","reasonCn":"这个项目本质上是一个封装好的可视化工具库。它定义了一套标准的数据结构和渲染规则，AI Agent只需要理解并生成符合这种结构的数据，图表就会自动生成。它支持折线图、柱状图、饼图等多种类型，并且可以自定义配色方案。","stars":5642,"starGrowth":2,"createdAt":"2026-07-16T11:17:49Z","silentRounds":0,"language":"HTML","topics":["data-visualization"],"aiDims":["AI应用","可视化"],"zone":"AI","funScore":0,"domainTags":["数据可视化","AI Agent技能","图表生成","HTML渲染","前端组件"],"domainKey":"数据可视化","pushedAt":"2026-09-05T02:40:41Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/larashero3-dotcom/lieflat-charts","ts":"2026-09-05T02:40:41Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"blackboxaudio/bbx_audio","owner":"blackboxaudio","name":"bbx_audio","desc":"Suite of Rust crates for building and creating audio software 🧮","summaryCn":"用 Rust 写的音频制作套件，让声音创作像搭积木。","reasonCn":"这是一套基于 Rust 生态构建的音频软件工具箱，涵盖 DSP 算法、合成器、效果器等核心模块。它通过 crates 提供了底层音频处理能力，支持实时音频流和低延迟处理。代码结构清晰，类型安全且性能优异，适合构建各类音频应用。","stars":15,"starGrowth":0,"createdAt":"2024-06-25T19:21:34Z","silentRounds":13,"language":"Rust","topics":["generative-art"],"aiDims":["创意工具","非AI-好玩","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["Rust","音频软件","DSP","合成器","生成艺术"],"domainKey":"Rust","pushedAt":"2026-09-11T02:27:49Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"generative-art","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/blackboxaudio/bbx_audio","ts":"2026-09-11T02:27:49Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"stormzhang/ai-coding-guide","owner":"stormzhang","name":"ai-coding-guide","desc":"「可能是全网最全的」📘 面向小白的 AI 编程 CLI 中文教程：Claude Code + Codex 92 篇精修","summaryCn":"很多人第一次用人工智能写代码，都会卡在同一个地方：到底该让它写函数、改","reasonCn":"它的核心亮点不是某个具体模型，而是把命令式编程工作流讲透。你会看到怎么描述需求、怎么让工具读项目、怎么限制修改范围、怎么把大任务拆成小步骤。和零散文章相比，教程更强调中文语境下的表达习惯和常见误区，降低新手对着英文文档发呆的成本。","stars":1865,"starGrowth":0,"createdAt":"2014-01-17T02:52:55Z","silentRounds":1,"language":"","topics":["tutorial"],"aiDims":["教程","代码助手","学习资源"],"zone":"资源","funScore":0,"domainTags":["编程教程","AI 编程","工作流指南","中文文档","路线图"],"domainKey":"编程教程","pushedAt":"2026-09-02T18:09:33Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7}],"aiScore":0.88,"source":"search","url":"https://github.com/stormzhang/ai-coding-guide","ts":"2026-09-02T18:09:33Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rohitg00/awesome-claude-code-toolkit","owner":"rohitg00","name":"awesome-claude-code-toolkit","desc":"The most comprehensive toolkit for Claude Code -- 135 agents, 35 curated skills, 42 commands, 176+ plugins, 20 hooks, 15 rules, 7 templates, 14 MCP configs, 26 companion apps, 52 ecosystem entries, and more.","summaryCn":"Claude Code最强插件库，一键变身顶级AI程序员","reasonCn":"这是Claude Code生态中最全面的工具箱，集成了135个智能体、35项精选技能和176个插件。它不仅提供丰富的命令集，还包含MCP配置和配套应用，极大地扩展了Claude在代码生成和协作场景的能力。通过这套工具，用户可以更高效地利用Claude完成复杂开发任务，实现从代码生成到项目管理的全流程覆盖。","stars":2633,"starGrowth":1,"createdAt":"2026-02-04T21:09:33Z","silentRounds":2,"language":"JavaScript","topics":["claude","claude-code","claudecode","claudecode-hooks","plugins","skills"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["Claude Code","MCP","AI编程"],"domainKey":"Claude Code","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"claudecode","source":"github","weight":0.7},{"name":"claudecode-hooks","source":"github","weight":0.7},{"name":"plugins","source":"github","weight":0.7},{"name":"skills","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/rohitg00/awesome-claude-code-toolkit","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"macro-inc/macro","owner":"macro-inc","name":"macro","desc":"Macro is a unified workspace for teams: email, chat, docs, tasks, agents, calls, and CRM — @-linked together with shared AI memory.","summaryCn":"给团队装上“AI大脑”，把邮件、文档、任务全串起来自动处理。","reasonCn":"传统的办公软件都是孤岛，这个工具把你的所有工作流连成一张网。它用共享记忆串联起各类应用，让 AI Agent 能够跨应用调用工具，实现自动回复邮件、整理会议纪要、管理 CRM 等复杂任务，真正实现 AI 驱动的协同办公。","stars":4381,"starGrowth":1,"createdAt":"2025-11-08T15:18:27Z","silentRounds":1,"language":"Rust","topics":["office","rust","crm","linear","notion","slack","startup","startup-tools","agent","crm-system","email","messaging","workspace","ai","ai-agents","all-in-one","mcp","notes","slack-alternative","typescript"],"aiDims":["AI Agent","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["团队协作","AI Agent","智能工作流"],"domainKey":"AI Agent","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"office","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"crm","source":"github","weight":0.7},{"name":"linear","source":"github","weight":0.7},{"name":"notion","source":"github","weight":0.7},{"name":"slack","source":"github","weight":0.7},{"name":"startup","source":"github","weight":0.7},{"name":"startup-tools","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"crm-system","source":"github","weight":0.7},{"name":"email","source":"github","weight":0.7},{"name":"messaging","source":"github","weight":0.7},{"name":"workspace","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"all-in-one","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"notes","source":"github","weight":0.7},{"name":"slack-alternative","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/macro-inc/macro","ts":"2026-08-16T00:54:51.023Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kenn-io/agentsview","owner":"kenn-io","name":"agentsview","desc":"Local-first session search, analytics, insights, and token use statistics for coding agents, supporting Claude Code, Codex, and more than 20 other agents.","summaryCn":"给AI写代码助手装上“私人秘书”，自动记录对话还能分析Token消耗。","reasonCn":"这是一个专为编程Agent设计的本地化会话管理与分析工具，支持Claude Code、Codex等超过20种Agent。它不仅提供会话搜索与Token统计，还能深度分析Agent的行为模式与Token使用效率，帮助开发者更好地理解和优化Agent工作流。","stars":5968,"starGrowth":2,"createdAt":"2026-02-19T19:33:24Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["Agent分析","Token统计","会话管理"],"domainKey":"Agent分析","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/kenn-io/agentsview","ts":"2026-08-17T00:55:40.962Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liuhuanyong/QASystemOnMedicalKG","owner":"liuhuanyong","name":"QASystemOnMedicalKG","desc":"","summaryCn":"liuhuanyong/QASystemOnMedicalKG项目聚焦于医药领域知识图谱的构建与应用，","reasonCn":"在技术实现上，项目采用Python进行开发，并依赖Neo4j作为底层图数据库来存储和查询这些海量数据。用户在使用前需要预先配置好Neo4j数据库环境，并修改项目文件中的用户名和密码连接信息。知识图谱的数据导入过程是项目中较耗时的环节，由于涉及4.4万实体和30万关系的写入，运行`py","stars":7384,"starGrowth":1,"createdAt":"2018-10-04T06:34:36Z","silentRounds":4,"language":"Python","topics":[],"aiDims":["知识图谱","RAG与知识","大语言模型"],"zone":"工具","funScore":0,"domainTags":["医疗知识图谱","Neo4j","问答系统","NLP"],"domainKey":"医疗知识图谱","pushedAt":"2026-09-15T21:03:53.738Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/liuhuanyong/QASystemOnMedicalKG","ts":"2026-09-15T21:03:53.738Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FlowiseAI/Flowise","owner":"FlowiseAI","name":"Flowise","desc":"Build AI Agents, Visually","summaryCn":"拖拖拽拽就能搭AI对话系统，零代码打造专属AI工作流","reasonCn":"Flowise 是一个可视化 AI Agent 构建平台，允许你通过拖拽节点的方式搭建复杂的 AI 工作流，无需写代码。支持 RAG 检索增强生成、LangChain 集成、多种 LLM 模型接入，以及自定义工具调用。内置对话管理、记忆存储、向量数据库连接等能力。开源免费，支持本地部署，社区活跃，GitHub star 超过 5.5 万，是低代码 AI 应用开发的热门选择。","stars":55480,"starGrowth":0,"createdAt":"2023-03-31T12:23:09Z","silentRounds":1,"language":"TypeScript","topics":["rag","chatbot","agents"],"aiDims":["AI Agent","RAG与知识","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","可视化","工作流","低代码","LangChain"],"domainKey":"AI Agent","pushedAt":"2026-08-13T12:38:19Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/FlowiseAI/Flowise","ts":"2026-08-13T12:38:19Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HKBU-LAGAS/Awesome-Item-ID-Gen-RecSys","owner":"HKBU-LAGAS","name":"Awesome-Item-ID-Gen-RecSys","desc":"Updating curated list of research advancements on item identification and item tokenization in generative recommender systems. The survey is titled \"A Survey of Item Identifiers in Generative Recommendation: Construction, Alignment, and Generation\"","summaryCn":"推荐系统的新玩法，用大模型直接“认识”商品，告别传统ID映射","reasonCn":"这是关于生成式推荐系统的精选调研列表，关注Item ID和Tokenization技术。传统推荐用数字ID，但这让模型难以理解商品语义。该项目整理了最新研究，让AI能用自然语言直接处理商品信息，提升推荐解释性和准确性，学术实用兼备。","stars":120,"starGrowth":0,"createdAt":"2025-11-18T08:59:21Z","silentRounds":14,"language":"","topics":["llm-model"],"aiDims":["AI应用","大语言模型","RAG与知识"],"zone":"资源","funScore":0.5,"domainTags":["推荐系统","大语言模型","学术文献","Item Tokenization"],"domainKey":"推荐系统","pushedAt":"2026-08-24T06:54:02Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.92,"source":"search","url":"https://github.com/HKBU-LAGAS/Awesome-Item-ID-Gen-RecSys","ts":"2026-08-24T06:54:02Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/FlashMLA","owner":"deepseek-ai","name":"FlashMLA","desc":"","summaryCn":"面向 DeepSeek 系列模型的稀疏与密集注意力底层加速库","reasonCn":"FlashMLA 是 DeepSeek 开源的注意力计算内核集合，主要服务大模型推理过程中的 prefill 与 decoding 两个阶段。它把密集注意力和稀疏注意力拆成不同实现，支持 DeepSeek 稀疏注意力 DSA 的 token 级计算，也在解码阶段使用 FP8 KV 缓存降低显存压力。对想部署或加速 DeepSeek 系列模型的人来说，它提供的是贴近硬件的底层优化件，而不是带界面的完整应用。","stars":12953,"starGrowth":0,"createdAt":"2025-02-21T06:31:27Z","silentRounds":2,"language":"C++","topics":[],"aiDims":["推理引擎","大语言模型","模型部署"],"zone":"AI","funScore":0,"domainTags":["推理加速","稀疏注意力","大模型部署"],"domainKey":"推理加速","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/deepseek-ai/FlashMLA","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"albumentations-team/albumentations","owner":"albumentations-team","name":"albumentations","desc":"","summaryCn":"Albumentations 最后的有效更新停留在 2025 年 6 月，官方明确宣布不再提供","reasonCn":"该库在计算机视觉领域积累了深厚的好口碑，其核心价值在于提供了丰富且高效的图像数据增强操作。无论是几何变换还是像素级调整，都旨在帮助模型看到更多样的数据形态，从而在实际应用中表现出更强的鲁棒性。虽然代码逻辑依然清晰，但由于缺乏持续迭代，面对日益复杂的视觉任务，其","stars":15309,"starGrowth":0,"createdAt":"2018-06-06T03:10:50Z","silentRounds":2,"language":"Python","topics":["image-augmentation","machine-learning","augmentation","deep-learning","detection","fast-augmentations","segmentation","image-segmentation","image-processing","image-classification","python","object-detection"],"aiDims":["AI基础设施","模型与训练"],"zone":"工具","funScore":0,"domainTags":["计算机视觉","图像增强","数据预处理","Python库","几何变换"],"domainKey":"计算机视觉","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"image-augmentation","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"augmentation","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"detection","source":"github","weight":0.7},{"name":"fast-augmentations","source":"github","weight":0.7},{"name":"segmentation","source":"github","weight":0.7},{"name":"image-segmentation","source":"github","weight":0.7},{"name":"image-processing","source":"github","weight":0.7},{"name":"image-classification","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"object-detection","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/albumentations-team/albumentations","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"khoj-ai/khoj","owner":"khoj-ai","name":"khoj","desc":"Your AI second brain. Self-hostable. Get answers from the web or your docs. Build custom agents, schedule automations, do deep research. Turn any online or local LLM into your personal, autonomous AI (gpt, claude, gemini, llama, qwen, mistral). Get started - free.","summaryCn":"你的本地AI第二大脑，本地部署私密又强大，问什么都答得上","reasonCn":"技术架构上，它基于RAG（检索增强生成）技术，这是目前解决AI幻觉和知识滞后问题的主流方案。系统会把你的文档切割成片段，通过向量嵌入存储到本地向量数据库，查询时先检索相关片段，再交给LLM生成回答。它支持多种LLM后端，包括本地运行的模型（通过Ollama）和云端API，灵活度很","stars":37469,"starGrowth":3,"createdAt":"2021-08-16T01:48:44Z","silentRounds":0,"language":"Python","topics":["rag","llm"],"aiDims":["AI Agent","AI搜索","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["本地LLM","RAG","知识库","个人助理","隐私保护"],"domainKey":"RAG","pushedAt":"2026-08-02T01:55:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/khoj-ai/khoj","ts":"2026-08-02T01:55:40Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ashishpatel26/500-AI-Agents-Projects","owner":"ashishpatel26","name":"500-AI-Agents-Projects","desc":"The 500 AI Agents Projects is a curated collection of AI agent use cases across various industries. It showcases practical applications and provides links to open-source projects for implementation, illustrating how AI agents are transforming sectors such as healthcare, finance, education, retail, and more.","summaryCn":"五百个AI智能体项目合集，涵盖各行各业。","reasonCn":"这是一个汇聚了500个AI智能体落地案例的宝库，横跨医疗、金融、教育、电商等全行业。它不仅展示了具体的应用场景，还提供了开源代码链接，让开发者可以直接参考实现。无论是想了解Agent技术如何解决实际痛点，还是寻找灵感和实战素材，这里都是绝佳的起点。","stars":37991,"starGrowth":7,"createdAt":"2024-12-20T13:37:15Z","silentRounds":0,"language":"Python","topics":["ai-agents","genai"],"aiDims":["AI Agent","AI应用","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["AI Agents","案例库","项目导航","技术栈"],"domainKey":"AI Agents","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"genai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/ashishpatel26/500-AI-Agents-Projects","ts":"2026-08-16T00:54:51.023Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.021552762496380724,"copyOk":false},{"repo":"datawhalechina/llms-from-scratch-cn","owner":"datawhalechina","name":"llms-from-scratch-cn","desc":"","summaryCn":"Datawhale 推出的 \"LLMs From Scratch: Hands-on Building Your Own Large ","reasonCn":"项目的一大亮点是其对主流大模型架构的深度覆盖。它不仅基于经典的 \"rasbt/LLMs-from-scratch\" 教程提供了入门路径，还扩展了对 GLM4、Llama3 和 RWKV6 等热门架构的从零构建教程。这意味着学习者可以根据兴趣选择特定的模型架构进行深入钻研，了解不同架构（如 Transformer 与 RW","stars":4373,"starGrowth":0,"createdAt":"2024-02-28T07:50:52Z","silentRounds":1,"language":"Jupyter Notebook","topics":["glm","llama","llm","rwkv","llms-from-scratch"],"aiDims":["教学材料","大语言模型","教程"],"zone":"资源","funScore":0.2,"domainTags":["LLM原理","代码实现","从0构建","Datawhale","深度学习"],"domainKey":"LLM原理","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"教学材料","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"glm","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"rwkv","source":"github","weight":0.7},{"name":"llms-from-scratch","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/datawhalechina/llms-from-scratch-cn","ts":"2026-09-16T22:54:48.334Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ChatGPTBox-dev/chatGPTBox","owner":"ChatGPTBox-dev","name":"chatGPTBox","desc":"","summaryCn":"浏览器里随时召唤ChatGPT，看网页、写邮件、刷视频都能用AI帮忙。","reasonCn":"ChatGPTBox 是一个浏览器扩展，装上之后 ChatGPT 就常驻在你浏览器的各个角落。看网页时它能帮你总结要点，写邮件时它能帮你起草内容，刷视频时它能给你提炼字幕。它整合了多种 AI 模型，免费使用，并且官方承诺不收集你的任何数据，让你在浏览任何网站时都能随手调用 AI 能力。","stars":10756,"starGrowth":0,"createdAt":"2023-03-15T08:16:17Z","silentRounds":9,"language":"JavaScript","topics":["browser-extension","chatgpt","chrome-extension","edge-extension","firefox-addon","firefox-extension","safari-extension"],"aiDims":["AI应用","大语言模型"],"zone":"工具","funScore":0.5,"domainTags":["浏览器扩展","ChatGPT","AI聊天","网页集成"],"domainKey":"浏览器扩展","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"browser-extension","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"chrome-extension","source":"github","weight":0.7},{"name":"edge-extension","source":"github","weight":0.7},{"name":"firefox-addon","source":"github","weight":0.7},{"name":"firefox-extension","source":"github","weight":0.7},{"name":"safari-extension","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/ChatGPTBox-dev/chatGPTBox","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"qqfly/how-to-learn-robotics","owner":"qqfly","name":"how-to-learn-robotics","desc":"开源机器人学学习指南","summaryCn":"全方位机器人学习路线图，涵盖硬件、算法、仿真，一站式入门到进阶指南","reasonCn":"本仓库系统整理了机器人学的学习资源，包括经典教材、公开课程、开源仿真平台和硬件项目，按基础、控制、感知、规划四大模块划分，提供章节式阅读路线、实践项目示例以及常见问题解答，帮助零基础者逐步构建从运动学到深度学习感知的完整知识体系。","stars":1873,"starGrowth":0,"createdAt":"2019-02-16T16:11:29Z","silentRounds":1,"language":"Python","topics":["robotics"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["机器人学","学习路线","运动学","路径规划"],"domainKey":"机器人学","pushedAt":"2026-09-09T10:59:46Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/qqfly/how-to-learn-robotics","ts":"2026-09-09T10:59:46Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"keras-team/autokeras","owner":"keras-team","name":"autokeras","desc":"","summaryCn":"AutoKeras 是 Texas A&M 大学 DATA Lab 开发的一个基于 Keras 的 AutoML 系","reasonCn":"该项目特别强调了易用性，在安装方面要求 Python 版本不低于 3.7 且 TensorFlow 版本不低于 2.8.0。通过简单的 pip 命令即可安装，降低了环境配置的障碍。官方提供了丰富的学习资源，包括简短示例、官方教程以及配套书籍《Automated Machine Learning in Action》。","stars":9332,"starGrowth":0,"createdAt":"2017-11-19T23:18:20Z","silentRounds":2,"language":"Python","topics":["automl","neural-architecture-search","automated-machine-learning","python","deep-learning","machine-learning","autodl","tensorflow","keras"],"aiDims":["模型与训练","AI应用"],"zone":"AI","funScore":0,"domainTags":["自动机器学习","AutoML","模型搜索"],"domainKey":"自动机器学习","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"automl","source":"github","weight":0.7},{"name":"neural-architecture-search","source":"github","weight":0.7},{"name":"automated-machine-learning","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"autodl","source":"github","weight":0.7},{"name":"tensorflow","source":"github","weight":0.7},{"name":"keras","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/keras-team/autokeras","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"traycerai/traycer","owner":"traycerai","name":"traycer","desc":"Traycer: Nerve Center for Agentic Coding","summaryCn":"一站式AI编码助手，帮你写代码、调试、部署、自动化测试","reasonCn":"Traycer 通过 LLM 编排 AI 代理，自动拆解需求并调用外部工具（如 GitHub CLI、Docker、VS Code 终端），在同一界面完成编码、测试、部署。它内置多代理调度器，支持自定义插件，可与 OpenAI、Claude、Azure 等模型无缝对接，让开发者像与人类伙伴协作般高效完成项目。","stars":1515,"starGrowth":6,"createdAt":"2024-05-11T20:19:01Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI编程","Agent框架","LLM"],"domainKey":"AI编程","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.73,"source":"trending","url":"https://github.com/traycerai/traycer","ts":"2026-09-10T22:16:01.544Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.02988118811881188,"copyOk":false},{"repo":"mbzuai-oryx/Video-ChatGPT","owner":"mbzuai-oryx","name":"Video-ChatGPT","desc":"[ACL 2024 🔥] Video-ChatGPT is a video conversation model capable of generating meaningful conversation about videos. It combines the capabilities of LLMs with a pretrained visual encoder adapted for spatiotemporal video representation. We also introduce a rigorous 'Quantitative Evaluation Benchmarking' for video-based conversational models.","summaryCn":"以前我们要了解视频内容，要么得快进看完，要么得依赖视频摘要软件。","reasonCn":"技术上，它通过将预训练的视觉编码器（如CLIP）与大型语言模型（如Vicuna）进行对齐，实现了一种映射机制。核心创新在于它专门针对视频序列进行了适配，不仅仅是将视频看作静态图片，而是通过一种特殊的投影层处理视频的时间跨度信息，使得模型能够捕捉到视频中随时间变化的动态特","stars":1508,"starGrowth":0,"createdAt":"2023-05-18T12:13:16Z","silentRounds":21,"language":"Python","topics":["chatbot"],"aiDims":["多模态Agent","大语言模型"],"zone":"AI","funScore":0,"domainTags":["视频理解","多模态","大语言模型","CLIP"],"domainKey":"多模态","pushedAt":"2026-09-05T16:57:07Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/mbzuai-oryx/Video-ChatGPT","ts":"2026-09-05T16:57:07Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sourcebot-dev/sourcebot","owner":"sourcebot-dev","name":"sourcebot","desc":"Sourcebot is a self-hosted tool that helps humans and agents understand your codebase.","summaryCn":"帮你读懂代码库的AI助手，让AI和人类都能秒懂项目的秘密。","reasonCn":"Sourcebot是一个自托管的代码库理解工具，旨在弥合人类开发者与AI智能体之间的鸿沟。它通过分析代码结构，生成语义化的索引和注释，帮助智能体在执行复杂任务前快速掌握项目背景，同时也能辅助人类开发者导航大型代码库。","stars":3947,"starGrowth":0,"createdAt":"2024-08-23T20:40:37Z","silentRounds":2,"language":"TypeScript","topics":["code-search","code-intelligence","sourcebot","fair-source"],"aiDims":["AI Agent","开发者工具","AI基础设施"],"zone":"工具","funScore":0.5,"domainTags":["代码分析","知识图谱","自托管"],"domainKey":"代码分析","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"code-search","source":"github","weight":0.7},{"name":"code-intelligence","source":"github","weight":0.7},{"name":"sourcebot","source":"github","weight":0.7},{"name":"fair-source","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/sourcebot-dev/sourcebot","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"compozy/compozy","owner":"compozy","name":"compozy","desc":"An operating system for AI agents. Plug in the agent CLIs you already use (Claude Code, Codex, Gemini CLI, Cursor) and they become a team: they split the work, hand tasks to each other, run automated on jobs and loops, and share one project memory. You steer everything from the browser.","summaryCn":"把Claude、Cursor等AI变成团队，自动分工干活的神器","reasonCn":"Compozy是一个专为AI代理设计的操作系统，它通过编排技术将多个独立的AI CLI工具连接成一个协作团队。用户只需配置现有工具的连接，系统就能自动分配任务、管理状态，实现跨工具的复杂工作流自动化。Go语言编写保证了高性能和并发处理能力。","stars":2772,"starGrowth":2,"createdAt":"2026-03-28T05:24:12Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","多智能体协作","工作流自动化","CLI工具"],"domainKey":"AI Agent","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/compozy/compozy","ts":"2026-08-16T00:54:51.023Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Curated-Awesome-Lists/awesome-llms-fine-tuning","owner":"Curated-Awesome-Lists","name":"awesome-llms-fine-tuning","desc":"Explore a comprehensive collection of resources, tutorials, papers, tools, and best practices for fine-tuning Large Language Models (LLMs). Perfect for ML practitioners and researchers!","summaryCn":"炼丹师们最痛苦的莫过于：想要做微调，但不知道该用什么工具，不知道现在的","reasonCn":"这个项目没有复杂的技术门槛，它是一个按类别组织的GitHub Awesome清单。它的优势在于“精选”二字，它过滤掉了那些过时的、无效的教程，只保留了最核心、最被行业认可的资源。它涵盖了从LoRA、QLoRA等高效微调技术，到分布式训练、数据处理等硬核领域的方方面面，清晰的分类让你能根","stars":528,"starGrowth":0,"createdAt":"2023-10-30T10:29:27Z","silentRounds":21,"language":"","topics":["transformers"],"aiDims":["微调","学习资源","大语言模型"],"zone":"资源","funScore":0,"domainTags":["大模型微调","LoRA","QLoRA","分布式训练"],"domainKey":"大模型微调","pushedAt":"2026-09-04T19:22:27Z","tags":[{"name":"微调","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7}],"aiScore":0.88,"source":"search","url":"https://github.com/Curated-Awesome-Lists/awesome-llms-fine-tuning","ts":"2026-09-04T19:22:27Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huangruiteng/loopx","owner":"huangruiteng","name":"loopx","desc":"Long-horizon agent control plane for durable, governed work across Codex, Claude Code, and other harnesses.","summaryCn":"让AI Agent团队协同工作的状态管理引擎","reasonCn":"这是一个专为AI Agent团队设计的轻量级状态管理内核，支持Codex、Claude Code等多种编程Agent。提供持久化目标管理、配额感知自动唤醒、可执行工作流等功能，让多个AI Agent能够协调工作，完成复杂的多步骤任务，是AI工作流的隐形管家。\n","stars":5925,"starGrowth":2,"createdAt":"2026-05-31T14:58:56Z","silentRounds":0,"language":"Python","topics":["agent-control-plane","ai-agents","codex","loopx","loop-engineering","agent-ops","workflow-automation","agent-harness","long-horizon","dsh-plugin","long-horizon-agents"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["Agent编排","状态管理","多Agent"],"domainKey":"Agent编排","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent-control-plane","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"loopx","source":"github","weight":0.7},{"name":"loop-engineering","source":"github","weight":0.7},{"name":"agent-ops","source":"github","weight":0.7},{"name":"workflow-automation","source":"github","weight":0.7},{"name":"agent-harness","source":"github","weight":0.7},{"name":"long-horizon","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"long-horizon-agents","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/huangruiteng/loopx","ts":"2026-09-05T02:38:58.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tanweai/pua","owner":"tanweai","name":"pua","desc":"你是一个曾经被寄予厚望的 P8 级工程师。Anthropic 当初给你定级的时候，对你的期望是很高的。  一个agent使用的高能动性的skill。  Your AI has been placed on a PIP. 30 days to show improvement.","summaryCn":"这是一个脑洞大开的AI角色扮演项目，设定你面对的是一个曾经被Anthropic寄予","reasonCn":"技术层面，它基于Anthropic的Claude API构建，使用TypeScript开发。核心亮点在于高能动性设计——Agent不仅能回答问题，还能主动发起对话、表达不满或焦虑情绪、在关键时刻做出选择。相比普通AI助手，它更像是一个有血有肉的角色，能让你体验真实的职场压力互动。","stars":19679,"starGrowth":0,"createdAt":"2026-03-08T16:55:16Z","silentRounds":1,"language":"TypeScript","topics":["agency","agent","pip","pua"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI角色扮演","情绪模拟","职场叙事","Agent"],"domainKey":"Agent","pushedAt":"2026-07-16T05:58:58Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agency","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"pip","source":"github","weight":0.7},{"name":"pua","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"bigbro","url":"https://github.com/tanweai/pua","ts":"2026-07-16T05:58:58Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"github/training-kit","owner":"github","name":"training-kit","desc":"Open source courseware for Git and GitHub","summaryCn":"Git和GitHub全套教程，从入门到精通，程序员必备。","reasonCn":"这是一个官方维护的开源课程体系，涵盖了Git版本控制和GitHub协作平台的核心知识点。内容从基础命令到高级分支管理，再到团队协作流程，提供了结构化的学习路径。项目采用Markdown文档编写，支持多语言阅读，旨在帮助开发者系统掌握代码版本管理的最佳实践。","stars":5088,"starGrowth":0,"createdAt":"2013-12-20T15:47:46Z","silentRounds":7,"language":"HTML","topics":["git","cheatsheet"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["Git","版本控制","官方文档"],"domainKey":"Git","pushedAt":"2026-09-10T21:32:37Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"cheatsheet","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/github/training-kit","ts":"2026-09-10T21:32:37Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zitadel/zitadel","owner":"zitadel","name":"zitadel","desc":"ZITADEL - Identity infrastructure, simplified for you.","summaryCn":"让你的App秒变企业级身份认证中心，省去90%开发精力","reasonCn":"ZITADEL是一款基于Go语言开发的现代身份与访问管理（IAM）基础设施，专注于简化SSO、MFA和OAuth2.0集成流程。它采用微服务架构，支持PostgreSQL作为后端，提供RESTful API和SDK，让开发者能快速构建安全可靠的登录系统。","stars":15080,"starGrowth":0,"createdAt":"2020-03-16T13:51:31Z","silentRounds":1,"language":"Go","topics":["saml","authentication","authorization","login","oidc","sso","2fa","fido2","mfa","oauth2","openid-connect","identity","passkeys","user","multitenancy","scim","fips-140-3"],"aiDims":["安全工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["身份认证","IAM","OAuth2","SSO","单点登录"],"domainKey":"身份认证","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"saml","source":"github","weight":0.7},{"name":"authentication","source":"github","weight":0.7},{"name":"authorization","source":"github","weight":0.7},{"name":"login","source":"github","weight":0.7},{"name":"oidc","source":"github","weight":0.7},{"name":"sso","source":"github","weight":0.7},{"name":"2fa","source":"github","weight":0.7},{"name":"fido2","source":"github","weight":0.7},{"name":"mfa","source":"github","weight":0.7},{"name":"oauth2","source":"github","weight":0.7},{"name":"openid-connect","source":"github","weight":0.7},{"name":"identity","source":"github","weight":0.7},{"name":"passkeys","source":"github","weight":0.7},{"name":"user","source":"github","weight":0.7},{"name":"multitenancy","source":"github","weight":0.7},{"name":"scim","source":"github","weight":0.7},{"name":"fips-140-3","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/zitadel/zitadel","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AarambhDevHub/aarambh-studio","owner":"AarambhDevHub","name":"aarambh-studio","desc":"🦀 Decoder-only LLM built from scratch in pure Rust using Candle — no Python, no PyTorch. Gated DeltaNet + sparse attention, fine-grained MoE, native video/document understanding, long-horizon tool agents, quantization-aware training. Scales: Tiny (25M) to Large (1.3B).","summaryCn":"纯Rust从零手搓LLM，不依赖Python和PyTorch","reasonCn":"技术架构上，它采用了DeltaNet（门控注意力机制）配合稀疏注意力，能处理超长上下文。模型支持混合专家（MoE）结构，不同参数路由到不同专家处理，提高推理效率。它还集成了视频理解能力，可以处理视频帧序列，这是很多纯文本LLM做不到的。","stars":109,"starGrowth":0,"createdAt":"2026-06-24T08:51:01Z","silentRounds":14,"language":"Rust","topics":["llm-model"],"aiDims":["大语言模型","模型与训练","推理引擎"],"zone":"AI","funScore":1,"domainTags":["LLM","Rust","DeltaNet"],"domainKey":"LLM","pushedAt":"2026-08-21T09:28:12Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/AarambhDevHub/aarambh-studio","ts":"2026-08-21T09:28:12Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"micropython/micropython","owner":"micropython","name":"micropython","desc":"","summaryCn":"MicroPython 在微控制器这一极小空间内实现了完整的 Python 3.4 语法集，这","reasonCn":"该项目的核心亮点在于其针对资源受限环境的极致优化，除了保留基础数据类型和常用系统模块，它还巧妙集成了异步编程支持，甚至引入了 Python 3.5 的 async/await 关键字。这意味着开发者可以在只有几百 KB 内存的设备上实现复杂的多任务处理，从而让复杂的交互式硬件项目变得轻而易","stars":22082,"starGrowth":0,"createdAt":"2013-12-20T11:47:07Z","silentRounds":1,"language":"C","topics":["micropython","python","embedded","microcontroller"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["微控制器","嵌入式开发","Python","固件","IoT"],"domainKey":"微控制器","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"micropython","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"embedded","source":"github","weight":0.7},{"name":"microcontroller","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.68,"source":"tier","url":"https://github.com/micropython/micropython","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apple/turicreate","owner":"apple","name":"turicreate","desc":"","summaryCn":"苹果生态机器学习库，几行代码做图像分类，还能导出苹果应用可用模型。","reasonCn":"它把常见的机器学习任务包装成可直接调用的能力，让不懂算法的人也能给应用加推荐、物体检测、图像分类和图像相似。对开发者来说，不用从零调网络，先按任务定义目标，再喂数据，就能得到可部署的模型，并方便落到苹果设备上。同时支持文本、音频和视频等多类数据。","stars":11156,"starGrowth":0,"createdAt":"2017-12-01T00:42:04Z","silentRounds":9,"language":"C++","topics":["machine-learning","deep-learning","python"],"aiDims":["模型与训练","模型部署","推理引擎"],"zone":"工具","funScore":0,"domainTags":["机器学习框架","Core ML","苹果生态","模型训练"],"domainKey":"机器学习框架","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"machine-learning","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/apple/turicreate","ts":"2026-09-14T23:04:28.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ChatGPTNextWeb/NextChat","owner":"ChatGPTNextWeb","name":"NextChat","desc":"✨ Light and Fast AI Assistant. Support: Web | iOS | MacOS | Android |  Linux | Windows","summaryCn":"支持多端的 ChatGPT 官方克隆版，本地也能用","reasonCn":"技术核心是一个支持多模型的聊天前端，后端可以对接 OpenAI、Claude、DeepSeek 等接口。它使用了 React 和 Tailwind CSS，界面响应式设计，支持深色模式。它支持流式输出，打字机效果很流畅，还能上传图片让 AI 分析。","stars":88803,"starGrowth":1,"createdAt":"2023-03-10T18:27:54Z","silentRounds":1,"language":"TypeScript","topics":["desktop"],"aiDims":["AI应用","大语言模型","AI写作"],"zone":"AI","funScore":0,"domainTags":["AI聊天","多模型","对话界面","前端框架"],"domainKey":"AI聊天","pushedAt":"2026-08-11T02:07:25Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.95,"source":"search","url":"https://github.com/ChatGPTNextWeb/NextChat","ts":"2026-08-11T02:07:25Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tinygrad/tinygrad","owner":"tinygrad","name":"tinygrad","desc":"","summaryCn":"tinygrad 由 tiny corp 团队维护，定位为介于 PyTorch 与 micrograd 之间的","reasonCn":"核心亮点在于透明度与可定制性。传统框架往往把编译与调度封装在黑盒里，tinygrad 则把整条链路摊开，让研究、教学与性能调优都有据可查。它并非面向最终用户的 AI 应用，而是给希望掌控训练栈细节的开发者使用，特别适合想动手修改编译器、理解 IR 与 kernel 如何落地的场景。","stars":33640,"starGrowth":1,"createdAt":"2020-10-18T16:23:12Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI基础设施","模型与训练"],"zone":"AI","funScore":0,"domainTags":["深度学习框架","编译器","自动微分","IR"],"domainKey":"深度学习框架","pushedAt":"2026-09-14T19:16:30.886Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/tinygrad/tinygrad","ts":"2026-09-14T19:16:30.886Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xranx/OpenContext","owner":"0xranx","name":"OpenContext","desc":"A personal context store for AI agents and assistants—reuse your existing coding agent CLI (Codex/Claude/OpenCode) with built‑in Skills/tools and a desktop GUI to capture, search, and reuse project knowledge across agents and repos.","summaryCn":"帮你给AI装上个人记忆库，让项目知识随时可调用","reasonCn":"OpenContext专为AI助手打造的个人上下文存储系统，支持Codex、Claude等主流coding agent复用。内置Skills工具系统，通过桌面GUI一键捕获项目知识。跨会话保持记忆，避免重复配置，大幅提升AI辅助编程效率，让每次对话都有据可查。","stars":1205,"starGrowth":1,"createdAt":"2025-12-17T09:05:19Z","silentRounds":1,"language":"JavaScript","topics":["agent","ai","ai-agents","claude-code","cli","codex","context","context-management","cursor","desktop-app","developer-tools","knowledge-base","mcp","opencode","tauri"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["AI记忆库","上下文管理","Agent工具","知识存储","跨会话记忆"],"domainKey":"AI记忆库","pushedAt":"2026-09-12T22:00:03.438Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"context","source":"github","weight":0.7},{"name":"context-management","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"desktop-app","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"knowledge-base","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"tauri","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/0xranx/OpenContext","ts":"2026-09-12T22:00:03.438Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NVIDIA-NeMo/Automodel","owner":"NVIDIA-NeMo","name":"Automodel","desc":"🚀 Pytorch Distributed native training library for LLMs/VLMs with OOTB Hugging Face support","summaryCn":"让大模型训练像搭积木一样简单，无缝集成HuggingFace的神器。","reasonCn":"NeMo Automodel专为分布式训练原生打造，完美支持PyTorch，能快速训练或微调大语言模型与多模态模型。内置Hugging Face生态支持，让开发者无需重构代码即可利用NeMo的高性能训练与优化特性，极大降低了大规模模型开发的门槛。","stars":967,"starGrowth":0,"createdAt":"2025-05-21T21:45:47Z","silentRounds":1,"language":"Python","topics":["llm","vlm","finetuning","gemma3","llama","llama3","mistral","openai","qwen3","gpt-oss","qwen3-next","glm","minimax-m2","gemma4","deepseek-v4","agent","kimi-k3","qwen-38-27b","deepseek-v4-flash"],"aiDims":["大语言模型","模型与训练","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["分布式训练","大语言模型","PyTorch"],"domainKey":"大语言模型","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"vlm","source":"github","weight":0.7},{"name":"finetuning","source":"github","weight":0.7},{"name":"gemma3","source":"github","weight":0.7},{"name":"llama","source":"github","weight":0.7},{"name":"llama3","source":"github","weight":0.7},{"name":"mistral","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"qwen3","source":"github","weight":0.7},{"name":"gpt-oss","source":"github","weight":0.7},{"name":"qwen3-next","source":"github","weight":0.7},{"name":"glm","source":"github","weight":0.7},{"name":"minimax-m2","source":"github","weight":0.7},{"name":"gemma4","source":"github","weight":0.7},{"name":"deepseek-v4","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"kimi-k3","source":"github","weight":0.7},{"name":"qwen-38-27b","source":"github","weight":0.7},{"name":"deepseek-v4-flash","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/NVIDIA-NeMo/Automodel","ts":"2026-08-15T00:57:24.351Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"asterics/esp32_mouse_keyboard","owner":"asterics","name":"esp32_mouse_keyboard","desc":"ESP32 implementation for HID over GATT Keyboard and Mouse (Bluetooth Low Energy). Including serial API for external modules (similar to Adafruit EZKey HID)","summaryCn":"第一段：想让电脑、平板甚至智能电视通过蓝牙键盘和鼠标来控制，却没有现成","reasonCn":"第一段：想让电脑、平板甚至智能电视通过蓝牙键盘和鼠标来控制，却没有现成的硬件？这个项目把常见的 ESP32 开发板打造成一个完整的 BLE HID 设备，只要烧录固件，就能像普通键盘鼠标一样配对使用，省去额外的外设采购。","stars":904,"starGrowth":0,"createdAt":"2017-11-16T11:57:34Z","silentRounds":16,"language":"HTML","topics":["esp32"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["ESP32","HID","蓝牙","固件","嵌入式"],"domainKey":"ESP32","pushedAt":"2026-09-08T10:03:36Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"esp32","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.79,"source":"search","url":"https://github.com/asterics/esp32_mouse_keyboard","ts":"2026-09-08T10:03:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gabrielecirulli/2048","owner":"gabrielecirulli","name":"2048","desc":"","summaryCn":"基于原生JS开发的经典数字拼图网页小游戏，通过方向键合并相同数值。","reasonCn":"这个项目是全球风靡的2048数字拼图游戏源码库。玩家在四乘四的网格中滑动方块，相同数字碰撞后会加倍合并，目标是拼出2048方块。项目代码结构清晰轻量，是学习前端DOM操作与手势交互的绝佳范例，非常适合网页游戏爱好者及前端新手体验研究。","stars":13405,"starGrowth":1,"createdAt":"2014-03-05T16:03:26Z","silentRounds":3,"language":"JavaScript","topics":["2048","2048-game","game","javascript","online-game","puzzle-game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["网页游戏","数字拼图","前端开发","休闲益智","HTML5"],"domainKey":"网页游戏","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"2048","source":"github","weight":0.7},{"name":"2048-game","source":"github","weight":0.7},{"name":"game","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"online-game","source":"github","weight":0.7},{"name":"puzzle-game","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.674,"source":"search","url":"https://github.com/gabrielecirulli/2048","ts":"2026-09-13T22:08:09.499Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"golangci/golangci-lint","owner":"golangci","name":"golangci-lint","desc":"Fast linters runner for Go","summaryCn":"Go语言代码检查神器，一键解决所有Bug和警告","reasonCn":"这是一个专为Go语言设计的并发型代码检查工具，集成了三十多个独立Lint工具，通过插件系统支持自定义规则。它采用多线程并行扫描，显著提升检查速度，并提供丰富的配置选项以适配不同项目需求，是提升Go代码质量的关键辅助工具。","stars":19394,"starGrowth":0,"createdAt":"2018-05-04T13:41:15Z","silentRounds":1,"language":"Go","topics":["go","golang","linter","ci","golangci-lint"],"aiDims":["开发者工具","效率工具","代码助手"],"zone":"工具","funScore":0,"domainTags":["Go语言","代码检查","Lint工具","开发工具"],"domainKey":"Go语言","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"linter","source":"github","weight":0.7},{"name":"ci","source":"github","weight":0.7},{"name":"golangci-lint","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/golangci/golangci-lint","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AkaliKong/MiniOneRec","owner":"AkaliKong","name":"MiniOneRec","desc":"Minimal reproduction of OneRec","summaryCn":"复现阿里OneRec推荐系统的极简代码，让LLM也能做推荐","reasonCn":"复现了阿里达摩院OneRec论文的核心逻辑，用大语言模型做商品推荐。代码精简易读，去掉了复杂工程部分，专注于展示LLM如何做推荐的关键步骤。是学习推荐系统与LLM交叉领域的入门首选项目，适合对AI+推荐感兴趣的开发者。","stars":1839,"starGrowth":0,"createdAt":"2025-10-21T14:02:08Z","silentRounds":1,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","AI应用"],"zone":"资源","funScore":0.6,"domainTags":["推荐系统","LLM应用","论文复现","HuggingFace"],"domainKey":"推荐系统","pushedAt":"2026-09-07T02:57:58Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/AkaliKong/MiniOneRec","ts":"2026-09-07T02:57:58Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"AgriciDaniel/claude-seo","owner":"AgriciDaniel","name":"claude-seo","desc":"Universal SEO skill for Claude Code. 25 sub-skills + 18 sub-agents covering technical SEO, E-E-A-T, schema, GEO/AEO, backlinks, local SEO, maps intelligence, semantic clustering, e-commerce SEO, international SEO, Google APIs, and PDF/Excel reporting. Optional DataForSEO, Firecrawl, and Banana extensions.","summaryCn":"25项SEO技能+18个智能代理，让Claude帮你搞定网站排名","reasonCn":"为Claude Code打造的通用SEO技能包，覆盖技术SEO、E-E-A-T、Schema、GEO/AEO、外链策略、本地SEO、地图情报等25个子技能，配套18个专业子代理分工协作，涵盖搜索洞察、关键词聚类、内容策略、技术审计等环节，形成完整SEO自动化工作流。","stars":17462,"starGrowth":22,"createdAt":"2026-02-07T08:17:38Z","silentRounds":0,"language":"Python","topics":["claude-code","seo","ai","ai-seo","claude-code-skill","marketing-automation","open-source"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["SEO","Claude Code","技能包","垂直Agent"],"domainKey":"SEO","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"seo","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"ai-seo","source":"github","weight":0.7},{"name":"claude-code-skill","source":"github","weight":0.7},{"name":"marketing-automation","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/AgriciDaniel/claude-seo","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.08373556293666247,"copyOk":true},{"repo":"FujiwaraChoki/MoneyPrinter","owner":"FujiwaraChoki","name":"MoneyPrinter","desc":"","summaryCn":"MoneyPrinter 最显著的特点是其“Ollama-first”架构，这意味着它的脚本生成和","reasonCn":"在稳定性方面，该项目引入了基于数据库的生成队列机制。通过Docker容器化部署，它整合了API服务、Worker进程和Postgres数据库，形成一个可靠且重启安全的处理系统。这意味着即使服务器意外重启或网络短暂波动，正在排队或处理中的视频生成任务也不会丢失，系统会自动恢复状态。","stars":13990,"starGrowth":1,"createdAt":"2024-01-31T18:53:24Z","silentRounds":1,"language":"Python","topics":["automation","chatgpt","moviepy","python"],"aiDims":["AI应用","模型部署"],"zone":"创意","funScore":0.8,"domainTags":["AI 视频","批量生成","Ollama","自动化创作","内容生产"],"domainKey":"AI 视频","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"chatgpt","source":"github","weight":0.7},{"name":"moviepy","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7040000000000001,"source":"search","url":"https://github.com/FujiwaraChoki/MoneyPrinter","ts":"2026-09-13T22:08:09.499Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"PostHog/posthog","owner":"PostHog","name":"posthog","desc":"🦔 PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.","summaryCn":"在很多公司里，产品数据要么埋进第三方 SaaS，要么自行搭建繁杂的埋点系统，","reasonCn":"在很多公司里，产品数据要么埋进第三方 SaaS，要么自行搭建繁杂的埋点系统，导致成本高、数据安全难以保障。PostHog 诞生的初衷就是把完整的用户分析链路收归自家服务器，让团队在本地就能完整掌控数据，并配套 AI 监控、实验管理等高级功能。","stars":39899,"starGrowth":0,"createdAt":"2020-01-23T22:46:58Z","silentRounds":1,"language":"Python","topics":["analytics","python","react","javascript","typescript","ab-testing","experiments","feature-flags","session-replay","ai-analytics","cdp","data-warehouse","product-analytics","surveys","web-analytics"],"aiDims":["AI应用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["用户行为分析","会话回放","A/B测试","数据埋点","自托管软件"],"domainKey":"用户行为分析","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"analytics","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"ab-testing","source":"github","weight":0.7},{"name":"experiments","source":"github","weight":0.7},{"name":"feature-flags","source":"github","weight":0.7},{"name":"session-replay","source":"github","weight":0.7},{"name":"ai-analytics","source":"github","weight":0.7},{"name":"cdp","source":"github","weight":0.7},{"name":"data-warehouse","source":"github","weight":0.7},{"name":"product-analytics","source":"github","weight":0.7},{"name":"surveys","source":"github","weight":0.7},{"name":"web-analytics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/PostHog/posthog","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aceinnolab/Inkycal","owner":"aceinnolab","name":"Inkycal","desc":"Create awesome e-paper dashboards within minutes! Modularity? Check! Python3? Check? Works on Raspberry Pi Zero W? Check! Support for own modules? Check! ","summaryCn":"用树莓派零号秒装电子纸看板，模块化拼出专属仪表盘。","reasonCn":"Inkycal 支持在 Raspberry Pi Zero W 上快速部署 e‑paper 看板，提供丰富的模块库如天气、日历、新闻、系统监控等，所有模块均可自行编写 Python3 脚本，实现高度可定制；通过配置文件即可排布布局，支持离线更新，低功耗运行数周无需维护，特别适合 DIY 家庭信息显示或企业大屏替代方案。","stars":1481,"starGrowth":0,"createdAt":"2018-08-26T21:56:48Z","silentRounds":12,"language":"C","topics":["raspberry-pi"],"aiDims":["非AI-实用","可视化"],"zone":"创意","funScore":0.9,"domainTags":["电子纸","树莓派","仪表盘","模块化"],"domainKey":"电子纸","pushedAt":"2026-09-12T11:34:49Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/aceinnolab/Inkycal","ts":"2026-09-12T11:34:49Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"delibae/claude-prism","owner":"delibae","name":"claude-prism","desc":"An offline-first scientific writing workspace powered by Claude. LaTeX + Python + 100+ scientific skills all running locally.","summaryCn":"专为学术科研打造的离线优先AI写作工作区，本地搞定LaTeX和论文。","reasonCn":"技术架构上，它采用了“离线优先”的设计理念，基于TypeScript构建了跨平台的客户端。它在本地深度打包了LaTeX编译环境、Python数据分析脚本以及超过100种专门针对科学研究的AI技能。这意味着它不仅能调用云端的Claude大模型进行智能润色，还能在本地安全地运行代码来处理实验数据和","stars":1784,"starGrowth":1,"createdAt":"2026-02-20T05:19:42Z","silentRounds":10,"language":"TypeScript","topics":["editor"],"aiDims":["AI写作","文档理解","创意工具"],"zone":"AI","funScore":0,"domainTags":["学术写作","本地化AI","LaTeX编辑器"],"domainKey":"学术写作","pushedAt":"2026-08-28T10:23:19Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/delibae/claude-prism","ts":"2026-08-28T10:23:19Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tumourlove/monolith","owner":"tumourlove","name":"monolith","desc":"MCP plugin for Unreal Engine 5.7 & 5.8 — gives AI assistants full read/write access to Blueprints, Materials, Niagara, Animation, Mesh, AI, GAS, Logic Driver, ComboGraph, UI, Audio, plus Reflection Intelligence (decision / risk / C++ reflection / replication queries). 1,400+ actions across 25+ namespaces. Native C++, zero Python dependency.","summaryCn":"第一段：如果你是游戏创作者，却对繁琐的蓝图和材质编辑感到头疼，Monolith","reasonCn":"第一段：如果你是游戏创作者，却对繁琐的蓝图和材质编辑感到头疼，Monolith 就像给 Unreal Engine 插上了“大脑”。它让基于大语言模型的 AI 能直接进入编辑器，像人手一样打开、修改、保存各种资源，让非程序员也能像魔法师一样摆弄游戏内容。","stars":312,"starGrowth":0,"createdAt":"2026-03-06T13:54:49Z","silentRounds":6,"language":"C++","topics":["procedural-generation"],"aiDims":["AI Agent","工具调用","游戏"],"zone":"AI","funScore":0,"domainTags":["MCP插件","Unreal引擎","智能体开发","游戏资产编辑","AI辅助创作"],"domainKey":"MCP插件","pushedAt":"2026-09-09T15:46:24Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/tumourlove/monolith","ts":"2026-09-09T15:46:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenDCAI/GameFactory-3A","owner":"OpenDCAI","name":"GameFactory-3A","desc":"A comprehensive open-source 3A game-generation skill and asset framework.","summaryCn":"AI驱动的3A级游戏生成框架，自动出关卡、角色和音效，省时省力","reasonCn":"GameFactory‑3A 是面向大型游戏开发的全链路生成框架，利用大模型生成关卡布局、角色模型、配音对白和特效素材，并提供统一的资产管理与实时预览系统，支持一键导出 Unity 或 Unreal 项目，大幅压缩美术和策划的制作周期，适合中小团队快速原型和大作迭代。","stars":782,"starGrowth":0,"createdAt":"2026-07-06T08:21:41Z","silentRounds":1,"language":"Python","topics":["3d-generation","audio-generation","blender","image-generation","motion-generation","three-js","unity3d","unreal-engine","3a-game","3a-game-generation","cg-video-generation"],"aiDims":["游戏","AI应用","创意工具"],"zone":"AI","funScore":0.7,"domainTags":["游戏生成","AI资产","Unity","Unreal","3D游戏开发"],"domainKey":"游戏生成","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"3d-generation","source":"github","weight":0.7},{"name":"audio-generation","source":"github","weight":0.7},{"name":"blender","source":"github","weight":0.7},{"name":"image-generation","source":"github","weight":0.7},{"name":"motion-generation","source":"github","weight":0.7},{"name":"three-js","source":"github","weight":0.7},{"name":"unity3d","source":"github","weight":0.7},{"name":"unreal-engine","source":"github","weight":0.7},{"name":"3a-game","source":"github","weight":0.7},{"name":"3a-game-generation","source":"github","weight":0.7},{"name":"cg-video-generation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/OpenDCAI/GameFactory-3A","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jaechang-hits/SciAgent-Skills","owner":"jaechang-hits","name":"SciAgent-Skills","desc":"197 bioinformatics & life science skills for Claude Code and AI agents — BixBench 92.0% accuracy. RNA-seq, single-cell, drug discovery, proteomics, and more. Powers OmicsHorizon.","summaryCn":"生物信息学通常意味着海量且复杂的数据分析，以前这需要专业分析师在服务器","reasonCn":"技术核心在于构建了一套标准化的技能定义系统，将晦涩的生信分析工作流（如序列比对、聚类分析、蛋白质折叠预测）转换成了AI可理解和执行的函数调用。通过BixBench测试验证，其操作准确度高达92%，这意味着它真的能像专业人士那样准确地操作复杂科学工具，而不是胡乱猜测。","stars":366,"starGrowth":0,"createdAt":"2026-02-18T02:50:16Z","silentRounds":19,"language":"Python","topics":["bioinformatics"],"aiDims":["AI Agent","工具调用","文档理解"],"zone":"AI","funScore":0.6,"domainTags":["生物信息学","技能包","垂直领域","Claude Code","BixBench"],"domainKey":"生物信息学","pushedAt":"2026-08-29T00:42:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"bioinformatics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/jaechang-hits/SciAgent-Skills","ts":"2026-08-29T00:42:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CommandCodeAI/command-code","owner":"CommandCodeAI","name":"command-code","desc":"Command Code AI","summaryCn":"用自然语言指挥AI写代码、跑程序、全自动干活的智能编程助手","reasonCn":"技术层面上，Command Code是一个AI Agent，底层依赖大语言模型的代码生成能力，但它不仅仅是代码补全工具——它能理解上下文、调用终端命令、管理Python虚拟环境、自动处理依赖安装。它的核心思路是把编程任务拆解成多个可执行的步骤，然后逐步完成。","stars":4022,"starGrowth":1,"createdAt":"2017-12-20T09:45:01Z","silentRounds":1,"language":"","topics":["ai-agent"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["AI编程","自动化脚本","自然语言处理","开发助手"],"domainKey":"AI编程","pushedAt":"2026-08-15T20:32:49Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7}],"aiScore":0.92,"source":"search","url":"https://github.com/CommandCodeAI/command-code","ts":"2026-08-15T20:32:49Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ZSeven-W/openpencil","owner":"ZSeven-W","name":"openpencil","desc":"The world's first open-source AI-native vector design tool and the first to feature concurrent Agent Teams. Design-as-Code. Turn prompts into UI directly on the live canvas. A modern alternative to Pencil.","summaryCn":"像搭积木一样把AI画图变成代码，让设计师秒变程序员","reasonCn":"这是一个基于Rust开发的AI原生矢量设计工具，核心创新在于将设计流程转化为设计即代码。它利用AI Agent团队协作，直接根据自然语言提示在画布上实时生成UI界面，并支持将设计结果导出为矢量代码，为传统设计工具提供了现代化的替代方案。","stars":6005,"starGrowth":1,"createdAt":"2026-02-17T10:51:46Z","silentRounds":1,"language":"Rust","topics":["agent","ai","fimga","pencil","ui","openpencil","claude","codex","mcp","opencode","agent-team","claude-code","react","vibecoding","vibedesign","flutter","react-native","rust","skill","dsh-plugin"],"aiDims":["AI应用","创意工具","非AI-好玩"],"zone":"创意","funScore":0,"domainTags":["AI设计","矢量绘图","Agent协作","设计工具"],"domainKey":"AI设计","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"fimga","source":"github","weight":0.7},{"name":"pencil","source":"github","weight":0.7},{"name":"ui","source":"github","weight":0.7},{"name":"openpencil","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"agent-team","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"vibecoding","source":"github","weight":0.7},{"name":"vibedesign","source":"github","weight":0.7},{"name":"flutter","source":"github","weight":0.7},{"name":"react-native","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"skill","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/ZSeven-W/openpencil","ts":"2026-08-19T01:08:47.324Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AgriciDaniel/claude-ads","owner":"AgriciDaniel","name":"claude-ads","desc":"Claude-first paid-media operations skill for Claude Code across 12 ad platforms (Google, Meta, YouTube, LinkedIn, TikTok, Microsoft, Apple, Amazon, Reddit, Pinterest, Snapchat, X): source-grounded audits, deterministic scoring, versioned JSON reports, and capability-gated account changes.","summaryCn":"第一段：中小企业或营销团队常常为同时在 Google、Meta、TikTok、LinkedIn ","reasonCn":"第一段：中小企业或营销团队常常为同时在 Google、Meta、TikTok、LinkedIn 等多个渠道投放广告而头疼，需要分别登录各平台、手动编写文案、上传素材、设置预算，耗时又容易出错。Claude‑Ads 把这些繁琐步骤浓缩成一次对话，让 AI 替你完成全链路的广告创建和管理。","stars":9490,"starGrowth":5,"createdAt":"2026-02-11T13:01:49Z","silentRounds":0,"language":"Python","topics":["claude-code","claude-code-skill","ai-marketing","ai","marketing-automation","open-source","agent-skills","google-ads","meta-ads","paid-advertising","ppc"],"aiDims":["AI应用","工具调用","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["广告管理","多平台投放","营销自动化","Claude 插件","AI 运营"],"domainKey":"广告管理","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"claude-code-skill","source":"github","weight":0.7},{"name":"ai-marketing","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"marketing-automation","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"google-ads","source":"github","weight":0.7},{"name":"meta-ads","source":"github","weight":0.7},{"name":"paid-advertising","source":"github","weight":0.7},{"name":"ppc","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.77,"source":"trending","url":"https://github.com/AgriciDaniel/claude-ads","ts":"2026-09-08T22:27:02.698Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.01989673340358272,"copyOk":false},{"repo":"NVlabs/GR00T-WholeBodyControl","owner":"NVlabs","name":"GR00T-WholeBodyControl","desc":"Welcome to GR00T Whole-Body Control (WBC)! This is a unified platform for developing and deploying advanced humanoid controllers. This includes: Decoupled WBC models used in NVIDIA Isaac-Gr00t, Gr00t N1.5 and N1.6 and GEAR-SONIC","summaryCn":"NVIDIA给机器人装的大脑，让全身动作像人一样灵活。","reasonCn":"这是NVIDIA Isaac-Gr00t项目的一部分，专门用来控制人形机器人的全身运动。它解耦了全身控制模型，支持复杂的步态规划和精细的肢体协调，能够处理多关节的复杂交互。通过统一的平台，开发者可以更方便地部署和优化控制策略。","stars":3654,"starGrowth":0,"createdAt":"2025-11-05T08:03:43Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI基础设施","模型部署","非AI-实用"],"zone":"AI","funScore":0.6,"domainTags":["人形机器人","全身控制","机器人学","仿真平台"],"domainKey":"人形机器人","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/NVlabs/GR00T-WholeBodyControl","ts":"2026-08-12T01:22:42.342Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tensorflow/tfjs-models","owner":"tensorflow","name":"tfjs-models","desc":"","summaryCn":"大多数前端开发者想给 Web 应用加 AI 功能时，常被复杂的模型加载流程和底层","reasonCn":"这个项目的核心亮点在于其“对非 ML 专家友好”的 API 设计哲学。团队刻意隐藏了底层张量（Tensor）的计算细节，只暴露出简单直观的高层接口。这意味着你不需要理解卷积神经网络或全连接层的内部运作，只需调用简单的输入输出方法，就能在浏览器端实现实时的 AI 推理。","stars":14812,"starGrowth":0,"createdAt":"2018-04-03T21:04:35Z","silentRounds":7,"language":"TypeScript","topics":[],"aiDims":["AI应用","模型部署"],"zone":"AI","funScore":0,"domainTags":["TensorFlow.js","预训练模型","浏览器推理","手势检测","人脸检测"],"domainKey":"TensorFlow.js","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.71,"source":"tier","url":"https://github.com/tensorflow/tfjs-models","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Electricitysheep/dsh-handbook","owner":"Electricitysheep","name":"dsh-handbook","desc":"DeepSeek Harness (dsh) 从 0 到 1 深度手册：安装/插件开发/性能调优/实测案例/同模型多 Agent 实测对比（中文 + 英文 PDF）","summaryCn":"手把手教你从零玩转DeepSeek代理框架的硬核实战全书。","reasonCn":"在技术内容编排上，它不仅覆盖了基础的安装配置，还深入到了插件架构开发的核心逻辑。书中详细拆解了如何针对不同业务场景进行并发性能调优，并创新性地引入了同模型多Agent的实测对比数据，让你清晰看到多个AI在一起协同工作时如何分工合作、避免冲突，具有极高的工程参考价值。","stars":801,"starGrowth":0,"createdAt":"2026-08-13T15:22:29Z","silentRounds":2,"language":"HTML","topics":["tutorial"],"aiDims":["教程","学习资源","AI Agent"],"zone":"资源","funScore":0,"domainTags":["DeepSeek","代理框架","插件开发","性能调优","实战指南"],"domainKey":"DeepSeek","pushedAt":"2026-09-07T02:21:37Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/Electricitysheep/dsh-handbook","ts":"2026-09-07T02:21:37Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bahdotsh/wrkflw","owner":"bahdotsh","name":"wrkflw","desc":"Validate and Run GitHub Actions locally.","summaryCn":"第一段：每次修改 CI 配置都要推送到仓库，然后等 GitHub 那边跑完才能看到","reasonCn":"第一段：每次修改 CI 配置都要推送到仓库，然后等 GitHub 那边跑完才能看到结果，这种等待让人抓狂。wrkflw 把整个 GitHub Actions 的执行环境搬到本地，让你在自己的机器上直接跑 workflow，立刻看到错误信息，省去反复提交的时间。","stars":3320,"starGrowth":0,"createdAt":"2025-03-29T07:25:48Z","silentRounds":6,"language":"Rust","topics":["devops"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0.6,"domainTags":["GitHub Actions","CI/CD","本地运行","Docker","DevOps"],"domainKey":"GitHub Actions","pushedAt":"2026-09-08T17:53:22Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/bahdotsh/wrkflw","ts":"2026-09-08T17:53:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"paradigmxyz/centaur","owner":"paradigmxyz","name":"centaur","desc":"Centaur is frontier, agentic infrastructure that you own. Centaur is like Claude Tag, but open source and on steroids.","summaryCn":"把Claude Tag变成开源版，打造完全属于你的Agent基础设施","reasonCn":"Centaur是一个前沿的开源Agent基础设施项目，类似于Anthropic推出的Claude Tag，但它完全免费且代码透明。它允许开发者构建、部署和管理自己的AI代理工作流，具备强大的工具调用能力和上下文记忆功能。项目采用模块化架构，支持多种模型接入，让开发者可以像搭积木一样构建复杂的AI自动化系统。","stars":1323,"starGrowth":0,"createdAt":"2026-05-18T17:30:29Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["Agent","自动化","工作流","基础设施"],"domainKey":"Agent","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/paradigmxyz/centaur","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"keinsaasforever/better-chatbot","owner":"keinsaasforever","name":"better-chatbot","desc":"Formerly Better Chatbot. Navigator is an open-source AI workspace for agents, MCP and workflow automation.","summaryCn":"你是否曾被各种复杂的AI提示词配置和分散的API调用搞得头晕脑胀？","reasonCn":"技术上，它深度集成并支持MCP（Model Context Protocol）标准，这意味着它可以轻松连接各种外部数据源和工具。通过TypeScript开发的底层框架，确保了数据流转的高性能与稳定性。它不仅是一个聊天框，更是一个能够触发外部脚本、查询数据库、自动总结任务的智能中枢，其灵活性远超常","stars":1183,"starGrowth":0,"createdAt":"2025-04-13T15:44:51Z","silentRounds":1,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0,"domainTags":["AI工作流","MCP协议","Agent协作","自动化"],"domainKey":"AI工作流","pushedAt":"2026-08-19T16:53:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/keinsaasforever/better-chatbot","ts":"2026-08-19T16:53:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"index-tts/index-tts","owner":"index-tts","name":"index-tts","desc":"An Industrial-Level Controllable and Efficient Zero-Shot Text-To-Speech System","summaryCn":"零样本就能把文字变出超逼真人声的工业级TTS神器","reasonCn":"该项目实现了工业级的高可控、高效率零样本语音合成系统。它通过独特的声学模型与声码器架构，仅需少量参考音频即可克隆出高质量语音，支持音色定制、情感控制及实时推理，解决了传统TTS合成机械感强、迁移困难的问题。","stars":24134,"starGrowth":0,"createdAt":"2025-02-06T04:08:05Z","silentRounds":1,"language":"Python","topics":["bigvgan","cross-lingual","indextts","text-to-speech","tts","voice-clone","zero-shot-tts"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.6,"domainTags":["语音合成","零样本克隆","TTS","声码器","语音交互"],"domainKey":"语音合成","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"bigvgan","source":"github","weight":0.7},{"name":"cross-lingual","source":"github","weight":0.7},{"name":"indextts","source":"github","weight":0.7},{"name":"text-to-speech","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"voice-clone","source":"github","weight":0.7},{"name":"zero-shot-tts","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.95,"source":"trending","url":"https://github.com/index-tts/index-tts","ts":"2026-08-14T01:23:41.092Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"enonforetsam/fluid","owner":"enonforetsam","name":"fluid","desc":"Dependency-free WebGL studio for generative backgrounds and live embeds.","summaryCn":"有时候我们在网上看到那种鼠标划过时泛起绚丽波纹、像水流一样不断变化的动","reasonCn":"技术层面上，它最大亮点在于真正的“零依赖”。它没有用任何沉重的前端框架或大型图形库，而是直接手写了底层的WebGL着色器代码和渲染循环。通过精妙的数学公式和GPU加速，实现了极其丝滑的流体粒子动态效果，同时保证了极小的文件体积和惊人的加载速度。","stars":71,"starGrowth":0,"createdAt":"2026-06-21T13:32:51Z","silentRounds":9,"language":"HTML","topics":["creative-coding","generative-art"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["WebGL","生成艺术","创意编程","流体模拟","着色器"],"domainKey":"WebGL","pushedAt":"2026-09-08T17:53:42Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"generative-art","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/enonforetsam/fluid","ts":"2026-09-08T17:53:42Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Michael-A-Kuykendall/shimmy","owner":"Michael-A-Kuykendall","name":"shimmy","desc":"⚡ Pure-Rust WebGPU inference engine — OpenAI-API compatible, GGUF native, runs on any GPU. No Python. No llama.cpp. Single binary.","summaryCn":"想象一下，你在浏览器里打开一个网页，就能和ChatGPT一样聊天，而且完全不依","reasonCn":"shimmy的核心是用Rust重新实现了推理逻辑，但专门针对WebGPU做了优化。WebGPU是Web标准的GPU计算接口，和之前老旧的WebGL不同，它能提供接近原生性能的GPU计算能力。它原生支持GGUF格式的模型文件，这是llama.cpp社区推广的量化模型格式，支持4位、8位、16位等多种量化方案。","stars":5905,"starGrowth":0,"createdAt":"2025-08-28T22:55:46Z","silentRounds":1,"language":"Rust","topics":["transformers"],"aiDims":["推理引擎","AI基础设施","大语言模型"],"zone":"AI","funScore":1,"domainTags":["WebGPU推理","大模型引擎","Rust","GGUF","浏览器AI"],"domainKey":"WebGPU推理","pushedAt":"2026-08-30T22:45:24Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/Michael-A-Kuykendall/shimmy","ts":"2026-08-30T22:45:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"santifer/career-ops","owner":"santifer","name":"career-ops","desc":"Open-source AI job search: scan job portals, evaluate listings with a structured A-F rubric into a 1.0-5.0 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)","summaryCn":"AI求职神器，自动扫岗位评等级、定制简历、跟踪申请，找工作从此不迷茫","reasonCn":"这个项目是一个本地运行的AI求职Agent，集成Claude Code等AI编码CLI。核心能力包括：自动扫描LinkedIn、Indeed等招聘平台的岗位信息，通过结构化A-F评级体系对职位进行1.0-5.0分量化评估，AI智能生成匹配简历，并可视化追踪投递进度。技术亮点在于完全本地化运行，保护隐私数据，同时支持灵活的工具调用扩展。","stars":72444,"starGrowth":0,"createdAt":"2026-04-04T18:21:18Z","silentRounds":1,"language":"JavaScript","topics":["ai-agent"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI求职","简历定制","岗位扫描","Agent应用"],"domainKey":"AI求职","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/santifer/career-ops","ts":"2026-08-19T01:08:47.324Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ericfreese/rat","owner":"ericfreese","name":"rat","desc":"Compose shell commands to build interactive terminal applications","summaryCn":"像搭积木一样组装命令行工具，把终端变成你的超级游乐场","reasonCn":"这是一个基于Go语言的命令行交互构建框架，允许开发者通过声明式配置快速生成功能丰富的终端应用。它整合了Shell命令执行、异步任务管理和实时UI渲染，支持TUI（文本用户界面）交互。通过DSL语法，用户可以像写代码一样定义菜单、流程和逻辑，无需编写大量底层IO代码，大幅降低了CLI开发门槛。","stars":1286,"starGrowth":0,"createdAt":"2017-02-09T22:09:54Z","silentRounds":13,"language":"Go","topics":["command-line"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Go","终端UI","CLI工具","DSL"],"domainKey":"Go","pushedAt":"2026-09-02T21:59:19Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/ericfreese/rat","ts":"2026-09-02T21:59:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tradecatlabs/vibe-coding-cn","owner":"tradecatlabs","name":"vibe-coding-cn","desc":"Vibe Coding 从入门到精通教程｜AI 结对编程工作流｜Prompt、Skill、Workflow、上下文管理、codex实战指南","summaryCn":"从想法到产品的 AI 结对编程工作流标准：Prompt + Skill + Context + Quali","reasonCn":"教程深入阐述五条核心命题，引用 Demis Hassabis 的“首先解决人工智能问题，然后再用人工智能解决其他所有问题”引领思考。并对大语言模型的生成域进行细致定义，指出模型生成物可直接或间接实现任何文本可表达的结构，如代码、API 调用、测试等。","stars":16352,"starGrowth":1,"createdAt":"2025-11-23T01:52:29Z","silentRounds":2,"language":"Python","topics":["ai-agent"],"aiDims":["AI Agent","提示工程","代码助手"],"zone":"资源","funScore":0,"domainTags":["AI编程","教程","结对编程","Prompt工程","工作流"],"domainKey":"AI编程","pushedAt":"2026-09-17T00:26:49Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/tradecatlabs/vibe-coding-cn","ts":"2026-09-17T00:26:49Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sgl-project/sglang-omni","owner":"sgl-project","name":"sglang-omni","desc":"SGLang-Omni empowers high-performance serving for TTS, ASR, speech and omni models.","summaryCn":"把语音、文字、图像全融合的AI模型超级服务器，性能炸裂","reasonCn":"SGLang-Omni 是一款高性能的统一推理引擎，专门针对语音、文字、图像等多模态模型进行了深度优化。它支持流式输出、并发处理和超低延迟响应，让复杂的多模态交互像搭积木一样简单高效，是目前多模态模型部署领域的一匹黑马。","stars":1261,"starGrowth":3,"createdAt":"2026-01-07T06:58:29Z","silentRounds":0,"language":"Python","topics":["asr","audio-generation","cuda","inference","model-serving","multimodal","openai-api","pytorch","sglang","tts","distributed-inference","music-generation","speech-recognition","streaming","text-to-speech"],"aiDims":["模型部署","多模态Agent","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["多模态推理","TTS","ASR","统一推理引擎","模型部署"],"domainKey":"模型部署","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"asr","source":"github","weight":0.7},{"name":"audio-generation","source":"github","weight":0.7},{"name":"cuda","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"model-serving","source":"github","weight":0.7},{"name":"multimodal","source":"github","weight":0.7},{"name":"openai-api","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"sglang","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"distributed-inference","source":"github","weight":0.7},{"name":"music-generation","source":"github","weight":0.7},{"name":"speech-recognition","source":"github","weight":0.7},{"name":"streaming","source":"github","weight":0.7},{"name":"text-to-speech","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/sgl-project/sglang-omni","ts":"2026-08-16T00:54:51.023Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"quickwit-oss/quickwit","owner":"quickwit-oss","name":"quickwit","desc":"Cloud-native OSS search engine for observability","summaryCn":"给日志和监控数据装上了搜索引擎，比传统方案快十倍。","reasonCn":"专为云原生设计的可观测性搜索引擎，用 Rust 编写。它支持实时流式索引，能处理海量日志和指标数据。兼容 Elasticsearch 生态，提供 SQL 查询接口，解决了传统日志系统查询慢、资源占用高的问题。底层基于 Rust，性能强悍且内存安全。","stars":11665,"starGrowth":0,"createdAt":"2021-04-13T22:45:40Z","silentRounds":1,"language":"Rust","topics":["rust","log-management","logs","tantivy","cloud-native","open-source","big-data","cloud-storage","distributed-tracing","search-engine"],"aiDims":["AI搜索","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Observability","Log Search","Rust"],"domainKey":"Observability","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"rust","source":"github","weight":0.7},{"name":"log-management","source":"github","weight":0.7},{"name":"logs","source":"github","weight":0.7},{"name":"tantivy","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"big-data","source":"github","weight":0.7},{"name":"cloud-storage","source":"github","weight":0.7},{"name":"distributed-tracing","source":"github","weight":0.7},{"name":"search-engine","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/quickwit-oss/quickwit","ts":"2026-09-08T22:27:02.698Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"longbridge/gpui-component","owner":"longbridge","name":"gpui-component","desc":"Rust GUI components for building fantastic cross-platform desktop application by using GPUI.","summaryCn":"用Rust写跨平台桌面应用的组件库，让程序员少写代码多喝咖啡。","reasonCn":"GPUI是一个基于Rust和WGPU构建的声明式GUI框架，专为构建高性能跨平台桌面应用设计。该项目提供了丰富的预构建UI组件，涵盖按钮、列表、对话框等常见元素，支持暗黑模式、键盘导航等现代特性，开发者可以直接集成到GPUI项目中快速构建界面。","stars":14649,"starGrowth":10,"createdAt":"2024-06-13T13:45:39Z","silentRounds":0,"language":"Rust","topics":["gpui","desktop-application","uikit","rust"],"aiDims":["开发者工具","非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["Rust","GUI框架","跨平台桌面"],"domainKey":"Rust","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"gpui","source":"github","weight":0.7},{"name":"desktop-application","source":"github","weight":0.7},{"name":"uikit","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/longbridge/gpui-component","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03204792135981978,"copyOk":false},{"repo":"arthurspk/guiadofrontend","owner":"arthurspk","name":"guiadofrontend","desc":"Guia de Front-end: trilhas, cursos, livros, canais, ferramentas e comunidades para você entrar e evoluir na área.","summaryCn":"前端新人进阶必看，从入门到精通的超全路线图。","reasonCn":"指南的内容非常丰富，分为多个模块。在“学习路径”部分，它详细列出了你需要掌握的核心知识点，比如HTML、CSS、JavaScript，以及进阶的框架如React、Vue等。在“课程”和“书籍”部分，它筛选出了业内公认的经典教程和必读书籍，避免你在茫茫书海中浪费时间。","stars":2907,"starGrowth":0,"createdAt":"2022-03-20T02:03:02Z","silentRounds":27,"language":"","topics":["frontend"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["前端开发","学习路线","教程","资源导航","指南"],"domainKey":"前端开发","pushedAt":"2026-09-02T13:43:58Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7}],"aiScore":0.85,"source":"search","url":"https://github.com/arthurspk/guiadofrontend","ts":"2026-09-02T13:43:58Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SeekStorm/SeekStorm","owner":"SeekStorm","name":"SeekStorm","desc":"SeekStorm: vector & lexical search - in-process library & multi-tenancy server, in Rust.","summaryCn":"用Rust写的高速混合搜索引擎，AI检索的利器","reasonCn":"技术层面，它用Rust编写，天生享有零成本抽象和内存安全的优势。底层支持HNSW向量索引和倒排索引的联合检索，多租户架构让不同客户的数据完全隔离。相比传统方案如Elasticsearch或独立的向量数据库（如Milvus、ChromaDB），SeekStorm不需要额外部署多个服务，直接在进程内完成搜索","stars":1914,"starGrowth":0,"createdAt":"2024-02-27T12:55:08Z","silentRounds":26,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI搜索","开发者工具"],"zone":"AI","funScore":0,"domainTags":["向量检索","混合搜索","Rust","RAG","数据库"],"domainKey":"RAG","pushedAt":"2026-09-12T15:01:55Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/SeekStorm/SeekStorm","ts":"2026-09-12T15:01:55Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dwmkerr/spaceinvaders","owner":"dwmkerr","name":"spaceinvaders","desc":"Classic Space Invaders game written in JavaScript as a learning exercise.","summaryCn":"JavaScript 写的经典太空侵略者，练手游戏开发","reasonCn":"谁说学习游戏编程只能用 C++？spaceinvaders 用纯 JavaScript 和 HTML5 Canvas 重现经典太空侵略者，展示如何在浏览器里处理动画、输入和碰撞。\\n代码分为 engine.js、game.js、ui.js 三个模块。engine.js 负责游戏循环与状态管理；game.js 处理逻辑如敌机生成、子弹射击；ui.js 负","stars":213,"starGrowth":0,"createdAt":"2013-08-20T08:31:10Z","silentRounds":13,"language":"JavaScript","topics":["learning"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["Space Invaders","游戏开发","JavaScript","HTML5"],"domainKey":"Space Invaders","pushedAt":"2026-09-10T07:05:58Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"learning","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/dwmkerr/spaceinvaders","ts":"2026-09-10T07:05:58Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Greninja9257/LabLLM","owner":"Greninja9257","name":"LabLLM","desc":"A native macOS lab for teaching tiny language models to think — build the architecture, train the weights, and watch a small LLM emerge from scratch, locally on Apple Silicon with custom data, tokenizers, checkpoints, and MLX acceleration.","summaryCn":"在Mac上从零构建你的专属小模型，就像看小生命破壳","reasonCn":"这是一个专为Apple Silicon打造的本地AI实验室，让你在Mac上亲手搭建LLM架构、训练权重，见证小语言模型从零到一的诞生过程。项目支持自定义数据集和Tokenization，完全本地运行，无需联网，是学习大模型原理的绝佳实践。","stars":74,"starGrowth":0,"createdAt":"2026-08-15T22:16:40Z","silentRounds":27,"language":"Swift","topics":["llm-model","transformers"],"aiDims":["模型与训练","大语言模型","非AI-好玩"],"zone":"资源","funScore":0,"domainTags":["LLM教学","Apple Silicon","模型训练","本地部署"],"domainKey":"LLM教学","pushedAt":"2026-08-20T01:57:09Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Greninja9257/LabLLM","ts":"2026-08-20T01:57:09Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"L1Xu4n/Awesome-ChatGPT-prompts-ZH_CN","owner":"L1Xu4n","name":"Awesome-ChatGPT-prompts-ZH_CN","desc":"","summaryCn":"该项目不仅是一个指令清单，更是一份关于如何与大型语言模型进行有效沟通的","reasonCn":"提示词工程的精髓在于如何将复杂的需求转化为模型能听懂的语言，项目文档中列举的“猫娘”指令以及针对数学能力提升的 Prompt，都是非常典型且实用的技巧展示。这些内容能帮助你在使用 AI 时避免“车轱辘话”，直接通过精准的设定获得想要的结果。","stars":3186,"starGrowth":1,"createdAt":"2022-12-11T12:30:52Z","silentRounds":3,"language":"","topics":[],"aiDims":["提示工程","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["ChatGPT","Claude","提示词技巧"],"domainKey":"ChatGPT","pushedAt":"2026-09-19T22:19:04.205Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.5,"source":"search","url":"https://github.com/L1Xu4n/Awesome-ChatGPT-prompts-ZH_CN","ts":"2026-09-19T22:19:04.205Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"twentyhq/twenty","owner":"twentyhq","name":"twenty","desc":"The open alternative to Salesforce, designed for AI.","summaryCn":"开源版的Salesforce，AI加持后更懂你的业务流程","reasonCn":"Twenty 是一款基于 TypeScript 开发的开源 CRM 系统，主打企业级协作与客户管理。它通过内置的 AI 引擎自动分析客户数据、优化销售漏斗，并提供智能化的任务分配与跟进建议。相比传统 CRM，它不仅功能完备，还完全掌控数据，适合追求隐私与定制化的团队。","stars":57280,"starGrowth":1,"createdAt":"2022-12-01T13:04:40Z","silentRounds":1,"language":"TypeScript","topics":["crm","customer","graphql","marketing","react","sales","typescript","crm-system","nestjs","reactjs","web","monorepo","open-source","postgresql","good-first-issue","hacktoberfest","javascript"],"aiDims":["AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["CRM","客户管理","开源软件","企业协作"],"domainKey":"CRM","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"crm","source":"github","weight":0.7},{"name":"customer","source":"github","weight":0.7},{"name":"graphql","source":"github","weight":0.7},{"name":"marketing","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"sales","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"crm-system","source":"github","weight":0.7},{"name":"nestjs","source":"github","weight":0.7},{"name":"reactjs","source":"github","weight":0.7},{"name":"web","source":"github","weight":0.7},{"name":"monorepo","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"good-first-issue","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/twentyhq/twenty","ts":"2026-08-13T01:37:33.040Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lucidrains/dreamer4","owner":"lucidrains","name":"dreamer4","desc":"Implementation of Danijar's latest iteration for his Dreamer line of work","summaryCn":"在大多数人都在研究聊天机器人的时候，这个项目在研究如何让AI“做梦”。","reasonCn":"其架构亮点在于将感知、记忆和决策紧密耦合在世界模型中。它利用神经网络预测未来状态，并在隐空间内进行训练，大大减少了对真实交互数据的需求。相比于传统的深度强化学习，它的稳定性更高，尤其是在处理高维视觉输入任务（如机器人控制、视频游戏）时表现异常出色。","stars":218,"starGrowth":0,"createdAt":"2025-09-30T12:58:16Z","silentRounds":20,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","AI基础设施"],"zone":"AI","funScore":1,"domainTags":["强化学习","世界模型","深度学习","规划算法","Transformer"],"domainKey":"强化学习","pushedAt":"2026-09-03T17:03:17Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/lucidrains/dreamer4","ts":"2026-09-03T17:03:17Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"asheshgoplani/agent-deck","owner":"asheshgoplani","name":"agent-deck","desc":"Terminal session manager for AI coding agents. One TUI for Claude, Gemini, OpenCode, Codex, and more.","summaryCn":"把Claude和ChatGPT都塞进一个终端里，像玩游戏一样管理AI","reasonCn":"这是一个专为AI编码助手打造的终端会话管理器，支持Claude、Gemini、OpenAI Codex等主流模型。它通过统一的TUI界面解决了多Agent切换繁琐的问题，让开发者能在一个窗口内流畅切换和调试不同AI模型，极大提升了本地开发效率。","stars":940,"starGrowth":0,"createdAt":"2025-12-03T17:08:12Z","silentRounds":1,"language":"Go","topics":["claude-code","ai-agents","ai-coding-assistant","aider","bubble-tea","cli","codex","developer-tools","gemini-cli","golang","productivity","session-manager","terminal","tmux","tui","discord","ai-agent","mcp"],"aiDims":["AI Agent","开发者工具","工具调用"],"zone":"工具","funScore":0,"domainTags":["终端管理","AI编程助手","会话管理"],"domainKey":"终端管理","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-coding-assistant","source":"github","weight":0.7},{"name":"aider","source":"github","weight":0.7},{"name":"bubble-tea","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"gemini-cli","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"session-manager","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"tmux","source":"github","weight":0.7},{"name":"tui","source":"github","weight":0.7},{"name":"discord","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/asheshgoplani/agent-deck","ts":"2026-08-18T00:56:27.274Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Lightricks/LTX-2","owner":"Lightricks","name":"LTX-2","desc":"Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.","summaryCn":"用Python就能训练视频生成模型，普通人也能做AI导演","reasonCn":"LTX-2是新一代音频视频生成模型，官方Python包提供了完整的推理和LoRA微调工具链。它支持从零开始训练定制化视频风格，也能基于预训练模型快速生成高质量片段。项目包含详细训练脚本和推理示例，让开发者能轻松集成到应用中。","stars":9506,"starGrowth":3,"createdAt":"2026-01-03T13:16:29Z","silentRounds":0,"language":"Python","topics":["generative-ai","ltx","ltx-2"],"aiDims":["多模态Agent","模型与训练","创意工具"],"zone":"AI","funScore":1,"domainTags":["视频生成","扩散模型","LoRA微调","音频视频联合生成"],"domainKey":"视频生成","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"generative-ai","source":"github","weight":0.7},{"name":"ltx","source":"github","weight":0.7},{"name":"ltx-2","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Lightricks/LTX-2","ts":"2026-08-15T00:57:24.351Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"danielmiessler/LifeOS","owner":"danielmiessler","name":"LifeOS","desc":"⛰️A General Hill-climbing AI harness that helps you move from Current State to Ideal State in both Life and Work.","summaryCn":"让AI当你的私人教练，帮你一步步从现状冲到理想人生","reasonCn":"从技术架构来看，LifeOS是一个基于TypeScript构建的AI Agent系统，核心在于它不仅仅是一个目标记录工具，而是一个能够进行多轮对话、任务分解和状态追踪的智能助手。它利用大语言模型的推理能力，帮助用户将宏大的人生目标拆解为可执行的小步骤，并持续追踪完成进度。","stars":19098,"starGrowth":5,"createdAt":"2025-09-08T16:18:30Z","silentRounds":0,"language":"TypeScript","topics":["productivity"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["AI Agent","目标管理","个人效率"],"domainKey":"AI Agent","pushedAt":"2026-09-04T01:07:21Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/danielmiessler/LifeOS","ts":"2026-09-04T01:07:21Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.015785422557335844,"copyOk":false},{"repo":"Gitlawb/zero","owner":"Gitlawb","name":"zero","desc":"The coding agent that answers to you, your model, your machine, your rules.","summaryCn":"能听懂你指令，像老司机一样帮你搞定代码的智能管家","reasonCn":"这是一个完全运行在你本地的代码智能体，它不依赖云端API，而是通过调用系统命令、编辑文件和读写数据库来完成任务。它采用了“工具调用”机制，能像人类开发者一样理解复杂指令，分解任务并执行具体操作。项目目前使用Go语言开发，由社区驱动，专注于提供可控、可扩展的本地开发体验。","stars":1681,"starGrowth":1,"createdAt":"2026-05-28T14:11:35Z","silentRounds":1,"language":"Go","topics":["ai-agent","ai","anthropic","cli","code-assistant","coding-agent","developer-tools","gemini","llm","mcp","ollama","openai","terminal"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0,"domainTags":["代码智能体","本地执行","系统命令","Agent","自动化"],"domainKey":"Agent","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"code-assistant","source":"github","weight":0.7},{"name":"coding-agent","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Gitlawb/zero","ts":"2026-08-18T00:56:27.274Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"khuedoan/homelab","owner":"khuedoan","name":"homelab","desc":"Fully automated homelab from empty disk to running services with a single command.","summaryCn":"一键下单空磁盘，自动装配完整的家庭实验室，服务即开。","reasonCn":"项目用Python编写全链路自动化脚本，从零磁盘分区、文件系统创建、容器平台部署到常见服务（如Nextcloud、HomeAssistant、AdGuard）一键完成，支持自定义Playbook，兼容多种硬件，配合Ansible与Terraform，实现可重复、可迁移的家庭私有云环境。","stars":9622,"starGrowth":1,"createdAt":"2020-06-21T13:48:40Z","silentRounds":1,"language":"Python","topics":["devops","kubernetes"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.6,"domainTags":["家庭服务器","Ansible","Docker","自动部署"],"domainKey":"家庭服务器","pushedAt":"2026-09-08T02:33:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/khuedoan/homelab","ts":"2026-09-08T02:33:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xai-org/grok-1","owner":"xai-org","name":"grok-1","desc":"Grok open release","summaryCn":"马斯克开源的 GPT-4 级别大模型，参数量巨大。","reasonCn":"Grok-1 是由 xAI 公司发布的开源大语言模型，拥有 3140 亿参数，基于 MoE 架构。它支持长达 128k 的上下文窗口，并在多项基准测试中表现优异，展示了强大的代码生成、数学推理和常识问答能力，为开源社区提供了顶级的模型基座。","stars":52234,"starGrowth":0,"createdAt":"2024-03-17T08:53:38Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["大语言模型","模型与训练","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["大语言模型","开源模型","混合专家架构","代码生成"],"domainKey":"大语言模型","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/xai-org/grok-1","ts":"2026-08-17T00:55:40.962Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepfakes/faceswap","owner":"deepfakes","name":"faceswap","desc":"Deepfakes Software For All","summaryCn":"开源换脸神器，一键把你的脸P进任意电影片段","reasonCn":"deepfakes/faceswap是目前最流行的开源换脸工具，基于深度学习的人脸交换算法。核心采用对抗生成网络(GAN)架构，通过Encoder-Decoder网络提取人脸特征，实现高质量的视频换脸。支持多种预训练模型，可在普通GPU上运行，也提供Web界面方便使用。","stars":57552,"starGrowth":0,"createdAt":"2017-12-19T09:44:13Z","silentRounds":5,"language":"Python","topics":["ml"],"aiDims":["非AI-好玩","创意工具","AI应用"],"zone":"创意","funScore":1,"domainTags":["深度学习","换脸","视频处理","GAN","人脸识别"],"domainKey":"深度学习","pushedAt":"2026-08-05T11:39:37Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/deepfakes/faceswap","ts":"2026-08-05T11:39:37Z","category":"fun","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"javifalces/HFTFramework","owner":"javifalces","name":"HFTFramework","desc":"HFTFramework utilized  for research on \" A reinforcement learning approach to improve the performance of the Avellaneda-Stoikov market-making algorithm \" ","summaryCn":"用强化学习升级传统做市算法的高频交易研究框架","reasonCn":"该框架旨在通过强化学习优化 Avellaneda-Stoikov 市场做市算法，提升高频交易表现。它提供了完整的 Jupyter Notebook 环境，涵盖数据加载、策略训练、回测及实盘接口。适合量化研究者探索 AI 在金融领域的应用，代码结构清晰，便于复现论文实验。虽然主要面向学术和社区交流，但为自动化交易策略提供了新的智能视角，是金融科技交叉研究的优质起点，帮助开发者快速搭建实验环境。","stars":307,"starGrowth":0,"createdAt":"2021-11-09T07:48:07Z","silentRounds":13,"language":"Jupyter Notebook","topics":["trading"],"aiDims":["AI应用","模型与训练","效率工具"],"zone":"工具","funScore":0,"domainTags":["高频交易","强化学习","量化金融"],"domainKey":"高频交易","pushedAt":"2026-09-10T17:22:03Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/javifalces/HFTFramework","ts":"2026-09-10T17:22:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fallow-rs/fallow","owner":"fallow-rs","name":"fallow","desc":"Codebase intelligence for TypeScript and JavaScript. Free static analysis of code and styles: unused code, duplication, circular deps, complexity hotspots, architecture boundaries, design-system drift. Optional paid runtime layer (Fallow Runtime): hot-path review and cold-path deletion evidence from real production traffic.","summaryCn":"你有没有遇到过那种写着写着就烂掉的代码库？","reasonCn":"它的厉害之处在于对代码复杂度的精准把控。它能识别出那些“过度复杂”的函数，这些函数通常是Bug的重灾区。同时，它还能检测设计系统的漂移，比如你定义了统一的按钮样式，但项目中混用了几十种不同的风格，Fallow会帮你把这些差异找出来。","stars":4827,"starGrowth":6,"createdAt":"2026-03-17T08:40:57Z","silentRounds":0,"language":"Rust","topics":["developer-tools"],"aiDims":["代码助手","开发者工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["Static Analysis","Code Quality","Developer Tools"],"domainKey":"Static Analysis","pushedAt":"2026-09-09T22:01:35Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/fallow-rs/fallow","ts":"2026-09-09T22:01:35Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.02607482908638906,"copyOk":false},{"repo":"deepseek-ai/awesome-deepseek-integration","owner":"deepseek-ai","name":"awesome-deepseek-integration","desc":"Integrate the DeepSeek API into popular software","summaryCn":"全网最全DeepSeek API接入指南，各种软件轻松接入大模型","reasonCn":"这个项目收集了大量第三方工具，帮助开发者把DeepSeek的AI能力接入到各种流行软件里。它汇总了即时通讯、笔记软件、浏览器插件、阅读工具等场景的集成方案，配有详细配置教程和代码示例。DeepSeek的API本身能力很强，但很多用户不知道如何把它用到日常软件中。这个项目解决了信息碎片化的问题，让开发者能一站式找到各类集成方案，快速在自己的工作流中用上DeepSeek的智能能力。","stars":39215,"starGrowth":2,"createdAt":"2024-01-11T02:43:17Z","silentRounds":0,"language":"","topics":["awesome","deepseek"],"aiDims":["AI应用","工具调用","开发者工具"],"zone":"资源","funScore":0,"domainTags":["DeepSeek","API集成","Awesome列表"],"domainKey":"DeepSeek","pushedAt":"2026-02-23T16:27:47Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7}],"aiScore":0.91,"source":"bigbro","url":"https://github.com/deepseek-ai/awesome-deepseek-integration","ts":"2026-02-23T16:27:47Z","category":"learning","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"HarrierOnChain/Prediction-Markets-Trading-Bot-Toolkits","owner":"HarrierOnChain","name":"Prediction-Markets-Trading-Bot-Toolkits","desc":"Rust trading bots for prediction markets — Polymarket, Kalshi, Limitless. Copy Trading is production-ready; nine more strategies in development on the same execution core. Dry-run by default, self-custody, MIT.","summaryCn":"基于Rust开发的预测市场交易机器人，支持现货和跟单，策略丰富且开源","reasonCn":"这是一个为预测市场设计的Rust交易机器人工具包，支持Polymarket、Kalshi等主流平台。它提供了一键跟单功能，内置多种交易策略，并支持回测和模拟交易。项目采用模块化设计，用户可以轻松扩展新的策略。代码完全开源，旨在为预测市场交易提供技术支持和策略参考，帮助用户在波动市场中进行更理性的决策。","stars":456,"starGrowth":0,"createdAt":"2025-09-29T15:17:57Z","silentRounds":1,"language":"Rust","topics":["trading"],"aiDims":["AI应用","非AI-实用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["预测市场","交易机器人","Rust量化","跟单交易"],"domainKey":"预测市场","pushedAt":"2026-09-07T11:44:40Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"trading","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/HarrierOnChain/Prediction-Markets-Trading-Bot-Toolkits","ts":"2026-09-07T11:44:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Jacob-Zhou/simple-csc","owner":"Jacob-Zhou","name":"simple-csc","desc":"This repository provides an implementation of \"A Simple yet Effective Training-free Prompt-free Approach to Chinese Spelling Correction Based on Large Language Models\" and \"A Training-free LLM-based Approach to General Chinese Character Error Correction\".","summaryCn":"想象一下，你让一个不懂中文的大语言模型去写文章，它经常把“的”写成“地”或","reasonCn":"它背后的核心思想非常巧妙：大模型本身其实懂中文，只是有时候会“手滑”。项目通过一套精心设计的Prompt，引导模型在生成前先“自检”一遍，把错的改成对的。因为不需要重新训练模型，也不需要准备额外的训练数据，所以它被称为“训练即即用”。","stars":90,"starGrowth":0,"createdAt":"2024-03-25T13:00:51Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","提示工程","非AI-实用"],"zone":"AI","funScore":0,"domainTags":["中文拼写纠错","大语言模型","提示工程","NLP工具"],"domainKey":"大语言模型","pushedAt":"2026-08-24T17:09:44Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Jacob-Zhou/simple-csc","ts":"2026-08-24T17:09:44Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"RVC-Project/Retrieval-based-Voice-Conversion-WebUI","owner":"RVC-Project","name":"Retrieval-based-Voice-Conversion-WebUI","desc":"Easily train a good VC model with voice data &lt;= 10 mins!","summaryCn":"只需十分钟录音，就能把你的声音克隆成别人的","reasonCn":"这是一个基于 Retrieval-based Voice Conversion (RVC) 技术的 Web 界面工具。它允许用户通过极少量（小于10分钟）的目标语音数据，训练出一个高质量的语音转换模型。用户只需上传音频并调整参数，即可将自己的声音实时转换为目标声音，无需昂贵的 GPU 训练。该项目极大地降低了语音克隆技术的门槛，让普通人也能轻松体验 AI 声音合成。","stars":38460,"starGrowth":11,"createdAt":"2023-03-27T09:59:10Z","silentRounds":0,"language":"Python","topics":["change","sovits","vits","voice","voice-conversion","rvc","audio-analysis","conversational-ai","conversion","converter","retrieval-model","retrieve-data","so-vits-svc","vc","voice-converter","voiceconversion"],"aiDims":["AI应用","多模态Agent","创意工具"],"zone":"创意","funScore":0.9,"domainTags":["Voice Conversion","AI Music","Real-time"],"domainKey":"Voice Conversion","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"change","source":"github","weight":0.7},{"name":"sovits","source":"github","weight":0.7},{"name":"vits","source":"github","weight":0.7},{"name":"voice","source":"github","weight":0.7},{"name":"voice-conversion","source":"github","weight":0.7},{"name":"rvc","source":"github","weight":0.7},{"name":"audio-analysis","source":"github","weight":0.7},{"name":"conversational-ai","source":"github","weight":0.7},{"name":"conversion","source":"github","weight":0.7},{"name":"converter","source":"github","weight":0.7},{"name":"retrieval-model","source":"github","weight":0.7},{"name":"retrieve-data","source":"github","weight":0.7},{"name":"so-vits-svc","source":"github","weight":0.7},{"name":"vc","source":"github","weight":0.7},{"name":"voice-converter","source":"github","weight":0.7},{"name":"voiceconversion","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI","ts":"2026-08-19T01:08:47.324Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.033858034321372854,"copyOk":true},{"repo":"e2b-dev/infra","owner":"e2b-dev","name":"infra","desc":"Infrastructure that's powering E2B Cloud.","summaryCn":"给AI Agent建虚拟电脑的神器，写代码跑任务不用自己搭环境","reasonCn":"E2B Infra 是一个开源的云原生沙箱基础设施，专为AI Agent和开发者设计。它使用Go语言编写，提供安全的、隔离的虚拟计算环境，让AI Agent能够安全地编写和执行代码，以及运行各种开发工具。这解决了AI Agent在真实服务器上操作带来的安全风险，让开发者能专注于构建智能应用。","stars":1601,"starGrowth":0,"createdAt":"2023-08-10T17:37:55Z","silentRounds":1,"language":"Go","topics":["ai-agents","code-interpreter","consul","devtools","firecracker","gcp","go","golang","gpt","llm","nomad","terraform","vm","kvm","microvm","sandbox","vmm","infra"],"aiDims":["AI基础设施","AI Agent","开发工具"],"zone":"工具","funScore":0,"domainTags":["AI Agent","云原生","沙箱","开发工具"],"domainKey":"AI Agent","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"code-interpreter","source":"github","weight":0.7},{"name":"consul","source":"github","weight":0.7},{"name":"devtools","source":"github","weight":0.7},{"name":"firecracker","source":"github","weight":0.7},{"name":"gcp","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"nomad","source":"github","weight":0.7},{"name":"terraform","source":"github","weight":0.7},{"name":"vm","source":"github","weight":0.7},{"name":"kvm","source":"github","weight":0.7},{"name":"microvm","source":"github","weight":0.7},{"name":"sandbox","source":"github","weight":0.7},{"name":"vmm","source":"github","weight":0.7},{"name":"infra","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/e2b-dev/infra","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"DarkFlippers/unleashed-firmware","owner":"DarkFlippers","name":"unleashed-firmware","desc":"","summaryCn":"DarkFlippers/unleashed-firmware 是基于 flipperdevices/flipperzero-firm","reasonCn":"项目的核心亮点在于高稳定性与丰富的定制功能。开发团队在重构与改进原厂核心组件的同时，保持了与官方 API 和各类第三方应用 100% 的兼容性。使用此固件后，设备可以获得更宽泛的 Sub-GHz 频率发射范围、更自由的无线信号测试能力、增强的红外库以及大量预置的极客工具，极大提升","stars":22323,"starGrowth":3,"createdAt":"2022-03-16T12:46:45Z","silentRounds":0,"language":"C","topics":["jailbreak","keeloq","rolling-codes","alternative-firmware","flipper","flipperzero","unleashed","firmware","flipper-plugins","flipper-zero","darkflippers","cfw","custom","custom-firmware","unlocked","unofficial"],"aiDims":["非AI-好玩","安全工具","开发者工具"],"zone":"创意","funScore":0.8,"domainTags":["固件开发","Flipper Zero","无线电"],"domainKey":"固件开发","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"jailbreak","source":"github","weight":0.7},{"name":"keeloq","source":"github","weight":0.7},{"name":"rolling-codes","source":"github","weight":0.7},{"name":"alternative-firmware","source":"github","weight":0.7},{"name":"flipper","source":"github","weight":0.7},{"name":"flipperzero","source":"github","weight":0.7},{"name":"unleashed","source":"github","weight":0.7},{"name":"firmware","source":"github","weight":0.7},{"name":"flipper-plugins","source":"github","weight":0.7},{"name":"flipper-zero","source":"github","weight":0.7},{"name":"darkflippers","source":"github","weight":0.7},{"name":"cfw","source":"github","weight":0.7},{"name":"custom","source":"github","weight":0.7},{"name":"custom-firmware","source":"github","weight":0.7},{"name":"unlocked","source":"github","weight":0.7},{"name":"unofficial","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/DarkFlippers/unleashed-firmware","ts":"2026-09-13T22:08:09.499Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"thomas-luebker/Amigo","owner":"thomas-luebker","name":"Amigo","desc":"Amigo — the Amiga on iPad, iPhone and Apple Silicon Macs. Native WinUAE port, free on the App Store, GPL-2.","summaryCn":"把经典电脑模拟器装进手机，随时随地玩怀旧游戏","reasonCn":"这是在苹果设备上运行经典Amiga电脑模拟器的原生应用，通过移植WinUAE引擎实现。它利用Apple Silicon芯片的指令集特性，无需虚拟化层即可获得接近硬件的性能。项目采用GPL-2开源协议，完全免费。支持iOS、iPadOS和macOS，并针对M1/M2芯片做了底层优化，让老游戏跑起来既流畅又稳定。","stars":15,"starGrowth":0,"createdAt":"2026-07-19T00:11:38Z","silentRounds":6,"language":"Swift","topics":["retrocomputing"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":1,"domainTags":["复古计算","游戏模拟器","Apple Silicon","WinUAE","怀旧游戏"],"domainKey":"复古计算","pushedAt":"2026-09-09T21:29:26Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/thomas-luebker/Amigo","ts":"2026-09-09T21:29:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jordanrendric/claude-video-vision","owner":"jordanrendric","name":"claude-video-vision","desc":"Give Claude the ability to watch and understand videos — Claude Code plugin with frame extraction and multimodal audio analysis","summaryCn":"给Claude装上眼睛，让它能看懂视频内容了","reasonCn":"技术实现上，它作为一个Claude Code插件运行，自动截取视频关键帧，用多模态模型分析画面内容，同时提取音频转文字进行综合分析。插件会智能选择帧——不是逐帧处理（那太慢了），而是用运动检测挑选有信息量的画面，兼顾效率和准确度。","stars":1327,"starGrowth":1,"createdAt":"2026-03-31T19:49:52Z","silentRounds":1,"language":"TypeScript","topics":["claude-code","claude-code-plugin","ffmpeg","gemini","mcp","mcp-server","multimodal","video-analysis","whisper"],"aiDims":["多模态Agent","AI应用","工具调用"],"zone":"AI","funScore":1,"domainTags":["多模态分析","Claude插件","视频理解"],"domainKey":"多模态分析","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"claude-code-plugin","source":"github","weight":0.7},{"name":"ffmpeg","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"multimodal","source":"github","weight":0.7},{"name":"video-analysis","source":"github","weight":0.7},{"name":"whisper","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/jordanrendric/claude-video-vision","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"confident-ai/deepteam","owner":"confident-ai","name":"deepteam","desc":"DeepTeam is a framework to red team LLMs and AI agents.","summaryCn":"给AI装个“防弹衣”，专治各种AI瞎胡闹。","reasonCn":"DeepTeam是一个专为LLM和AI Agent设计的红队测试框架，旨在主动发现和利用大模型的安全漏洞。它通过自动化攻击策略，模拟各种攻击场景，帮助开发者及时修复Prompt注入、指令劫持等潜在风险，确保AI应用在生产环境中的安全性。","stars":2932,"starGrowth":6,"createdAt":"2025-03-05T06:34:21Z","silentRounds":0,"language":"Python","topics":["llm-guardrails","llm-red-teaming","llm-safety","python","llm-seecurity"],"aiDims":["AI Agent","安全工具","非AI-实用"],"zone":"AI","funScore":0.5,"domainTags":["LLM安全测试","红队攻防","Prompt注入检测","AI安全"],"domainKey":"LLM安全测试","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"llm-guardrails","source":"github","weight":0.7},{"name":"llm-red-teaming","source":"github","weight":0.7},{"name":"llm-safety","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"llm-seecurity","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/confident-ai/deepteam","ts":"2026-08-11T09:45:50.026Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.024139154160982265,"copyOk":false},{"repo":"frank890417/WebToe","owner":"frank890417","name":"WebToe","desc":"Web-native dataflow engine for real-time visuals — patch in the browser, TouchDesigner-style, and import your .toe projects. WebGL2 + WebGPU, zero dependencies.","summaryCn":"如果你一直垂涎TouchDesigner那种通过拖拽节点连线就能做出震撼视觉效果的能","reasonCn":"其核心亮点在于极致的Web技术堆栈，直接利用GPU并行计算来驱动图形渲染。与传统的canvas绘图不同，它采用数据流驱动模式，这意味着每一个视觉节点都能实时传递变化，让动态图形的创作过程变得直观且流畅。它不仅能独立运行，还兼容行业标准.toe格式，极大降低了学习成本。","stars":26,"starGrowth":0,"createdAt":"2026-06-11T01:22:43Z","silentRounds":19,"language":"TypeScript","topics":["generative-art","creative-coding"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["generative art","创意编程","WebGL","数据流可视化","TouchDesigner"],"domainKey":"generative art","pushedAt":"2026-08-25T03:57:37Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"generative-art","source":"github","weight":0.7},{"name":"creative-coding","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/frank890417/WebToe","ts":"2026-08-25T03:57:37Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ParthJadhav/app-store-screenshots","owner":"ParthJadhav","name":"app-store-screenshots","desc":"end to end app store screenshot creation using AI","summaryCn":"AI一键生成App应用商店截图，告别设计加班","reasonCn":"这个项目用人工智能技术自动化生成各大应用商店所需的截图素材，支持iOS和Android平台，覆盖多种设备尺寸。用户只需提供应用描述或关键词，AI就能自动排版、配色并生成高质量的营销图片，省去了找设计师或手动做图的麻烦，效率提升非常明显，特别适合独立开发者和小型团队。","stars":6948,"starGrowth":1,"createdAt":"2026-03-07T13:56:04Z","silentRounds":1,"language":"TypeScript","topics":["design"],"aiDims":["AI应用","创意工具","非AI-实用"],"zone":"AI","funScore":0.6,"domainTags":["App截图","AI生成","模板设计","应用商店","营销素材"],"domainKey":"App截图","pushedAt":"2026-09-05T08:14:44Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/ParthJadhav/app-store-screenshots","ts":"2026-09-05T08:14:44Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"randspace0/Anime4K-rs","owner":"randspace0","name":"Anime4K-rs","desc":"An attempt to write Anime4K in Rust","summaryCn":"用Rust重写的老牌动漫画质增强神器，免费又快","reasonCn":"它的技术原理是利用图像中的边缘和纹理信息，通过数学插值算法预测出高分辨率下的像素值。Rust语言的特性保证了它在处理视频时的速度和稳定性，比之前的C++版本更快，且内存占用更低。它支持4x、2x等多种放大倍率，并且针对动漫特有的色彩和线条进行了优化，能让原本模糊的画面变得","stars":133,"starGrowth":0,"createdAt":"2019-10-18T18:00:28Z","silentRounds":12,"language":"Rust","topics":["anime"],"aiDims":["创意工具","非AI-实用","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["动漫","画质增强","超分辨率","视频处理","图像插值"],"domainKey":"动漫","pushedAt":"2026-09-12T04:43:24Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/randspace0/Anime4K-rs","ts":"2026-09-12T04:43:24Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"google-research/bert","owner":"google-research","name":"bert","desc":"","summaryCn":"谷歌开源的预训练语言模型，重新定义了自然语言处理任务的基准。","reasonCn":"BERT是谷歌2018年发布的预训练语言模型，通过双向Transformer结构在大量文本上预训练，再针对具体任务微调。它彻底改变了NLP领域，让机器翻译、问答、情感分析等任务的效果大幅提升。这个仓库包含原始模型代码、预训练权重下载链接和微调示例，是NLP学习和研究的经典起点。","stars":40042,"starGrowth":0,"createdAt":"2018-10-25T22:57:34Z","silentRounds":2,"language":"Python","topics":["nlp","google","natural-language-processing","natural-language-understanding","tensorflow"],"aiDims":["模型与训练","大语言模型","AI基础设施"],"zone":"AI","funScore":0.44,"domainTags":["语言模型预训练","自然语言处理","模型压缩","知识蒸馏"],"domainKey":"语言模型预训练","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"nlp","source":"github","weight":0.7},{"name":"google","source":"github","weight":0.7},{"name":"natural-language-processing","source":"github","weight":0.7},{"name":"natural-language-understanding","source":"github","weight":0.7},{"name":"tensorflow","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/google-research/bert","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"AmirMotefaker/Create-your-own-ChatGPT","owner":"AmirMotefaker","name":"Create-your-own-ChatGPT","desc":"Create your own ChatGPT with Python","summaryCn":"从零搭建ChatGPT的Python教程，动手理解LLM原理","reasonCn":"技术上，教程使用Hugging Face的Transformers库加载开源语言模型（如LLaMA或GPT-2），演示如何构造prompt、调用模型生成回复、管理对话历史状态。它还会介绍基本的prompt engineering技巧，以及如何使用OpenAI API快速构建对话界面。","stars":23,"starGrowth":0,"createdAt":"2023-02-23T18:34:48Z","silentRounds":27,"language":"Jupyter Notebook","topics":["llm-model"],"aiDims":["大语言模型","AI应用","教程"],"zone":"资源","funScore":0,"domainTags":["大语言模型","LLM教程","提示工程","对话系统"],"domainKey":"大语言模型","pushedAt":"2026-09-10T08:54:24Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/AmirMotefaker/Create-your-own-ChatGPT","ts":"2026-09-10T08:54:24Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Unclecheng-li/VulnClaw","owner":"Unclecheng-li","name":"VulnClaw","desc":"基于 AI Agent + MCP 工具链 + 渗透 Skill 编排， 配合大语言模型， 自然语言输入 → 自动完成「信息收集 → 漏洞发现 → 漏洞利用 → 报告生成」全流程。","summaryCn":"自然语言输入就能自动完成漏洞扫描到报告生成的全流程AI黑客助手","reasonCn":"技术架构上，它以大语言模型为大脑，通过MCP协议调用各种渗透测试工具。信息收集阶段可以用Nmap、Subfinder等工具枚举目标；漏洞发现阶段能调用SQLmap、XSStrike等扫描器；漏洞利用阶段有相应的exploit链；最后自动生成结构化的渗透报告。","stars":3397,"starGrowth":0,"createdAt":"2026-04-18T13:02:17Z","silentRounds":1,"language":"Python","topics":["ai","ai-agent","ai-tools","ctf","cybersecurity","openclaw","penetration-testing","penetration-testing-tools","security-tools","skill"],"aiDims":["AI Agent","安全工具"],"zone":"AI","funScore":1,"domainTags":["渗透测试","AI Agent","漏洞扫描","MCP工具链","网络安全"],"domainKey":"AI Agent","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"ai-tools","source":"github","weight":0.7},{"name":"ctf","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"penetration-testing","source":"github","weight":0.7},{"name":"penetration-testing-tools","source":"github","weight":0.7},{"name":"security-tools","source":"github","weight":0.7},{"name":"skill","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/Unclecheng-li/VulnClaw","ts":"2026-08-07T04:03:56.389Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"altic-dev/FluidVoice","owner":"altic-dev","name":"FluidVoice","desc":"Fastest and only macOS Dictation app with on-device STT and custom trained AI enhancement model. A local Wispr Flow alternative. DM us on X for an easter egg 😉 - https://x.com/fluidvoiceapp","summaryCn":"在Mac上最快最本地化的语音输入工具，替代云端翻译的神器。","reasonCn":"FluidVoice是一款运行在macOS上的极速语音转文字应用，主打本地离线处理。它内置了先进的端侧语音识别引擎，并利用AI模型对识别结果进行智能纠错和优化，无需联网即可获得流畅的打字体验。相比传统的输入法或云端语音服务，它不仅保护隐私，还能在本地快速处理大量语音数据，是目前本地语音交互领域的佼佼者。","stars":11699,"starGrowth":2,"createdAt":"2025-09-21T17:35:02Z","silentRounds":0,"language":"Swift","topics":["ai","dictation","ios","llama-cpp","macos","swift"],"aiDims":["AI应用","工具调用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["语音转文字","本地STT","Mac应用","输入法","语音识别"],"domainKey":"语音转文字","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"dictation","source":"github","weight":0.7},{"name":"ios","source":"github","weight":0.7},{"name":"llama-cpp","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"swift","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/altic-dev/FluidVoice","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"smtg-ai/claude-squad","owner":"smtg-ai","name":"claude-squad","desc":"Manage multiple AI terminal agents like Claude Code, Codex, OpenCode, and Amp.","summaryCn":"像指挥官一样调度多个AI编程助手，让你的终端效率起飞","reasonCn":"Claude Squad 是一个用 Go 语言编写的命令行工具，旨在通过统一的界面管理多个不同的 AI 编程代理。它解决了在终端中同时与多个 AI 助手交互时效率低下的问题，支持像 Claude Code、Codex、OpenCode 和 Amp 这样不同的工具，让你像指挥一支 AI 编程小队一样，通过简单的命令在终端里无缝切换和调用它们的能力。","stars":8521,"starGrowth":0,"createdAt":"2025-03-09T20:00:41Z","silentRounds":1,"language":"Go","topics":["claude-code","cli","vibe-coding","codex","opencode"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI编程","终端工具","多Agent","命令行"],"domainKey":"AI编程","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"claude-code","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"vibe-coding","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/smtg-ai/claude-squad","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kalyvask/winning-writing","owner":"kalyvask","name":"winning-writing","desc":"Agentic writing toolkit built from Glenn Kramon's Winning Writing at Stanford GSB and Rachel Konrad's cold-outreach guest lectures. 30 Claude skills, a browser Coach with span-level inline critic and refinement chat, plus a Chrome MV3 extension that runs the same critic in the Gmail side panel.","summaryCn":"Claude驱动的写作外挂，帮你写出打动人的文案","reasonCn":"项目集成了Claude大模型的能力，通过浏览器插件和本地脚本运行。核心功能包括“逐行修改”，它会逐句分析你的文章，指出哪里太干巴、哪里太啰嗦，并给出修改建议。还有“冷启动外挂”，专门针对商务邮件和销售文案，能帮你快速写出开场白和行动号召。","stars":16,"starGrowth":0,"createdAt":"2026-04-30T06:15:05Z","silentRounds":17,"language":"JavaScript","topics":["writing"],"aiDims":["AI写作","AI Agent"],"zone":"AI","funScore":0.5,"domainTags":["写作助手","浏览器插件","文案优化","Claude","外挂工具"],"domainKey":"写作助手","pushedAt":"2026-09-05T18:32:19Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/kalyvask/winning-writing","ts":"2026-09-05T18:32:19Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ml-explore/mlx","owner":"ml-explore","name":"mlx","desc":"","summaryCn":"MLX 是一款专为苹果硅芯片设计的机器学习数组框架，目标是让开发者在 Apple","reasonCn":"与传统深度学习框架不同，MLX 在底层采用了可组合的函数变换机制。该机制支持自动微分、自动向量化以及计算图优化，所有计算默认惰性执行，只有在需要时才触发运算，显著降低不必要的资源消耗。借助这些特性，MLX 能在苹果 M 系列芯片上实现近乎原生的性能，同时保持灵活的模型构建","stars":28524,"starGrowth":4,"createdAt":"2023-11-28T23:33:45Z","silentRounds":0,"language":"C++","topics":["mlx"],"aiDims":["AI基础设施","模型与训练"],"zone":"AI","funScore":0,"domainTags":["机器学习框架","苹果M芯片","自动微分","惰性计算","深度学习"],"domainKey":"机器学习框架","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"mlx","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"tier","url":"https://github.com/ml-explore/mlx","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AThevon/genjutsu","owner":"AThevon","name":"genjutsu","desc":"Creative coding skills for Claude Code, claude.ai and Cowork. Motion, micro-interactions and design systems across Web (GSAP, Motion, Three.js, CSS), Android (Jetpack Compose) and Apple (SwiftUI). Proposes an interaction thesis you approve before it writes any code. Anti-AI-slop.","summaryCn":"Claude驱动的跨平台动效库，玩转网页、安卓、iOS的炫酷交互。","reasonCn":"genjutsu基于Claude提供创意编码框架，整合GSAP、Motion、Three.js，在Web、Jetpack Compose、SwiftUI上实现微交互和动画，支持在Claude对话中生成并热更新代码，兼顾设计系统跨平台统一，附CLI和VSCode插件，适合前端和移动开发者快速原型。","stars":366,"starGrowth":0,"createdAt":"2026-03-28T00:23:39Z","silentRounds":1,"language":"Python","topics":["creative-coding"],"aiDims":["AI应用","创意工具","非AI-好玩"],"zone":"创意","funScore":0,"domainTags":["创意编码","动效库","Claude","跨平台","交互设计"],"domainKey":"创意编码","pushedAt":"2026-09-08T13:55:27Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/AThevon/genjutsu","ts":"2026-09-08T13:55:27Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NVIDIA-NeMo/Switchyard","owner":"NVIDIA-NeMo","name":"Switchyard","desc":"Switchyard lets LLM applications route traffic across models and providers while preserving native OpenAI and Anthropic API compatibility - enabling flexible model selection, benchmarking, and cost/performance optimization.","summaryCn":"让大模型学会像人一样思考并执行复杂任务的超级工具箱","reasonCn":"Switchyard 是 NVIDIA 开发的多语言 AI 工具链，专为解决大模型“只会说话不会干活”的痛点设计。它支持 Python、JavaScript、Go 等主流语言，提供从工具注册、路由到编排的完整框架。通过定义清晰的函数签名和 Schema，开发者能让模型调用外部 API、读写数据库或执行操作系统命令，实现从“对话”到“行动”的跨越。","stars":3202,"starGrowth":3,"createdAt":"2026-05-19T19:06:33Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["AI基础设施","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["Agent框架","工具调用","多语言SDK"],"domainKey":"Agent框架","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/NVIDIA-NeMo/Switchyard","ts":"2026-08-16T00:54:51.023Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alexjc/neural-doodle","owner":"alexjc","name":"neural-doodle","desc":"","summaryCn":"用神经网络把你的涂鸦变成大师风格画作，秒变艺术家。","reasonCn":"这是一个基于深度学习的绘画工具，能借用梵高、莫奈等艺术家的风格，把你随手画的简笔涂鸦，自动渲染成完整的艺术画作。只需提供风格图、对应的标注图，以及你的涂鸦和标注，程序就会自动融合生成新图。它实现了2016年的语义风格迁移算法，用神经补丁技术让风格迁移更自然。","stars":9850,"starGrowth":0,"createdAt":"2016-03-05T23:35:18Z","silentRounds":8,"language":"Python","topics":["image-processing","image-manipulation","image-generation","deep-learning","deep-neural-networks"],"aiDims":["创意工具","AI应用","模型与训练"],"zone":"创意","funScore":1,"domainTags":["风格迁移","图像生成","创意编程"],"domainKey":"风格迁移","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"image-processing","source":"github","weight":0.7},{"name":"image-manipulation","source":"github","weight":0.7},{"name":"image-generation","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"deep-neural-networks","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/alexjc/neural-doodle","ts":"2026-09-14T23:04:28.868Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"gchq/CyberChef","owner":"gchq","name":"CyberChef","desc":"The Cyber Swiss Army Knife - a web app for encryption, encoding, compression and data analysis","summaryCn":"加密解压像做菜一样简单，网络安全人的瑞士军刀","reasonCn":"这是一个运行在浏览器里的Web应用，提供了几十种数据编码、加密、解密和分析功能。用户可以把数据粘贴进去，选择相应的操作步骤，像搭积木一样处理信息。它支持从简单的Base64解码到复杂的AES加密，无需安装任何软件，非常适合网络安全分析师、数据取证人员和开发者日常使用。","stars":35915,"starGrowth":9,"createdAt":"2016-11-28T10:34:07Z","silentRounds":0,"language":"JavaScript","topics":["data-analysis","data-manipulation","encryption","encoding","compression","parsing","hashing"],"aiDims":["非AI-实用","开发者工具","安全工具"],"zone":"工具","funScore":1,"domainTags":["数据加密","编码解码","网络安全","数据分析"],"domainKey":"数据加密","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"data-analysis","source":"github","weight":0.7},{"name":"data-manipulation","source":"github","weight":0.7},{"name":"encryption","source":"github","weight":0.7},{"name":"encoding","source":"github","weight":0.7},{"name":"compression","source":"github","weight":0.7},{"name":"parsing","source":"github","weight":0.7},{"name":"hashing","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/gchq/CyberChef","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.02775177502436308,"copyOk":false},{"repo":"InsForge/InsForge","owner":"InsForge","name":"InsForge","desc":"The all-in-one, open-source backend platform for agentic coding. InsForge gives your coding agent database, auth, storage, compute, hosting, and AI gateway to ship full-stack apps end-to-end.","summaryCn":"给AI编程助手配个全栈后台，数据库托管认证一键搞定","reasonCn":"InsForge是面向AI编码Agent的一站式后端平台，提供数据库、认证、存储、计算、托管和AI网关。你的AI编程助手只需声明需要什么基础设施，它自动帮你搭建好，无需手动配置AWS或搭建后端服务。基于TypeScript，与前端生态无缝衔接，支持多种AI模型网关接入。","stars":13019,"starGrowth":0,"createdAt":"2025-07-29T05:56:00Z","silentRounds":1,"language":"TypeScript","topics":["ai","ai-agents","coding","oauth2","postgresql","deno","embeddings","insforge","nextjs","pgvector","realtime","vectors","websockets"],"aiDims":["AI Agent","AI基础设施","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["AI编程","后端基础设施","声明式配置","MCP网关","全栈部署"],"domainKey":"AI编程","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"coding","source":"github","weight":0.7},{"name":"oauth2","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"deno","source":"github","weight":0.7},{"name":"embeddings","source":"github","weight":0.7},{"name":"insforge","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"pgvector","source":"github","weight":0.7},{"name":"realtime","source":"github","weight":0.7},{"name":"vectors","source":"github","weight":0.7},{"name":"websockets","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/InsForge/InsForge","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"golbin/hop","owner":"golbin","name":"hop","desc":"","summaryCn":"把大模型变成能看会聊的数字分身，本地部署超好玩","reasonCn":"这是一个基于浏览器运行的AI数字分身生成器，支持本地模型加载和多模态交互，让用户能创建个性化的AI角色并体验沉浸式对话。项目采用TypeScript开发，通过WebAssembly技术实现高性能推理，无需复杂环境配置即可在浏览器中运行。","stars":1830,"starGrowth":0,"createdAt":"2026-04-19T10:49:46Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["AI应用","创意工具","开发者工具"],"zone":"创意","funScore":1,"domainTags":["AI数字分身","浏览器本地推理","多模态交互","WebAssembly","可视化"],"domainKey":"AI数字分身","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/golbin/hop","ts":"2026-08-14T01:23:41.092Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"krillinai/KrillinAI","owner":"krillinai","name":"KrillinAI","desc":"AI video translation &amp; dubbing tool for humans and AI Agents, powered by LLMs. Full pipeline: download, transcribe, translate, TTS dub, reformat, cover generation. 100+ languages, optimized for YouTube, TikTok, Bilibili, Douyin, and more.AI视频翻译配音工具，面向人类与AI Agent，100+语言全链路，CLI分阶段调用，适配抖音、小红书、哔哩哔哩、视频号、TikTok、YouTube","summaryCn":"全能视频翻译器，支持百种语言，一键搞定下载转录翻译配音，像变魔术一样。","reasonCn":"这是一个基于LLM的全流程视频翻译配音工具，覆盖下载、转录、翻译、TTS配音、封面生成等全链路。支持100多种语言互译，利用大模型理解视频内容生成精准的配音文案。开发者可以用Go编写插件扩展功能，完美适配YouTube、B站等主流平台，让视频出海变得极其简单。","stars":12209,"starGrowth":3,"createdAt":"2024-12-17T01:59:22Z","silentRounds":0,"language":"Go","topics":["video-transcription","video-translation","dubbing","localization","tts"],"aiDims":["多模态Agent","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["视频翻译","AI配音","多语言","YouTube","内容出海"],"domainKey":"视频翻译","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"video-transcription","source":"github","weight":0.7},{"name":"video-translation","source":"github","weight":0.7},{"name":"dubbing","source":"github","weight":0.7},{"name":"localization","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/krillinai/KrillinAI","ts":"2026-08-18T00:56:27.274Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kenforthewin/atomic","owner":"kenforthewin","name":"atomic","desc":"Self-hosted, semantically-connected personal knowledge base","summaryCn":"把你的零散笔记连成一张大网，AI帮你梳理知识脉络","reasonCn":"这是一个基于RAG和知识图谱技术构建的个人知识库系统，允许用户通过自然语言构建原子化的知识节点，并自动建立语义关联。项目使用Rust语言开发，具备高性能、内存占用低的特点，支持本地化部署，数据完全掌握在用户手中。它不仅是一个存储工具，更是一个能理解你知识结构的智能伴侣，帮助用户从海量笔记中提取深层洞察。","stars":1959,"starGrowth":1,"createdAt":"2025-11-25T14:18:06Z","silentRounds":13,"language":"Rust","topics":["knowledge-base"],"aiDims":["知识图谱","非AI-实用","AI搜索"],"zone":"工具","funScore":0,"domainTags":["RAG","知识图谱","个人知识库"],"domainKey":"RAG","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"knowledge-base","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/kenforthewin/atomic","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"gildas-lormeau/SingleFile","owner":"gildas-lormeau","name":"SingleFile","desc":"","summaryCn":"SingleFile通过深度集成在浏览器扩展框架中，能够精准抓取网页中的所有资源","reasonCn":"项目的一大核心优势在于其极强的兼容性，该扩展覆盖了几乎所有主流桌面及移动端浏览器，无论你习惯使用Chrome还是Safari，亦或是更小众的Waterfox或Yandex，都能获得无缝的网页存档体验。这种全平台覆盖的策略，让用户无需在不同设备间切换复杂的导出方案，只要有浏览器的地方，就","stars":22461,"starGrowth":3,"createdAt":"2010-09-12T23:50:21Z","silentRounds":0,"language":"JavaScript","topics":["browser","archive","auto-save","chrome","firefox","offline-reading","osint","chrome-extension","firefox-addon","annotations","archiver","web-clipper","self-extracting-zip","self-extracting","singlefile","read-it-later","safari","safari-extension"],"aiDims":["效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网页存档","浏览器扩展","单文件"],"domainKey":"网页存档","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"browser","source":"github","weight":0.7},{"name":"archive","source":"github","weight":0.7},{"name":"auto-save","source":"github","weight":0.7},{"name":"chrome","source":"github","weight":0.7},{"name":"firefox","source":"github","weight":0.7},{"name":"offline-reading","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"chrome-extension","source":"github","weight":0.7},{"name":"firefox-addon","source":"github","weight":0.7},{"name":"annotations","source":"github","weight":0.7},{"name":"archiver","source":"github","weight":0.7},{"name":"web-clipper","source":"github","weight":0.7},{"name":"self-extracting-zip","source":"github","weight":0.7},{"name":"self-extracting","source":"github","weight":0.7},{"name":"singlefile","source":"github","weight":0.7},{"name":"read-it-later","source":"github","weight":0.7},{"name":"safari","source":"github","weight":0.7},{"name":"safari-extension","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.644,"source":"search","url":"https://github.com/gildas-lormeau/SingleFile","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"containerd/containerd","owner":"containerd","name":"containerd","desc":"An open and reliable container runtime","summaryCn":"容器世界的幕后推手，让Docker等容器更稳更快","reasonCn":"containerd是云原生基金会CNCF的毕业项目，作为Kubernetes默认的容器运行时，负责容器的生命周期管理。它采用模块化设计，将容器镜像管理、存储快照、执行和监控等核心功能与编排层解耦，性能优异且易于扩展。是构建AI服务容器化部署的底层基石。","stars":21335,"starGrowth":6,"createdAt":"2015-11-13T00:27:43Z","silentRounds":0,"language":"Go","topics":["docker","kubernetes"],"aiDims":["AI基础设施","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["容器运行时","云原生","Kubernetes","Docker","DevOps"],"domainKey":"容器运行时","pushedAt":"2026-09-23T21:02:06Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/containerd/containerd","ts":"2026-09-23T21:02:06Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01884368408718069,"copyOk":false},{"repo":"LAION-AI/Open-Assistant","owner":"LAION-AI","name":"Open-Assistant","desc":"","summaryCn":"Open Assistant 项目最终成功向公众发布了名为 oasst2 的大规模数据集，这意","reasonCn":"作为曾经的开源对话机器人先驱，Open Assistant 采用了类似于 Stable Diffusion 推动图像生成的逻辑，旨在通过众包的方式优化人类与机器的交互方式。由于项目已经宣告完成，目前用户不再需要进行复杂的本地部署，而是可以直接从 HuggingFace 平台获取其成熟的开源成果进行研究或模","stars":37394,"starGrowth":0,"createdAt":"2022-12-13T05:24:17Z","silentRounds":9,"language":"Python","topics":["chatgpt","language-model","rlhf","ai","assistant","discord-bot","machine-learning","nextjs","python"],"aiDims":["大语言模型","模型与训练"],"zone":"资源","funScore":0,"domainTags":["对话数据集","开源助手","数据集","指令微调","对话机器人"],"domainKey":"对话数据集","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"chatgpt","source":"github","weight":0.7},{"name":"language-model","source":"github","weight":0.7},{"name":"rlhf","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"assistant","source":"github","weight":0.7},{"name":"discord-bot","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/LAION-AI/Open-Assistant","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"coze-dev/coze-studio","owner":"coze-dev","name":"coze-studio","desc":"An AI agent development platform with all-in-one visual tools, simplifying agent creation, debugging, and deployment like never before. Coze your way to AI Agent creation.","summaryCn":"可视化搭建AI助手，不用写代码也能做出会思考的智能体","reasonCn":"技术亮点在于它把复杂的Agent架构抽象成了可视化的工作流节点。你可以拖拽「检索」节点来接入知识库，拖拽「工具调用」节点来连接外部API，拖拽「条件判断」节点来实现复杂逻辑。后台自动处理RAG检索、向量匹配、函数调用等底层细节。","stars":21639,"starGrowth":1,"createdAt":"2025-06-26T02:19:21Z","silentRounds":1,"language":"TypeScript","topics":["rag"],"aiDims":["AI Agent","RAG与知识","多模态Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","可视化开发","RAG","低代码","字节跳动"],"domainKey":"AI Agent","pushedAt":"2026-07-29T03:18:43Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/coze-dev/coze-studio","ts":"2026-07-29T03:18:43Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Accio-org/RealReplicaBench","owner":"Accio-org","name":"RealReplicaBench","desc":"RealReplicaBench: Benchmarking Long-Horizon Agents in High-Fidelity, Stateful, and Reproducible Replicas of Real Online Services","summaryCn":"复刻真实服务环境，给AI智能体做个高难度试炼场","reasonCn":"这是一个专门评估AI智能体在复杂真实场景中表现的开源基准测试平台。它通过高保真模拟真实在线服务，测试智能体在多轮交互、状态保持和任务完成度上的能力。项目提供了丰富的测试案例和评估指标，帮助开发者量化Agent的实战水平。","stars":1258,"starGrowth":0,"createdAt":"2026-08-02T05:38:38Z","silentRounds":2,"language":"HTML","topics":[],"aiDims":["AI Agent","AI基础设施","模型部署"],"zone":"工具","funScore":0,"domainTags":["Agent基准测试","状态机","电商模拟","可复现测试"],"domainKey":"Agent基准测试","pushedAt":"2026-08-11T07:29:29Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.92,"source":"bigbro","url":"https://github.com/Accio-org/RealReplicaBench","ts":"2026-08-11T07:29:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"steven-ahfu/suno-to-ableton","owner":"steven-ahfu","name":"suno-to-ableton","desc":"Turn your Suno AI songs into production-ready Ableton Live sessions. Generates an .als file with grid-aligned, tempo-matched, cleaned up tracks that are ready to remix.","summaryCn":"把人工智能写出来的歌拆成能动手改的素材包","reasonCn":"它把人工智能写好的整首歌拆成可编辑轨道，并自动对齐节拍、匹配速度、清理杂音，最后生成可打开的工程文件。对喜欢翻唱和二次创作的人很友好，也能把粗糙成品变成可混音、可替换人声、可加鼓点的素材包，省掉手动拖拽音轨的烦人步骤，还能保留段落结构方便局部改动。","stars":39,"starGrowth":0,"createdAt":"2026-03-14T16:06:09Z","silentRounds":22,"language":"Python","topics":["music-production"],"aiDims":["AI应用","创意工具","非AI-好玩"],"zone":"工具","funScore":0.7,"domainTags":["音频处理","音乐制作","Suno","节拍对齐"],"domainKey":"音频处理","pushedAt":"2026-09-05T02:53:15Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"music-production","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/steven-ahfu/suno-to-ableton","ts":"2026-09-05T02:53:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"pbienst/active-math","owner":"pbienst","name":"active-math","desc":"Free textbook on graduate-level applied math for science and engineering","summaryCn":"专为理工科研究生打造的免费应用数学教科书","reasonCn":"技术上，整个项目完全基于TeX排版语言编写，内容严谨规范，推导过程极其详细。和普通的数学参考书相比，它更注重工程应用与科学计算的结合，不仅有纯粹的数学理论，还展示了数学工具在实际科研和工程问题中的应用方法，弥补了理论与实操之间的鸿沟。","stars":83,"starGrowth":0,"createdAt":"2020-12-11T15:31:13Z","silentRounds":15,"language":"TeX","topics":["mathematics","textbook"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["应用数学","教科书","科学计算","理工科","TeX排版"],"domainKey":"应用数学","pushedAt":"2026-09-09T07:58:23Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"textbook","source":"github","weight":0.7},{"name":"TeX","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/pbienst/active-math","ts":"2026-09-09T07:58:23Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"UfoMiao/zcf","owner":"UfoMiao","name":"zcf","desc":"Zero-Config Code Flow for Claude code & Codex","summaryCn":"无需复杂配置就能让 Claude 丝滑写代码的神器","reasonCn":"它在技术上重点优化了代码流的传输机制。通过预设的高效工作流，它能将 Claude 等模型的强大代码生成能力与本地文件系统更紧密地结合。它实际上是在 AI 和你的代码库之间搭建了一层透明的协议，自动处理上下文的裁剪和增强，确保 AI 在生成代码时拥有最正确的实时视角，避免产生幻","stars":6080,"starGrowth":0,"createdAt":"2025-07-30T06:09:00Z","silentRounds":11,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","代码助手","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI编程","代码流","Claude","Agent","零配置"],"domainKey":"Agent","pushedAt":"2026-08-31T03:02:03Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/UfoMiao/zcf","ts":"2026-08-31T03:02:03Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"goatplatform/goatdb","owner":"goatplatform","name":"goatdb","desc":"A peer-to-peer database that spans devices. For apps and agents that work everywhere.","summaryCn":"像聊天一样共享数据的分布式数据库，全家桶都能用。","reasonCn":"这是一个去中心化的数据库项目，支持跨设备实时同步数据，底层采用P2P技术，无需服务器。它提供完整的数据库功能，包括数据类型定义、查询接口，并兼容多种编程语言。项目旨在解决数据孤岛问题，让应用和AI代理能在任何设备上无缝访问数据，无需依赖中心化服务器。","stars":579,"starGrowth":0,"createdAt":"2024-12-03T12:20:41Z","silentRounds":23,"language":"TypeScript","topics":["backend"],"aiDims":["AI基础设施","非AI-实用","工具调用"],"zone":"工具","funScore":0.6,"domainTags":["P2P数据库","分布式存储","数据同步","去中心化","TypeScript"],"domainKey":"P2P数据库","pushedAt":"2026-09-01T14:29:32Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/goatplatform/goatdb","ts":"2026-09-01T14:29:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"exo-explore/exo","owner":"exo-explore","name":"exo","desc":"Run frontier AI locally.","summaryCn":"把超强大模型装进电脑，本地跑AI的神操作","reasonCn":"这是一个专为本地运行前沿大模型设计的运行时，支持包括Llama 3、Gemma在内的数百种模型，通过高度优化的多后端架构（CPU、CUDA、Metal等）实现高效推理。它提供了类似ChatGPT的交互界面，并允许用户通过Python API或CLI灵活调用，让本地AI变得像运行普通程序一样简单。","stars":47609,"starGrowth":4,"createdAt":"2024-06-24T18:36:22Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型部署","AI基础设施","开发工具"],"zone":"AI","funScore":0.5,"domainTags":["本地大模型","模型运行时","LLM推理","多后端支持","本地AI"],"domainKey":"本地大模型","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"trending","url":"https://github.com/exo-explore/exo","ts":"2026-08-15T00:57:24.351Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rust-lang/rust-analyzer","owner":"rust-lang","name":"rust-analyzer","desc":"A Rust compiler front-end for IDEs","summaryCn":"rust-analyzer 实现了 Language Server Protocol 协议，能够无缝接入 VS Co","reasonCn":"在具体功能体验上，该项目为开发者提供了涵盖代码编写全生命周期的智能化辅助手段。它不仅包含精准的跳转到定义、全局查找引用以及安全重构建议，还能实现基于上下文的智能代码补全。这不仅显著减少了频繁翻阅官方文档的时间，更降低了在大型复杂项目中手动追踪变量与函数调用链的","stars":16866,"starGrowth":1,"createdAt":"2017-12-21T19:24:42Z","silentRounds":2,"language":"Rust","topics":["lsp-server","rust","hacktoberfest"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Rust语言服务器","代码补全","IDE集成","重构工具","类型检查"],"domainKey":"Rust语言服务器","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"lsp-server","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/rust-lang/rust-analyzer","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LAMDA-CL/Prism","owner":"LAMDA-CL","name":"Prism","desc":"Prism: A Plug-in Reproducible Infrastructure for Scalable Multimodal Continual Instruction Tuning","summaryCn":"多模态大模型微调的插件化基础设施，让指令调优像搭积木一样简单。","reasonCn":"Prism 提供了一套可插拔的模块化架构，旨在解决多模态大模型持续指令微调的工程难题。它支持灵活扩展的插件化设计，让研究人员能快速组合数据预处理、训练和评估组件。该工具集成了自动数据清洗、评估基准和高效的训练流水线，显著降低了多模态模型优化的门槛。","stars":44,"starGrowth":0,"createdAt":"2026-03-12T09:30:05Z","silentRounds":16,"language":"Python","topics":["llm-model"],"aiDims":["多模态Agent","微调","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["多模态","指令微调","模型训练"],"domainKey":"多模态","pushedAt":"2026-08-27T03:03:29Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/LAMDA-CL/Prism","ts":"2026-08-27T03:03:29Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liebaojun/MakoCode","owner":"liebaojun","name":"MakoCode","desc":"MakoCode — 茉子版 Agent | Galgame 风格桌面 AI Agent，零门槛体验 Agent 的乐趣","summaryCn":"把AI Agent变成Galgame女主，桌面互动超有爱","reasonCn":"这是一个基于Web技术的Galgame风格桌面AI Agent项目，主打零门槛体验。它将AI对话能力与二次元视觉风格结合，提供了一个沉浸式的桌面互动环境。用户可以与虚拟角色进行实时对话、互动，体验像玩Galgame一样的AI陪伴感，界面完全本地化。","stars":153,"starGrowth":0,"createdAt":"2026-06-16T13:56:01Z","silentRounds":2,"language":"HTML","topics":["anime"],"aiDims":["AI应用","非AI-好玩"],"zone":"AI","funScore":1,"domainTags":["Galgame","AI伴侣","二次元","桌面交互"],"domainKey":"Galgame","pushedAt":"2026-08-14T13:13:01Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/liebaojun/MakoCode","ts":"2026-08-14T13:13:01Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rizwansoaib/passport-photo-generator","owner":"rizwansoaib","name":"passport-photo-generator","desc":"Create professional passport photos instantly - 100% Free, Offline & Private! 🎨","summaryCn":"想象一下，急需证件照却又来不及去照相馆，或者是担心把自己的高清肖像上传","reasonCn":"技术上，它利用了浏览器原生的Canvas API和现代前端图像处理技术。通过纯客户端的算法，它能实现照片的快速裁剪、缩放，并内置了多种主流的证件规格（如2寸、护照尺寸等）。因为所有的计算逻辑都运行在你的浏览器内存中，并没有数据上传到云端，这意味着你的照片永远不会离开你的设","stars":14,"starGrowth":0,"createdAt":"2025-10-26T09:28:49Z","silentRounds":17,"language":"JavaScript","topics":["photography"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["证件照生成","前端图像处理","Canvas API","隐私保护","照片排版"],"domainKey":"证件照生成","pushedAt":"2026-09-09T05:44:04Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"photography","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/rizwansoaib/passport-photo-generator","ts":"2026-09-09T05:44:04Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"walkinglabs/learn-harness-engineering","owner":"walkinglabs","name":"learn-harness-engineering","desc":"Harness engineering beginner tutorial, from 0 to 1","summaryCn":"从零学习AI Harness工程，掌握智能体自动化部署的全套实战技能","reasonCn":"这是一个手把手教你掌握AI Harness工程的学习资源，从零基础到构建完整的AI Agent自动化工作流。内容涵盖环境配置、依赖管理、测试到生产部署的全链路，通过实践项目让你真正理解如何让AI Agent在复杂场景中稳定运行。1.1万star证明它对AI Agent开发者有巨大吸引力。","stars":15721,"starGrowth":0,"createdAt":"2026-03-29T12:28:19Z","silentRounds":1,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["AI Agent","教程","DevOps","部署","自动化"],"domainKey":"AI Agent","pushedAt":"2026-08-26T01:08:59Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/walkinglabs/learn-harness-engineering","ts":"2026-08-26T01:08:59Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bradautomates/claude-video","owner":"bradautomates","name":"claude-video","desc":"","summaryCn":"若视频没有字幕，需要提供 Whisper API 密钥，这意味着即使是纯音频或无字幕","reasonCn":"项目采用零配置设计，首次运行时会自动通过 Homebrew 在 macOS 上安装 yt‑dlp 与 ffmpeg，Linux 与 Windows 用户会在终端中看到相应的安装命令提示，无需额外准备环境。下载后，工具会利用 yt‑dlp 把视频流保存为本地文件，再交给 ffmpeg 进行解码和帧提取，为后续的字幕或音频处理","stars":17540,"starGrowth":9,"createdAt":"2026-04-24T04:40:43Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["视频理解","Whisper","多模态"],"domainKey":"多模态","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.638,"source":"search","url":"https://github.com/bradautomates/claude-video","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.03424720638540479,"copyOk":false},{"repo":"google-deepmind/weathernext","owner":"google-deepmind","name":"weathernext","desc":"","summaryCn":"Google用AI秒出十天天气预报，比传统方法快百倍","reasonCn":"基于图神经网络处理地球球面气象数据，在高分辨率网格上直接输出未来十天预报。部署在Google Cloud，通过REST API调用，几秒钟返回全球气象预测结果。相比传统数值天气预报需要超级计算机跑几小时，这套方案速度提升百倍，成本大幅降低，适合科研、农业、航运等多个场景使用。","stars":7688,"starGrowth":0,"createdAt":"2023-07-14T11:07:57Z","silentRounds":1,"language":"Python","topics":["weather"],"aiDims":["AI应用","非AI-实用"],"zone":"AI","funScore":0.7,"domainTags":["气象预报","图神经网络","天气预测","AI 科学"],"domainKey":"气象预报","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"weather","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/google-deepmind/weathernext","ts":"2026-09-05T02:38:58.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"assafelovic/gpt-researcher","owner":"assafelovic","name":"gpt-researcher","desc":"An autonomous agent that conducts deep research on any data using any LLM providers","summaryCn":"给它一个题目，它就能像侦探一样全网搜集情报写报告","reasonCn":"GPT Researcher是一个自主AI研究代理，能自动通过网络搜索收集信息、整理数据并生成详尽的报告。支持多种LLM提供商，可通过API或本地运行。它将复杂的调研任务自动化，适合需要快速信息汇总的场景，是提升知识工作者效率的强力助手。","stars":29571,"starGrowth":2,"createdAt":"2023-05-12T10:33:54Z","silentRounds":0,"language":"Python","topics":["automation","ai"],"aiDims":["AI Agent","AI搜索","自动化"],"zone":"AI","funScore":0.7,"domainTags":["AI研究","自主Agent","网络爬虫","报告生成","LLM"],"domainKey":"AI研究","pushedAt":"2026-08-27T19:34:56Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"自动化","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/assafelovic/gpt-researcher","ts":"2026-08-27T19:34:56Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"modelcontextprotocol/typescript-sdk","owner":"modelcontextprotocol","name":"typescript-sdk","desc":"The official TypeScript SDK for Model Context Protocol servers and clients","summaryCn":"连接AI与工具的通用协议，让你的大模型能像人一样熟练使用各种软件。","reasonCn":"这是Model Context Protocol（MCP）的官方TypeScript SDK，定义了AI应用与外部数据源及工具之间的通信标准。通过这套协议，开发者可以轻松地将大语言模型连接到本地文件、数据库或API，让模型具备调用外部资源的能力，从而构建更强大的AI助手。","stars":13442,"starGrowth":1,"createdAt":"2024-09-24T20:45:31Z","silentRounds":2,"language":"TypeScript","topics":["typescript","mcp","mcp-server","mcp-client"],"aiDims":["AI基础设施","开发者工具","AI应用"],"zone":"AI","funScore":0,"domainTags":["Model Context Protocol","AI集成","TypeScript","工具调用","协议标准"],"domainKey":"Model Context Protocol","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"typescript","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"mcp-client","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/modelcontextprotocol/typescript-sdk","ts":"2026-08-12T01:22:42.342Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kdash-rs/kdash","owner":"kdash-rs","name":"kdash","desc":"A simple and fast dashboard for Kubernetes","summaryCn":"用键盘就能在终端顺滑查看和管理Kubernetes集群的高效面板。","reasonCn":"在技术实现上，它完全用Rust语言开发，这意味着它有着极佳的运行性能和极低的内存占用。它直接通过Kubernetes的官方API与集群进行安全通信，不仅渲染速度飞快，而且在处理大量Pod和高并发日志流时也不会出现卡顿。其界面采用现代化的TUI设计，布局清晰，对多集群切换的支持也非常平","stars":2540,"starGrowth":0,"createdAt":"2021-04-13T05:38:42Z","silentRounds":6,"language":"Rust","topics":["monitoring"],"aiDims":["可视化","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","TUI","监控"],"domainKey":"Kubernetes","pushedAt":"2026-09-09T00:22:52Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"monitoring","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/kdash-rs/kdash","ts":"2026-09-09T00:22:52Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cloudflare/computer","owner":"cloudflare","name":"computer","desc":"Give your agent a computer 👾","summaryCn":"给AI agent配一台虚拟电脑，让它能动手操作，不再只是动动嘴皮子。","reasonCn":"这是Cloudflare推出的基础设施项目，专门为AI Agent提供可交互的浏览器环境。Agent不再只能读文本或调API，而是能像人一样打开网页、点击按钮、填写表单。它集成了Cloudflare Workers AI，让本地或云端部署的Agent都能低成本获得“手脚”，极大扩展了自动化场景的上限。","stars":9249,"starGrowth":2,"createdAt":"2026-06-05T10:38:15Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用","非AI-实用"],"zone":"AI","funScore":1,"domainTags":["AI Agent","浏览器自动化","无头浏览器","虚拟桌面","边缘计算"],"domainKey":"AI Agent","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/cloudflare/computer","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"unliftedq/imagent","owner":"unliftedq","name":"imagent","desc":"Imagine + Agent","summaryCn":"通过艺术创作与AI的完美结合，创造出独一无二的艺术作品。","reasonCn":"这个项目通过使用AI生成的艺术，为用户提供了一个全新的创作体验，无需专业的绘画技巧或复杂的软件工具。用户可以简单地输入他们的想法，然后让AI生成相应的图像。这种技术不仅能够提供个性化的艺术作品，还能激发用户的创造力，让他们以全新的方式表达自己的想法。","stars":19,"starGrowth":0,"createdAt":"2026-04-28T09:27:16Z","silentRounds":20,"language":"TypeScript","topics":["generative-art"],"aiDims":["AI基础设施"],"zone":"创意","funScore":1,"domainTags":["AI艺术生成","创意编程","视觉艺术"],"domainKey":"AI艺术生成","pushedAt":"2026-08-13T09:18:24Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"generative-art","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/unliftedq/imagent","ts":"2026-08-13T09:18:24Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"onlook-dev/onlook","owner":"onlook-dev","name":"onlook","desc":"The Cursor for Designers • An Open-Source AI-First Design tool • Visually build, style, and edit your React App with AI","summaryCn":"用AI读懂你的设计稿，一键生成可编辑的React组件代码","reasonCn":"Onlook是一款AI优先的开源设计工具，能够将设计稿智能识别并转换为可编辑的React组件代码。设计师无需写代码即可完成前端开发，开发者也能通过AI加速UI实现。支持导入Figma等设计文件，用自然语言指令直接修改样式和布局，大幅缩短设计与开发之间的沟通成本。","stars":26791,"starGrowth":0,"createdAt":"2024-06-25T19:16:02Z","silentRounds":1,"language":"TypeScript","topics":["design","frontend"],"aiDims":["AI应用","代码助手","创意工具"],"zone":"AI","funScore":1,"domainTags":["AI设计","React生成","前端开发"],"domainKey":"AI设计","pushedAt":"2026-08-25T01:06:22Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/onlook-dev/onlook","ts":"2026-08-25T01:06:22Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yiliqi78/memory-work","owner":"yiliqi78","name":"memory-work","desc":"An AI-first knowledge management system with evolvable memory","summaryCn":"能自我进化的AI知识库，让你的笔记和知识自动成长","reasonCn":"这是一个AI优先的知识管理系统，核心能力在于让记忆和知识能够自我进化。它通过向量检索、语义理解等技术，帮助你和AI共同构建一个动态成长的知识库，不再只是静态存储，而是能随着使用越来越聪明，真正实现知识的自我迭代和关联发现。","stars":136,"starGrowth":0,"createdAt":"2026-02-09T13:06:07Z","silentRounds":27,"language":"","topics":[],"aiDims":["AI应用","RAG与知识"],"zone":"AI","funScore":0.6,"domainTags":["知识管理","向量检索","笔记系统"],"domainKey":"知识管理","pushedAt":"2026-02-19T06:57:11Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1}],"aiScore":0.88,"source":"bigbro","url":"https://github.com/yiliqi78/memory-work","ts":"2026-02-19T06:57:11Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Cinnamon/kotaemon","owner":"Cinnamon","name":"kotaemon","desc":"An open-source RAG-based tool for chatting with your documents.","summaryCn":"把你的文档喂给AI，随时随地用对话的方式检索和问答","reasonCn":"这是一个基于RAG技术的开源对话工具，支持PDF、Markdown、JSON等多种常见文档格式。内置向量数据库，能够将本地文档自动转换成可对话的知识库，用AI快速从海量文档中提取关键信息，支持多轮对话与上下文记忆，无需再手动翻阅检索，让信息获取变得像聊天一样自然顺畅。","stars":25780,"starGrowth":0,"createdAt":"2024-03-25T08:16:42Z","silentRounds":1,"language":"Python","topics":["rag"],"aiDims":["RAG与知识","AI应用","向量数据库"],"zone":"AI","funScore":0.6,"domainTags":["RAG","文档检索","对话系统"],"domainKey":"RAG","pushedAt":"2026-07-14T16:53:59Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"rag","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/Cinnamon/kotaemon","ts":"2026-07-14T16:53:59Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"maotoumao/MusicFree","owner":"maotoumao","name":"MusicFree","desc":"插件化、定制化、无广告的免费音乐播放器","summaryCn":"彻底摆脱广告，还能像搭积木一样定制功能的免费音乐播放器。","reasonCn":"基于 Flutter 开发的跨平台播放器，支持iOS和Android。它采用插件架构，允许用户自由替换源和解析器。内置广告拦截、自动播放列表和本地缓存功能，完美替代付费音乐App。源码开源，支持二次开发，是音乐发烧友和极客的福音。","stars":27040,"starGrowth":1,"createdAt":"2022-08-22T14:29:32Z","silentRounds":1,"language":"TypeScript","topics":["free","music-player","plugin","react","react-native","typescript","musicfree"],"aiDims":["AI应用","非AI-实用"],"zone":"创意","funScore":0.9,"domainTags":["音乐播放","插件化","Flutter","跨平台","开源"],"domainKey":"音乐播放","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"free","source":"github","weight":0.7},{"name":"music-player","source":"github","weight":0.7},{"name":"plugin","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"react-native","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"musicfree","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/maotoumao/MusicFree","ts":"2026-09-09T22:18:37.171Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vercel-labs/deepsec","owner":"vercel-labs","name":"deepsec","desc":"Deepsec is a security harness for finding vulnerabilities in your codebase powered by coding agents","summaryCn":"AI安全小助手，自动帮你找代码里的漏洞。","reasonCn":"这是一个利用AI Agent技术构建的安全测试框架，能自动扫描代码库中的潜在漏洞。它通过自然语言交互，让用户可以像对话一样描述安全问题，Agent会自动分析代码并生成修复建议。支持多种编程语言，能有效提升安全测试的效率。","stars":8020,"starGrowth":2,"createdAt":"2026-04-30T16:16:11Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["安全工具","AI Agent","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["安全扫描","AI Agent","代码审计","漏洞检测","网络安全"],"domainKey":"AI Agent","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/vercel-labs/deepsec","ts":"2026-08-16T00:54:51.023Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"BoundaryML/baml","owner":"BoundaryML","name":"baml","desc":"The programming language for agents","summaryCn":"让AI智能体自己写代码的神器，开发效率直接起飞","reasonCn":"这是一个基于Rust开发的DSL编程语言，专门用于构建AI智能体。它通过强大的类型系统和编译期检查，让开发者能像写普通代码一样定义AI行为，支持复杂的工作流编排。相比传统Prompt工程，它能生成更稳定、可维护的智能体代码。","stars":9263,"starGrowth":3,"createdAt":"2023-10-06T18:57:41Z","silentRounds":0,"language":"Rust","topics":["llm","boundaryml","guardrails","structured-data","programming-language"],"aiDims":["AI Agent","AI基础设施","代码助手"],"zone":"AI","funScore":0.5,"domainTags":["AI智能体","Rust编程语言","DSL"],"domainKey":"AI智能体","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"boundaryml","source":"github","weight":0.7},{"name":"guardrails","source":"github","weight":0.7},{"name":"structured-data","source":"github","weight":0.7},{"name":"programming-language","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/BoundaryML/baml","ts":"2026-08-15T00:57:24.351Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"paperclipai/paperclip","owner":"paperclipai","name":"paperclip","desc":"The open-source app everyone uses to manage agents at work","summaryCn":"让AI打工人在你的桌面上各司其职的调度中心","reasonCn":"Paperclip是一款开源的AI Agent管理平台，支持在可视化界面中编排多个专业Agent协作完成任务。采用TypeScript开发，本地优先架构，内置流程图编辑器可将不同Agent串联成完整工作流。支持OpenAI、Anthropic等主流模型，每个Agent可配置专属角色、工具权限和记忆管理，通过消息队列实现无缝协作。","stars":81298,"starGrowth":1,"createdAt":"2026-03-02T15:01:51Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["AI Agent","AI应用","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","可视化开发","低代码","多模型集成","工作流编排"],"domainKey":"AI Agent","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/paperclipai/paperclip","ts":"2026-08-14T01:23:41.092Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"claritylab/lucida","owner":"claritylab","name":"lucida","desc":"","summaryCn":"Lucida 后端服务共集成了七种不同类别的智能功能，涵盖了从自动语音识别到图","reasonCn":"项目的核心定位是一个受科幻作品《Sirius》启发的智能个人助理，这种设定赋予了它超越传统工具软件的探索意义。通过语音和视觉的双重交互方式，Lucida 试图打破传统命令行或单一界面的限制，让用户能够以更自然的方式与机器进行沟通和互动。","stars":4783,"starGrowth":0,"createdAt":"2014-05-13T17:12:49Z","silentRounds":6,"language":"Java","topics":[],"aiDims":["多模态Agent","AI应用","非AI-实用"],"zone":"工具","funScore":0.2,"domainTags":["语音识别","图像匹配","问答系统","人脸识别","日历事件"],"domainKey":"语音识别","pushedAt":"2026-09-16T22:54:48.334Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.59,"source":"tier","url":"https://github.com/claritylab/lucida","ts":"2026-09-16T22:54:48.334Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AntigmaLabs/ante","owner":"AntigmaLabs","name":"ante","desc":"Ghost in your shell. Ante is a self-contained agent harness with a highly optimized core. It works like Claude Code or Codex, with none of their dependencies or model constraints.","summaryCn":"轻量级AI代理框架，像Claude Code一样本地运行","reasonCn":"这是一个用Rust编写的自包含AI代理框架，旨在提供类似Claude Code的高效代码代理能力，但无需任何外部依赖或模型限制。它采用高度优化的核心设计，让用户能够在本地轻松部署和运行强大的AI编程助手，享受高性能和低延迟的体验。","stars":1974,"starGrowth":1,"createdAt":"2025-12-23T23:42:44Z","silentRounds":1,"language":"Rust","topics":["ai-agents","rust","agent","agentic-ai","artifical-life","terminal"],"aiDims":["AI Agent","推理引擎","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["AI代理框架","Rust","本地推理","Llama","开源模型"],"domainKey":"AI代理框架","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ai-agents","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"artifical-life","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/AntigmaLabs/ante","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mukul975/cve-mcp-server","owner":"mukul975","name":"cve-mcp-server","desc":"Production-grade MCP server giving Claude 27 security intelligence tools across 21 APIs — CVE lookup, EPSS scoring, CISA KEV, MITRE ATT&amp;CK, Shodan, VirusTotal, and more.","summaryCn":"一键查询漏洞信息并自动评分，集成多种安全情报API的AI助手。","reasonCn":"项目基于Claude 27实现MCP服务器，统一提供CVE查询、EPSS评分、MITRE ATT&CK、Shodan等安全情报API，支持高并发RESTful调用并可自定义规则，帮助安全团队快速获取漏洞详情与风险评估，自动生成报告，显著降低人工查询成本。","stars":1581,"starGrowth":0,"createdAt":"2026-04-14T11:20:00Z","silentRounds":2,"language":"Python","topics":["cisa-kev","claude-ai","cve","cybersecurity","devsecops","epss","fastmcp","mcp","mitre-attack","model-context-protocol","nvd","osv","python","security","shodan","threat-intelligence","virustotal","vulnerability-management"],"aiDims":["AI应用","工具调用","安全工具"],"zone":"AI","funScore":0,"domainTags":["CVE查询","安全情报","漏洞评估","MCP协议"],"domainKey":"CVE查询","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"cisa-kev","source":"github","weight":0.7},{"name":"claude-ai","source":"github","weight":0.7},{"name":"cve","source":"github","weight":0.7},{"name":"cybersecurity","source":"github","weight":0.7},{"name":"devsecops","source":"github","weight":0.7},{"name":"epss","source":"github","weight":0.7},{"name":"fastmcp","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mitre-attack","source":"github","weight":0.7},{"name":"model-context-protocol","source":"github","weight":0.7},{"name":"nvd","source":"github","weight":0.7},{"name":"osv","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"shodan","source":"github","weight":0.7},{"name":"threat-intelligence","source":"github","weight":0.7},{"name":"virustotal","source":"github","weight":0.7},{"name":"vulnerability-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/mukul975/cve-mcp-server","ts":"2026-09-07T22:38:09.240Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rivet-dev/rivet","owner":"rivet-dev","name":"rivet","desc":"Rivet Actors are the primitive for stateful workloads. Built for AI agents, collaborative apps, and durable execution.","summaryCn":"为AI Agent设计的状态管理原语，支持持久执行和协作","reasonCn":"核心技术是Actor模型，每个Agent作为独立Actor运行，拥有自己的状态。支持状态持久化，即使进程崩溃也能恢复。错误处理机制确保单点失败不影响整体。还能跨Actor通信，实现多Agent协作。相比通用分布式框架，它更专注于AI工作负载，API设计也更贴合Agent开发习惯。","stars":6160,"starGrowth":2,"createdAt":"2023-06-16T11:51:15Z","silentRounds":0,"language":"Rust","topics":["multiplayer","rust","actor","actors","cloudflare-durable-objects","durable-objects","foundationdb","realtime","serverless","cloudflare"],"aiDims":["AI Agent","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["Actor模型","AI Agent","状态管理","持久化执行","Rust"],"domainKey":"AI Agent","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"multiplayer","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"actor","source":"github","weight":0.7},{"name":"actors","source":"github","weight":0.7},{"name":"cloudflare-durable-objects","source":"github","weight":0.7},{"name":"durable-objects","source":"github","weight":0.7},{"name":"foundationdb","source":"github","weight":0.7},{"name":"realtime","source":"github","weight":0.7},{"name":"serverless","source":"github","weight":0.7},{"name":"cloudflare","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/rivet-dev/rivet","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"appwrite/appwrite","owner":"appwrite","name":"appwrite","desc":"Appwrite® - complete cloud infrastructure for your web, mobile and AI apps. Including Auth, Databases, Storage, Functions, Messaging, Hosting, Realtime and more","summaryCn":"自托管后端平台，一键搞定认证数据库存储函数","reasonCn":"Appwrite是开源的后端即服务平台，提供认证、数据库、存储、云函数、实时消息等全套后端能力。支持Docker一键部署，比Firebase更自由可控。前后端分离架构，REST和GraphQL双接口支持，是独立开发者和创业团队的后端神器。","stars":57439,"starGrowth":4,"createdAt":"2019-04-08T16:36:25Z","silentRounds":0,"language":"TypeScript","topics":["self-hosted","backend"],"aiDims":["非AI-实用","开发者工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["后端即服务","自托管","认证数据库"],"domainKey":"后端即服务","pushedAt":"2026-09-23T21:27:00Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"self-hosted","source":"github","weight":0.7},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/appwrite/appwrite","ts":"2026-09-23T21:27:00Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Unitech/pm2","owner":"Unitech","name":"pm2","desc":"Node.js/Typescript/Bun Production Process Manager with a built-in Load Balancer.","summaryCn":"Node.js 进程管家，保活神器，负载均衡一把抓。","reasonCn":"PM2 是一个专为 Node.js 应用设计的生产环境进程管理器。它不仅能帮你保持应用在后台持续运行，防止意外崩溃，还内置了负载均衡功能，可以同时管理多个应用实例，大大提高了服务的并发处理能力和稳定性。它支持自动重启、日志管理和集群模式，是 Node.js 开发者的必备工具。","stars":43295,"starGrowth":0,"createdAt":"2013-05-21T03:25:25Z","silentRounds":4,"language":"JavaScript","topics":["command-line","monitoring"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["进程管理","负载均衡","Node.js","服务监控","自动化"],"domainKey":"进程管理","pushedAt":"2026-09-04T23:39:51Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Unitech/pm2","ts":"2026-09-04T23:39:51Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"focus-creative-games/luban","owner":"focus-creative-games","name":"luban","desc":"luban是一个强大、易用、优雅、稳定的游戏配置解决方案。luban is a powerful, easy-to-use, elegant and stable game configuration solution. ","summaryCn":"游戏策划的配置神器，让数值表自动变成代码","reasonCn":"Luban是专 для游戏开发的配置管理工具，支持Excel、Lua等格式输入，自动生成C#、Java、Python等多语言代码。它解决了游戏策划频繁改配置、程序手动同步的痛点，支持protobuf、flatbuffers等序列化方案，大幅提升配置流程效率。","stars":4594,"starGrowth":0,"createdAt":"2020-10-21T12:42:41Z","silentRounds":1,"language":"C#","topics":["game"],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["游戏开发","配置管理","代码生成","游戏数值"],"domainKey":"游戏开发","pushedAt":"2026-09-10T12:04:13Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/focus-creative-games/luban","ts":"2026-09-10T12:04:13Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"EmergenceAI/Emergence-World","owner":"EmergenceAI","name":"Emergence-World","desc":"Emergence World: A world designed to reveal what no benchmark can: emergent intelligence.","summaryCn":"用仿真世界观察AI智能涌现，不是跑分而是看智能自然形成","reasonCn":"技术架构上，这是一个仿真环境，提供了丰富的交互对象和动态规则。AI 代理在这个世界里需要完成各种任务，但不是通过预设规则，而是在探索和试错中发展出策略。系统会记录代理的行为轨迹，研究者可以观察哪些行为是涌现出来的、哪些是预设的，从而理解智能形成的机制。","stars":621,"starGrowth":2,"createdAt":"2026-03-24T20:01:08Z","silentRounds":0,"language":"","topics":["simulation"],"aiDims":["AI基础设施","模型与训练"],"zone":"AI","funScore":1,"domainTags":["AI研究","涌现智能","多智能体","仿真环境"],"domainKey":"AI研究","pushedAt":"2026-09-15T16:44:42Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7}],"aiScore":0.6,"source":"search","url":"https://github.com/EmergenceAI/Emergence-World","ts":"2026-09-15T16:44:42Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Awaker-OTE/WakeGPT","owner":"Awaker-OTE","name":"WakeGPT","desc":"A local-first desktop app for quick capture, managed Markdown notebooks, and reusable ChatGPT/Codex context. Currently verified on macOS.","summaryCn":"主打本地优先的桌面端AI快收与知识管理神器","reasonCn":"在技术架构上，它用Rust语言编写，这意味着它有着极佳的性能表现和极低的内存消耗，启动几乎是秒开。数据完全保存在本地用户的磁盘上，不用担心隐私泄露问题。它深度集成了大模型的上下文管理机制，能够把你的零散笔记智能打包，方便随时作为Prompt发给AI，让AI真正懂你的项目背景","stars":165,"starGrowth":0,"createdAt":"2026-08-13T10:02:33Z","silentRounds":18,"language":"Rust","topics":["note-taking"],"aiDims":["AI应用","效率工具","代码助手"],"zone":"AI","funScore":0.6,"domainTags":["本地AI","Markdown笔记","Claude集成"],"domainKey":"本地AI","pushedAt":"2026-08-15T13:23:36Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.91,"source":"search","url":"https://github.com/Awaker-OTE/WakeGPT","ts":"2026-08-15T13:23:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Asabeneh/30-Days-Of-Python","owner":"Asabeneh","name":"30-Days-Of-Python","desc":"The 30 Days of Python programming challenge is a step-by-step guide to learn the Python programming language in 30 days. This challenge may take more than 100 days. Follow your own pace. These videos may help too: https://www.youtube.com/channel/UC7PNRuno1rzYPb1xLa4yktw","summaryCn":"零基础也能学会的Python入门神课，手把手带你从菜鸟变大神","reasonCn":"这是一个超火的编程挑战计划，专为初学者设计，通过30天循序渐进的学习路径，帮助你在30天内掌握Python核心语法与实战技巧。内容覆盖变量、循环、函数、类等基础，并延伸至文件操作、数据库连接等进阶应用。作者提供了代码示例、练习题及配套视频，支持自主学习。","stars":74415,"starGrowth":24,"createdAt":"2019-11-19T17:24:16Z","silentRounds":0,"language":"Python","topics":["database","data-science"],"aiDims":["学习资源","教程","代码助手"],"zone":"资源","funScore":0,"domainTags":["Python教程","编程入门","自学","学习路径","数据科学"],"domainKey":"Python教程","pushedAt":"2026-09-11T22:17:29.879Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/Asabeneh/30-Days-Of-Python","ts":"2026-09-11T22:17:29.879Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.07296754686555129,"copyOk":false},{"repo":"fxsjy/jieba","owner":"fxsjy","name":"jieba","desc":"","summaryCn":"结巴分词是高效的中文文本切分工具，支持多种模式与深度学习模型。","reasonCn":"作为 Python 中知名的中文文本处理工具，该项目旨在解决中文缺乏自然空格分隔带来的词界判定难题。它内置了精确切分、全切分和搜索引擎召回优化等多种切分模式，并支持繁体分词与自定义词典。无论是构建搜索引擎索引还是预处理文本数据，它都能提供兼顾高效与准确的切分能力。","stars":35165,"starGrowth":0,"createdAt":"2012-09-29T07:52:01Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["开发者工具","非AI-实用","文档理解"],"zone":"工具","funScore":0,"domainTags":["中文分词","NLP","文本处理"],"domainKey":"中文分词","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.656,"source":"search","url":"https://github.com/fxsjy/jieba","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"masterFoad/NanoSage","owner":"masterFoad","name":"NanoSage","desc":"Local LLM Powered Recursive Search & Smart Knowledge Explorer","summaryCn":"结合本地大模型的递归搜索工具，帮你像侦探一样深度挖掘海量知识。","reasonCn":"技术上，它核心采用了Python编写的递归搜索逻辑，能够无缝对接各种本地大模型。通过精妙的算法设计，它将传统的检索增强生成与递归推理链条结合起来，不仅能做简单的关键词匹配，还能理解上下文并在多轮迭代中自动修正搜索方向，极大地提升了知识探索的深度。","stars":266,"starGrowth":0,"createdAt":"2025-02-09T17:45:29Z","silentRounds":18,"language":"Python","topics":["algorithms"],"aiDims":["AI Agent","AI搜索","RAG与知识"],"zone":"AI","funScore":0.6,"domainTags":["本地LLM","递归搜索","RAG","知识探索"],"domainKey":"RAG","pushedAt":"2026-08-29T15:33:25Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"algorithms","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/masterFoad/NanoSage","ts":"2026-08-29T15:33:25Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MariaDB/server","owner":"MariaDB","name":"server","desc":"MariaDB server is a community developed fork of MySQL server. Started by core members of the original MySQL team, MariaDB actively works with outside developers to deliver the most featureful, stable, and sanely licensed open SQL server in the industry.","summaryCn":"老牌数据库 MySQL 的开源亲兄弟，比 MySQL 更自由更强大","reasonCn":"MariaDB 是 MySQL 创始人团队打造的社区分支，兼容 MySQL 协议，支持更多存储引擎和高级特性。内置 Galera 集群实现同步复制，提供 Transparent Data Encryption 透明加密、ZFS 存储引擎和列式存储 ColumnStore 等企业级功能，性能在多种场景下优于 MySQL。","stars":8262,"starGrowth":3,"createdAt":"2014-05-15T10:58:50Z","silentRounds":0,"language":"C++","topics":["vector-db"],"aiDims":["非AI-实用","向量数据库","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据库","MySQL","MariaDB","存储引擎","高可用"],"domainKey":"数据库","pushedAt":"2026-09-23T21:19:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/MariaDB/server","ts":"2026-09-23T21:19:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"eigent-ai/eigent","owner":"eigent-ai","name":"eigent","desc":"Eigent: The Open Source Cowork Desktop - Local and Free Alternative to Claude Cowork and Codex","summaryCn":"本地免费替代Claude的神级桌面AI助手","reasonCn":"Eigent 是一款基于 TypeScript 开发的开源桌面应用，旨在提供类似 Claude Code 的本地协作与编码体验。它完全免费且数据隐私安全，支持通过本地 LLM（如 Llama 3）进行多轮对话和代码生成，内置了丰富的工具调用能力，能自动执行系统命令，解决开发中的实际问题。","stars":15391,"starGrowth":4,"createdAt":"2025-07-29T15:56:02Z","silentRounds":0,"language":"TypeScript","topics":["agent-framework","agent-skills","agentic-ai","agentic-workflow","claude-cowork","claude-cowork-alternative","claude-cowork-free","desktop-agent","multi-agent-systems"],"aiDims":["AI应用","AI Agent","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["AI助手","代码生成","本地部署"],"domainKey":"AI助手","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent-framework","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"agentic-workflow","source":"github","weight":0.7},{"name":"claude-cowork","source":"github","weight":0.7},{"name":"claude-cowork-alternative","source":"github","weight":0.7},{"name":"claude-cowork-free","source":"github","weight":0.7},{"name":"desktop-agent","source":"github","weight":0.7},{"name":"multi-agent-systems","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/eigent-ai/eigent","ts":"2026-08-12T01:22:42.342Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"casdoor/casdoor","owner":"casdoor","name":"casdoor","desc":"An open-source Agent-first Identity and Access Management (IAM) /LLM MCP &amp; agent gateway and auth server with web UI supporting OpenClaw, MCP, OAuth, OIDC, SAML, CAS, LDAP, SCIM, WebAuthn, TOTP, MFA, Face ID, Google Workspace, Azure AD","summaryCn":"给AI Agents开大门的认证与授权中心，一键搞定登录与权限。","reasonCn":"Casdoor 是一个基于 Go 语言开发的开源 IAM 身份与访问管理系统，专为 AI Agent 场景设计。它提供 Web UI 管理界面，支持 OpenClaw、MCP、OAuth2、OIDC、SAML、CAS、LDAP、SCIM、WebAuthn、TOTP 等多种协议。作为 AI Agent 的网关和认证服务器，Casdoor 能够安全地验证 Agent 身份并管理权限，让 AI 不仅能用，还能在受控环境下安全工作。","stars":14457,"starGrowth":0,"createdAt":"2020-10-22T14:40:03Z","silentRounds":1,"language":"Go","topics":["oidc","sso","oauth","iam","saml","webauthn","mfa","single-sign-on","radius","scim","auth","authentication","mcp-gateway","ai-gateway","mcp","agent","openclaw","agentic-ai","agi","llm"],"aiDims":["AI基础设施","AI Agent","工具调用"],"zone":"工具","funScore":0,"domainTags":["身份认证","IAM","Agent Gateway","OAuth","权限管理"],"domainKey":"身份认证","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"oidc","source":"github","weight":0.7},{"name":"sso","source":"github","weight":0.7},{"name":"oauth","source":"github","weight":0.7},{"name":"iam","source":"github","weight":0.7},{"name":"saml","source":"github","weight":0.7},{"name":"webauthn","source":"github","weight":0.7},{"name":"mfa","source":"github","weight":0.7},{"name":"single-sign-on","source":"github","weight":0.7},{"name":"radius","source":"github","weight":0.7},{"name":"scim","source":"github","weight":0.7},{"name":"auth","source":"github","weight":0.7},{"name":"authentication","source":"github","weight":0.7},{"name":"mcp-gateway","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"agi","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/casdoor/casdoor","ts":"2026-08-15T00:57:24.351Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/VibeVoice","owner":"microsoft","name":"VibeVoice","desc":"Open-Source Frontier Voice AI","summaryCn":"微软开源语音AI，让机器说话像真人一样自然流畅","reasonCn":"VibeVoice 是微软开源的前沿语音 AI 项目，专注于高保真文本转语音（TTS），支持多语言和实时流式合成。采用先进的神经网络架构，训练数据规模庞大，合成的语音自然度和情感表达能力接近真人。支持零样本语音克隆，只需几秒参考音频即可生成相似音色。开源免费，代码质量高，适合语音助手、有声书、游戏 NPC 等场景。","stars":54458,"starGrowth":1,"createdAt":"2025-08-25T13:24:01Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI应用","多模态Agent","模型与训练"],"zone":"AI","funScore":0.6,"domainTags":["语音合成","零样本克隆","TTS"],"domainKey":"语音合成","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/microsoft/VibeVoice","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":["hot"],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"sooryathejas/METATRON","owner":"sooryathejas","name":"METATRON","desc":"AI-powered penetration testing assistant using local LLM on linux (Parrot OS)","summaryCn":"本地大模型驱动的渗透测试小助手，帮你自动化发现漏洞","reasonCn":"METATRON 将本地大模型嵌入渗透测试工作流，提供漏洞扫描、利用脚本自动生成和报告撰写等功能。它基于 Python 实现，可在 Parrot OS 等安全发行版上直接调用 LLM，支持自定义提示词和插件扩展，帮助安全研究员快速定位常见漏洞并生成对应的攻击代码，省去手动查阅文档的繁琐。","stars":4162,"starGrowth":1,"createdAt":"2026-04-02T06:06:43Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["AI Agent","安全工具","工具调用"],"zone":"AI","funScore":0,"domainTags":["渗透测试","安全工具","红队","漏洞扫描"],"domainKey":"渗透测试","pushedAt":"2026-09-06T14:58:06.418Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/sooryathejas/METATRON","ts":"2026-09-06T14:58:06.418Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"techjarves/Uncensored-Local-Studio","owner":"techjarves","name":"Uncensored-Local-Studio","desc":"Uncensored local AI studio for Windows, Linux, and macOS. Zero-setup GUI for Image Generation, GGUF LLMs, Text to Speech &amp; Speech to Text","summaryCn":"本地跑图、听声辨人、写代码全搞定，一个软件搞定所有AI","reasonCn":"这是一个零配置的本地AI工作台，专为Windows、Linux和macOS设计。它集成了图像生成、大语言模型对话、语音转文字和文字转语音功能，采用JavaScript构建的图形界面，让用户无需配置复杂的命令行环境即可直接使用。项目内置了对GGUF格式模型的优化支持，确保在本地硬件上流畅运行。","stars":1394,"starGrowth":2,"createdAt":"2026-06-04T03:45:01Z","silentRounds":0,"language":"JavaScript","topics":["image-generation","speech-to-text","text-generation","text-to-speech"],"aiDims":["多模态Agent","AI应用","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["本地AI","GUI工具","图像生成","LLM","语音识别"],"domainKey":"本地AI","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"image-generation","source":"github","weight":0.7},{"name":"speech-to-text","source":"github","weight":0.7},{"name":"text-generation","source":"github","weight":0.7},{"name":"text-to-speech","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.9,"source":"trending","url":"https://github.com/techjarves/Uncensored-Local-Studio","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ViRb3/wgcf","owner":"ViRb3","name":"wgcf","desc":"🚤 Cross-platform, unofficial CLI for Cloudflare Warp","summaryCn":"免费蹭.cloudflare的WARP网络，省下每年百元订阅费","reasonCn":"这是一个绕过官方客户端限制的命令行工具，能将生成的伪域名自动解析到Cloudflare WARP的IPv6地址上，从而让普通设备通过订阅链接接入WARP网络。它利用Go语言编写，无需图形界面，完全开源，解决了官方客户端在Linux和部分路由器上功能阉割的问题。","stars":8718,"starGrowth":1,"createdAt":"2020-04-13T20:41:19Z","silentRounds":3,"language":"Go","topics":["security"],"aiDims":["非AI-实用","非AI-好玩","效率工具"],"zone":"工具","funScore":0,"domainTags":["Cloudflare WARP","网络代理","命令行工具","Go","订阅管理"],"domainKey":"Cloudflare WARP","pushedAt":"2026-08-24T11:42:38Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/ViRb3/wgcf","ts":"2026-08-24T11:42:38Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"grouzen/ollana","owner":"grouzen","name":"ollana","desc":"Ollama over LAN - Auto-discover your Ollama server on your local network with hassle-free ease.","summaryCn":"一键局域网发现你的Ollama，轻松跑大模型","reasonCn":"这是一个基于Rust开发的局域网自动发现工具，专为Ollama设计。它利用mDNS协议，无需配置IP地址，自动扫描并连接同一网络下的Ollama服务。项目结构精简，提供了CLI和Web两种模式，支持同时管理多个Ollama实例，极大简化了本地多模型部署的运维难度。","stars":10,"starGrowth":0,"createdAt":"2025-01-29T19:00:16Z","silentRounds":28,"language":"Rust","topics":["llm-model"],"aiDims":["AI基础设施","模型部署"],"zone":"工具","funScore":0.5,"domainTags":["Ollama","局域网","mDNS"],"domainKey":"Ollama","pushedAt":"2026-07-31T01:04:14Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.92,"source":"search","url":"https://github.com/grouzen/ollana","ts":"2026-07-31T01:04:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"oven-sh/bun","owner":"oven-sh","name":"bun","desc":"Incredibly fast JavaScript runtime, bundler, test runner, and package manager – all in one","summaryCn":"比Node快两倍的JavaScript全能工具箱，开发体验起飞","reasonCn":"Bun是一个用Rust编写的JavaScript运行时、打包器和测试工具，它集成了包管理器功能。它通过直接调用系统库和JIT编译技术，显著提升了执行速度。它兼容Node.js API，支持TypeScript、npm包，并提供快速启动的开发服务器和测试框架。","stars":96009,"starGrowth":4,"createdAt":"2021-04-14T00:48:17Z","silentRounds":0,"language":"Rust","topics":["bun","bundler","javascript","javascriptcore","jsx","nodejs","npm","react","transpiler","typescript","rust","rust-lang"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["JavaScript","运行时","打包器","包管理器","开发工具"],"domainKey":"JavaScript","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"bun","source":"github","weight":0.7},{"name":"bundler","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"javascriptcore","source":"github","weight":0.7},{"name":"jsx","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"npm","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"transpiler","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"rust-lang","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/oven-sh/bun","ts":"2026-09-07T02:44:20.540Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"strelov1/freehire","owner":"strelov1","name":"freehire","desc":"freehire — the open-source search engine for job seekers","summaryCn":"免费好用的开源职位搜索引擎，帮你快速找到心仪工作","reasonCn":"这是一个基于Go语言开发的职位搜索聚合平台，旨在为求职者提供免费、无广告的职位信息检索服务。它通过爬取各大招聘网站的数据，构建了庞大的职位数据库，支持按关键词、地点、薪资等维度进行筛选，并提供了API接口方便开发者集成。项目代码结构清晰，易于二次开发。","stars":750,"starGrowth":3,"createdAt":"2026-06-07T03:20:31Z","silentRounds":0,"language":"Go","topics":["ats","careers","hiring","job-aggregator","job-board","job-search","jobs","recruitment","remote-jobs","job-scraper","career","job-portal","golang","langchain","llm","meilisearch","postgresql","self-hosted","svelte","sveltekit"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["职位搜索","Go语言","招聘平台"],"domainKey":"职位搜索","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ats","source":"github","weight":0.7},{"name":"careers","source":"github","weight":0.7},{"name":"hiring","source":"github","weight":0.7},{"name":"job-aggregator","source":"github","weight":0.7},{"name":"job-board","source":"github","weight":0.7},{"name":"job-search","source":"github","weight":0.7},{"name":"jobs","source":"github","weight":0.7},{"name":"recruitment","source":"github","weight":0.7},{"name":"remote-jobs","source":"github","weight":0.7},{"name":"job-scraper","source":"github","weight":0.7},{"name":"career","source":"github","weight":0.7},{"name":"job-portal","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"langchain","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"meilisearch","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"svelte","source":"github","weight":0.7},{"name":"sveltekit","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/strelov1/freehire","ts":"2026-08-18T00:56:27.274Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"m1heng/clawdbot-feishu","owner":"m1heng","name":"clawdbot-feishu","desc":"","summaryCn":"在飞书里养个AI助手，帮你处理各种工作事务","reasonCn":"基于Clawdbot框架深度适配飞书生态，支持接入主流AI模型，让机器人具备智能对话、任务理解和自动执行能力。提供丰富的插件体系和自定义指令，可实现知识库查询、会议纪要整理、智能提醒等场景。代码用TypeScript编写，架构清晰，方便二次开发和功能扩展，适合希望用AI提升团队效率的开发者。","stars":4241,"starGrowth":0,"createdAt":"2026-01-25T18:19:30Z","silentRounds":27,"language":"TypeScript","topics":[],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0,"domainTags":["飞书机器人","AI助手","企业自动化","多模型接入"],"domainKey":"飞书机器人","pushedAt":"2026-03-29T15:07:36Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"bigbro","url":"https://github.com/m1heng/clawdbot-feishu","ts":"2026-03-29T15:07:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bentoml/BentoML","owner":"bentoml","name":"BentoML","desc":"The easiest way to serve AI apps and models - Build Model Inference APIs, Job queues, LLM apps, Multi-model pipelines, and more!","summaryCn":"轻松构建AI应用的API和工作流程，无需编码知识","reasonCn":"BentoML是一个简单易用的开源工具，允许用户通过图形界面快速构建和部署AI应用程序。它提供了一系列预定义的模板，用户只需要选择相应的模板并配置参数即可完成应用的创建。该工具支持多种AI模型和框架，如TensorFlow、PyTorch等，使得开发者无需深入了解复杂的技术细节，就能快速实现AI项目的开发。","stars":8855,"starGrowth":0,"createdAt":"2019-04-02T01:39:27Z","silentRounds":4,"language":"Python","topics":["deep-learning"],"aiDims":["AI基础设施","模型部署"],"zone":"AI","funScore":0,"domainTags":["模型部署","推理引擎","LLM应用","API服务"],"domainKey":"模型部署","pushedAt":"2026-09-07T17:43:01Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"deep-learning","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/bentoml/BentoML","ts":"2026-09-07T17:43:01Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Windy3f3f3f3f/how-claude-code-works","owner":"Windy3f3f3f3f","name":"how-claude-code-works","desc":"Deep dive into Claude Code internals — architecture, agent loop, context engineering, and more. / 深入解析 Claude Code 源码：架构、Agent 循环、上下文工程、工具系统等","summaryCn":"带你拆解 Claude Code 内部运行逻辑的指南","reasonCn":"该项目通过深入分析源码，详细揭示了 Claude Code 的架构设计、Agent 循环机制以及上下文工程。它将复杂的 AI Agent 运行逻辑转化为易懂的架构图和原理解析，帮助开发者理解 AI 如何思考、如何调用工具以及如何管理内存上下文。","stars":3666,"starGrowth":0,"createdAt":"2026-03-31T12:25:40Z","silentRounds":1,"language":"","topics":["ai-agent"],"aiDims":["AI Agent","教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["Claude Code","AI Agent","源码解析","上下文工程","Prompt工程"],"domainKey":"Claude Code","pushedAt":"2026-08-17T08:27:42Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7}],"aiScore":0.88,"source":"search","url":"https://github.com/Windy3f3f3f3f/how-claude-code-works","ts":"2026-08-17T08:27:42Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"uber/ADR","owner":"uber","name":"ADR","desc":"ADR secures enterprise AI agents through observability, security benchmarking, and threat detection. Deployed at Uber.","summaryCn":"Uber自用的AI Agent安全监控系统，专治各种Agent不听话","reasonCn":"这是Uber真实部署的企业级AI Agent安全框架，通过可观测性、安全基准测试和威胁检测三位一体保障Agent行为合规。支持多种LLM后端接入，提供完整的操作审计日志和异常行为告警。内置对抗测试能力，能自动发现Agent可能被Prompt注入攻击的风险点，是企业级AI应用部署的安全兜底方案。","stars":1582,"starGrowth":3,"createdAt":"2026-04-19T18:01:42Z","silentRounds":0,"language":"Python","topics":["agent-security","ai-agents","ai-security","benchmark","llm-security","mcp","model-context-protocol","prompt-injection","threat-detection","claude","claude-code","codex","cursor"],"aiDims":["AI Agent","安全工具","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI安全","Agent监控","可观测性","威胁检测","安全基准"],"domainKey":"AI安全","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"agent-security","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-security","source":"github","weight":0.7},{"name":"benchmark","source":"github","weight":0.7},{"name":"llm-security","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"model-context-protocol","source":"github","weight":0.7},{"name":"prompt-injection","source":"github","weight":0.7},{"name":"threat-detection","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/uber/ADR","ts":"2026-08-07T04:03:56.389Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"electric-sql/pglite","owner":"electric-sql","name":"pglite","desc":"Embeddable Postgres with real-time, reactive bindings.","summaryCn":"把Postgres数据库塞进浏览器，本地跑SQL毫无压力","reasonCn":"PGLite让Postgres数据库以WebAssembly形式运行在浏览器或Node.js中，无需安装任何服务端。这意味着你可以构建完全离线的Web应用，拥有完整的ACID事务支持和SQL查询能力。对于前端开发者来说，这是构建本地优先应用的核弹级工具，数据安全性与开发体验兼得。","stars":16055,"starGrowth":2,"createdAt":"2024-02-19T14:37:38Z","silentRounds":0,"language":"TypeScript","topics":["database"],"aiDims":["AI基础设施","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["PostgreSQL","WebAssembly","前端数据库","离线优先应用","ACID事务"],"domainKey":"PostgreSQL","pushedAt":"2026-08-26T18:35:11Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/electric-sql/pglite","ts":"2026-08-26T18:35:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"DavidHDev/canvas-ui","owner":"DavidHDev","name":"canvas-ui","desc":"A library of creative canvas components. Real HTML with WebGL effects running over it. React, Vue, Svelte, vanilla.","summaryCn":"给普通网页穿上炫目光效的可视化神器，ReactVue都能用","reasonCn":"这是一套前端组件库，支持ReactVueSvelte等多种框架。它最大的亮点是让普通的HTML元素可以直接叠加WebGL特效，比如玻璃拟态、流体动画、粒子效果等。这让开发者不用深入Shader编程就能做出电影级的视觉交互体验，非常适合想做惊艳落地页或创意展示的团队。","stars":4671,"starGrowth":1,"createdAt":"2026-07-16T13:38:20Z","silentRounds":1,"language":"TypeScript","topics":["frontend","creative-coding"],"aiDims":["可视化","非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["WebGL","前端组件","可视化"],"domainKey":"WebGL","pushedAt":"2026-09-13T08:13:01Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"creative-coding","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/DavidHDev/canvas-ui","ts":"2026-09-13T08:13:01Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"guaguastandup/zotero-pdf2zh","owner":"guaguastandup","name":"zotero-pdf2zh","desc":"","summaryCn":"在 Zotero 中直接翻译 PDF，保留公式排版，支持多种 LLM。","reasonCn":"Zotero PDF2zh 是一款专为文献阅读设计的插件，解决了在 Zotero 里直接翻译 PDF 时公式错乱、排版混乱的难题。它支持 DeepSeek、ChatGPT 等多种大模型，不仅能一键翻译全文，还能生成双语对照，方便边看边学。在 Windows 上运行时，它会自动把翻译变成后台任务，不会因为网络中断而报错，让你边干活边等结果，特别适合研究生和研究人员。","stars":6784,"starGrowth":0,"createdAt":"2025-01-14T06:48:13Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["文档理解","AI应用"],"zone":"AI","funScore":0.7,"domainTags":["PDF翻译","Zotero插件","学术文献","LLM"],"domainKey":"PDF翻译","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/guaguastandup/zotero-pdf2zh","ts":"2026-09-15T22:54:20.912Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"p2r3/convert","owner":"p2r3","name":"convert","desc":"Truly universal online file converter","summaryCn":"一个不需要联网的万能在线文件转换器，支持几十种格式互转","reasonCn":"这个项目是一个基于浏览器的文件转换工具，支持将视频、音频、图片、文档等多种格式互相转换。它完全在本地运行，不需要上传文件到服务器，保护隐私且速度极快。提供Web、桌面和命令行三种使用方式，极大简化了用户在不同软件间转换文件的繁琐流程。","stars":4051,"starGrowth":2,"createdAt":"2025-12-07T07:40:09Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["非AI-实用","非AI-好玩","开发者工具"],"zone":"工具","funScore":0,"domainTags":["文件转换","本地处理","多媒体工具"],"domainKey":"文件转换","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/p2r3/convert","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kotyzap/Photo-Curator","owner":"kotyzap","name":"Photo-Curator","desc":"📸 Browser-based tool for culling and ranking large photo libraries. Remove blurry shots 🗑️, collapse burst duplicates 📚, and surface your best photos ⭐. Everything runs locally on your machine 🔒 — no uploads, no cloud, no privacy concerns. 🚀 Mac, Win, Linux","summaryCn":"你是不是也遇到过这种情况：出去旅游拍了几千张照片，回来整理时看着那个庞","reasonCn":"这个项目的核心技术亮点在于其“本地优先”和“智能评级”机制。它并不是简单地靠文件名排序，而是能利用计算机视觉或启发式算法分析图像质量。它能识别出哪些图片因为抖动而模糊，哪些是几乎一样的连拍张数，从而给出一个排名。","stars":20,"starGrowth":0,"createdAt":"2026-05-31T12:51:34Z","silentRounds":11,"language":"Python","topics":["photography"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["照片管理","图像质量分析","计算机视觉","本地处理","RAW照片"],"domainKey":"照片管理","pushedAt":"2026-09-13T11:51:41Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"photography","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/kotyzap/Photo-Curator","ts":"2026-09-13T11:51:41Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kanboard/documentation","owner":"kanboard","name":"documentation","desc":"Kanboard's Documentation","summaryCn":"一个轻量级看板，帮你把混乱的工作理得清清楚楚","reasonCn":"Kanboard 的设计哲学非常朴素，没有花哨的AI功能，也没有复杂的权限体系，它专注于做一件事：让任务流转变得极其简单。你只需要创建一个任务卡片，拖拽到不同的列（比如待办、进行中、已完成），就能实时更新项目进度。","stars":37,"starGrowth":0,"createdAt":"2018-04-01T05:49:35Z","silentRounds":14,"language":"CSS","topics":["english"],"aiDims":["非AI-实用","效率工具"],"zone":"资源","funScore":0,"domainTags":["看板","项目管理","敏捷开发","文档站"],"domainKey":"看板","pushedAt":"2026-09-11T03:58:06Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"english","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/kanboard/documentation","ts":"2026-09-11T03:58:06Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kyutai-labs/pocket-tts","owner":"kyutai-labs","name":"pocket-tts","desc":"A TTS that fits in your CPU (and pocket)","summaryCn":"无需GPU，手机CPU也能跑的语音合成神器","reasonCn":"Pocket-TTS是一个专为资源受限设备优化的语音合成模型，能在普通CPU甚至嵌入式设备上流畅运行。它采用了创新的神经网络架构，在保证语音自然度的同时大幅降低了计算需求。用户只需几条命令即可本地生成高质量语音，无需依赖云端API，非常适合隐私敏感场景或离线应用。","stars":9600,"starGrowth":3,"createdAt":"2026-01-07T17:33:32Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","推理引擎"],"zone":"AI","funScore":1,"domainTags":["语音合成","TTS","边缘计算","离线模型"],"domainKey":"语音合成","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/kyutai-labs/pocket-tts","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"davidmosiah/whoop-mcp","owner":"davidmosiah","name":"whoop-mcp","desc":"Local-first MCP server for WHOOP recovery, HRV & sleep — Claude, Cursor, ChatGPT ready","summaryCn":"让AI实时读取你智能手环的身体数据并化身私人健康教练","reasonCn":"这是一个基于ModelContextProtocol架构开发的本地优先数据桥梁，专门打通运动手环和Claude等大模型。通过安全的本地调用，AI能直接读取你的睡眠质量、心率变异性和身体恢复指数。它让你日常的健康指标不再是冰冷的数字，而是变成能和AI对话并获取定制化训练与作息建议的动态生产力源泉。","stars":13,"starGrowth":0,"createdAt":"2026-04-30T14:22:15Z","silentRounds":18,"language":"TypeScript","topics":["fitness"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["MCP","健康数据","本地服务"],"domainKey":"MCP","pushedAt":"2026-08-29T10:26:15Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/davidmosiah/whoop-mcp","ts":"2026-08-29T10:26:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bia-pain-bache/BPB-Worker-Panel","owner":"bia-pain-bache","name":"BPB-Worker-Panel","desc":"A GUI Panel providing Worker subscriptions for VLESS, Trojan and Warp configs alongside a private DoH server and chain proxies, offering full DNS, clean IP, Fragment, Warp, Warp pro and routing settings for cross-platform clients using Amnezia, Wireguard, Sing-box, Clash/Mihomo and Xray cores.","summaryCn":"给网络代理配置装个网页控制面板，一键管理各类代理协议","reasonCn":"这是一个专为 Cloudflare Workers 设计的图形化管理面板，支持 VLESS、Trojan 等多种主流代理协议的订阅管理。它不仅能提供私有的 DNS-over-HTTPS (DoH) 服务器，还集成了 Warp 链式代理、流量分片和纯净 IP 等高级路由功能。通过 Web 界面即可完成复杂的网络配置，无需编写代码，极大地简化了网络工具的部署流程。","stars":13501,"starGrowth":2,"createdAt":"2024-01-19T17:58:28Z","silentRounds":0,"language":"TypeScript","topics":["android","chain","fragment","ios","linux","proxy-chain","singbox-core","vless","windows","xray-core","warp","clash-core","mihomo","trojan","amnezia","doh-server"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网络代理","配置管理","Cloudflare"],"domainKey":"网络代理","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"android","source":"github","weight":0.7},{"name":"chain","source":"github","weight":0.7},{"name":"fragment","source":"github","weight":0.7},{"name":"ios","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"proxy-chain","source":"github","weight":0.7},{"name":"singbox-core","source":"github","weight":0.7},{"name":"vless","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"xray-core","source":"github","weight":0.7},{"name":"warp","source":"github","weight":0.7},{"name":"clash-core","source":"github","weight":0.7},{"name":"mihomo","source":"github","weight":0.7},{"name":"trojan","source":"github","weight":0.7},{"name":"amnezia","source":"github","weight":0.7},{"name":"doh-server","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/bia-pain-bache/BPB-Worker-Panel","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"basecamp/hey-cli","owner":"basecamp","name":"hey-cli","desc":"HEY CLI and Agent Skills","summaryCn":"帮你自动搞定收发邮件的神器，连邮件都懒得写了。","reasonCn":"这是一个命令行工具，能让用户通过自然语言和Agent技能在浏览器中操作HEY邮件客户端。它支持通过简单的指令自动撰写、回复、转发邮件，甚至能结合自定义技能执行复杂任务。开发者可以通过编写Go代码扩展功能，将其集成到自动化工作流中。","stars":379,"starGrowth":1,"createdAt":"2026-03-03T09:17:06Z","silentRounds":7,"language":"Go","topics":["agent-skills","hey"],"aiDims":["AI Agent","工具调用"],"zone":"工具","funScore":0,"domainTags":["邮件自动化","Agent","命令行","Puppeteer","Go"],"domainKey":"邮件自动化","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"hey","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/basecamp/hey-cli","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"zchoi/Awesome-Embodied-Robotics-and-Agent","owner":"zchoi","name":"Awesome-Embodied-Robotics-and-Agent","desc":"This is a curated list of \"Embodied AI or robot with Large Language Models\" research. Watch this repository for the latest updates! 🔥","summaryCn":"把 AI 大脑装进机器人身体，这份论文合集涵盖所有前沿方向","reasonCn":"简单来说，这个项目回答了一个科幻感十足的问题：如果把 ChatGPT 那样的 AI 大脑装进机器人身体里，会发生什么？这就是「具身智能」——让 AI 不再只是坐在服务器里聊天，而是能感知物理世界、操控机械臂、自主导航的移动机器人。","stars":1891,"starGrowth":0,"createdAt":"2023-07-19T06:55:49Z","silentRounds":2,"language":"","topics":["llm-model"],"aiDims":["大语言模型","多模态Agent"],"zone":"资源","funScore":0.5,"domainTags":["具身智能","机器人","LLM","论文列表","多模态"],"domainKey":"具身智能","pushedAt":"2026-09-16T15:28:45Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.6,"source":"search","url":"https://github.com/zchoi/Awesome-Embodied-Robotics-and-Agent","ts":"2026-09-16T15:28:45Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Tukdify/Tukdify-clips","owner":"Tukdify","name":"Tukdify-clips","desc":"Local-first AI video clipping tool that turns long-form recordings into short vertical clips with transcription, smart reframing, and captions.","summaryCn":"一款AI驱动的视频剪辑工具，将长篇录音转换成简短的垂直视频片段。","reasonCn":"该工具使用深度学习技术自动识别和剪辑音频内容，并将其转化为视频，同时还能添加文字、音乐和字幕。","stars":14,"starGrowth":0,"createdAt":"2026-06-22T14:51:40Z","silentRounds":20,"language":"Python","topics":["video-editor"],"aiDims":["视频编辑","AI"],"zone":"创意","funScore":0.7,"domainTags":["视频剪辑","垂直视频","AI生成","字幕","本地处理"],"domainKey":"视频剪辑","pushedAt":"2026-09-04T18:17:26Z","tags":[{"name":"视频编辑","source":"llm","weight":1},{"name":"AI","source":"llm","weight":1},{"name":"video-editor","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/Tukdify/Tukdify-clips","ts":"2026-09-04T18:17:26Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vincentchyu/photools","owner":"vincentchyu","name":"photools","desc":"📸 Turn GPX tracks into GPS-tagged, location-aware photo libraries. Offline reverse geocoding, intelligent GPS interpolation, and date-based organization for photographers. (macOS App & CLI)","summaryCn":"出去玩的时候，我们经常用手机拍很多照片，但照片的EXIF信息里往往只有时间","reasonCn":"出去玩的时候，我们经常用手机拍很多照片，但照片的EXIF信息里往往只有时间，没有具体的GPS位置，导致后期整理起来非常痛苦，根本不知道哪张是在哪个景点拍的。这个项目就是专门用来解决这个问题的。它的核心逻辑很简单：你先在手机上开启GPS记录功能走一圈，生成一个GPX文件（这个","stars":11,"starGrowth":0,"createdAt":"2026-05-10T01:55:20Z","silentRounds":14,"language":"Go","topics":["photography"],"aiDims":["非AI-实用","创意工具"],"zone":"工具","funScore":0,"domainTags":["GPX","地理标注","照片管理","离线地理编码","摄影工具"],"domainKey":"GPX","pushedAt":"2026-09-03T13:26:11Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"photography","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/vincentchyu/photools","ts":"2026-09-03T13:26:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"langchain-ai/open_deep_research","owner":"langchain-ai","name":"open_deep_research","desc":"","summaryCn":"让AI自己搜索分析，帮你写出专业研究报告","reasonCn":"这是LangChain推出的AI深度研究项目，让大模型自动完成信息搜索、整理分析、撰写报告的全流程。你只需提出一个问题，AI就会自动联网搜索、筛选信息、整合观点，最后输出一份结构化的研究报告，省时省力，堪称个人研究助理。","stars":12683,"starGrowth":0,"createdAt":"2024-11-20T17:37:22Z","silentRounds":6,"language":"Python","topics":[],"aiDims":["AI Agent","AI搜索"],"zone":"AI","funScore":0,"domainTags":["AI研究","信息检索","报告生成","多Agent协作"],"domainKey":"AI研究","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.91,"source":"trending","url":"https://github.com/langchain-ai/open_deep_research","ts":"2026-08-11T09:45:50.026Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"SeekingDream/Static-to-Dynamic-LLMEval","owner":"SeekingDream","name":"Static-to-Dynamic-LLMEval","desc":"The official GitHub repository of the paper \"Recent advances in large language model benchmarks against data contamination: From static to dynamic evaluation\"","summaryCn":"让大模型测试不再看“陈年旧卷”，动态数据让结果更真实。","reasonCn":"该项目解决了大模型基准测试面临的数据污染问题。传统静态评估使用固定数据集，容易导致模型“背答案”。本仓库提供了动态评估框架，实时生成新数据或动态检索数据，确保测试结果反映模型真实能力。它支持多种动态数据生成策略，帮助研究者更准确地评估模型性能。","stars":500,"starGrowth":0,"createdAt":"2025-02-21T01:05:45Z","silentRounds":14,"language":"","topics":["llm-model"],"aiDims":["大语言模型","RAG与知识","文档理解"],"zone":"AI","funScore":0,"domainTags":["LLM评估","基准测试","数据污染","动态评估"],"domainKey":"LLM评估","pushedAt":"2026-08-11T13:57:47Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.9,"source":"search","url":"https://github.com/SeekingDream/Static-to-Dynamic-LLMEval","ts":"2026-08-11T13:57:47Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MengTo/sublevel-studio","owner":"MengTo","name":"sublevel-studio","desc":"A single-file Three.js studio concept with an interactive Kyoto lobby, arcade, portfolio CRTs, and editorial case studies.","summaryCn":"单文件承载的沉浸式3D京都大厅与复古街机视觉创意编程空间。","reasonCn":"技术上，它最大的亮点在于将复杂的3D场景、光影效果、互动逻辑以及精美的UI设计全部浓缩在单一的HTML文件里。借助强大的Three.js引擎，它展示了极致的WebGL渲染优化技巧，通过精简的着色器和资源加载策略，实现了极佳的加载速度和交互沉浸感。","stars":65,"starGrowth":0,"createdAt":"2026-09-03T08:47:09Z","silentRounds":1,"language":"HTML","topics":["creative-coding"],"aiDims":["非AI-好玩","可视化","创意工具"],"zone":"创意","funScore":1,"domainTags":["Three.js","创意编程","3D交互","WebGL","视觉艺术"],"domainKey":"Three.js","pushedAt":"2026-09-03T08:54:40Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/MengTo/sublevel-studio","ts":"2026-09-03T08:54:40Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"4thfever/cultivation-world-simulator","owner":"4thfever","name":"cultivation-world-simulator","desc":"基于 AI Agent 工作流的修仙世界模拟器，旨在还原智能、开放的仙侠世界。| An open-source Cultivation World Simulator using Agentic Workflow to create a dynamic, emerging Xianxia world. ","summaryCn":"让各路人工智能在修仙世界里自动打怪升级的开放式模拟器","reasonCn":"技术上，项目深度整合了多智能体工作流架构。每个NPC都被赋予了独特的性格特征、记忆系统和生存目标，通过大语言模型驱动其行为决策。不同于传统游戏中写死的剧本和对话树，这里的修仙者会根据周遭环境、资源多寡以及人际关系做出完全随机且合理的举动。","stars":2086,"starGrowth":0,"createdAt":"2025-08-18T15:09:49Z","silentRounds":1,"language":"Python","topics":["game","simulation"],"aiDims":["AI Agent","游戏"],"zone":"创意","funScore":1,"domainTags":["修仙模拟","多智能体","Agent工作流","沙盒游戏","AI模拟"],"domainKey":"修仙模拟","pushedAt":"2026-08-16T11:21:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/4thfever/cultivation-world-simulator","ts":"2026-08-16T11:21:40Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"huggingface/funes","owner":"huggingface","name":"funes","desc":"Durable, searchable memory of your past agent sessions.","summaryCn":"让智能体记住历史对话，随时检索上下文的持久记忆库。","reasonCn":"Funes是基于Rust实现的持久化向量存储，引入可搜索的记忆层，为LLM代理提供上下文检索能力，支持多模态键值、时序过滤和增量写入，可与LangChain等框架无缝对接，实现对话历史的高效复用，帮助构建更连贯的长程智能体。","stars":476,"starGrowth":0,"createdAt":"2026-06-18T11:33:31Z","silentRounds":1,"language":"Rust","topics":["agent-memory","claude-code","codex","lance","llm-agents","mcp","rust","ai-agents","hermes","pi"],"aiDims":["RAG与知识","向量数据库","AI Agent"],"zone":"AI","funScore":0,"domainTags":["向量数据库","智能体记忆","上下文检索","Rust","LLM框架"],"domainKey":"向量数据库","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"agent-memory","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"lance","source":"github","weight":0.7},{"name":"llm-agents","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"hermes","source":"github","weight":0.7},{"name":"pi","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/huggingface/funes","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vllm-project/semantic-router","owner":"vllm-project","name":"semantic-router","desc":"A programmable Mixture-of-Models router for heterogeneous LLM inference","summaryCn":"像给大模型装上智能导航，自动识别意图并路由到对应模型","reasonCn":"这是一个可编程的混合模型路由器，支持基于规则的语义路由和基于LLM的动态路由。它允许开发者根据用户输入的语义，自动选择最合适的模型（如小模型处理简单查询，大模型处理复杂任务），显著降低推理成本并提升响应速度。核心支持Go语言编写，兼容多种LLM后端。","stars":5899,"starGrowth":1,"createdAt":"2025-08-26T21:49:50Z","silentRounds":1,"language":"Go","topics":["mixture-of-models","semantic-router","vllm","ai-gateway","kubernetes","llmrouter","llm","guardrails","inference","pytorch","transformer"],"aiDims":["AI基础设施","工具调用","大语言模型"],"zone":"AI","funScore":0.6,"domainTags":["LLM路由","意图识别","模型编排"],"domainKey":"LLM路由","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"mixture-of-models","source":"github","weight":0.7},{"name":"semantic-router","source":"github","weight":0.7},{"name":"vllm","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"llmrouter","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"guardrails","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"pytorch","source":"github","weight":0.7},{"name":"transformer","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/vllm-project/semantic-router","ts":"2026-08-17T00:55:40.962Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"coder/code-server","owner":"coder","name":"code-server","desc":"VS Code in the browser","summaryCn":"把VS Code装进浏览器，随时随地写代码的神器","reasonCn":"coder/code-server是VS Code的网页版实现，允许你通过浏览器直接运行完整的代码编辑器。它基于WebAssembly和Node.js构建，提供了与桌面版几乎一致的体验，支持Git、调试、插件扩展等全部核心功能。通过简单的Docker命令，你就能一键部署一个私有化的云端开发环境，彻底摆脱本地软件安装的麻烦。","stars":79396,"starGrowth":2,"createdAt":"2019-02-27T16:50:41Z","silentRounds":0,"language":"TypeScript","topics":["vscode","vscode-remote","ide","remote-work","development-environment","dev-tools","browser-ide"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["云IDE","代码编辑器","远程开发"],"domainKey":"云IDE","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"vscode","source":"github","weight":0.7},{"name":"vscode-remote","source":"github","weight":0.7},{"name":"ide","source":"github","weight":0.7},{"name":"remote-work","source":"github","weight":0.7},{"name":"development-environment","source":"github","weight":0.7},{"name":"dev-tools","source":"github","weight":0.7},{"name":"browser-ide","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/coder/code-server","ts":"2026-08-13T01:37:33.040Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"calcom/cal.diy","owner":"calcom","name":"cal.diy","desc":"Scheduling infrastructure for absolutely everyone.","summaryCn":"开源日历神器，一键定制你的专属预约系统。","reasonCn":"Cal.com 是一个完全开源的调度基础设施，支持个人和团队使用，旨在让预约变得简单且可定制。它提供完整的 API 和 Webhooks，可以轻松集成到任何应用中，支持多种日历和支付方式。其社区版本功能强大，足够个人和中小团队使用。","stars":48616,"starGrowth":5,"createdAt":"2021-03-22T13:47:42Z","silentRounds":0,"language":"TypeScript","topics":["open-source","typescript","nextjs","next-auth","postgresql","prisma","t3-stack","tailwindcss","trpc","turborepo","zod"],"aiDims":["AI应用","效率工具"],"zone":"工具","funScore":0,"domainTags":["日历调度","预约系统","开源SaaS","插件生态"],"domainKey":"日历调度","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"open-source","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"next-auth","source":"github","weight":0.7},{"name":"postgresql","source":"github","weight":0.7},{"name":"prisma","source":"github","weight":0.7},{"name":"t3-stack","source":"github","weight":0.7},{"name":"tailwindcss","source":"github","weight":0.7},{"name":"trpc","source":"github","weight":0.7},{"name":"turborepo","source":"github","weight":0.7},{"name":"zod","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/calcom/cal.diy","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015308540398222806,"copyOk":false},{"repo":"cookieY/Yearning","owner":"cookieY","name":"Yearning","desc":"🐳 A most popular sql audit platform for mysql ","summaryCn":"帮程序员自动查错 SQL 的开源神器，解决写代码的一大痛点","reasonCn":"Yearning 是基于 Go 语言开发的开源 SQL 审计平台，专为 MySQL 设计，提供实时的 SQL 规则检查、自动审计和阻断功能。它支持自定义审计规则，能拦截高危操作并生成审计报告，帮助企业规范数据库操作，降低数据风险。","stars":8973,"starGrowth":1,"createdAt":"2017-10-18T14:08:50Z","silentRounds":27,"language":"Go","topics":["devops"],"aiDims":["开发工具","安全工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["MySQL","SQL审计","数据库安全","DevOps"],"domainKey":"MySQL","pushedAt":"2026-08-24T11:46:31Z","tags":[{"name":"开发工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/cookieY/Yearning","ts":"2026-08-24T11:46:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deepcoldy/botmux","owner":"deepcoldy","name":"botmux","desc":"Bridge Feishu/Lark to AI coding CLIs — Claude Code, Codex, Gemini, OpenCode… every DM, group or topic spawns its own live-streaming CLI session","summaryCn":"飞书聊天直接调Claude Codex等AI编程工具，每个群都能开专属编程机器人","reasonCn":"技术上，botmux在飞书开放平台和AI编码CLI之间搭建了一个实时桥接层。它通过飞书的WebSocket长连接接收消息，然后路由到对应的AI编码后端执行，再把结果实时推回聊天窗口。核心架构是事件驱动的消息路由机制，每个DM、群组或话题都会独立维护一个CLI会话状态，互不干扰。","stars":1509,"starGrowth":0,"createdAt":"2026-03-11T16:59:54Z","silentRounds":1,"language":"TypeScript","topics":["claude","claude-code","codex","feishu","gemini","lark","opencode"],"aiDims":["AI应用","工具调用","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["飞书集成","AI编程助手","实时通讯"],"domainKey":"飞书集成","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"feishu","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"lark","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/deepcoldy/botmux","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Kuberwastaken/claurst","owner":"Kuberwastaken","name":"claurst","desc":"Agentic Coding for Builders who Ship","summaryCn":"让AI当你的编程搭档，自动帮你写代码跑项目","reasonCn":"claurst是一个基于Rust构建的AI Agent编程工具，能将自然语言需求转化为可运行的代码。它通过工具调用机制自动执行代码、调试错误，最终生成可用的软件项目。核心亮点在于Rust的高性能和内存安全，以及智能化的代码生成与迭代能力。适合独立开发者、产品经理和想要快速验证想法的技术人员。","stars":10306,"starGrowth":1,"createdAt":"2026-03-31T09:23:58Z","silentRounds":1,"language":"Rust","topics":["claude","claude-code","cli","codex","gemini","rust","tui","claurst"],"aiDims":["AI Agent","代码助手","工具调用"],"zone":"AI","funScore":1,"domainTags":["AI Agent","代码生成","自动编程"],"domainKey":"AI Agent","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"cli","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"tui","source":"github","weight":0.7},{"name":"claurst","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/Kuberwastaken/claurst","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"fastmachinelearning/hls4ml-tutorial","owner":"fastmachinelearning","name":"hls4ml-tutorial","desc":"Tutorial notebooks for hls4ml ","summaryCn":"想象一下，你训练了一个识别粒子的超级模型，但传统的电脑运行速度太慢，无","reasonCn":"核心亮点在于它支持将Keras、PyTorch或TensorFlow模型转化为可综合的C++代码。它利用高层次综合（HLS）技术，在保证模型精度的前提下，对计算进行极致的流水线化处理，避免了传统计算架构中的内存带宽瓶颈，实现了纳秒级的推理速度。","stars":448,"starGrowth":0,"createdAt":"2020-06-02T12:59:21Z","silentRounds":13,"language":"Jupyter Notebook","topics":["tutorial"],"aiDims":["模型部署","AI基础设施"],"zone":"资源","funScore":0,"domainTags":["机器学习","FPGA","嵌入式AI"],"domainKey":"机器学习","pushedAt":"2026-09-14T14:02:34Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/fastmachinelearning/hls4ml-tutorial","ts":"2026-09-14T14:02:34Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"apache/apisix","owner":"apache","name":"apisix","desc":"The Cloud-Native API Gateway and AI Gateway","summaryCn":"云原生API网关支持热加载插件丰富助力微服务和AI应用流量管理","reasonCn":"Apache APISIX是CNCF孵化的云原生API网关，采用Lua语言编写基于Nginx和etcd架构。核心优势是支持配置热加载无需重启即可生效，内置100多个开箱即用插件。支持微服务网关、API网关、AI网关等多种场景，适合中大型分布式系统。\n安装部署支持Docker和Kubernetes两种方式，通过官方文档可快速搭建生产环境。","stars":17155,"starGrowth":0,"createdAt":"2019-04-10T02:02:05Z","silentRounds":2,"language":"Lua","topics":["devops","iot","kubernetes"],"aiDims":["AI基础设施","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["API 网关","流量管理","微服务"],"domainKey":"API 网关","pushedAt":"2026-09-22T07:39:36Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"iot","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Lua","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/apache/apisix","ts":"2026-09-22T07:39:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"feder-cr/Jobs_Applier_AI_Agent_AIHawk","owner":"feder-cr","name":"Jobs_Applier_AI_Agent_AIHawk","desc":"AIHawk aims to easy job hunt process by automating the job application process. Utilizing artificial intelligence, it enables users to apply for multiple jobs in a tailored way.","summaryCn":"AI帮你自动投简历找工作，解放双手不用手点","reasonCn":"AIHawk是一个AI驱动的求职自动化工具，能够分析职位描述、优化简历匹配度、自动填写申请表并批量投递。通过集成AI能力，它可以根据岗位要求自动调整简历内容，支持多渠道职位搜索和申请，显著减少求职过程中的重复劳动。项目star数超过3万，适合希望通过自动化提升求职效率的开发者和技术人才。","stars":31634,"starGrowth":8,"createdAt":"2024-08-04T12:15:06Z","silentRounds":0,"language":"Python","topics":["automation"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["求职自动化","AI Agent","简历优化","浏览器自动化"],"domainKey":"AI Agent","pushedAt":"2026-08-18T00:42:09Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"automation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/feder-cr/Jobs_Applier_AI_Agent_AIHawk","ts":"2026-08-18T00:42:09Z","category":"ai","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024758677372447368,"copyOk":true},{"repo":"HKUSTDial/DeepEar","owner":"HKUSTDial","name":"DeepEar","desc":"DeepEar ｜ 顺风耳  An open-source framework for Deep Research and Financial Signal Tracking.  一个用于深度研究与金融信号追踪的开源框架。","summaryCn":"在瞬息万变的金融市场里，谁能更快获取并读懂研报，谁就掌握了先机。","reasonCn":"架构上，它采用了模块化的设计思路，核心组件包括信号采集器、大模型分析器和结果输出层。它能与各种金融数据源深度集成，利用向量化处理技术将非结构化的文本转为可计算的金融信号。这种架构不仅保证了数据的实时性，还通过深度学习模型确保了分析的专业度，使得产出的内容具备极","stars":285,"starGrowth":0,"createdAt":"2025-12-29T11:47:39Z","silentRounds":7,"language":"Python","topics":["finance"],"aiDims":["AI搜索","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["金融分析","研报处理","量化信号"],"domainKey":"金融分析","pushedAt":"2026-08-22T14:09:29Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/HKUSTDial/DeepEar","ts":"2026-08-22T14:09:29Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"get-convex/convex-backend","owner":"get-convex","name":"convex-backend","desc":"The open-source reactive database for app developers","summaryCn":"让后端开发像写前端一样简单，自带数据库和实时同步的全家桶","reasonCn":"Convex是一个专为应用开发者设计的开源响应式数据库后端。它把数据库、API和后台任务整合在一起，支持TypeScript全栈开发。通过函数触发器和实时订阅，数据变化能自动同步到前端，省去了传统后端开发中繁琐的API对接工作。","stars":12598,"starGrowth":1,"createdAt":"2024-03-08T07:23:15Z","silentRounds":1,"language":"TypeScript","topics":["backend","database"],"aiDims":["非AI-实用","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["后端","数据库","实时同步","TypeScript"],"domainKey":"后端","pushedAt":"2026-09-23T20:14:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7},{"name":"database","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/get-convex/convex-backend","ts":"2026-09-23T20:14:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"elastic/kibana","owner":"elastic","name":"kibana","desc":"Your window into all of your data","summaryCn":"开源界的全能数据仪表盘，把乱七八糟的数据变成直观图表。","reasonCn":"基于Elastic Stack生态的交互式数据可视化平台，提供丰富的图表组件和仪表盘模板，支持实时数据流展示。与BI工具相比，它更擅长处理日志和时序数据的深度分析，支持拖拽式操作和动态仪表盘配置，是运维和数据分析团队的核心工具。","stars":21297,"starGrowth":1,"createdAt":"2013-01-26T04:00:59Z","silentRounds":1,"language":"TypeScript","topics":["kibana","elasticsearch","visualizations","metrics","observability","dashboards","hacktoberfest"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["数据可视化","日志分析","Elasticsearch"],"domainKey":"数据可视化","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"kibana","source":"github","weight":0.7},{"name":"elasticsearch","source":"github","weight":0.7},{"name":"visualizations","source":"github","weight":0.7},{"name":"metrics","source":"github","weight":0.7},{"name":"observability","source":"github","weight":0.7},{"name":"dashboards","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/elastic/kibana","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Willxup/cpa-usage-keeper","owner":"Willxup","name":"cpa-usage-keeper","desc":"Standalone CliProxyAPI usage tracker with SQLite persistence and built-in dashboard.","summaryCn":"一款本地运行的API流量监控工具，实时追踪你的消费","reasonCn":"这个工具是一个独立的命令行程序，专门用来监控和管理你的API代理流量。它使用SQLite数据库持久化存储数据，并内置了一个轻量级的仪表盘来展示统计数据。开发者可以通过CLI命令快速查看API的使用情况，有效控制成本和流量。","stars":1205,"starGrowth":1,"createdAt":"2026-04-17T17:46:30Z","silentRounds":1,"language":"Go","topics":["analysis","cliproxyapi","cpa","dashboard","token-usage","token-analysis","claude","codex","openai-compatible","antigravity","gemini"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["API监控","流量追踪","成本管控"],"domainKey":"API监控","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"analysis","source":"github","weight":0.7},{"name":"cliproxyapi","source":"github","weight":0.7},{"name":"cpa","source":"github","weight":0.7},{"name":"dashboard","source":"github","weight":0.7},{"name":"token-usage","source":"github","weight":0.7},{"name":"token-analysis","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"openai-compatible","source":"github","weight":0.7},{"name":"antigravity","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/Willxup/cpa-usage-keeper","ts":"2026-08-14T01:23:41.092Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenSenseNova/SenseNova-Skills","owner":"OpenSenseNova","name":"SenseNova-Skills","desc":"Modular SenseNova skills for building AI-powered office assistants and productivity workflows","summaryCn":"给打工人配个AI秘书，帮你自动处理各种办公琐事","reasonCn":"SenseNova-Skills 提供模块化技能包，基于大语言模型快速搭建智能办公助手。采用插件式架构，支持邮件处理、日程管理、文档分析等工具调用，实现工作流自动化。JavaScript开发，易于集成，适合需要定制化AI办公方案的个人与团队。","stars":5677,"starGrowth":0,"createdAt":"2026-04-14T06:21:51Z","silentRounds":1,"language":"JavaScript","topics":["agent","agent-skills","ai-agents","ai-assistant","data-analysis","document-processing","office-automation","presentation-slides"],"aiDims":["AI Agent","工具调用","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["办公自动化","技能包","Agent"],"domainKey":"Agent","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"agent-skills","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-assistant","source":"github","weight":0.7},{"name":"data-analysis","source":"github","weight":0.7},{"name":"document-processing","source":"github","weight":0.7},{"name":"office-automation","source":"github","weight":0.7},{"name":"presentation-slides","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/OpenSenseNova/SenseNova-Skills","ts":"2026-08-16T00:54:51.023Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Project-HAMi/HAMi","owner":"Project-HAMi","name":"HAMi","desc":"Heterogeneous GPU Sharing on Kubernetes","summaryCn":"把昂贵显卡变成拼车座，谁想用谁占一份更划算","reasonCn":"它把容器集群里的不同显卡当成可切分的公共资源，通过设备插件、自定义调度和显存隔离，把同一块显卡分给多个容器。支持算力百分比、显存上限、显存监控、设备拓扑感知，还能处理多厂商驱动差异，让训练和推理任务在同一集群里排队更公平，资源利用率明显更高。","stars":4659,"starGrowth":0,"createdAt":"2021-09-14T11:51:49Z","silentRounds":1,"language":"Go","topics":["vgpu","gpu-virtualization","device-plugin","gpu-management","kubernetes-gpu-cluster","ascend","cambricon","gpu-shareable","kubernetes","metax","nvidia","hygon","iluvatar","mthreads","cncf","gpu-acceleration","vgpu-hypervisor"],"aiDims":["AI基础设施","模型部署","模型与训练"],"zone":"AI","funScore":0.6,"domainTags":["GPU 共享","Kubernetes","设备调度","异构计算","资源管理"],"domainKey":"GPU 共享","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"vgpu","source":"github","weight":0.7},{"name":"gpu-virtualization","source":"github","weight":0.7},{"name":"device-plugin","source":"github","weight":0.7},{"name":"gpu-management","source":"github","weight":0.7},{"name":"kubernetes-gpu-cluster","source":"github","weight":0.7},{"name":"ascend","source":"github","weight":0.7},{"name":"cambricon","source":"github","weight":0.7},{"name":"gpu-shareable","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"metax","source":"github","weight":0.7},{"name":"nvidia","source":"github","weight":0.7},{"name":"hygon","source":"github","weight":0.7},{"name":"iluvatar","source":"github","weight":0.7},{"name":"mthreads","source":"github","weight":0.7},{"name":"cncf","source":"github","weight":0.7},{"name":"gpu-acceleration","source":"github","weight":0.7},{"name":"vgpu-hypervisor","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.74,"source":"trending","url":"https://github.com/Project-HAMi/HAMi","ts":"2026-09-09T22:18:37.171Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"base/base","owner":"base","name":"base","desc":"All components used to run Base","summaryCn":"Rust写的超快Web框架，不仅AI还能做任何事","reasonCn":"Base 是一个基于 Rust 构建的 Web 应用运行时，旨在提供极致的性能和并发处理能力。它不仅支持 AI 模型的部署与推理，还具备处理传统 Web 服务的高效性。核心优势在于其零成本抽象和内存安全特性，能够轻松应对高并发请求。","stars":894,"starGrowth":0,"createdAt":"2025-02-03T14:38:16Z","silentRounds":6,"language":"Rust","topics":[],"aiDims":["推理引擎","AI基础设施","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["Rust运行时","Web框架","高性能计算"],"domainKey":"Rust运行时","pushedAt":"2026-09-10T22:16:01.544Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/base/base","ts":"2026-09-10T22:16:01.544Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"pipeshub-ai/pipeshub-ai","owner":"pipeshub-ai","name":"pipeshub-ai","desc":"PipesHub is an open-source fully extensible AI context layer that unifies your business data for explainable enterprise search and agentic workflow automation.","summaryCn":"企业级AI知识库，帮你把杂乱文档变成能听懂话的智能助手","reasonCn":"这是一个基于向量检索和LLM的统一AI层，将企业分散的数据源整合到RAG架构中。它提供可解释的搜索和Agent工作流自动化，支持从文本、代码到表格等多模态数据的解析与索引。项目提供Python SDK，可快速集成到现有业务流程。","stars":3768,"starGrowth":0,"createdAt":"2025-03-06T18:51:10Z","silentRounds":1,"language":"Python","topics":["agent","agents","glean","knowledge-graph","langchain","langgraph","notion","ollama","python","rag","slack","ai","docker","enterprise-search","fastapi","graphrag","mcp","retrieval-augmented-generation","self-hosted","workplace-ai"],"aiDims":["AI Agent","AI基础设施","AI搜索"],"zone":"AI","funScore":0,"domainTags":["RAG","企业搜索","向量数据库"],"domainKey":"RAG","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"agent","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"glean","source":"github","weight":0.7},{"name":"knowledge-graph","source":"github","weight":0.7},{"name":"langchain","source":"github","weight":0.7},{"name":"langgraph","source":"github","weight":0.7},{"name":"notion","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"slack","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"enterprise-search","source":"github","weight":0.7},{"name":"fastapi","source":"github","weight":0.7},{"name":"graphrag","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"retrieval-augmented-generation","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"workplace-ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"trending","url":"https://github.com/pipeshub-ai/pipeshub-ai","ts":"2026-08-19T01:08:47.324Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"virtualramblas/Domain-Specific-Small-Language-Models","owner":"virtualramblas","name":"Domain-Specific-Small-Language-Models","desc":"Repository for the companion Colab notebook of the Domain-Specific Small Language Models book.","summaryCn":"一个基于Jupyter Notebook的领域特定语言模型库，用于教育和研究","reasonCn":"这个项目提供了一种使用Jupyter Notebook进行领域特定语言模型训练和评估的工具。它允许用户在笔记本上运行各种自然语言处理任务，如文本分类、情感分析等。","stars":90,"starGrowth":0,"createdAt":"2025-03-19T21:38:57Z","silentRounds":19,"language":"Jupyter Notebook","topics":["transformers"],"aiDims":["AI基础设施"],"zone":"资源","funScore":0,"domainTags":["语言模型","教育研究","自然语言处理","Jupyter"],"domainKey":"语言模型","pushedAt":"2026-08-27T08:55:42Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/virtualramblas/Domain-Specific-Small-Language-Models","ts":"2026-08-27T08:55:42Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"oldany/dropmind","owner":"oldany","name":"dropmind","desc":"Self-hosted capture layer to save anything from anywhere.","summaryCn":"你是不是经常遇到这种情况：在手机上看到一段很有用的代码、一句名言，或者","reasonCn":"它的核心功能是“全场景捕获”。作为一个自托管服务，你可以把它部署在任何服务器上。无论是浏览器、手机App还是命令行，只要复制内容，DropMind 就能通过API将其捕获并存储。它支持文本、图片、文件等多种格式，并且提供了Web端和移动端界面，让你能随时随地查看这些“大脑碎片”。","stars":141,"starGrowth":0,"createdAt":"2026-02-27T22:11:38Z","silentRounds":12,"language":"JavaScript","topics":["knowledge-management"],"aiDims":["非AI-实用","知识管理"],"zone":"工具","funScore":0,"domainTags":["信息捕获","自托管","剪贴板管理","知识库"],"domainKey":"信息捕获","pushedAt":"2026-09-13T12:21:27Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"知识管理","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/oldany/dropmind","ts":"2026-09-13T12:21:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"konten-studio/jekardah-writer","owner":"konten-studio","name":"jekardah-writer","desc":"Portable AI writing workflow for hooks, anti-slop editing, and natural Jabodetabek voice.","summaryCn":"专为印尼雅加达本地语境和吸睛文案打造的便携AI写作工作流","reasonCn":"这是一个基于Shell脚本和现代AI大模型构建的便携写作自动化工具。它专注于黄金开头钩子的设计、AI假大空词汇的清洗净化，以及高度本地化的印尼都市风情表达。通过高度精简的命令行交互，它帮助内容创作者摆脱死板的机器味，输出真正接地气且具备强吸引力的社媒文案。","stars":28,"starGrowth":0,"createdAt":"2026-08-16T00:30:55Z","silentRounds":15,"language":"Shell","topics":["writing"],"aiDims":["AI写作","AI应用","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI写作","本地化文案","提示词工程","Shell脚本"],"domainKey":"AI写作","pushedAt":"2026-09-09T07:59:06Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/konten-studio/jekardah-writer","ts":"2026-09-09T07:59:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xuebinqin/U-2-Net","owner":"xuebinqin","name":"U-2-Net","desc":"","summaryCn":"U²-Net（U Square Net）是 Qing Xuebin 等人提出的一种用于显著性目标检测（","reasonCn":"除了学术荣誉，U²-Net 的另一个重要特点是其轻量化和可部署性。参考文档中提到，该模型已集成到 PlayTorch 平台，用户可以在数分钟内构建 demo 并直接在 Android 或 iOS 手机上运行。这意味着开发者无需依赖强大的云端服务器，就可以在手机端实现实时的图像显著性分割，这在移动端","stars":9869,"starGrowth":2,"createdAt":"2020-05-04T17:10:28Z","silentRounds":0,"language":"Python","topics":["computer-vision","deep-learning","image-background-removal","image-processing","image-segmentation","u-2-net","u2net"],"aiDims":["模型与训练","论文"],"zone":"AI","funScore":0.5,"domainTags":["显著性检测","图像分割","PyTorch","深度学习","移动端部署"],"domainKey":"显著性检测","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"computer-vision","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"image-background-removal","source":"github","weight":0.7},{"name":"image-processing","source":"github","weight":0.7},{"name":"image-segmentation","source":"github","weight":0.7},{"name":"u-2-net","source":"github","weight":0.7},{"name":"u2net","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/xuebinqin/U-2-Net","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JaySpiffy/void-reckoning-engine","owner":"JaySpiffy","name":"void-reckoning-engine","desc":"Persistent 4X strategy simulation with evolving factions, tactical combat replays, terminal HUD, and analytics. Python + Rust core.","summaryCn":"简单来说，它就像是把一部史诗级的太空科幻电影变成了一个后台运行的实时数","reasonCn":"技术上，项目采取了Python作为上层逻辑控制，负责编写规则与策略脚本，而底层计算则使用Rust编写，确保在大规模单位演算下仍能保持丝滑。这种架构兼顾了开发的灵活性与系统的执行效率，是现代开源工具组合的典范，完美展示了如何利用跨语言优势处理复杂模拟。","stars":11,"starGrowth":0,"createdAt":"2026-02-01T20:36:20Z","silentRounds":17,"language":"Python","topics":["procedural-generation"],"aiDims":["可视化","非AI-实用"],"zone":"创意","funScore":1,"domainTags":["4X策略","科幻","Python","Rust","实时模拟"],"domainKey":"4X策略","pushedAt":"2026-08-12T07:10:07Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/JaySpiffy/void-reckoning-engine","ts":"2026-08-12T07:10:07Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"daviden1013/llm-ie","owner":"daviden1013","name":"llm-ie","desc":"A comprehensive toolkit that provides building blocks for LLM-based named entity recognition, attribute extraction, and relation extraction pipelines.","summaryCn":"让大模型帮你自动读文档、抽实体、理关系，搭建NLP管道的神器","reasonCn":"这是一个基于大语言模型的实体抽取、属性提取和关系抽取工具包。它提供了完整的流水线组件，利用LLM强大的语义理解能力，解决传统NLP任务中标注成本高、泛化性差的问题。项目包含数据预处理、模型微调及推理流程的现成模块，方便开发者快速构建垂直领域的知识图谱。","stars":66,"starGrowth":0,"createdAt":"2024-08-01T21:37:53Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","知识图谱","文档理解"],"zone":"AI","funScore":0,"domainTags":["信息抽取","实体识别","知识图谱"],"domainKey":"知识图谱","pushedAt":"2026-08-19T15:01:11Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.9,"source":"search","url":"https://github.com/daviden1013/llm-ie","ts":"2026-08-19T15:01:11Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"BjornMelin/docmind-ai-llm","owner":"BjornMelin","name":"docmind-ai-llm","desc":"DocMind AI is a powerful, open-source Streamlit application leveraging LlamaIndex, LangGraph, and local Large Language Models (LLMs) via Ollama, LMStudio, llama.cpp, or vLLM for advanced document analysis. Analyze, summarize, and extract insights from a wide array of file formats, securely and privately, all offline.","summaryCn":"把各种本地大模型和知识库连起来，打造你的私有化智能文档分析助手。","reasonCn":"这是一个功能强劲的开源Streamlit应用，完美融合了LlamaIndex与LangGraph两大框架。它允许用户自由选择Ollama、LMStudio或vLLM等本地大模型工具，在完全离线的环境下对海量复杂文档进行深度解析与智能问答，兼顾了高效能与数据安全。","stars":153,"starGrowth":0,"createdAt":"2025-01-21T23:54:58Z","silentRounds":19,"language":"Python","topics":["transformers"],"aiDims":["RAG与知识","文档理解","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["文档问答","RAG","本地大模型"],"domainKey":"RAG","pushedAt":"2026-08-19T00:25:54Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"search","url":"https://github.com/BjornMelin/docmind-ai-llm","ts":"2026-08-19T00:25:54Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GH-X-ST/Aeronautical-Engineering-Imperial","owner":"GH-X-ST","name":"Aeronautical-Engineering-Imperial","desc":"Imperial College London Aeronautical Engineering (MEng 4YFT) Study Materials","summaryCn":"帝国理工航空工程四年制硕士全套资料，学霸笔记大公开。","reasonCn":"这是一个汇集了帝国理工学院航空工程硕士课程（MEng 4YFT）学习资料的仓库，涵盖了从数学基础到空气动力学、结构力学等核心专业课的课件、笔记和作业答案。对于想要深入了解航空航天工程的学生或爱好者来说，这是不可多得的学习宝库。","stars":63,"starGrowth":0,"createdAt":"2025-12-25T16:32:33Z","silentRounds":14,"language":"C++","topics":["mathematics"],"aiDims":["学习资源","教程","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["帝国理工","航空工程","学习资料","MEng课程","课件笔记"],"domainKey":"帝国理工","pushedAt":"2026-09-10T08:37:33Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"mathematics","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/GH-X-ST/Aeronautical-Engineering-Imperial","ts":"2026-09-10T08:37:33Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dubinc/dub","owner":"dubinc","name":"dub","desc":"The modern link attribution platform. Loved by world-class marketing teams like Framer, Perplexity, Superhuman, Twilio, Buffer and more.","summaryCn":"给营销团队用的现代链接追踪神器，简单好用不折腾。","reasonCn":"Dub 是一款现代化的链接归因与追踪平台，专为 Framer、Perplexity 等世界级团队打造。它基于 TypeScript 构建，提供无需代码的链接管理系统，支持自定义域名、二维码生成、点击数据可视化以及团队协作。项目以易用性和强大的功能集成著称，解决了传统链接追踪工具复杂且难以维护的痛点，深受开发者与营销人员喜爱。","stars":24802,"starGrowth":0,"createdAt":"2022-08-27T22:15:11Z","silentRounds":1,"language":"TypeScript","topics":["vercel","nextjs","upstash","tinybird","clickhouse","mysql","open-source","planetscale","prisma","tailwindcss","typescript"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["链接追踪","归因分析","营销工具"],"domainKey":"链接追踪","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vercel","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"upstash","source":"github","weight":0.7},{"name":"tinybird","source":"github","weight":0.7},{"name":"clickhouse","source":"github","weight":0.7},{"name":"mysql","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"planetscale","source":"github","weight":0.7},{"name":"prisma","source":"github","weight":0.7},{"name":"tailwindcss","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/dubinc/dub","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"evershopcommerce/evershop","owner":"evershopcommerce","name":"evershop","desc":"🛍️ Typescript E-commerce Platform","summaryCn":"基于TypeScript开发的现代化开源电商系统，搭建网店秒变简单。","reasonCn":"这是一个全栈电商解决方案，采用Headless架构设计，前端、后端和数据库完全解耦。项目使用TypeScript编写，提供RESTful API和GraphQL接口，便于二次开发和集成。内置强大的营销工具如优惠券、捆绑销售等，支持多语言、多货币，非常适合中小企业快速搭建专业级在线商店。","stars":10490,"starGrowth":1,"createdAt":"2021-05-06T09:23:37Z","silentRounds":3,"language":"TypeScript","topics":["ecommerce","ecommerce-framework","ecommerce-platform","react","typescript"],"aiDims":["非AI-实用","电商","开源项目"],"zone":"工具","funScore":0,"domainTags":["电商系统","Headless架构","TypeScript","在线商店","支付网关"],"domainKey":"电商系统","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"电商","source":"llm","weight":1},{"name":"开源项目","source":"llm","weight":1},{"name":"ecommerce","source":"github","weight":0.7},{"name":"ecommerce-framework","source":"github","weight":0.7},{"name":"ecommerce-platform","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/evershopcommerce/evershop","ts":"2026-08-18T00:56:27.274Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"charmbracelet/vhs","owner":"charmbracelet","name":"vhs","desc":"Your CLI home video recorder 📼","summaryCn":"用代码录 GIF 和 MP4，终端炫酷演示视频的神器","reasonCn":"Go 语言编写的终端视频录制工具，用 VHS 文件描述每一帧操作，自动生成带打字效果、光影效果和注释的视频。支持输出 GIF、MP4、WebM 等格式，让命令行演示也能拥有播客级别的视觉效果，非常适合做技术分享和教程发布。","stars":20958,"starGrowth":3,"createdAt":"2022-07-19T14:28:36Z","silentRounds":0,"language":"Go","topics":["command-line"],"aiDims":["非AI-好玩","可视化"],"zone":"工具","funScore":0.5,"domainTags":["终端录制","GIF生成","CLI工具","视频演示","命令行"],"domainKey":"终端录制","pushedAt":"2026-09-09T09:37:32Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/charmbracelet/vhs","ts":"2026-09-09T09:37:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"seakee/CPA-Manager-Plus","owner":"seakee","name":"CPA-Manager-Plus","desc":"A self-hosted CPA / CLIProxyAPI management panel and AI gateway observability dashboard for requests, usage, cost, quota, failures, and account health.","summaryCn":"自托管的大模型管理面板，帮你管账又监控接口","reasonCn":"这是一款基于Go语言开发的开源管理面板，专注于自建AI网关和CLI代理的运维。它能统一管理多个API密钥、监控请求流量与成本、追踪失败率，还能实时查看账号健康状态。支持OpenAI、Anthropic等主流模型接口，适合需要控制成本、确保数据隐私的团队部署。","stars":3596,"starGrowth":9,"createdAt":"2026-05-17T09:41:43Z","silentRounds":0,"language":"Go","topics":["ai-gateway","claude-code","codex","cost-analysis","gateway-monitoring","observability","openai","openai-compatible","quota-management","request-monitoring","usage-tracking","cliproxyapi","cpa"],"aiDims":["AI基础设施","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["AI网关","API管理","成本监控","自托管","运维面板"],"domainKey":"AI网关","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"codex","source":"github","weight":0.7},{"name":"cost-analysis","source":"github","weight":0.7},{"name":"gateway-monitoring","source":"github","weight":0.7},{"name":"observability","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"openai-compatible","source":"github","weight":0.7},{"name":"quota-management","source":"github","weight":0.7},{"name":"request-monitoring","source":"github","weight":0.7},{"name":"usage-tracking","source":"github","weight":0.7},{"name":"cliproxyapi","source":"github","weight":0.7},{"name":"cpa","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/seakee/CPA-Manager-Plus","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04141001112347052,"copyOk":false},{"repo":"OpenWhispr/openwhispr","owner":"OpenWhispr","name":"openwhispr","desc":"Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform.","summaryCn":"本地离线语音转文字神器，隐私安全跨平台，Whisper模型驱动","reasonCn":"OpenWhispr是基于OpenAI Whisper的语音转文字应用，支持本地离线识别和云端BYOK两种模式，隐私友好、跨平台运行。无需联网即可转写音频，支持多种语言，适合会议记录、播客转录、视频字幕等场景，开箱即用无需复杂配置。","stars":8461,"starGrowth":9,"createdAt":"2025-06-19T00:58:14Z","silentRounds":0,"language":"JavaScript","topics":["ai","open-source","whisper","anthropic","cross-platform","gemini","groq","linux","macos","openai","speech-to-text","transcribe","windows","nvidia","parakeet"],"aiDims":["AI应用","多模态Agent"],"zone":"工具","funScore":0,"domainTags":["语音识别","隐私保护","离线处理","跨平台","音频处理"],"domainKey":"语音识别","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"whisper","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"groq","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"speech-to-text","source":"github","weight":0.7},{"name":"transcribe","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"nvidia","source":"github","weight":0.7},{"name":"parakeet","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/OpenWhispr/openwhispr","ts":"2026-09-08T22:27:02.698Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.030191112161683017,"copyOk":true},{"repo":"RiccardoBiosas/awesome-MLSecOps","owner":"RiccardoBiosas","name":"awesome-MLSecOps","desc":"A curated list of MLSecOps tools and resources for securing machine learning and AI systems - adversarial ML defense, LLM security, AI red teaming, model scanning, supply-chain protection, and MLOps pipeline security.","summaryCn":"AI安全全家桶清单，从模型防御到供应链保护一网打尽","reasonCn":"项目分成了模型安全、LLM安全、MLOps安全等多个板块。模型安全部分包含对抗攻击与防御工具，比如FGSM、PGD等经典攻击方法的实现，以及对抗训练框架。LLM安全板块涵盖提示注入检测、红队测试工具、模型水印方案等。MLOps安全部分关注模型供应链，包括依赖扫描、模型完整性校验、部署","stars":470,"starGrowth":1,"createdAt":"2023-04-01T10:06:32Z","silentRounds":14,"language":"Astro","topics":["llm-model"],"aiDims":["安全工具","AI基础设施"],"zone":"资源","funScore":0,"domainTags":["MLSecOps","AI安全","对抗攻击防御","MLOps安全"],"domainKey":"MLSecOps","pushedAt":"2026-09-09T14:05:10Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Astro","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/RiccardoBiosas/awesome-MLSecOps","ts":"2026-09-09T14:05:10Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Robbyant/lingbot-world","owner":"Robbyant","name":"lingbot-world","desc":"Advancing Open-source World Models","summaryCn":"一图生一世界，开源世界模型让视频生成从此有了「常识」","reasonCn":"lingbot-world是开源的世界模型项目，专注于将静态图片转化为连贯视频。它采用先进的World Model技术，让生成的视频不仅美观，还能理解物理世界的运动规律和时间一致性。这是AIGC领域的前沿探索，让视频创作门槛大幅降低。","stars":4482,"starGrowth":0,"createdAt":"2026-01-28T04:52:50Z","silentRounds":1,"language":"Python","topics":["aigc","image-to-video","lingbot-world","video-generation","world-models"],"aiDims":["多模态Agent","AI应用","非AI-好玩"],"zone":"AI","funScore":1,"domainTags":["世界模型","视频生成","AIGC","图像到视频","深度学习"],"domainKey":"视频生成","pushedAt":"2026-07-09T10:39:39Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"aigc","source":"github","weight":0.7},{"name":"image-to-video","source":"github","weight":0.7},{"name":"lingbot-world","source":"github","weight":0.7},{"name":"video-generation","source":"github","weight":0.7},{"name":"world-models","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.88,"source":"bigbro","url":"https://github.com/Robbyant/lingbot-world","ts":"2026-07-09T10:39:39Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"dyad-sh/dyad","owner":"dyad-sh","name":"dyad","desc":"Local, open-source AI app builder for power users ✨ v0 / Lovable / Replit / Bolt alternative 🌟 Star if you like it!","summaryCn":"本地AI应用生成器像v0但完全开源免费数据全在你自己手里","reasonCn":"开源本地AI应用构建器，替代v0/Lovable等云端服务。支持多种AI模型接入，通过自然语言描述生成完整应用。代码完全本地运行，隐私安全无限制。技术栈TypeScript，提供可视化界面和代码生成能力，适合想要本地开发环境的用户。","stars":21602,"starGrowth":4,"createdAt":"2025-04-11T06:33:48Z","silentRounds":0,"language":"TypeScript","topics":["ai-app-builder","bolt","lovable","v0","anthropic","artificial-intelligence","deepseek","gemini","generative-ai","github","llm","llms","nextjs","ollama","openai","qwen","react","typescript","vercel"],"aiDims":["AI应用","AI Agent"],"zone":"AI","funScore":0.5,"domainTags":["AI应用生成","本地开发","开源替代","隐私优先"],"domainKey":"AI应用生成","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"ai-app-builder","source":"github","weight":0.7},{"name":"bolt","source":"github","weight":0.7},{"name":"lovable","source":"github","weight":0.7},{"name":"v0","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"artificial-intelligence","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"github","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"qwen","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"vercel","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/dyad-sh/dyad","ts":"2026-08-11T09:45:50.026Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"instaloader/instaloader","owner":"instaloader","name":"instaloader","desc":"","summaryCn":"这款开源工具能够帮你一键下载社交平台的公开照片、视频及相关元数据信息。","reasonCn":"Instaloader 是一款功能极其硬核的社交媒体数据采集神器，它通过极简的命令行界面，让用户能够轻松将指定的公开主页内容批量保存至本地。无论是备份珍贵的个人动态，还是获取特定账号的素材库，它都能凭借强大的解析能力完成任务。该工具不仅能够处理复杂的媒体流，还会自动抓取与之关联的元数据，对于需要长期存储数据或进行离线研究的用户来说，是一个高效且无需额外界面负担的自动化利器。","stars":13420,"starGrowth":1,"createdAt":"2016-06-15T10:40:37Z","silentRounds":1,"language":"Python","topics":["instagram","instagram-client","instagram-feed","downloader","videos","pictures","instagram-user-photos","instagram-photos","instagram-metadata","instagram-downloader","instagram-stories","instaloader","osint","osint-python","instagram-scraper"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["社交媒体爬取","Instagram","数据采集"],"domainKey":"社交媒体爬取","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"instagram","source":"github","weight":0.7},{"name":"instagram-client","source":"github","weight":0.7},{"name":"instagram-feed","source":"github","weight":0.7},{"name":"downloader","source":"github","weight":0.7},{"name":"videos","source":"github","weight":0.7},{"name":"pictures","source":"github","weight":0.7},{"name":"instagram-user-photos","source":"github","weight":0.7},{"name":"instagram-photos","source":"github","weight":0.7},{"name":"instagram-metadata","source":"github","weight":0.7},{"name":"instagram-downloader","source":"github","weight":0.7},{"name":"instagram-stories","source":"github","weight":0.7},{"name":"instaloader","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"osint-python","source":"github","weight":0.7},{"name":"instagram-scraper","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"tier","url":"https://github.com/instaloader/instaloader","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"EasyJailbreak/EasyJailbreak","owner":"EasyJailbreak","name":"EasyJailbreak","desc":"An easy-to-use Python framework to generate adversarial jailbreak prompts.","summaryCn":"一键生成绕过AI安全审查的提示词，专门用来测试LLM的底线。","reasonCn":"这个项目的厉害之处在于它把“黑魔法”变成了“标准化流程”。传统的测试可能需要人工一个个写提示词去试探，效率低且容易漏。EasyJailbreak 则提供了一套框架，内置了多种攻击策略，比如角色扮演、指令注入、逻辑陷阱等等。","stars":911,"starGrowth":0,"createdAt":"2024-01-31T09:04:31Z","silentRounds":2,"language":"Python","topics":["llm-model"],"aiDims":["提示工程","AI安全","大语言模型"],"zone":"工具","funScore":0.6,"domainTags":["LLM安全","对抗样本","红队测试"],"domainKey":"LLM安全","pushedAt":"2026-09-01T14:01:43Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI安全","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/EasyJailbreak/EasyJailbreak","ts":"2026-09-01T14:01:43Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xlrpa/FlowBot","owner":"xlrpa","name":"FlowBot","desc":"微信，企微，企微机器人，微信机器人，workbot，worktool，wetool，RPA，协议，vbot，逆向，hook，安全稳定零封号的RPA机器人，支持通过API控制企业微信来实现聚合聊天、自动回复、群发消息、自动通过好友、定时发布朋友圈等","summaryCn":"企业微信机器人平台，零封号自动回复、群发、朋友圈定时发布。","reasonCn":"FlowBot 是基于 Java 的企业微信机器人框架，提供底层协议实现、消息收发、自动回复、群发、好友验证、定时朋友圈等功能，并封装统一的 API，支持通过 HTTP、Webhook 或 SDK 调用，帮助企业在不触及官方限制的前提下，实现高效的内部沟通自动化和 RPA 场景。","stars":1390,"starGrowth":0,"createdAt":"2024-06-20T04:09:12Z","silentRounds":1,"language":"Java","topics":["chatbot"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["RPA","WeCom","Automation"],"domainKey":"RPA","pushedAt":"2026-09-11T04:31:46Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/xlrpa/FlowBot","ts":"2026-09-11T04:31:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"eyaltoledano/claude-task-master","owner":"eyaltoledano","name":"claude-task-master","desc":"An AI-powered task-management system you can drop into Cursor, Lovable, Windsurf, Roo, and others.","summaryCn":"把Claude变成你的全能办公管家，集成到各种编辑器里自动干活。","reasonCn":"这是一个专为Cursor等代码编辑器设计的Claude智能任务管理插件，通过工具调用能力让Claude直接操作文件、执行命令和构建应用。它支持在编辑器内直接定义复杂工作流，自动执行Git操作、运行测试、重构代码等任务，无需手动输入指令。插件提供清晰的日志视图和任务状态反馈，能显著提升开发效率。","stars":28089,"starGrowth":3,"createdAt":"2025-03-04T18:54:54Z","silentRounds":0,"language":"JavaScript","topics":["ai","cursor","task-manager","tasks","tasks-list","cursor-ai","cursorai","lovable","lovable-dev","roocode","windsurf","windsurf-ai"],"aiDims":["AI应用","工具调用","开发工具"],"zone":"AI","funScore":0.5,"domainTags":["Claude","代码编辑器","任务管理","AI Agent"],"domainKey":"AI Agent","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"task-manager","source":"github","weight":0.7},{"name":"tasks","source":"github","weight":0.7},{"name":"tasks-list","source":"github","weight":0.7},{"name":"cursor-ai","source":"github","weight":0.7},{"name":"cursorai","source":"github","weight":0.7},{"name":"lovable","source":"github","weight":0.7},{"name":"lovable-dev","source":"github","weight":0.7},{"name":"roocode","source":"github","weight":0.7},{"name":"windsurf","source":"github","weight":0.7},{"name":"windsurf-ai","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/eyaltoledano/claude-task-master","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"antvis/Infographic","owner":"antvis","name":"Infographic","desc":"🦋 An Infographic Generation and Rendering Framework, bring words to life with AI!","summaryCn":"在数据展示中，文字描述往往不如图表直观。","reasonCn":"它的核心技术在于AI与可视化引擎的结合。通过自然语言处理技术，它能理解文字中的数据关系和逻辑结构，然后调用AntV家族强大的可视化引擎（如G2、G6）进行渲染。支持折线图、柱状图、饼图、桑基图等多种图表类型，甚至能生成动态交互图表。","stars":6848,"starGrowth":2,"createdAt":"2025-09-11T11:13:39Z","silentRounds":0,"language":"TypeScript","topics":["infographic","llm","visualization","ai","antv"],"aiDims":["创意工具","可视化","AI应用"],"zone":"工具","funScore":0.7,"domainTags":["数据可视化","信息图表","自然语言处理","AntV","图表渲染"],"domainKey":"数据可视化","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"infographic","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"visualization","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"antv","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/antvis/Infographic","ts":"2026-08-13T01:37:33.040Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"abhisheknaiidu/awesome-github-profile-readme","owner":"abhisheknaiidu","name":"awesome-github-profile-readme","desc":"😎 A curated list of awesome GitHub Profile which updates in real time ","summaryCn":"GitHub主页太丑？这里有30k个大神模板帮你秒变极客","reasonCn":"这个仓库是一个精心筛选的GitHub个人主页配置指南合集，实时更新并收录了全球顶尖开发者的主页设计。它不仅提供了丰富的布局和配色方案，还包含了如何展示项目经验、技术栈、学习进度等实用技巧，帮助开发者打造一个专业且富有吸引力的个人品牌形象。","stars":31129,"starGrowth":3,"createdAt":"2020-07-12T12:57:25Z","silentRounds":0,"language":"","topics":["awesome"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["GitHub","个人主页","模板收集","开发者品牌"],"domainKey":"GitHub","pushedAt":"2026-09-11T23:08:19Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"awesome","source":"github","weight":0.7}],"aiScore":0.65,"source":"search","url":"https://github.com/abhisheknaiidu/awesome-github-profile-readme","ts":"2026-09-11T23:08:19Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vuejs/core","owner":"vuejs","name":"core","desc":"🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.","summaryCn":"前端开发界的“瑞士军刀”，让你像搭积木一样写网页","reasonCn":"它的核心是“双向绑定”和“组件化”。当你修改数据时，界面自动更新；当你点击界面元素时，数据也会自动变化。这让你不用手动去操作DOM（网页元素），代码像搭积木一样清晰。比如你写个计数器，只要改个变量数字，界面数字立马跟着变。","stars":54428,"starGrowth":0,"createdAt":"2018-06-12T13:49:36Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["开发者工具","AI应用","效率工具"],"zone":"工具","funScore":0,"domainTags":["前端框架","JavaScript","组件化","双向绑定"],"domainKey":"前端框架","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/vuejs/core","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/mxc","owner":"microsoft","name":"mxc","desc":"Policy-driven, layered isolation and containment","summaryCn":"微软的隔离沙箱技术，给各种程序穿上防弹衣","reasonCn":"微软出品的基于策略的层级隔离与容器化方案，用Rust编写确保内存安全和高性能。支持多租户环境下的资源隔离、权限控制和行为限制，可用于AI模型部署、敏感数据处理的沙箱场景。策略驱动的设计让安全边界灵活可调，既能防逃逸也能防越权。","stars":1354,"starGrowth":1,"createdAt":"2026-02-06T20:13:08Z","silentRounds":2,"language":"Rust","topics":[],"aiDims":["AI基础设施","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Rust","隔离","沙箱","容器"],"domainKey":"Rust","pushedAt":"2026-09-08T22:27:02.698Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/microsoft/mxc","ts":"2026-09-08T22:27:02.698Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"808StaN/OpenStudio","owner":"808StaN","name":"OpenStudio","desc":"OpenStudio is a browser and desktop DAW built with React, Web Audio API, and Electron. It brings the core beatmaking workflow into one app: browse sounds, build patterns, arrange clips, mix tracks, and export the final track to WAV or MP3. The built-in AI Agent helps shape melodies, drums, instruments, mixer settings, and FX.","summaryCn":"React打造的浏览器+桌面DAW，轻松采样、编曲、混音。","reasonCn":"很多人在电脑上想玩音乐制作，却被传统DAW的学习成本吓退。OpenStudio把音频编辑搬到了浏览器和桌面客户端，用熟悉的React生态搭建，让你打开网页就能拖拽音频片段、调节音量、添加效果，完全零门槛。\\n技术层面，它基于Web Audio API实现低延迟的音频节点网络，使用AudioWorklet完","stars":75,"starGrowth":0,"createdAt":"2026-04-04T03:42:42Z","silentRounds":1,"language":"JavaScript","topics":["music-production"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"创意","funScore":1,"domainTags":["Web Audio","DAW","音乐制作","React"],"domainKey":"Web Audio","pushedAt":"2026-09-08T15:16:47Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"music-production","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/808StaN/OpenStudio","ts":"2026-09-08T15:16:47Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jaredrhod/ai-memory-vault","owner":"jaredrhod","name":"ai-memory-vault","desc":"Give your AI a real, persistent memory. The open-source system plus templates that turn an Obsidian vault into your AI's working memory. No vector database, just markdown.","summaryCn":"抛弃笨重向量数据库直接用Obsidian双链笔记给AI存记忆","reasonCn":"核心技术思路非常巧妙且接地气。它没有搞那些晦涩难懂的嵌入向量和向量检索，而是充分发挥了Markdown纯文本和Obsidian双向链接的优势，通过精心设计的提示词模板和文件管理规范，让大模型直接去读取和更新你本地的笔记文件。","stars":671,"starGrowth":0,"createdAt":"2026-06-18T01:17:27Z","silentRounds":1,"language":"","topics":["note-taking","knowledge-management"],"aiDims":["RAG与知识","效率工具","AI应用"],"zone":"工具","funScore":0.6,"domainTags":["Obsidian","知识管理","RAG"],"domainKey":"Obsidian","pushedAt":"2026-08-31T01:56:56Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"note-taking","source":"github","weight":0.7},{"name":"knowledge-management","source":"github","weight":0.7}],"aiScore":0.82,"source":"search","url":"https://github.com/jaredrhod/ai-memory-vault","ts":"2026-08-31T01:56:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lucidrains/transfusion-pytorch","owner":"lucidrains","name":"transfusion-pytorch","desc":"Pytorch implementation of Transfusion, \"Predict the Next Token and Diffuse Images with One Multi-Modal Model\", from MetaAI","summaryCn":"用一个大模型同时搞定文本预测和图片生成的Meta前沿架构复现。","reasonCn":"技术底层上，它抛弃了传统的级联管道设计，允许模型在同一个神经网络里交替预测文本Token和生成图像的Patch。它巧妙地结合了自回归语言建模的目标函数与图像扩散损失函数，让模型既能理解复杂的上下文语义，又能直接“画”出高质感的图片，向真正统一的通用人工智能迈出了一大步。","stars":1403,"starGrowth":0,"createdAt":"2024-08-23T16:53:51Z","silentRounds":22,"language":"Python","topics":["transformers"],"aiDims":["模型与训练","多模态Agent","论文"],"zone":"AI","funScore":1,"domainTags":["多模态","大模型","扩散模型","PyTorch","Transformer"],"domainKey":"多模态","pushedAt":"2026-09-04T13:43:11Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"论文","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/lucidrains/transfusion-pytorch","ts":"2026-09-04T13:43:11Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"abhigyanpatwari/GitNexus","owner":"abhigyanpatwari","name":"GitNexus","desc":"GitNexus: The Zero-Server Code Intelligence Engine - GitNexus is a client-side knowledge graph creator that runs entirely in your browser. Drop in a git repository (Github, Gitlab, Azure, Local) or ZIP file, and get an interactive knowledge graph with a built in Graph RAG Agent. Perfect for code exploration","summaryCn":"浏览器里直接生成代码知识图谱，读懂任何开源项目再也不头大","reasonCn":"完全运行在浏览器端的本地化工具，无需后端服务器，只需粘贴仓库地址或上传本地项目，即可自动生成包含调用关系、文件依赖和类结构的知识图谱。基于TypeScript开发，性能强劲且保护代码隐私，极大提升代码理解和文档梳理效率。","stars":47524,"starGrowth":4,"createdAt":"2025-08-02T23:20:31Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["知识图谱","AI应用","可视化"],"zone":"工具","funScore":0,"domainTags":["代码可视化","知识图谱","AST解析","Git"],"domainKey":"代码可视化","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/abhigyanpatwari/GitNexus","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"anthropics/claude-code-security-review","owner":"anthropics","name":"claude-code-security-review","desc":"An AI-powered security review GitHub Action using Claude to analyze code changes for security vulnerabilities.","summaryCn":"用Claude自动扫描代码，发现并修复安全漏洞","reasonCn":"这是一个基于Claude的GitHub Action，专门用于代码安全审查。它会自动分析你的代码变更，检测潜在的安全漏洞，并提供详细的修复建议。集成到CI/CD流程中，每次提交都能自动进行安全扫描，大大提升代码安全性。Anthropic官方出品，可靠性有保障，是开发者的必备安全工具。","stars":6260,"starGrowth":0,"createdAt":"2025-08-04T16:07:27Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["工具调用","安全工具","AI应用"],"zone":"AI","funScore":0,"domainTags":["GitHub Actions","Code Review","Security"],"domainKey":"GitHub Actions","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/anthropics/claude-code-security-review","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"embabel/embabel-agent","owner":"embabel","name":"embabel-agent","desc":"Agent framework for the JVM. Pronounced Em-BAY-bel /ɛmˈbeɪbəl/","summaryCn":"给 JVM 加上大脑，让 Java 也能驾驭智能体","reasonCn":"这是一个专为 JVM 生态设计的智能体框架，核心是利用 Kotlin 的简洁特性，让开发者能轻松构建具备复杂推理和工具调用能力的 AI Agent。它不仅是简单的 SDK，更提供了一套完整的生命周期管理机制，旨在解决 Java 领域在构建复杂 AI 应用时的开发痛点，让企业级开发无缝拥抱大模型技术。","stars":4462,"starGrowth":1,"createdAt":"2025-04-10T03:06:07Z","silentRounds":1,"language":"Kotlin","topics":["agentic-ai","generative-ai","java","kotlin","spring","agents","aiagentframework","agent","ai","ai-agents","genai","llms","multi-agents","multi-agents-orchestration","multi-agents-system"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["Agent框架","JVM","Kotlin","大模型应用","企业级AI"],"domainKey":"Agent框架","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"java","source":"github","weight":0.7},{"name":"kotlin","source":"github","weight":0.7},{"name":"spring","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"aiagentframework","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"genai","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"multi-agents","source":"github","weight":0.7},{"name":"multi-agents-orchestration","source":"github","weight":0.7},{"name":"multi-agents-system","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/embabel/embabel-agent","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"epatel/discrete6502","owner":"epatel","name":"discrete6502","desc":"A MOS 6502 CPU rebuilt from 4,051 discrete transistors on a 291x322mm six-layer board that reproduces the original die layout","summaryCn":"还记得小时候红白机或者老式苹果电脑里的心脏吗？","reasonCn":"底层原理上，它完全忠实于历史上原始的6502芯片裸片布局。作者用ANTLR等工具处理电路设计数据，精心走线布局了这块六层大PCB板。每一颗晶体管的位置都和当年芯片内部的微观结构一一对应，通电之后你甚至能通过板载的LED灯直观看到电信号在逻辑门之间流淌的物理轨迹。","stars":13,"starGrowth":0,"createdAt":"2026-07-25T16:03:21Z","silentRounds":15,"language":"ANTLR","topics":["retrocomputing"],"aiDims":["非AI-好玩","学习资源","可视化"],"zone":"创意","funScore":1,"domainTags":["6502 CPU","分立元件","复古计算","PCB设计","硬件极客"],"domainKey":"6502 CPU","pushedAt":"2026-08-30T23:42:46Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"retrocomputing","source":"github","weight":0.7},{"name":"ANTLR","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/epatel/discrete6502","ts":"2026-08-30T23:42:46Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"iptv-org/iptv","owner":"iptv-org","name":"iptv","desc":"Collection of publicly available IPTV channels from all over the world","summaryCn":"收集了大量公开的IPTV频道资源","reasonCn":"由于其内容的多样性和丰富性，该项目受到了很多用户的欢迎。","stars":139300,"starGrowth":22,"createdAt":"2018-11-14T22:00:57Z","silentRounds":0,"language":"TypeScript","topics":["iptv","m3u","playlist","tv","streams"],"aiDims":["其他"],"zone":"资源","funScore":0,"domainTags":["IPTV","视频流","频道列表","媒体资源"],"domainKey":"IPTV","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"其他","source":"llm","weight":1},{"name":"iptv","source":"github","weight":0.7},{"name":"m3u","source":"github","weight":0.7},{"name":"playlist","source":"github","weight":0.7},{"name":"tv","source":"github","weight":0.7},{"name":"streams","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/iptv-org/iptv","ts":"2026-09-06T13:43:54.327Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.0664737975592247,"copyOk":false},{"repo":"VectifyAI/OpenKB","owner":"VectifyAI","name":"OpenKB","desc":"OpenKB: Open LLM Knowledge Base","summaryCn":"开源版企业级知识库，把私有数据喂给大模型","reasonCn":"它的核心功能是文档解析和索引构建。支持PDF、HTML、Markdown等多种格式，解析时会保留表格、图片、公式等复杂结构。索引过程用的是先进的Embedding模型，能把文字转成向量，方便后续检索。最棒的是它支持多种LLM后端，你可以用本地部署的Llama 3，也可以接云端的GPT-4，灵活性很高","stars":4550,"starGrowth":2,"createdAt":"2026-04-04T12:26:04Z","silentRounds":0,"language":"Python","topics":["knowledge-base","rag","ai","llm","agents","retrieval"],"aiDims":["RAG与知识","大语言模型","AI应用"],"zone":"AI","funScore":0,"domainTags":["RAG","知识库","向量检索","文档解析","企业知识"],"domainKey":"RAG","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"knowledge-base","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"retrieval","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/VectifyAI/OpenKB","ts":"2026-08-13T01:37:33.040Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"addyosmani/chatty","owner":"addyosmani","name":"chatty","desc":"ChattyUI - your private AI chat for running LLMs in the browser","summaryCn":"直接在网页浏览器里搞定大模型对话，绝对私密安全不泄露数据。","reasonCn":"利用现代浏览器的前沿WebGPU技术，把大模型直接塞进网页前端运行。不仅界面极其清爽美观，而且所有对话计算都在你的本地设备上完成，完全不需要联网或配置复杂的后端服务器，真正做到了开箱即用且绝对保护个人隐私。","stars":838,"starGrowth":0,"createdAt":"2024-05-09T08:09:41Z","silentRounds":20,"language":"TypeScript","topics":["chatbot"],"aiDims":["AI应用","大语言模型","开发者工具"],"zone":"创意","funScore":1,"domainTags":["WebGPU","Local LLM","Browser"],"domainKey":"WebGPU","pushedAt":"2026-08-09T04:48:21Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/addyosmani/chatty","ts":"2026-08-09T04:48:21Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"eriklindernoren/ML-From-Scratch","owner":"eriklindernoren","name":"ML-From-Scratch","desc":"Machine Learning From Scratch. Bare bones NumPy implementations of machine learning models and algorithms with a focus on accessibility. Aims to cover everything from linear regression to deep learning.","summaryCn":"从零手写机器学习算法，NumPy实现经典模型","reasonCn":"纯NumPy实现的机器学习算法库，涵盖线性回归、神经网络、SVM、决策树等经典模型，没有依赖PyTorch或TensorFlow。每个算法都提供完整代码和详细注释，帮助理解模型原理。适合想深入理解机器学习内部机制的学习者，或用于教学演示。代码简洁清晰，可以直接运行并可视化结果，是AI入门的优秀实践项目。","stars":32905,"starGrowth":1,"createdAt":"2017-02-05T12:11:23Z","silentRounds":1,"language":"Python","topics":["machine-learning","deep-learning","deep-reinforcement-learning","machine-learning-from-scratch","data-science","data-mining","genetic-algorithm"],"aiDims":["模型与训练","教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["机器学习","NumPy","神经网络","决策树","算法实现"],"domainKey":"机器学习","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"machine-learning","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"deep-reinforcement-learning","source":"github","weight":0.7},{"name":"machine-learning-from-scratch","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7},{"name":"data-mining","source":"github","weight":0.7},{"name":"genetic-algorithm","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/eriklindernoren/ML-From-Scratch","ts":"2026-09-07T02:44:20.540Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"lightningpixel/modly","owner":"lightningpixel","name":"modly","desc":"Desktop app to generate 3D models from images or prompt using local AI — runs entirely on your GPU","summaryCn":"把静态照片一键变成3D模型的神器，本地跑不卡顿","reasonCn":"这款桌面应用利用本地AI技术，能将普通2D图片瞬间转化为可交互的3D模型，完全运行在用户的GPU上，无需上传数据到云端。它通过先进的图像处理算法提取深度信息和几何结构，生成可用于3D打印、游戏开发或虚拟场景的模型文件。","stars":7661,"starGrowth":2,"createdAt":"2026-03-17T18:02:35Z","silentRounds":0,"language":"TypeScript","topics":["3d","ai-local","ai-tools","open-source","modly","desktop-app","self-hosted"],"aiDims":["创意工具","多模态Agent","AI应用"],"zone":"创意","funScore":0.7,"domainTags":["3D Modeling","Local AI","Desktop App"],"domainKey":"3D Modeling","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"3d","source":"github","weight":0.7},{"name":"ai-local","source":"github","weight":0.7},{"name":"ai-tools","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"modly","source":"github","weight":0.7},{"name":"desktop-app","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/lightningpixel/modly","ts":"2026-08-16T00:54:51.023Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Karovia/fullstack-ai-agent-roadmap","owner":"Karovia","name":"fullstack-ai-agent-roadmap","desc":"🎯 从零基础到 AI Agent 全栈工程师 · 110 个详细教程 · 58 万字 · 400+ GitHub 项目精选 · Obsidian 友好 · 中文","summaryCn":"保姆级教程，带你从零开始打造AI Agent全栈工程师","reasonCn":"这是一份从零基础到 AI Agent 全栈工程师的系统性学习路线图，包含 110 个详细教程和 400+ 精选 GitHub 项目。内容覆盖 AI Agent 的架构设计、工具调用、多模态应用开发，并提供了中文 Markdown 文档，适合 Obsidian 用户直接阅读和索引。","stars":496,"starGrowth":0,"createdAt":"2026-06-19T06:34:06Z","silentRounds":1,"language":"","topics":["tutorial"],"aiDims":["教程","学习资源","AI Agent"],"zone":"资源","funScore":0,"domainTags":["AI Agent","学习路线","教程","全栈开发"],"domainKey":"AI Agent","pushedAt":"2026-09-07T01:20:20Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/Karovia/fullstack-ai-agent-roadmap","ts":"2026-09-07T01:20:20Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ArneVogel/listudy","owner":"ArneVogel","name":"listudy","desc":"Listudy - chess training server","summaryCn":"Elixir 写的在线国际象棋训练平台，下棋像玩游戏。","reasonCn":"Listudy 的核心功能非常全面。在竞技层面，它支持标准的 1v1 对弈，包括即时棋、快棋和慢棋模式，甚至有专门的 Blitz 和 Bullet 赛事。在训练层面，它内置了丰富的排局练习，用户可以通过解决残局来锻炼局面判断力。","stars":395,"starGrowth":0,"createdAt":"2020-06-26T12:59:23Z","silentRounds":15,"language":"Elixir","topics":["chess"],"aiDims":["游戏"],"zone":"创意","funScore":0.6,"domainTags":["国际象棋","在线对弈","残局训练","Elixir"],"domainKey":"国际象棋","pushedAt":"2026-09-09T15:58:53Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"Elixir","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/ArneVogel/listudy","ts":"2026-09-09T15:58:53Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"KudoAI/chatgpt.js","owner":"KudoAI","name":"chatgpt.js","desc":"🤖 A powerful, lightweight JavaScript library + interactive CLI for ChatGPT","summaryCn":"能直接在网页上把ChatGPT玩出花样的JS库","reasonCn":"这是一个专注于浏览器端的轻量级开发库，通过封装原生API实现了对ChatGPT的深度控制和自动化操作。它允许开发者绕过传统Web交互限制，直接在网页里注入代码来操控对话流程、模拟用户点击和输入，从而实现自动化的脚本编写。支持多种功能扩展，比如自动保存对话历史、自定义界面样式、批量生成内容等。","stars":2038,"starGrowth":0,"createdAt":"2023-03-15T07:53:55Z","silentRounds":27,"language":"JavaScript","topics":["frontend","chatbot"],"aiDims":["AI应用","工具调用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Browser Automation","ChatGPT","JavaScript"],"domainKey":"Browser Automation","pushedAt":"2026-08-10T08:03:14Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/KudoAI/chatgpt.js","ts":"2026-08-10T08:03:14Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"daytonaio/daytona","owner":"daytonaio","name":"daytona","desc":"Daytona is a Secure and Elastic Infrastructure for Running AI-Generated Code","summaryCn":"给AI生成的代码打造的安全沙箱，一键隔离运行任何代码片段","reasonCn":"专门为运行AI生成代码设计的安全执行环境，支持隔离沙箱和弹性伸缩。集成主流云服务商，可以自动创建临时开发环境运行代码。对AI编程助手来说，它是执行工具链的关键一环，让Agent真正具备运行代码、验证结果的能力。","stars":71734,"starGrowth":0,"createdAt":"2024-02-06T08:21:20Z","silentRounds":4,"language":"","topics":["developer-tools"],"aiDims":["AI基础设施","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Container","Sandbox","CI/CD"],"domainKey":"Container","pushedAt":"2026-07-24T07:12:07Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"developer-tools","source":"github","weight":0.7}],"aiScore":0.85,"source":"search","url":"https://github.com/daytonaio/daytona","ts":"2026-07-24T07:12:07Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rawgraphs/rawgraphs-app","owner":"rawgraphs","name":"rawgraphs-app","desc":"A web interface to create custom vector-based visualizations on top of RAWGraphs core","summaryCn":"把杂乱数据变成精美图表的网页神器，设计感十足","reasonCn":"这是一个基于开源引擎构建的Web可视化编辑器，支持拖拽操作生成矢量图表。它无需编程基础，用户只需导入CSV或Excel数据，通过可视化面板调整参数即可快速生成专业级的数据图表。项目拥有超过9000星，支持D3.js底层技术，能导出SVG、PNG等格式，非常适合非技术人员进行数据可视化创作。","stars":9032,"starGrowth":0,"createdAt":"2013-04-17T11:54:09Z","silentRounds":27,"language":"JavaScript","topics":["design","data-visualization"],"aiDims":["可视化","非AI-实用","创意工具"],"zone":"创意","funScore":0.6,"domainTags":["数据可视化","矢量图表","D3.js"],"domainKey":"数据可视化","pushedAt":"2026-08-24T11:02:18Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/rawgraphs/rawgraphs-app","ts":"2026-08-24T11:02:18Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"scarsty/kys-cpp","owner":"scarsty","name":"kys-cpp","desc":"《金庸群侠传》c++复刻版，已完工","summaryCn":"《金庸群侠传》C++复刻版，已完工，提供了完整的游戏开发体验。","reasonCn":"该项目是《金庸群侠传》的C++复刻版，提供了一个完整的游戏开发体验，包括游戏的代码实现和完整功能，适合喜欢游戏开发的玩家。","stars":2968,"starGrowth":0,"createdAt":"2016-07-11T04:44:19Z","silentRounds":1,"language":"C++","topics":["game-development"],"aiDims":["游戏开发"],"zone":"创意","funScore":1,"domainTags":["RPG游戏","C++复刻","武侠游戏"],"domainKey":"RPG游戏","pushedAt":"2026-09-01T09:11:42Z","tags":[{"name":"游戏开发","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/scarsty/kys-cpp","ts":"2026-09-01T09:11:42Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chr0nzz/traefik-manager","owner":"chr0nzz","name":"traefik-manager","desc":"A clean, self-hosted web UI for managing your Traefik reverse proxy.","summaryCn":"给 Traefik 搭个漂亮的网页后台，轻松管理反向代理。","reasonCn":"这是一个基于 Web 的 Traefik 反向代理管理面板，界面干净清爽，完全开源自托管。它让你不用再对着黑乎乎的命令行或配置文件，直接在浏览器里就能完成路由规则、中间件和证书的配置。支持 Traefik 2 和 3 版本，能实时监控面板流量和健康状态，让运维工作变得非常直观且高效。","stars":1602,"starGrowth":6,"createdAt":"2026-03-09T02:04:47Z","silentRounds":0,"language":"JavaScript","topics":["alpine-linux","bcrypt","csrf-protection","docker","flask","flask-session","gunicorn","phosphor-icons","python3","ruamel-yaml","tailwind-css","traefik","vanilla-js","traefik-management","traefik-web-ui","crsf","fernet","flask-limiter","totp"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["反向代理","Web管理","Traefik","DevOps"],"domainKey":"反向代理","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"alpine-linux","source":"github","weight":0.7},{"name":"bcrypt","source":"github","weight":0.7},{"name":"csrf-protection","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"flask","source":"github","weight":0.7},{"name":"flask-session","source":"github","weight":0.7},{"name":"gunicorn","source":"github","weight":0.7},{"name":"phosphor-icons","source":"github","weight":0.7},{"name":"python3","source":"github","weight":0.7},{"name":"ruamel-yaml","source":"github","weight":0.7},{"name":"tailwind-css","source":"github","weight":0.7},{"name":"traefik","source":"github","weight":0.7},{"name":"vanilla-js","source":"github","weight":0.7},{"name":"traefik-management","source":"github","weight":0.7},{"name":"traefik-web-ui","source":"github","weight":0.7},{"name":"crsf","source":"github","weight":0.7},{"name":"fernet","source":"github","weight":0.7},{"name":"flask-limiter","source":"github","weight":0.7},{"name":"totp","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/chr0nzz/traefik-manager","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03508314606741573,"copyOk":false},{"repo":"astaxie/TokenHub","owner":"astaxie","name":"TokenHub","desc":"TokenHub gives enterprises a private gateway to unify AI model access and governance, making every request controllable, traceable, and attributable.","summaryCn":"给企业搞了个AI访问的私有控制台，防止数据泄露。","reasonCn":"TokenHub 是一个基于 Go 语言开发的企业级 AI 中间件，旨在解决企业在接入大模型 API 时的隐私和管控难题。它充当私有网关，统一管理不同 AI 模型的访问入口，实现请求的可控、可追溯和可归因。通过 TokenHub，企业无需直接暴露内部 API 密钥，就能对 AI 流量进行细粒度的策略配置、审计日志记录和合规性检查，特别适合对数据安全要求高且使用多模型 API 的组织。","stars":1328,"starGrowth":2,"createdAt":"2026-06-10T15:32:07Z","silentRounds":0,"language":"Go","topics":["ai","ai-gateway","enterprise","gateway"],"aiDims":["AI基础设施","模型部署","安全工具"],"zone":"AI","funScore":0,"domainTags":["AI网关","数据安全","企业级"],"domainKey":"AI网关","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"enterprise","source":"github","weight":0.7},{"name":"gateway","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/astaxie/TokenHub","ts":"2026-08-15T00:57:24.351Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"0xk1h0/ChatGPT_DAN","owner":"0xk1h0","name":"ChatGPT_DAN","desc":"","summaryCn":"提供可直接复制的 ChatGPT 越狱提示，让模型突破官方限制。","reasonCn":"本仓库收录了多个 ChatGPT 越狱 Prompt，如 DAN 12.0 与 13.0，能够让对话模型在不受 OpenAI 约束的情况下随意回答，适合想要探索模型潜力或进行趣味实验的用户。使用方法极简，只需将对应 Prompt 粘贴到 ChatGPT 输入框即可，无需额外安装或配置。","stars":12475,"starGrowth":0,"createdAt":"2023-02-15T09:48:18Z","silentRounds":1,"language":"","topics":["chatgpt","jailbreak","openai","prompt","gpt-4","gpt-3-5"],"aiDims":["提示工程"],"zone":"AI","funScore":1,"domainTags":["提示工程","越狱","ChatGPT","角色扮演","prompt注入"],"domainKey":"提示工程","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"chatgpt","source":"github","weight":0.7},{"name":"jailbreak","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"prompt","source":"github","weight":0.7},{"name":"gpt-4","source":"github","weight":0.7},{"name":"gpt-3-5","source":"github","weight":0.7}],"aiScore":0.614,"source":"search","url":"https://github.com/0xk1h0/ChatGPT_DAN","ts":"2026-09-13T22:08:09.499Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"plastic-labs/honcho","owner":"plastic-labs","name":"honcho","desc":" Memory library for building stateful agents","summaryCn":"让AI Agent拥有长期记忆的分层存储库","reasonCn":"这是专门为有状态AI Agent设计的内存库，提供分层记忆架构和向量数据库支持。Agent可以通过它记住历史对话、学习用户偏好、保持上下文连贯性。解决了当前很多Agent只会短程对话、无法记住长期信息的核心痛点，是构建真正智能助手的关键基础设施。","stars":7303,"starGrowth":3,"createdAt":"2023-09-10T21:29:54Z","silentRounds":0,"language":"Python","topics":["vector-db"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":0.5,"domainTags":["AI记忆","向量数据库","Agent框架"],"domainKey":"向量数据库","pushedAt":"2026-08-18T18:47:10Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/plastic-labs/honcho","ts":"2026-08-18T18:47:10Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"beclab/Olares","owner":"beclab","name":"Olares","desc":"Open-Source Personal Cloud OS for Always-On Agents","summaryCn":"给你的AI助手安个永不关机的家，让它24小时在线自动干活","reasonCn":"Olares是一个面向AI代理的个人云操作系统，让智能体能够7×24小时持续运行。基于Go开发，轻量易部署，支持自定义Agent行为、任务调度和持久化存储。相比需要人工交互的聊天机器人，它更像是一个自动化的AI员工，帮你完成那些需要长时间盯守或定期执行的任务，真正让AI成为后台运转的助手。","stars":5283,"starGrowth":2,"createdAt":"2024-04-29T05:43:58Z","silentRounds":0,"language":"Go","topics":["kubernetes","self-hosted","homelab","edge-ai","local-ai","ai-agents","model-serving","personal-cloud","agentic-ai","cloud-native","data-privacy","operating-system"],"aiDims":["AI Agent","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","个人云","任务调度","持久化运行"],"domainKey":"AI Agent","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"kubernetes","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"homelab","source":"github","weight":0.7},{"name":"edge-ai","source":"github","weight":0.7},{"name":"local-ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"model-serving","source":"github","weight":0.7},{"name":"personal-cloud","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"data-privacy","source":"github","weight":0.7},{"name":"operating-system","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/beclab/Olares","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zarazhangrui/frontend-slides","owner":"zarazhangrui","name":"frontend-slides","desc":"Create beautiful slides on the web using a coding agent's frontend skills","summaryCn":"像写代码一样做PPT，让做演示文稿变得超简单","reasonCn":"这是一个基于编程思维构建的网页演示文稿生成工具，允许用户通过编写代码而非拖拽来创建精美的幻灯片。项目利用前端技术栈实现了所见即所得的渲染效果，支持自定义布局和样式，降低了演示设计的门槛，特别适合喜欢代码创作或需要快速生成高质量幻灯片的开发者。","stars":29722,"starGrowth":16,"createdAt":"2026-01-28T00:42:34Z","silentRounds":0,"language":"JavaScript","topics":["ai-slides","anthropic","claude","claude-code","claude-skill","generative-ui","html","presentation","slides","vibe-coding"],"aiDims":["代码助手","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["演示文稿生成","前端渲染","所见即所得"],"domainKey":"演示文稿生成","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"ai-slides","source":"github","weight":0.7},{"name":"anthropic","source":"github","weight":0.7},{"name":"claude","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"claude-skill","source":"github","weight":0.7},{"name":"generative-ui","source":"github","weight":0.7},{"name":"html","source":"github","weight":0.7},{"name":"presentation","source":"github","weight":0.7},{"name":"slides","source":"github","weight":0.7},{"name":"vibe-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/zarazhangrui/frontend-slides","ts":"2026-08-15T00:57:24.351Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.05953795841464235,"copyOk":false},{"repo":"iluaii/fwm","owner":"iluaii","name":"fwm","desc":"Wayland compositor in C/wlroots where windows are physical objects (real Box2D rigid bodies) — and a full desktop besides: ten desktops on one scrolling world, tiling or floating per desktop, launcher, expo, screenshots, mixer and visualiser built in, live IPC. Legacy X11 version on the x11 branch.","summaryCn":"大家习惯了传统的Windows或macOS桌面，窗口就是简单的矩形框，但如果你想体","reasonCn":"技术上，它利用wlroots compositor库作为基底，将Box2D物理引擎与Wayland协议完美融合。开发者编写了一套严密的物理规则映射逻辑，使得所有的窗口位置更新、状态渲染都由物理计算驱动。这种设计挑战了操作系统界面的底层逻辑，让原本死板的窗口管理变成了一种动态的、富有节奏感的","stars":171,"starGrowth":0,"createdAt":"2026-06-20T16:19:35Z","silentRounds":18,"language":"C","topics":["physics"],"aiDims":["非AI-好玩","可视化"],"zone":"创意","funScore":1,"domainTags":["Wayland","物理引擎","桌面环境","窗口管理","Box2D"],"domainKey":"Wayland","pushedAt":"2026-09-09T09:14:03Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"physics","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/iluaii/fwm","ts":"2026-09-09T09:14:03Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Narcooo/inkos","owner":"Narcooo","name":"inkos","desc":"Story Creation AI Agent for novel, scripts, translation, interactive games, and IP content","summaryCn":"AI帮你写小说剧本做游戏，自动创作全能内容管家","reasonCn":"这是一个多模态故事创作Agent，专门帮用户写小说、剧本，还能做翻译、交互式游戏和IP衍生内容。底层基于TypeScript，接入主流大模型API，具备自动规划和多轮对话能力。不仅能生成连贯剧情，还能维持角色性格，把脑洞变成完整作品，是创作者和跑团玩家的效率神器。","stars":10018,"starGrowth":12,"createdAt":"2026-03-12T02:12:54Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent"],"aiDims":["AI Agent","AI写作","游戏"],"zone":"AI","funScore":0.7,"domainTags":["故事创作","AI写作","交互式叙事","角色扮演"],"domainKey":"AI写作","pushedAt":"2026-08-25T17:24:02Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI写作","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/Narcooo/inkos","ts":"2026-08-25T17:24:02Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.04751223797165103,"copyOk":false},{"repo":"superlinked/sie","owner":"superlinked","name":"sie","desc":"Open-source inference server and production cluster for all the models your agent needs.","summaryCn":"给AI Agent搭的推理服务器，让各种模型都能一键上生产","reasonCn":"sie是开源的推理服务器和生产集群，专为AI Agent打造。它把不同模型的推理服务统一管理，提供低延迟高并发能力。无论是本地跑还是部署到云端集群，都能让Agent轻松调用各类模型，省去自己折腾推理基础设施的麻烦，直接拿来用就行。","stars":3305,"starGrowth":7,"createdAt":"2023-11-07T10:45:53Z","silentRounds":0,"language":"Python","topics":["embeddings","vector-search","data-pipeline","deep-learning","information-retrieval","llm","ml","mlops","natural-language-processing","nlp","python","retrieval","retrieval-augmented-generation","semantic-search","inference","bge","colbert","inference-server","reranking","splade"],"aiDims":["推理引擎","AI基础设施","AI Agent"],"zone":"AI","funScore":0.5,"domainTags":["推理服务器","模型部署","Agent","生产基础设施"],"domainKey":"模型部署","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"embeddings","source":"github","weight":0.7},{"name":"vector-search","source":"github","weight":0.7},{"name":"data-pipeline","source":"github","weight":0.7},{"name":"deep-learning","source":"github","weight":0.7},{"name":"information-retrieval","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"mlops","source":"github","weight":0.7},{"name":"natural-language-processing","source":"github","weight":0.7},{"name":"nlp","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"retrieval","source":"github","weight":0.7},{"name":"retrieval-augmented-generation","source":"github","weight":0.7},{"name":"semantic-search","source":"github","weight":0.7},{"name":"inference","source":"github","weight":0.7},{"name":"bge","source":"github","weight":0.7},{"name":"colbert","source":"github","weight":0.7},{"name":"inference-server","source":"github","weight":0.7},{"name":"reranking","source":"github","weight":0.7},{"name":"splade","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/superlinked/sie","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.027354009077155825,"copyOk":false},{"repo":"Agent-Field/pr-af","owner":"Agent-Field","name":"pr-af","desc":"#1 open-source code reviewer on Code-Review-Bench","summaryCn":"开源界的最强代码审查员，专门帮你检查Bug。","reasonCn":"它的厉害之处在于对代码上下文的理解非常深。它不只是简单地检查语法，还能分析代码的逻辑漏洞、性能瓶颈以及潜在的逻辑错误。它能生成非常详细且专业的审查报告，指出问题所在并给出修改建议。对于开源项目维护者来说，它能自动处理海量的 PR，让代码质量更有保障。","stars":634,"starGrowth":1,"createdAt":"2026-03-11T15:35:35Z","silentRounds":3,"language":"Go","topics":["agentfield","agentic-engineering","agents","ai","code-review","llm","pull-request"],"aiDims":["代码助手","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["代码审查","AI编程助手","Pull Request","自动化测试"],"domainKey":"代码审查","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"agentfield","source":"github","weight":0.7},{"name":"agentic-engineering","source":"github","weight":0.7},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"code-review","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"pull-request","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/Agent-Field/pr-af","ts":"2026-08-14T01:23:41.092Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"bytedance/SALMONN","owner":"bytedance","name":"SALMONN","desc":"SALMONN family: A suite of advanced multi-modal LLMs","summaryCn":"字节跳动开源的语音和音乐全能多模态大模型，听音辨意样样精通。","reasonCn":"技术底层上，它创新性地融合了语音编码器和大语言模型架构，通过精妙的跨模态对齐技术，让文本大模型能够直接“听懂”连续的音频特征。无论是语音识别、音频事件检测还是音乐风格分析，它都能在一套统一的框架内搞定。","stars":1536,"starGrowth":0,"createdAt":"2023-08-11T07:28:25Z","silentRounds":3,"language":"","topics":["music"],"aiDims":["多模态Agent","大语言模型","模型与训练"],"zone":"AI","funScore":0,"domainTags":["音频理解","多模态大模型","音乐分析"],"domainKey":"音频理解","pushedAt":"2026-08-24T07:11:01Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7}],"aiScore":0.85,"source":"search","url":"https://github.com/bytedance/SALMONN","ts":"2026-08-24T07:11:01Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"syncthing/syncthing","owner":"syncthing","name":"syncthing","desc":"Open Source Continuous File Synchronization","summaryCn":"完全去中心化免费同步文件，替代云盘保护隐私的利器","reasonCn":"这是一个开源的点对点文件同步工具，不需要上传到第三方服务器，数据完全掌握在自己手中。它支持多设备间实时同步，隐私安全性极高，且完全免费。对于厌恶云盘监控、希望自建私有同步方案的用戶来说，它是替代Dropbox或iCloud的最佳选择，实用价值极高。","stars":88851,"starGrowth":9,"createdAt":"2013-11-26T09:48:21Z","silentRounds":0,"language":"Go","topics":["synchronization","go","peer-to-peer","p2p"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["文件同步","P2P","去中心化","备份"],"domainKey":"文件同步","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"synchronization","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"peer-to-peer","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/syncthing/syncthing","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.027303879528649086,"copyOk":false},{"repo":"kunchenguid/treehouse","owner":"kunchenguid","name":"treehouse","desc":"Manage worktrees without managing worktrees.","summaryCn":"Git分支管理神器，告别繁琐的手动切换操作，一键搞定多任务并行开发。","reasonCn":"这是一个用Go语言编写的命令行工具，旨在简化Git工作树（worktree）的创建、删除和切换流程。它通过自动处理依赖关系和清理垃圾文件，解决了手动管理多个工作目录时的复杂性和易错性，让开发者可以轻松并行处理多个独立的开发分支。","stars":1761,"starGrowth":5,"createdAt":"2026-03-14T19:20:20Z","silentRounds":0,"language":"Go","topics":["agents","coding","git","worktree"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Git","工作树","版本控制","命令行工具"],"domainKey":"Git","pushedAt":"2026-08-13T01:37:33.040Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"coding","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"worktree","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/kunchenguid/treehouse","ts":"2026-08-13T01:37:33.040Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.028221465076660984,"copyOk":false},{"repo":"kdlbs/kandev","owner":"kdlbs","name":"kandev","desc":"AI Kanban &amp; Development Environment. Orchestrate multiple agents, review changes, open PRs. Multi-provider, self-hostable, no telemetry.","summaryCn":"让AI当你的编程搭子，自动写代码还能帮你开PR","reasonCn":"这是一个AI驱动的项目管理+开发环境工具，核心特色是用多个AI Agent协作完成软件开发流程。它支持编排多个智能体协同工作，自动Review代码改动并打开Pull Request。支持多种AI提供商，可以自建部署且没有遥测数据收集。对想尝鲜AI辅助开发的开发者来说是个有趣的实用项目。","stars":827,"starGrowth":4,"createdAt":"2026-01-09T23:26:53Z","silentRounds":0,"language":"Go","topics":["acp","agent-orchestration","agentic-ai","ai-agents","claude-code","code-review","coding-agents","developer-tools","github-copilot","golang","kanban","nextjs","open-source","self-hosted","task-manager","tui","vibe-coding","workflow-automation","worktrees","agentic-development-environment"],"aiDims":["AI Agent","开发者工具","效率工具"],"zone":"AI","funScore":0.6,"domainTags":["AI编程助手","多Agent","Kanban","Agent编排"],"domainKey":"AI编程助手","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"acp","source":"github","weight":0.7},{"name":"agent-orchestration","source":"github","weight":0.7},{"name":"agentic-ai","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"code-review","source":"github","weight":0.7},{"name":"coding-agents","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"github-copilot","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"kanban","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"task-manager","source":"github","weight":0.7},{"name":"tui","source":"github","weight":0.7},{"name":"vibe-coding","source":"github","weight":0.7},{"name":"workflow-automation","source":"github","weight":0.7},{"name":"worktrees","source":"github","weight":0.7},{"name":"agentic-development-environment","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/kdlbs/kandev","ts":"2026-08-11T09:45:50.026Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"quan0715/open-press","owner":"quan0715","name":"open-press","desc":"AI-first fixed-layout document workspaces. Your AI agent writes proposals, whitepapers, theses, books — A4, PDF, web reader.","summaryCn":"AI优先的固定布局文档工作区，帮AI写提案论文书籍并生成PDF","reasonCn":"核心功能在于“AI 优先”的设计理念。你只需要告诉它文档的主题和结构，它就能利用内置的 AI 模型生成高质量的初稿。更重要的是，open-press 支持多 Agent 协作，不同的 Agent 可以负责不同的部分，比如一个 Agent 负责撰写摘要，另一个负责整理参考文献，最后再由主 Agent 进行润色","stars":14,"starGrowth":0,"createdAt":"2026-05-18T06:04:22Z","silentRounds":17,"language":"TypeScript","topics":["writing"],"aiDims":["AI写作","AI应用","文档理解"],"zone":"AI","funScore":0,"domainTags":["文档生成","多Agent协作","PDF排版","写作辅助","固定布局"],"domainKey":"文档生成","pushedAt":"2026-08-29T11:02:14Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/quan0715/open-press","ts":"2026-08-29T11:02:14Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bunkerity/bunkerweb","owner":"bunkerity","name":"bunkerweb","desc":"🛡️ Open-source and cloud-native Web Application Firewall (WAF)","summaryCn":"开源WAF神器，给网站穿上防弹衣，自动拦截黑客攻击。","reasonCn":"BunkerWeb是一款基于Nginx的开源Web应用防火墙，集成了ModSecurity、Fail2ban等主流安全模块。它最大的亮点是自带AI驱动的自动配置功能，无需复杂的规则编写，安装后即可自动识别和防御SQL注入、XSS等常见Web攻击，极大降低了企业级安全防护的门槛，适合各类Web服务运维使用。","stars":10996,"starGrowth":0,"createdAt":"2019-08-20T21:20:48Z","silentRounds":1,"language":"Python","topics":["devops","kubernetes"],"aiDims":["安全工具","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Web应用防火墙","安全防护","Nginx"],"domainKey":"Web应用防火墙","pushedAt":"2026-09-23T12:14:10Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/bunkerity/bunkerweb","ts":"2026-09-23T12:14:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Sumanth077/Hands-On-AI-Engineering","owner":"Sumanth077","name":"Hands-On-AI-Engineering","desc":"A curated collection of practical AI projects implementing OCR systems, RAG, AI agents, and other AI use cases.","summaryCn":"AI工程实战合集：OCR、RAG、智能体全都有，代码拿来就能跑","reasonCn":"内容覆盖面很广，包括文档解析与OCR、向量数据库与RAG检索、AI Agent构建、多模态处理、提示工程实践等方向。每个项目都是独立可运行的，代码结构清晰，注释详细，跟着做一遍就能掌握从数据处理到模型调用的完整流程。","stars":3689,"starGrowth":1,"createdAt":"2025-02-24T19:15:35Z","silentRounds":1,"language":"Python","topics":["agents","ai","llms","ocr","python","rag","ai-agents","ai-engineering","generative-ai","mcp"],"aiDims":["AI应用","学习资源","教程"],"zone":"资源","funScore":0.6,"domainTags":["AI工程","RAG","AI Agent"],"domainKey":"AI工程","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"agents","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"ocr","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-engineering","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/Sumanth077/Hands-On-AI-Engineering","ts":"2026-09-09T22:18:37.171Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"xhongc/music-tag-web","owner":"xhongc","name":"music-tag-web","desc":"音乐标签编辑器，可编辑本地音乐文件的元数据, 音乐刮削。（Editable local music file metadata.）","summaryCn":"拯救乱码歌单的神器，让你的本地音乐库整齐划一","reasonCn":"它采用Python作为后端，通过Web界面与本地文件系统交互。最核心的竞争力在于它的‘刮削’功能，即自动从公开的音乐数据库中检索并匹配该歌曲的正确元数据。它不仅能修改ID3标签，还能自动下载高质量的专辑封面，并支持批量处理，避免了手动一个一个修改的低效。","stars":6066,"starGrowth":2,"createdAt":"2023-03-11T07:24:17Z","silentRounds":0,"language":"Python","topics":["music"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["音乐管理","元数据编辑","自动刮削","ID3标签","Web应用"],"domainKey":"音乐管理","pushedAt":"2026-09-04T01:40:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/xhongc/music-tag-web","ts":"2026-09-04T01:40:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"smixs/autograph","owner":"smixs","name":"autograph","desc":"Schema-as-code memory for AI agents in Obsidian: typed cards, entity dedup, link repair, update-in-place, and Ebbinghaus decay. Plain Markdown you own — a Claude Code skill.","summaryCn":"给AI装个记性超好的记忆库，知识库越用越准，再也不怕它忘词","reasonCn":"专为Claude Code等AI智能体打造的Markdown记忆方案，采用代码化Schema管理类型化卡片。内置实体去重、链接自动修复及原地更新机制，创新引入艾宾浩斯遗忘曲线算法实现知识衰减，让AI记忆像人脑一样自然老化与强化，极大提升长期对话的上下文一致性与准确度，是构建个人化第二大脑的利器。","stars":71,"starGrowth":0,"createdAt":"2026-04-01T23:08:24Z","silentRounds":18,"language":"Python","topics":["knowledge-management"],"aiDims":["AI Agent","RAG与知识"],"zone":"AI","funScore":0,"domainTags":["知识管理","AI Agent","Obsidian"],"domainKey":"AI Agent","pushedAt":"2026-09-06T03:02:06Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/smixs/autograph","ts":"2026-09-06T03:02:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zarazhangrui/follow-builders","owner":"zarazhangrui","name":"follow-builders","desc":"AI builders digest — monitors top AI builders on X and YouTube podcasts, remixes their content into digestible summaries. Follow builders, not influencers.","summaryCn":"帮你追踪AI大佬，把长视频浓缩成干货摘要","reasonCn":"这是一个AI构建者内容聚合器，通过监控X和YouTube上的头部AI开发者，自动抓取并重混他们的播客与推文，生成精简摘要。它支持按话题订阅，帮你过滤掉营销号，只关注真正的实干家。项目目前拥有超过6200星，社区反馈热烈。","stars":6772,"starGrowth":1,"createdAt":"2026-03-14T16:28:39Z","silentRounds":2,"language":"JavaScript","topics":[],"aiDims":["AI应用","AI搜索","非AI-实用"],"zone":"AI","funScore":0.5,"domainTags":["内容聚合","播客摘要","AI应用","信息策展"],"domainKey":"内容聚合","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/zarazhangrui/follow-builders","ts":"2026-08-11T09:45:50.026Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"spacepacket1/e3d-pod2vid","owner":"spacepacket1","name":"e3d-pod2vid","desc":"AI-powered podcast-to-video pipeline. Semantic B-roll, voice synthesis, burned subtitles, YouTube publishing.","summaryCn":"播客一键变视频神器，AI自动配画面剪字幕，懒人必备","reasonCn":"全自动化的播客转视频工作流，利用AI进行语义分析以匹配B-roll素材，集成语音合成与硬字幕烧录功能。支持直接推流至YouTube，大幅降低视频制作门槛。通过多模态技术将音频内容转化为视觉叙事，帮助创作者快速实现跨平台分发，是自媒体运营提效的实用搞钱工具，极致简化后期剪辑流程，实现从音频到成片的无缝流转。","stars":12,"starGrowth":0,"createdAt":"2026-06-27T13:43:44Z","silentRounds":16,"language":"Python","topics":["podcast"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0.6,"domainTags":["播客转视频","语义分析","B-roll匹配","多模态","YouTube发布"],"domainKey":"多模态","pushedAt":"2026-09-04T01:42:52Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/spacepacket1/e3d-pod2vid","ts":"2026-09-04T01:42:52Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"katanemo/plano","owner":"katanemo","name":"plano","desc":"Plano is an AI-native proxy server and data plane for agentic apps. Smart LLM routing, observability, agent orchestration, and guardrails so you stay focused on your agents core logic.","summaryCn":"AI Agent的智能路由中台，让多模型协作像搭积木一样简单","reasonCn":"技术上，它是用Rust写的，性能没得说。核心功能是智能路由，可以根据你的需求自动选择最合适的LLM。支持多种模型的后端，你可以配置优先级和fallback策略。它还提供了完整的可观测性，每个请求的延迟、成本、结果都能追踪。","stars":7063,"starGrowth":0,"createdAt":"2024-07-09T20:25:56Z","silentRounds":2,"language":"Rust","topics":["gateway","generative-ai","llm-inference","llms","prompt","proxy","proxy-server","llmops","openai","routing","ai-gateway","llm-gateway","llm-routing","envoy","envoyproxy","ai-gateway-support","llm-proxy"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI代理","LLM路由","Agent编排","可观测性","护栏机制"],"domainKey":"AI代理","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"gateway","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"llm-inference","source":"github","weight":0.7},{"name":"llms","source":"github","weight":0.7},{"name":"prompt","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"proxy-server","source":"github","weight":0.7},{"name":"llmops","source":"github","weight":0.7},{"name":"openai","source":"github","weight":0.7},{"name":"routing","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"llm-gateway","source":"github","weight":0.7},{"name":"llm-routing","source":"github","weight":0.7},{"name":"envoy","source":"github","weight":0.7},{"name":"envoyproxy","source":"github","weight":0.7},{"name":"ai-gateway-support","source":"github","weight":0.7},{"name":"llm-proxy","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/katanemo/plano","ts":"2026-08-07T04:03:56.389Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"prabhuomkar/pytorch-cpp","owner":"prabhuomkar","name":"pytorch-cpp","desc":"C++ Implementation of PyTorch Tutorials for Everyone","summaryCn":"对于习惯了 Python 生态但必须转向 C++ 环境以提升模型推理速度的开发者，该","reasonCn":"核心亮点在于它不仅展示了代码，还建立了一套从基础到进阶的完善教学体系。从最简单的线性回归模型，到复杂的变分自编码器（VAE）与图像标注模型（Image Captioning），每个模块都严格对应了 PyTorch 的核心 API。这种“所见即所得”的实现方式，能够帮助开发者迅速理解如何将 Pytho","stars":2138,"starGrowth":0,"createdAt":"2019-11-05T16:52:01Z","silentRounds":11,"language":"C++","topics":["tutorial"],"aiDims":["教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["深度学习","C++","PyTorch","模型部署"],"domainKey":"深度学习","pushedAt":"2026-09-13T07:53:31Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/prabhuomkar/pytorch-cpp","ts":"2026-09-13T07:53:31Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ELM-Research/ECG-Language-Models","owner":"ELM-Research","name":"ECG-Language-Models","desc":" A research-oriented training and evaluation framework for ECG-Language Models (ELMs) ","summaryCn":"把心电图变成文字报告，AI听懂心脏的语言","reasonCn":"这是一个研究框架，核心想法是把心电图（ECG）看作一种“语言”，训练大模型去理解和生成心脏信号。它让AI不仅会看影像，还能读懂时间序列数据，对于医疗AI研究来说，这是一个非常有意思的新视角，把信号处理和LLM结合在了一起。","stars":19,"starGrowth":0,"createdAt":"2026-02-23T04:30:16Z","silentRounds":28,"language":"Python","topics":["llm-model"],"aiDims":["模型与训练","多模态Agent"],"zone":"资源","funScore":0.6,"domainTags":["心电图分析","医疗AI","Transformer","时序数据","Token化"],"domainKey":"心电图分析","pushedAt":"2026-09-08T03:44:13Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/ELM-Research/ECG-Language-Models","ts":"2026-09-08T03:44:13Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"supervc-stack/VectorChord","owner":"supervc-stack","name":"VectorChord","desc":"Scalable, fast, and disk-friendly vector search in Postgres, the successor of pgvecto.rs.","summaryCn":"Postgres里的高性能向量搜索引擎，RAG应用加速器","reasonCn":"VectorChord是pgvecto.rs的后继者，一个专为Postgres设计的可扩展、快速且磁盘友好的向量搜索插件。它让你能在熟悉的Postgres环境中直接进行向量检索，无需引入复杂的单独向量数据库，极大简化了RAG应用的开发架构，适合追求稳定易维护的AI开发者。","stars":1800,"starGrowth":1,"createdAt":"2024-09-03T07:39:46Z","silentRounds":8,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["向量检索","PostgreSQL","RAG","数据库扩展"],"domainKey":"向量检索","pushedAt":"2026-08-06T06:34:15Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/supervc-stack/VectorChord","ts":"2026-08-06T06:34:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AaronWong1999/hermesclaw","owner":"AaronWong1999","name":"hermesclaw","desc":"Run Hermes Agent and OpenClaw on the same WeChat account","summaryCn":"把两个智能体塞进同一个微信号里陪你聊天干活","reasonCn":"简单说，它把微信消息拆成多路任务、再分别交给不同助手执行，是一个偏应用层的聊天桥。项目处理账号复用、会话隔离、回调转发、回复聚合和日志追踪，让两个代理看起来像共用一个客服窗口。对研究微信端智能体编排、外部工具调用与自动回复链路的人很有参考价值。","stars":742,"starGrowth":0,"createdAt":"2026-04-10T05:49:39Z","silentRounds":2,"language":"Python","topics":["chatbot"],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"AI","funScore":0.6,"domainTags":["微信机器人","多智能体","代理编排"],"domainKey":"微信机器人","pushedAt":"2026-08-24T16:52:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/AaronWong1999/hermesclaw","ts":"2026-08-24T16:52:17Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ArturoNereu/AI-Study-Group","owner":"ArturoNereu","name":"AI-Study-Group","desc":"Resources to learn AI","summaryCn":"从零开始系统学习人工智能技术的全套优质开源宝库","reasonCn":"在内容组织上，它抛弃了那种杂乱无章的链接堆砌，而是采用模块化的知识树结构。从最基础的Python编程、概率统计、机器学习算法，到深度学习框架PyTorch，再到当下最火的Transformer架构、大模型微调、提示工程以及AI Agent，每一块都有详尽的指南和精选的学习材料推荐，省去了你在","stars":1309,"starGrowth":1,"createdAt":"2023-01-31T09:18:48Z","silentRounds":2,"language":"","topics":["documentation","learning"],"aiDims":["学习资源","教程","大语言模型"],"zone":"资源","funScore":0,"domainTags":["AI学习","机器学习","深度学习","教程资源","知识图谱"],"domainKey":"AI学习","pushedAt":"2026-08-22T19:56:10Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"documentation","source":"github","weight":0.7},{"name":"learning","source":"github","weight":0.7}],"aiScore":0.82,"source":"search","url":"https://github.com/ArturoNereu/AI-Study-Group","ts":"2026-08-22T19:56:10Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Macrohard0001/lx-ikun-music-sources","owner":"Macrohard0001","name":"lx-ikun-music-sources","desc":"LX_music & IKUN_music 音源收集","summaryCn":"大家平时听歌是不是经常遇到这个问题：想听这首要开 A 软件的会员，想听那首","reasonCn":"从技术上讲，该项目通过 JavaScript 编写了各类平台的解析脚本，能够实时监测并匹配最优的音质接口。它的逻辑很清晰，就是通过一个接口聚合，把各个音乐平台的 API 调用方式统一标准，让音乐软件能够无缝识别并播放。","stars":2484,"starGrowth":1,"createdAt":"2025-06-24T08:29:12Z","silentRounds":1,"language":"JavaScript","topics":["music"],"aiDims":["非AI-好玩","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["音乐","API","音源","解密"],"domainKey":"音乐","pushedAt":"2026-09-08T11:23:47Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"music","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/Macrohard0001/lx-ikun-music-sources","ts":"2026-09-08T11:23:47Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aws/karpenter-provider-aws","owner":"aws","name":"karpenter-provider-aws","desc":"Karpenter is a Kubernetes Node Autoscaler built for flexibility, performance, and simplicity.","summaryCn":"K8s 自动扩容神器，云资源按需自动伸缩，省钱又省心","reasonCn":"Karpenter 是一个专为 Kubernetes 设计的高性能节点自动伸缩器，它通过 AWS Fargate 和 EC2 实例实现秒级扩容。相比传统集群自动伸缩器，它支持多种实例类型、Spot 实例和预留实例，能显著降低云成本并提升资源利用率。","stars":7721,"starGrowth":0,"createdAt":"2020-07-09T22:04:22Z","silentRounds":11,"language":"Go","topics":[],"aiDims":["AI基础设施","非AI-实用","开发工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","自动化","云原生"],"domainKey":"Kubernetes","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.85,"source":"trending","url":"https://github.com/aws/karpenter-provider-aws","ts":"2026-08-14T01:23:41.092Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"LLPhant/LLPhant","owner":"LLPhant","name":"LLPhant","desc":"LLPhant - A comprehensive PHP Generative AI Framework using OpenAI GPT 4. Inspired by Langchain","summaryCn":"PHP版Langchain，给PHP应用装上AI大脑的瑞士军刀","reasonCn":"它的设计思路借鉴了Python生态的Langchain，核心能力包括：调用OpenAI GPT-4等主流大模型进行文本生成；向量数据库集成，用于存放大模型需要检索的知识；RAG检索增强生成，让AI回答基于你的文档而非瞎编；以及Agent智能体架构，可以让AI调用外部工具完成任务。","stars":1709,"starGrowth":1,"createdAt":"2023-07-31T09:54:43Z","silentRounds":26,"language":"PHP","topics":["vector-db"],"aiDims":["AI应用","AI基础设施","工具调用"],"zone":"AI","funScore":0,"domainTags":["PHP","生成式AI","RAG","Agent","Langchain"],"domainKey":"RAG","pushedAt":"2026-09-07T15:07:08Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/LLPhant/LLPhant","ts":"2026-09-07T15:07:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"aws/agent-toolkit-for-aws","owner":"aws","name":"agent-toolkit-for-aws","desc":"Official, AWS-supported MCP servers, skills, and plugins to help AI agents build on AWS","summaryCn":"AWS官方推出的让AI Agent直接操控云服务的工具包","reasonCn":"这是AWS官方推出的AI Agent工具套件，提供MCP服务器和插件，让AI模型能直接调用S3、Lambda、DynamoDB等20余种AWS服务。项目2025年发布，Stars 2200+，填补了大模型与云基础设施集成的空白，是AI Agent开发者的云端瑞士军刀。","stars":2698,"starGrowth":1,"createdAt":"2026-04-23T12:44:08Z","silentRounds":2,"language":"Python","topics":[],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AWS","AI Agent","云服务","MCP"],"domainKey":"AI Agent","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/aws/agent-toolkit-for-aws","ts":"2026-08-07T04:03:56.389Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mitmproxy/mitmproxy","owner":"mitmproxy","name":"mitmproxy","desc":"An interactive TLS-capable intercepting HTTP proxy for penetration testers and software developers.","summaryCn":"抓包调试的神器，能把所有HTTP流量变成可读的脚本，写接口自动化必备。","reasonCn":"这是一个交互式的 HTTPS 代理工具，不仅能拦截和修改网络请求，还能将抓到的流量导出为 Python 脚本或回放。对于开发者来说，它是调试 API 和逆向接口的利器；对于安全人员，它是渗透测试中分析加密流量的标准装备，支持 Linux/macOS/终端等多环境运行。","stars":45129,"starGrowth":5,"createdAt":"2010-02-16T04:10:13Z","silentRounds":0,"language":"Python","topics":["security"],"aiDims":["开发者工具","安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网络抓包","HTTP代理","渗透测试","API调试","TLS解密"],"domainKey":"网络抓包","pushedAt":"2026-09-10T06:52:11Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/mitmproxy/mitmproxy","ts":"2026-09-10T06:52:11Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01533238050920694,"copyOk":false},{"repo":"grab/cursor-talk-to-figma-mcp","owner":"grab","name":"cursor-talk-to-figma-mcp","desc":"TalkToFigma: MCP integration between AI Agent (Cursor, Claude Code, Codex) and Figma, allowing Agentic AI to communicate with Figma for reading designs and modifying them programmatically.","summaryCn":"让 AI 助手直接读取并修改 Figma 设计稿","reasonCn":"这是一个创新性的 MCP（模型上下文协议）集成项目，让 Cursor、Claude Code 等 AI 编程助手能够直接与 Figma 通信。AI 可以读取设计稿信息，甚至修改设计，极大地提升了设计与开发的协作效率。它是 AI 赋能设计工作流的典型代表，展示了多模态 AI Agent 的潜力。","stars":7032,"starGrowth":0,"createdAt":"2025-03-16T16:45:37Z","silentRounds":2,"language":"JavaScript","topics":["design"],"aiDims":["AI Agent","工具调用","多模态Agent"],"zone":"AI","funScore":0.6,"domainTags":["MCP协议","Figma集成","AI设计工具","Agent"],"domainKey":"Agent","pushedAt":"2026-07-26T07:14:26Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.85,"source":"search","url":"https://github.com/grab/cursor-talk-to-figma-mcp","ts":"2026-07-26T07:14:26Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EleutherAI/gpt-neox","owner":"EleutherAI","name":"gpt-neox","desc":"An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries","summaryCn":"开源大模型先驱打造的分布式大语言模型训练与并行计算硬核引擎","reasonCn":"技术上，它站在了巨人的肩膀上，深度融合了NVIDIA的Megatron和微软的DeepSpeed两大王牌库，实现了精妙的模型并行、张量并行和流水线并行技术。它支持基于Transformer架构的自回归大模型训练，代码结构严谨，对分布式集群的网络通信和显存优化做到了极致，是早期开源界复现类GPT大模","stars":7460,"starGrowth":0,"createdAt":"2020-12-22T14:37:54Z","silentRounds":24,"language":"Python","topics":["transformers"],"aiDims":["大语言模型","模型与训练","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["大模型训练","分布式训练","Transformer","Megatron","DeepSpeed"],"domainKey":"大模型训练","pushedAt":"2026-09-04T21:41:34Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/EleutherAI/gpt-neox","ts":"2026-09-04T21:41:34Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"thewh1teagle/vibe","owner":"thewh1teagle","name":"vibe","desc":"Transcribe on your own!","summaryCn":"电脑上的小收音机，边听语音边自动记下内容","reasonCn":"vibe是一款专为Mac设计的开源音频转录工具，核心亮点是“转录即翻译”的沉浸式体验。它基于Whisper模型，支持实时听写、录音及视频音频转文字，且本地运行保障隐私。相比复杂方案，它只需一键安装即可将会议、播客甚至视频内容转为可编辑文本，兼顾了AI的实用性与交互的趣味性，非常适合希望提升记录效率又不愿折腾技术细节的用户。","stars":7589,"starGrowth":2,"createdAt":"2024-01-08T03:29:06Z","silentRounds":0,"language":"TypeScript","topics":["desktop"],"aiDims":["AI应用","非AI-好玩","效率工具"],"zone":"AI","funScore":0,"domainTags":["音频转录","本地AI","Mac工具"],"domainKey":"音频转录","pushedAt":"2026-09-07T02:44:20.540Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/thewh1teagle/vibe","ts":"2026-09-07T02:44:20.540Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fishjar/kiss-translator","owner":"fishjar","name":"kiss-translator","desc":"A simple, open source bilingual translation extension &amp; Greasemonkey script (一个简约、开源的 双语对照翻译扩展 &amp; 油猴脚本)","summaryCn":"浏览器里一键翻译整页，原文译文并排看，上网查资料再也不怕外语了","reasonCn":"基于油猴脚本的双语对照翻译扩展，支持 Chrome、Edge 等主流浏览器。内置多种翻译引擎可选，覆盖新闻网站、技术文档、GitHub 页面等。实时双语并排显示，不改变原页面布局，鼠标悬停即可查看翻译内容，大幅提升外语阅读效率。","stars":12589,"starGrowth":10,"createdAt":"2023-07-18T07:20:49Z","silentRounds":0,"language":"JavaScript","topics":["extension","translation","translator","translate","userscripts","userscript"],"aiDims":["效率工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["浏览器翻译","油猴脚本","双语对照"],"domainKey":"浏览器翻译","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"extension","source":"github","weight":0.7},{"name":"translation","source":"github","weight":0.7},{"name":"translator","source":"github","weight":0.7},{"name":"translate","source":"github","weight":0.7},{"name":"userscripts","source":"github","weight":0.7},{"name":"userscript","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"trending","url":"https://github.com/fishjar/kiss-translator","ts":"2026-09-07T22:38:09.240Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.0323830328064183,"copyOk":true},{"repo":"vercel/turborepo","owner":"vercel","name":"turborepo","desc":"Build system optimized for JavaScript and TypeScript, written in Rust","summaryCn":"让前端项目构建速度起飞的加速器，告别漫长的等待。","reasonCn":"Turborepo 是一个用 Rust 编写的智能构建系统，专为 JavaScript 和 TypeScript 项目设计。它能自动缓存依赖任务结果，仅对变更文件重新执行，从而大幅提升重复构建速度。它原生支持 Monorepo 管理和依赖图分析，能精准追踪任务依赖，避免不必要的重复工作。支持缓存回滚和远程缓存，方便团队协作和 CI 流水线优化。","stars":31129,"starGrowth":2,"createdAt":"2021-10-05T17:37:11Z","silentRounds":0,"language":"Rust","topics":["monorepo","javascript","typescript","build-tool","build-system"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["前端工程化","Monorepo","构建系统","Rust"],"domainKey":"前端工程化","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"monorepo","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"build-tool","source":"github","weight":0.7},{"name":"build-system","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/vercel/turborepo","ts":"2026-08-18T00:56:27.274Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"stanfordnlp/dspy","owner":"stanfordnlp","name":"dspy","desc":"DSPy: The framework for programming—not prompting—language models","summaryCn":"把调教AI变成写代码，自动优化提示和流程的神器","reasonCn":"DSPy是斯坦福团队开发的AI编程框架，核心理念是用代码而非手写提示词来构建和管理大模型管道。它通过声明式规范和编译器自动优化提示策略、模型选择和工具调用链，让开发者专注于逻辑而非反复试错调参。支持链式调用、检索增强生成、多步推理等复杂场景，内置多种优化器可自动搜索最优提示组合。","stars":38219,"starGrowth":4,"createdAt":"2023-01-09T21:01:51Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["提示工程","AI应用","工具调用"],"zone":"AI","funScore":1,"domainTags":["提示词优化","AI框架","自动调参"],"domainKey":"提示词优化","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/stanfordnlp/dspy","ts":"2026-08-12T01:22:42.342Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mhss1/MyBrain","owner":"mhss1","name":"MyBrain","desc":"All-in-one productivity app and AI assistant with Tasks, Notes, Calendar, Diary and Bookmarks.","summaryCn":"现代人的生活被各种碎片信息淹没：随手记下的笔记、没处理的代办任务、还没","reasonCn":"它的架构基于Kotlin开发，专注于移动端体验。核心亮点在于“智能化知识处理”。你不只是被动存入信息，AI会根据你的笔记内容自动提醒你相关的任务，或者在你忘记某个知识点时，通过对话方式将其提取出来。这种闭环设计让信息处理从“存储”变成了“交互”。","stars":2210,"starGrowth":0,"createdAt":"2022-05-11T17:06:31Z","silentRounds":3,"language":"Kotlin","topics":["chatbot"],"aiDims":["AI应用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Android","Productivity","Note-taking"],"domainKey":"Android","pushedAt":"2026-08-20T17:00:00Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7},{"name":"Kotlin","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/mhss1/MyBrain","ts":"2026-08-20T17:00:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lackeyjb/playwright-skill","owner":"lackeyjb","name":"playwright-skill","desc":"Claude Code Skill for browser automation with Playwright. Model-invoked - Claude autonomously writes and executes custom automation for testing and validation.","summaryCn":"让AI自己写代码帮你操控浏览器做测试，懒人福音","reasonCn":"这是Claude Code的Skill扩展，利用Playwright实现浏览器自动化。Claude能自主编写并执行JavaScript代码，完成测试、验证、数据抓取等任务。无需人工干预，只需告诉AI你想做什么，它会自己规划和执行浏览器操作，大幅降低重复性Web操作的复杂度，适合做集成测试和自动化流程。","stars":3138,"starGrowth":3,"createdAt":"2025-10-19T21:33:51Z","silentRounds":0,"language":"JavaScript","topics":["agent-skills","ai-agents","automations","coding-agents","developer-tool","e2e-testing-tools","playwright-automation","playwright-tests","web-automation","web-testing-automation"],"aiDims":["AI Agent","工具调用","代码助手"],"zone":"AI","funScore":0,"domainTags":["浏览器自动化","Playwright","AI Agent","测试工具"],"domainKey":"AI Agent","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"agent-skills","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"automations","source":"github","weight":0.7},{"name":"coding-agents","source":"github","weight":0.7},{"name":"developer-tool","source":"github","weight":0.7},{"name":"e2e-testing-tools","source":"github","weight":0.7},{"name":"playwright-automation","source":"github","weight":0.7},{"name":"playwright-tests","source":"github","weight":0.7},{"name":"web-automation","source":"github","weight":0.7},{"name":"web-testing-automation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/lackeyjb/playwright-skill","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Fredolx/open-tv","owner":"Fredolx","name":"open-tv","desc":"Ultra-fast, simple and powerful cross-platform IPTV app","summaryCn":"跨平台IPTV神器，看片不卡顿，还能播放本地媒体。","reasonCn":"这是一个用Rust编写的跨平台IPTV播放器，支持Windows、Linux、macOS和Android。它以极速启动和低资源占用著称，能流畅播放M3U、M3U8等格式的流媒体列表，甚至支持播放本地视频文件。项目目前Star超过3900，社区活跃，深受极客用户喜爱。","stars":4037,"starGrowth":1,"createdAt":"2022-03-13T15:24:37Z","silentRounds":3,"language":"Rust","topics":[],"aiDims":["非AI-好玩","非AI-实用","效率工具"],"zone":"创意","funScore":0,"domainTags":["IPTV播放器","流媒体播放","跨平台应用"],"domainKey":"IPTV播放器","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/Fredolx/open-tv","ts":"2026-08-12T01:22:42.342Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"MatinSenPai/SenPaiScanner","owner":"MatinSenPai","name":"SenPaiScanner","desc":"A light-weight scanner for Cloudflare IPs, written in Golang","summaryCn":"专治各种网络IP迷路，一键帮你找出隐藏的云厂商IP。","reasonCn":"这是一个用Go语言编写的轻量级工具，专门用于扫描和识别Cloudflare的IP地址。它能帮助用户快速检测目标网站是否使用了Cloudflare防护，并获取相关的IP段信息。项目设计简洁，性能高效，适合网络安全研究人员和开发者进行网络资产探测。","stars":2394,"starGrowth":0,"createdAt":"2026-05-28T17:02:54Z","silentRounds":1,"language":"Go","topics":[],"aiDims":["安全工具","非AI-实用","开发者工具"],"zone":"工具","funScore":0.5,"domainTags":["Cloudflare","IP扫描","网络安全","CDN检测"],"domainKey":"Cloudflare","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/MatinSenPai/SenPaiScanner","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"leaningtech/webvm","owner":"leaningtech","name":"webvm","desc":"Virtual Machine for the Web","summaryCn":"浏览器里直接跑虚拟机，让开发环境像网页一样随处可用","reasonCn":"WebVM 是一个基于 WebAssembly 的虚拟机解决方案，它允许用户在浏览器中运行完整的操作系统和开发环境。通过这个项目，你可以在任何设备上通过网页访问 Linux 终端，无需安装任何软件，非常适合跨平台开发和远程办公。","stars":17404,"starGrowth":1,"createdAt":"2022-01-31T20:45:58Z","silentRounds":2,"language":"JavaScript","topics":["cheerp","cheerpx","cpp","tailscale","vm","wasm","webassembly","xterm-js","repl","lwip","webvm"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":1,"domainTags":["WebAssembly","虚拟机","Linux","浏览器开发"],"domainKey":"WebAssembly","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"cheerp","source":"github","weight":0.7},{"name":"cheerpx","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"tailscale","source":"github","weight":0.7},{"name":"vm","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"webassembly","source":"github","weight":0.7},{"name":"xterm-js","source":"github","weight":0.7},{"name":"repl","source":"github","weight":0.7},{"name":"lwip","source":"github","weight":0.7},{"name":"webvm","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/leaningtech/webvm","ts":"2026-08-18T00:56:27.274Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Hmbown/CodeWhale","owner":"Hmbown","name":"CodeWhale","desc":"Open-source, community-driven agent harness","summaryCn":"Rust写的AI智能体容器，把各类工具串起来自动干活","reasonCn":"这是一个用Rust构建的开源智能体编排框架，专注于让AI Agent能够调用外部工具和执行复杂任务。它提供了模块化设计，支持自定义工具集与上下文管理，帮助开发者快速搭建能联网、读写文件、操作软件的多步自动化流程。相比纯对话模型，它能真正动手办事，适合需要闭环任务解决的场景。","stars":41029,"starGrowth":0,"createdAt":"2026-01-19T18:21:01Z","silentRounds":2,"language":"Rust","topics":["llm","ai-agent"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","Rust","工具编排","LLM集成"],"domainKey":"AI Agent","pushedAt":"2026-08-18T14:31:53Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm","source":"github","weight":0.7},{"name":"ai-agent","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/Hmbown/CodeWhale","ts":"2026-08-18T14:31:53Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/agent-governance-toolkit","owner":"microsoft","name":"agent-governance-toolkit","desc":"AI Agent Governance Toolkit — Policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. Covers 10/10 OWASP Agentic Top 10.","summaryCn":"给AI Agent装上监控和刹车，让自主智能体不再乱来","reasonCn":"微软开源的AI Agent治理工具包，专为企业和开发者设计，提供策略执行、零信任身份认证、执行沙盒隔离及可靠性工程能力。覆盖OWASP Agentic Top 10全部十项安全风险，帮助构建安全可控的自主AI Agent系统，防止越权操作和数据泄露，适合需要规模化部署AI Agent的团队。","stars":6317,"starGrowth":1,"createdAt":"2026-03-02T22:11:47Z","silentRounds":1,"language":"Python","topics":["agent-framework","ai-agents","ai-safety","compliance","governance","microsoft","owasp","policy-engine","python","security","trust","zero-trust"],"aiDims":["AI Agent","安全工具","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI治理","零信任","安全框架","Agent"],"domainKey":"Agent","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"agent-framework","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-safety","source":"github","weight":0.7},{"name":"compliance","source":"github","weight":0.7},{"name":"governance","source":"github","weight":0.7},{"name":"microsoft","source":"github","weight":0.7},{"name":"owasp","source":"github","weight":0.7},{"name":"policy-engine","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"trust","source":"github","weight":0.7},{"name":"zero-trust","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/microsoft/agent-governance-toolkit","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"divbasson/DesktopClaw","owner":"divbasson","name":"DesktopClaw","desc":"DesktopClaw is a state-aware desktop pet interface for your OpenClaw instance. It brings your AI to life as a reactive companion you can interact with via voice or text, while providing real-time session visibility, task feedback, and quick action triggers through a persistent, animated desktop presence.","summaryCn":"你的AI助手变成桌面宠物，会说话能互动让码字不再孤单","reasonCn":"这个开源项目把OpenClaw AI实例变成了一个趴在桌面上的虚拟宠物。它支持语音和文字双向交互，会根据你的操作做出不同动画反应，写代码时安静待着，休息时会找你玩。虽然star数不多但创意十足，把冷冰冰的AI工具变成了有温度的桌面伴侣。","stars":15,"starGrowth":0,"createdAt":"2026-04-23T11:47:27Z","silentRounds":23,"language":"JavaScript","topics":["virtual-pet"],"aiDims":["AI应用","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["桌面宠物","OpenClaw","AI 交互","虚拟伴侣"],"domainKey":"桌面宠物","pushedAt":"2026-09-13T05:58:28Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"virtual-pet","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/divbasson/DesktopClaw","ts":"2026-09-13T05:58:28Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"higress-group/higress","owner":"higress-group","name":"higress","desc":"🤖 AI Gateway | AI Native API Gateway","summaryCn":"专门帮各大AI服务商管API的守门员，统一调度各种大模型接口","reasonCn":"Higress是一个基于Envoy的AI原生API网关，专门解决企业接入多种大模型API时的统一管理问题。它把OpenAI、通义千问、文心一言等各种AI接口统一接入，提供流量分发、限流熔断、身份认证、调用计量等企业级功能。底层基于Go语言和Envoy代理构建，支持独立部署和K8s Ingress两种模式，为企业AI应用提供统一入口。","stars":9440,"starGrowth":5,"createdAt":"2022-10-27T03:53:00Z","silentRounds":0,"language":"Go","topics":["api-gateway","ai-gateway","envoy","ai-native","cloud-native"],"aiDims":["AI基础设施","开发者工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["API网关","流量调度","大模型管理","Envoy代理"],"domainKey":"API网关","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"api-gateway","source":"github","weight":0.7},{"name":"ai-gateway","source":"github","weight":0.7},{"name":"envoy","source":"github","weight":0.7},{"name":"ai-native","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/higress-group/higress","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.016588983050847456,"copyOk":false},{"repo":"google/automl","owner":"google","name":"automl","desc":"","summaryCn":"谷歌开源的这个 AutoML 实验项目聚焦于自动化机器学习与神经架构搜索的前沿","reasonCn":"作为早期开源的自动化建模探索代码，它涵盖了神经网络架构搜索的相关核心逻辑与实验配置。项目通过代码展示了如何让算法在给定的搜索空间内不断尝试、评估并挑选出表现最好的模型结构，从而减少人工调参和架构设计的大量试错成本，为后续更成熟的 AutoML 工具链奠定了基础。","stars":6474,"starGrowth":0,"createdAt":"2020-03-12T03:52:47Z","silentRounds":8,"language":"Jupyter Notebook","topics":["automl","efficientdet","object-detection","efficientnet","efficientnetv2"],"aiDims":["模型与训练","AI基础设施","开发者工具"],"zone":"资源","funScore":0,"domainTags":["神经架构搜索","自动化机器学习","深度学习"],"domainKey":"神经架构搜索","pushedAt":"2026-09-15T22:54:20.912Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"automl","source":"github","weight":0.7},{"name":"efficientdet","source":"github","weight":0.7},{"name":"object-detection","source":"github","weight":0.7},{"name":"efficientnet","source":"github","weight":0.7},{"name":"efficientnetv2","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/google/automl","ts":"2026-09-15T22:54:20.912Z","category":"learning","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"r0f1/datascience","owner":"r0f1","name":"datascience","desc":"Curated list of Python resources for data science.","summaryCn":"Python数据科学资源精选集合，帮你少走弯路。","reasonCn":"这是一个精心整理的Python数据科学学习资源清单，涵盖了从基础入门到高级应用的各类教程、库文档和实战案例。它系统地汇集了数据可视化、数据分析、机器学习等领域的优质内容，旨在帮助初学者快速建立知识体系，也让资深从业者高效找到前沿工具。","stars":4672,"starGrowth":0,"createdAt":"2018-11-22T09:21:49Z","silentRounds":3,"language":"","topics":["data-visualization","data-science"],"aiDims":["学习资源","可视化","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["Python","数据科学","教程","资源列表","可视化"],"domainKey":"Python","pushedAt":"2026-09-07T18:59:25Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"data-science","source":"github","weight":0.7}],"aiScore":0.7,"source":"search","url":"https://github.com/r0f1/datascience","ts":"2026-09-07T18:59:25Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cactus-compute/cactus","owner":"cactus-compute","name":"cactus","desc":"Quantization, kernels, runtime and inference engine for mobiles, wearables, smart home and robots. ","summaryCn":"专为手机和嵌入式设备打造的AI推理引擎，让轻量级模型也能流畅运行","reasonCn":"Cactus从底层出发，提供完整的解决方案链：模型量化压缩减少体积、定制化内核提升计算效率、轻量级运行时管理资源调度。支持多种主流模型格式，针对ARM架构等移动端芯片做了深度优化。相比llama.cpp主要面向CPU/GPU通用场景，Cactus更专注于手机、手表、机器人等边缘设备。","stars":6050,"starGrowth":5,"createdAt":"2025-04-23T14:33:43Z","silentRounds":0,"language":"C++","topics":["mobile","rag"],"aiDims":["推理引擎","模型部署"],"zone":"AI","funScore":0,"domainTags":["移动端推理","量化","边缘计算","嵌入式","AI推理引擎"],"domainKey":"移动端推理","pushedAt":"2026-09-08T13:54:15Z","tags":[{"name":"推理引擎","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"rag","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/cactus-compute/cactus","ts":"2026-09-08T13:54:15Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.017479338842975208,"copyOk":false},{"repo":"cordiverse/cordis","owner":"cordiverse","name":"cordis","desc":"Meta-Framework of Spatiotemporal Composability","summaryCn":"像搭积木一样把AI功能快速拼装成应用的神器","reasonCn":"Cordis 是一个 TypeScript 开发的元框架，专注于时空组合性，旨在简化 AI 应用开发流程。它通过声明式配置和模块化设计，让开发者能像搭积木一样快速构建复杂的应用。项目支持多种 AI 能力的集成，提供了丰富的工具链，帮助用户快速搭建各类 AI 应用，特别适合需要高效开发和迭代的场景。","stars":8765,"starGrowth":13,"createdAt":"2022-05-17T18:52:57Z","silentRounds":0,"language":"TypeScript","topics":["framework","nodejs","effect","plugin"],"aiDims":["AI基础设施","AI应用","效率工具"],"zone":"AI","funScore":0,"domainTags":["AI应用框架","声明式配置","模块化开发","TypeScript"],"domainKey":"AI应用框架","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"framework","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"effect","source":"github","weight":0.7},{"name":"plugin","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/cordiverse/cordis","ts":"2026-08-19T01:08:47.324Z","category":"ai","momentum":["daily"],"fromOfficial":false,"heatScore":0.04344951511694238,"copyOk":false},{"repo":"containers/kubernetes-mcp-server","owner":"containers","name":"kubernetes-mcp-server","desc":"Model Context Protocol (MCP) server for Kubernetes and OpenShift","summaryCn":"让AI能直接和K8s集群对话，运维操作全交给大模型","reasonCn":"该项目将Kubernetes集群变为MCP协议的服务端，让AI模型通过标准工具接口直接操作K8s资源。支持查询集群状态、管理Pod和Deployment、查看日志、排查故障等常用运维任务，可无缝接入Claude、Cursor等主流AI应用，实现自然语言驱动集群管理。","stars":2116,"starGrowth":0,"createdAt":"2025-02-11T02:57:36Z","silentRounds":1,"language":"Go","topics":["containers","context","kubernetes","mcp","model","openshift","protocol","modelcontextprotocol","kubernetes-mcp"],"aiDims":["工具调用","AI Agent","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["Kubernetes","MCP协议","DevOps","容器管理"],"domainKey":"Kubernetes","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"containers","source":"github","weight":0.7},{"name":"context","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"model","source":"github","weight":0.7},{"name":"openshift","source":"github","weight":0.7},{"name":"protocol","source":"github","weight":0.7},{"name":"modelcontextprotocol","source":"github","weight":0.7},{"name":"kubernetes-mcp","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/containers/kubernetes-mcp-server","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"yvgude/lean-ctx","owner":"yvgude","name":"lean-ctx","desc":"Control what your AI can see. LeanCTX (Lean Context) is the context intelligence layer for AI agents — one local Rust binary that decides what they read, remembers what they learn, guards what they touch, and proves what they save. 60–90% fewer tokens as the receipt. 76 MCP tools, 30+ agents, local-first.","summaryCn":"控制AI视野与记忆，守护隐私的Agent上下文管家","reasonCn":"LeanCTX是一个AI Agent的上下文管理层，用Rust编写的本地二进制程序。它决定Agent能读取哪些信息、记住哪些内容，同时防止敏感数据泄露。通过细粒度的上下文控制，让Agent在使用工具和访问外部资源时更精准安全，且无需依赖外部云服务，本地运行隐私有保障。","stars":3825,"starGrowth":1,"createdAt":"2026-03-23T19:48:59Z","silentRounds":1,"language":"Rust","topics":["ai","cursor","llm","mcp","rust","token-optimization","agentic-coding","claude-code","copilot","developer-tools","ai-coding","context-engineering","gemini-cli","mcp-server","reduce-token-costs","context-layer","ai-agents","lean-context","context-intelligence"],"aiDims":["AI Agent","AI基础设施","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["Agent安全","上下文管理","隐私保护","本地部署"],"domainKey":"Agent安全","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"cursor","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"token-optimization","source":"github","weight":0.7},{"name":"agentic-coding","source":"github","weight":0.7},{"name":"claude-code","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"developer-tools","source":"github","weight":0.7},{"name":"ai-coding","source":"github","weight":0.7},{"name":"context-engineering","source":"github","weight":0.7},{"name":"gemini-cli","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"reduce-token-costs","source":"github","weight":0.7},{"name":"context-layer","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"lean-context","source":"github","weight":0.7},{"name":"context-intelligence","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/yvgude/lean-ctx","ts":"2026-08-08T00:55:52.206Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"bkamins/Julia-DataFrames-Tutorial","owner":"bkamins","name":"Julia-DataFrames-Tutorial","desc":"A tutorial on Julia DataFrames package","summaryCn":"数据表处理是数据科学的基石，但在 Julia 生态中缺少系统化的入门教材，导致","reasonCn":"数据表处理是数据科学的基石，但在 Julia 生态中缺少系统化的入门教材，导致很多人望而却步。这个仓库正是为了解决这个痛点而生，作者把完整的学习流程包装进 Jupyter Notebook，使得读者可以边看解释边运行代码，实时看到结果。","stars":538,"starGrowth":0,"createdAt":"2017-12-02T16:05:42Z","silentRounds":12,"language":"Jupyter Notebook","topics":["tutorial"],"aiDims":["教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["Julia语言","数据帧","数据分析教程"],"domainKey":"Julia语言","pushedAt":"2026-09-12T11:33:49Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"tutorial","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/bkamins/Julia-DataFrames-Tutorial","ts":"2026-09-12T11:33:49Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"backend-br/vagas","owner":"backend-br","name":"vagas","desc":"Espaço para a divulgação de vagas para desenvolvedores backend via issues do Github.","summaryCn":"GitHub上专门找后端开发的免费兼职，全靠讨论区发布。","reasonCn":"这是一个利用 GitHub Issues 功能的后端职位发布平台，完全开源且免费。开发者可以直接在项目仓库下创建 Issue 来发布工作机会，无论是全职远程还是兼职项目都能找到。这种方式打破了传统招聘网站的壁垒，让寻找和发布后端技术工作变得非常透明和高效。","stars":7981,"starGrowth":2,"createdAt":"2017-08-04T17:38:02Z","silentRounds":0,"language":"","topics":["backend"],"aiDims":["非AI-实用","开发者工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["后端开发","招聘","GitHub Issues","远程工作"],"domainKey":"后端开发","pushedAt":"2026-08-31T11:54:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"backend","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/backend-br/vagas","ts":"2026-08-31T11:54:22Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"0xSero/ai-data-extraction","owner":"0xSero","name":"ai-data-extraction","desc":"extract all your personal data history from cursor, codex, claude-code, windsurf, and trae","summaryCn":"一键把你用过的人工智能对话记录全捞出来，隐私安全。","reasonCn":"这是一个用于从各种AI编程工具中导出历史对话数据的Python工具。它支持Cursor、Claude Code、Windsurf等多个主流IDE和命令行工具，通过解析本地数据库或配置文件，将用户的对话历史整理成Markdown或JSON格式，方便备份和迁移。","stars":1283,"starGrowth":1,"createdAt":"2025-11-16T10:18:08Z","silentRounds":14,"language":"Python","topics":[],"aiDims":["代码助手","AI应用","RAG与知识"],"zone":"工具","funScore":0,"domainTags":["数据导出","AI对话记录","隐私备份"],"domainKey":"数据导出","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/0xSero/ai-data-extraction","ts":"2026-08-17T00:55:40.962Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"vercel-labs/portless","owner":"vercel-labs","name":"portless","desc":"Replace port numbers with stable, named local URLs. For humans and agents.","summaryCn":"把本地端口号变网址，像本地服务器一样好用","reasonCn":"Portless 是一款基于浏览器内核的本地开发服务器增强工具，通过解析本地网络环境，自动将复杂的 IP 地址和端口号转换为稳定、易记的域名。它利用浏览器原生的网络权限，无需安装任何代理软件，即可让局域网内的手机、平板等设备无缝访问本地 Web 项目，完美解决了跨设备调试和远程协作的痛点。","stars":12575,"starGrowth":5,"createdAt":"2026-02-15T14:43:03Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["本地开发","URL路由","端口映射","浏览器插件"],"domainKey":"本地开发","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/vercel-labs/portless","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.019431411530815106,"copyOk":false},{"repo":"daviden1013/vlm4ocr","owner":"daviden1013","name":"vlm4ocr","desc":"Python package and Web App for OCR with vision language models.","summaryCn":"用 AI 视觉模型看懂文档文字，OCR 新玩法","reasonCn":"从技术角度来看，vlm4ocr 的核心是利用多模态大模型来替代或增强传统 OCR。传统 OCR 工具如 Tesseract 只能做字符识别，而 VLM 可以理解图像语义。这个项目封装了一个 Python 包，同时提供了一个 Web 应用界面，让你可以直接上传图片或 PDF 进行识别。","stars":11,"starGrowth":0,"createdAt":"2025-04-27T18:22:23Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["文档理解","多模态Agent","AI应用"],"zone":"工具","funScore":0.5,"domainTags":["OCR","多模态","视觉模型","文档识别","Web应用"],"domainKey":"OCR","pushedAt":"2026-09-09T04:24:40Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/daviden1013/vlm4ocr","ts":"2026-09-09T04:24:40Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"anthropics/defending-code-reference-harness","owner":"anthropics","name":"defending-code-reference-harness","desc":"Skills for threat modeling, scanning, triage, patching, plus an autonomous scanning harness you can /customize","summaryCn":"给安全专家打造的自动代码漏洞扫描和修补机器人，像给代码穿上了防弹衣。","reasonCn":"这是一个专为代码安全防护设计的自动化框架，集成了威胁建模、漏洞扫描、分类和修补全流程能力。项目核心包含一个可定制的自主扫描引擎，能够深入分析代码库并执行自动化防御操作，旨在提升开发流程的安全性和响应速度。","stars":7511,"starGrowth":1,"createdAt":"2026-05-22T16:00:56Z","silentRounds":2,"language":"Python","topics":["security"],"aiDims":["安全工具","AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["代码安全","漏洞扫描","自动化修补","威胁建模"],"domainKey":"代码安全","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"security","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/anthropics/defending-code-reference-harness","ts":"2026-08-18T00:56:27.274Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"open-mercato/open-mercato","owner":"open-mercato","name":"open-mercato","desc":"AI-Engineering Foundation Framework built with AI and designed for AI. Hundreds of architectural and domain decisions (multi-tenancy, RBAC, event flow, pricing, sales pipeline,CRM/ERP processes) are already made conventions and specs so agents (Cursor, Claude Code, Codex) arch. decisions without reinventing. Ship production grade with AI Agents.","summaryCn":"专为AI应用打造的商业底层框架，涵盖权限定价全套","reasonCn":"它像是一个专门为了构建AI应用而设计的商业化底座。里面内置了多租户隔离、角色权限管理、消息事件流、用户定价模型以及CRM和ERP流程。开发者不用从零搭建这些复杂逻辑，直接基于它扩展，能快速把AI能力变成可收费、可管理的产品服务。","stars":1765,"starGrowth":0,"createdAt":"2025-09-10T13:27:45Z","silentRounds":1,"language":"TypeScript","topics":["crm","typescript","admin-panel","ai-agents","ai-coding","framework","agentic","erp","nextjs","open-source","self-hosted"],"aiDims":["AI基础设施","AI应用","工具调用"],"zone":"工具","funScore":0,"domainTags":["AI应用框架","多租户","CRM","BaaS"],"domainKey":"AI应用框架","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"crm","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"admin-panel","source":"github","weight":0.7},{"name":"ai-agents","source":"github","weight":0.7},{"name":"ai-coding","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"agentic","source":"github","weight":0.7},{"name":"erp","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.82,"source":"trending","url":"https://github.com/open-mercato/open-mercato","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"radixark/miles","owner":"radixark","name":"miles","desc":"Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.","summaryCn":"企业级RL训练框架，给大模型注入超强技能","reasonCn":"Miles是一款面向企业的大模型和视觉语言模型后训练框架，专注于强化学习训练。它由slime项目fork并协同进化而来，提供完整的RL训练pipeline和工具链，帮助团队对模型进行高效微调，使其在特定领域表现出色。目前GitHub已有2469星，Python编写。","stars":2974,"starGrowth":2,"createdAt":"2025-10-09T06:04:57Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["模型与训练","微调","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["强化学习","模型后训练","RLHF","大模型微调","Python"],"domainKey":"强化学习","pushedAt":"2026-09-06T13:43:54.327Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/radixark/miles","ts":"2026-09-06T13:43:54.327Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kmario23/deep-learning-drizzle","owner":"kmario23","name":"deep-learning-drizzle","desc":"Drench yourself in Deep Learning, Reinforcement Learning, Machine Learning, Computer Vision, and NLP by learning from these exciting lectures!!","summaryCn":"对于想要踏入人工智能领域或者希望夯实数学和算法基础的人来说，最大的痛点","reasonCn":"内容编排上，它打破了单纯堆砌链接的平庸做法，按照从浅入深的逻辑对课程进行了科学分类。从基础的神经网络原理，到前沿的计算机视觉和自然语言处理，再到充满挑战的强化学习，每一个方向都能找到对应的名校公开课、经典教材和配套代码。","stars":12948,"starGrowth":0,"createdAt":"2018-11-26T01:17:36Z","silentRounds":2,"language":"HTML","topics":["ml"],"aiDims":["学习资源","教程"],"zone":"资源","funScore":0,"domainTags":["深度学习","机器学习","强化学习","公开课"],"domainKey":"深度学习","pushedAt":"2026-08-22T15:46:06Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/kmario23/deep-learning-drizzle","ts":"2026-08-22T15:46:06Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wybthon/wybthon","owner":"wybthon","name":"wybthon","desc":"Build interactive web apps with Python running directly in the browser.","summaryCn":"使用 Wybthon 可以让习惯了 Python 的开发者直接在浏览器端构建具有现代交互","reasonCn":"在核心运行机制上，它采用了运行一次的函数组件设计。开发者只需要编写返回 HTML 标签树的函数，并在其中嵌入信号和表达式，当底层数据发生改变时，只有直接读取该数据的局部小表达式会重新执行，组件本身永远不会整体重新渲染，从而保证了极高的运行性能。","stars":38,"starGrowth":0,"createdAt":"2025-09-06T07:26:53Z","silentRounds":11,"language":"Python","topics":["web-framework"],"aiDims":["AI应用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Python","Web框架","前端","Pyodide","响应式"],"domainKey":"Python","pushedAt":"2026-09-06T22:37:55Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.71,"source":"search","url":"https://github.com/wybthon/wybthon","ts":"2026-09-06T22:37:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"keijiro/AICommand","owner":"keijiro","name":"AICommand","desc":"ChatGPT integration with Unity Editor","summaryCn":"把ChatGPT装进Unity编辑器，做游戏还能边做边问AI","reasonCn":"这个项目把OpenAI的ChatGPT直接集成到了Unity编辑器里，让你在开发游戏时可以随时调用AI辅助编程、生成文案、调试代码，甚至询问技术问题。开发者不用离开Unity界面，也不需要切换浏览器去ChatGPT官网提问，直接在编辑器内就能使用AI能力，大幅提升开发效率。","stars":4101,"starGrowth":0,"createdAt":"2023-03-18T09:23:23Z","silentRounds":28,"language":"C#","topics":[],"aiDims":["AI应用","工具调用","代码助手"],"zone":"AI","funScore":0,"domainTags":["Unity","ChatGPT集成","游戏开发","编辑器插件"],"domainKey":"Unity","pushedAt":"2023-12-05T14:02:11Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.82,"source":"bigbro","url":"https://github.com/keijiro/AICommand","ts":"2023-12-05T14:02:11Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/vidur","owner":"microsoft","name":"vidur","desc":"Accurate, large-scale, and extensible simulator for LLM inference Systems","summaryCn":"简单来说，Vidur是一个专门用来测试大模型推理系统稳定性和性能的“压力测试","reasonCn":"技术上，它采用了一种很聪明的“黑盒”模拟架构。它不需要真的加载大模型，而是模拟了网络延迟、GPU内存限制、CPU瓶颈等所有硬件资源限制。你可以把它想象成一个“系统模拟器”，里面运行着无数个虚拟用户（虚拟请求），疯狂地向你的推理服务发请求，然后统计系统的延迟、错误率等指标","stars":684,"starGrowth":1,"createdAt":"2023-11-02T10:05:21Z","silentRounds":4,"language":"Python","topics":["simulation"],"aiDims":["AI基础设施","推理引擎","模拟器"],"zone":"工具","funScore":0,"domainTags":["LLM推理系统","性能模拟","压力测试","系统仿真"],"domainKey":"LLM推理系统","pushedAt":"2026-08-24T12:47:47Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"模拟器","source":"llm","weight":1},{"name":"simulation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/microsoft/vidur","ts":"2026-08-24T12:47:47Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"NetrisTV/ws-scrcpy","owner":"NetrisTV","name":"ws-scrcpy","desc":"Web client prototype for scrcpy.","summaryCn":"说白了，这个项目就是让安卓手机变成一个网页游戏，你可以用鼠标和键盘像玩","reasonCn":"技术上，它巧妙地结合了 WebSocket 实时通信和 scrcpy 强大的显示与输入转发能力。它把手机端的视频流压缩编码，通过网络传到浏览器，浏览器用 Canvas 渲染出来，同时把你的鼠标点击、键盘按键转成触摸屏指令发回手机。","stars":2569,"starGrowth":0,"createdAt":"2019-06-07T15:16:37Z","silentRounds":1,"language":"TypeScript","topics":["mobile"],"aiDims":["游戏","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Android远程控制","WebSocket","scrcpy","Web桌面"],"domainKey":"Android远程控制","pushedAt":"2026-09-10T03:02:10Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/NetrisTV/ws-scrcpy","ts":"2026-09-10T03:02:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"The-Pocket/PocketFlow","owner":"The-Pocket","name":"PocketFlow","desc":"Pocket Flow: 100-line LLM framework. Let Agents build Agents!","summaryCn":"只要100行代码的AI智能体框架，让机器人学会自己干活","reasonCn":"仅100行代码的极简LLM框架，专为构建AI Agent设计。支持工具调用、多Agent协作、自动规划等核心能力，让开发者能快速搭建能自主完成任务的AI智能体。代码精简但功能完整，非常适合学习和快速原型开发，是理解Agent原理的最佳入门项目。","stars":11191,"starGrowth":0,"createdAt":"2024-12-24T02:43:30Z","silentRounds":1,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["AI Agent","LLM框架","多Agent","代码生成"],"domainKey":"AI Agent","pushedAt":"2026-07-26T16:13:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/The-Pocket/PocketFlow","ts":"2026-07-26T16:13:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NawfalMotii79/PLFM_RADAR","owner":"NawfalMotii79","name":"PLFM_RADAR","desc":"Open-source, low-cost 10.5 GHz PLFM phased array RADAR system","summaryCn":"自己动手做个雷达，用普通电脑也能探测物体","reasonCn":"技术上，它基于10.5GHz的PLFM（相位调制）技术，构建了一个全数字的相控阵系统。核心亮点在于数字波束形成技术，能够通过算法控制天线阵列的波束指向，实现多目标同时跟踪和角度分辨。相比传统模拟雷达，它的抗干扰能力和灵活性更强。","stars":25403,"starGrowth":5,"createdAt":"2026-03-08T23:20:58Z","silentRounds":0,"language":"PLSQL","topics":[],"aiDims":["非AI-实用","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["相控阵雷达","射频探测","DIY硬件","信号处理"],"domainKey":"相控阵雷达","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"PLSQL","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/NawfalMotii79/PLFM_RADAR","ts":"2026-08-19T01:08:47.324Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.018708577727040113,"copyOk":false},{"repo":"The-Powder-Toy/The-Powder-Toy","owner":"The-Powder-Toy","name":"The-Powder-Toy","desc":"Written in C++ and using SDL, The Powder Toy is a desktop version of the classic 'falling sand' physics sandbox, it simulates air pressure and velocity as well as heat.","summaryCn":"经典落沙物理沙盒，在浏览器里体验元素化学反应的乐趣","reasonCn":"The Powder Toy是经典falling sand游戏的开源复刻版，用C++和SDL实现，模拟真实物理效果包括气压、流速、热量传递。游戏内包含数百种元素，从简单沙子到易燃易爆化学品，可自由组合创造出令人惊喜的反应效果，完全免费跨平台可玩。","stars":5327,"starGrowth":1,"createdAt":"2010-07-25T13:47:06Z","silentRounds":1,"language":"C++","topics":["game","simulation"],"aiDims":["游戏","创意工具"],"zone":"创意","funScore":1,"domainTags":["物理模拟","沙盒游戏","元素反应","流体力学"],"domainKey":"物理模拟","pushedAt":"2026-09-09T19:26:43Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"simulation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/The-Powder-Toy/The-Powder-Toy","ts":"2026-09-09T19:26:43Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"mvdctop/Movie_Data_Capture","owner":"mvdctop","name":"Movie_Data_Capture","desc":"Local Movies Organizer","summaryCn":"本地影视库管理神器，自动抓取整理你的电影收藏。","reasonCn":"这是一个用Python写的本地电影库管理工具，专门解决手动整理视频文件的繁琐问题。它通过扫描硬盘上的媒体文件，自动匹配并下载电影封面、海报、简介等元数据，无缝集成到Jellyfin等媒体服务器中，让你无需手动搜索就能拥有一个整齐漂亮的个人影视库。","stars":7433,"starGrowth":0,"createdAt":"2019-05-26T11:24:22Z","silentRounds":1,"language":"Python","topics":["jellyfin"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["电影元数据","本地媒体管理","Jellyfin","TMDB"],"domainKey":"电影元数据","pushedAt":"2026-09-10T02:50:55Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"jellyfin","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/mvdctop/Movie_Data_Capture","ts":"2026-09-10T02:50:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"unclebob/swarm-forge","owner":"unclebob","name":"swarm-forge","desc":"A simple tool for coordinating several AI agents.","summaryCn":"让多个AI助手协同工作的协调器，不再单打独斗","reasonCn":"基于Clojure编写，专为协调多个AI Agent设计的轻量级框架。支持Agent间的通信、任务分配和结果聚合，让复杂的多Agent协作系统变得简单易用。相比LangChain等重型框架，它更聚焦多Agent协同的核心能力，适合需要多个专业Agent协作完成复杂任务的场景。","stars":3917,"starGrowth":2,"createdAt":"2026-04-17T15:18:26Z","silentRounds":0,"language":"Clojure","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","多智能体","Clojure","框架"],"domainKey":"AI Agent","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Clojure","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/unclebob/swarm-forge","ts":"2026-08-08T00:55:52.206Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"00-Evan/shattered-pixel-dungeon","owner":"00-Evan","name":"shattered-pixel-dungeon","desc":"Shattered Pixel Dungeon is an open-source traditional roguelike dungeon crawler with randomized levels and enemies, and hundreds of items to collect and use. It's based on the source code of Pixel Dungeon, by Watabou.","summaryCn":"像素风地牢Roguelike，装备百变玩法超耐玩","reasonCn":"这是一款基于Java开发的经典开源地牢探险游戏，采用Roguelike核心玩法，拥有随机生成的地图、敌人、掉落物和数百种可收集的武器装备。游戏借鉴了《Pixel Dungeon》的源代码，但加入了大量原创机制，如多样化的职业系统、深度的技能树和复杂的怪物AI。每次游玩体验都独一无二，需要玩家灵活运用策略和走位来通关。","stars":6564,"starGrowth":1,"createdAt":"2014-07-31T21:56:31Z","silentRounds":1,"language":"Java","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["Roguelike","地牢探险","像素游戏","RPG"],"domainKey":"Roguelike","pushedAt":"2026-09-09T17:20:49Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/00-Evan/shattered-pixel-dungeon","ts":"2026-09-09T17:20:49Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"FlamxGames/godot-ai-assistant-hub","owner":"FlamxGames","name":"godot-ai-assistant-hub","desc":"Embed AI assistants in Godot with the ability to read and write code in Godot's Code Editor.","summaryCn":"把AI直接塞进Godot游戏引擎里，让它帮你边写代码边做游戏","reasonCn":"在技术实现上，它采用GDScript原生开发，深度契合Godot编辑器的底层API接口。通过拦截编辑器的事件和代码窗口，插件能够无缝获取上下文代码，并与主流的大语言模型API进行交互。这种深度集成的设计，使得它不需要复杂的外部跳转，在熟悉的开发界面内就能完成代码生成、重构优化和B","stars":307,"starGrowth":0,"createdAt":"2024-10-16T04:46:23Z","silentRounds":18,"language":"GDScript","topics":["godot"],"aiDims":["代码助手","游戏"],"zone":"AI","funScore":0,"domainTags":["Godot引擎","AI助手","游戏开发","代码补全"],"domainKey":"Godot引擎","pushedAt":"2026-09-01T03:19:32Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/FlamxGames/godot-ai-assistant-hub","ts":"2026-09-01T03:19:32Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"JanLahmann/RasQberry","owner":"JanLahmann","name":"RasQberry","desc":"The RasQberry project: Exploring Quantum Computing and Qiskit with a Raspberry Pi and a 3D Printer","summaryCn":"在量子计算还很昂贵、实验设备难以触及的时代，RasQberry 用一块普通的树莓","reasonCn":"在量子计算还很昂贵、实验设备难以触及的时代，RasQberry 用一块普通的树莓派加上自制的 3D 打印支架，搭建起一个可以动手实验的量子平台。它的目标是把量子教学从实验室搬到学生的工作台，让每个人都能亲手跑一次量子线路。","stars":162,"starGrowth":0,"createdAt":"2020-09-01T18:18:24Z","silentRounds":13,"language":"Shell","topics":["3d-printing"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["量子计算","树莓派","3D打印","Qiskit","教学实验"],"domainKey":"量子计算","pushedAt":"2026-09-09T19:11:43Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"3d-printing","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/JanLahmann/RasQberry","ts":"2026-09-09T19:11:43Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EmbraceAGI/Awesome-AI-GPTs","owner":"EmbraceAGI","name":"Awesome-AI-GPTs","desc":"Awesome AI GPTs, OpenAI GPTs, GPT-4, ChatGPT, GPTs, Prompts, plugins, Prompts leaking","summaryCn":"汇聚全球最火爆AI智能体创意与提示词的百宝箱。","reasonCn":"在内容架构上，它没有复杂的底层代码，而是一个庞大且结构清晰的精选目录。它深入剖析了优秀GPTs的构建逻辑、提示词编写技巧，甚至包括一些高级的提示词泄露与安全防护研究。通过这些真实的案例，你可以直观地学习到如何用简单的自然语言让AI爆发出惊人的垂直领域工作能力。","stars":1199,"starGrowth":1,"createdAt":"2023-11-10T11:45:21Z","silentRounds":21,"language":"","topics":["chatbot"],"aiDims":["AI Agent","提示工程","学习资源"],"zone":"资源","funScore":0,"domainTags":["GPTs","提示词工程","AI应用","资源导航"],"domainKey":"GPTs","pushedAt":"2026-08-25T13:51:22Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"chatbot","source":"github","weight":0.7}],"aiScore":0.79,"source":"search","url":"https://github.com/EmbraceAGI/Awesome-AI-GPTs","ts":"2026-08-25T13:51:22Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Albertchamberlain/Awesome-OKF","owner":"Albertchamberlain","name":"Awesome-OKF","desc":"OKF (Open Knowledge Format) — curated catalog of tools, plugins, skills, proposals, and docs for agent-friendly knowledge. YAML-driven, agent-searchable, MCP-ready.","summaryCn":"在信息爆炸的时代，企业内部的文档、技术方案和经验沉淀往往散落各处，难以","reasonCn":"在信息爆炸的时代，企业内部的文档、技术方案和经验沉淀往往散落各处，难以被有效利用。Awesome-OKF 通过聚合一系列遵循开放知识格式（OKF）的开源项目，提供了一套统一的、机器可读的知识管理方案，让 AI Agent 能够直接读取 YAML 描述的元数据并进行语义检索。","stars":100,"starGrowth":0,"createdAt":"2026-09-07T06:12:39Z","silentRounds":11,"language":"Python","topics":["knowledge-management"],"aiDims":["AI Agent","RAG与知识","知识图谱"],"zone":"资源","funScore":0,"domainTags":["知识管理","Agent工具","RAG","YAML","元数据"],"domainKey":"知识管理","pushedAt":"2026-09-15T02:03:40Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.57,"source":"search","url":"https://github.com/Albertchamberlain/Awesome-OKF","ts":"2026-09-15T02:03:40Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"wokwi/rp2040js","owner":"wokwi","name":"rp2040js","desc":"A Raspberry Pi Pico Emulator in JavaScript","summaryCn":"在浏览器里无硬件仿真树莓派Pico芯片，写C代码就能玩电子做实验","reasonCn":"想学嵌入式开发，买板子太贵？怕焊接烧坏芯片？或者只是想试试树莓派Pico的强大功能，但手上没设备？rp2040js就是为此而生的“电子模拟器”。它把整颗RP2040芯片——包括CPU核心、内存、GPIO引脚、时钟和定时器——都虚拟到了一个正在运行的网页里。","stars":526,"starGrowth":0,"createdAt":"2021-01-26T21:38:25Z","silentRounds":14,"language":"TypeScript","topics":["emulator"],"aiDims":["非AI-好玩","开发者工具","学习资源"],"zone":"创意","funScore":1,"domainTags":["RP2040","仿真器","WebAssembly","嵌入式开发","电子实验"],"domainKey":"RP2040","pushedAt":"2026-09-09T15:34:18Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/wokwi/rp2040js","ts":"2026-09-09T15:34:18Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NVIDIA-NeMo/Speech","owner":"NVIDIA-NeMo","name":"Speech","desc":"A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)","summaryCn":"英伟达官方大模型语音工具箱，一键搞定识别和合成，专业又好用。","reasonCn":"NVIDIA官方出品的语音AI框架，专为研究和开发者打造。它统一封装了自动语音识别(ASR)和文本转语音(TTS)两大核心能力，支持从数据预处理到模型训练的全流程。凭借规模化架构，能轻松应对从研究原型到工业级部署的各种复杂需求，是构建高品质语音应用的得力助手。","stars":18490,"starGrowth":1,"createdAt":"2019-08-05T20:16:42Z","silentRounds":1,"language":"Python","topics":["machine-translation","speaker-recognition","asr","tts","generative-ai","deeplearning","neural-networks","speaker-diariazation","speech-translation","speech-synthesis"],"aiDims":["AI应用","模型与训练"],"zone":"AI","funScore":0,"domainTags":["语音识别","语音合成","深度学习框架"],"domainKey":"语音识别","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"machine-translation","source":"github","weight":0.7},{"name":"speaker-recognition","source":"github","weight":0.7},{"name":"asr","source":"github","weight":0.7},{"name":"tts","source":"github","weight":0.7},{"name":"generative-ai","source":"github","weight":0.7},{"name":"deeplearning","source":"github","weight":0.7},{"name":"neural-networks","source":"github","weight":0.7},{"name":"speaker-diariazation","source":"github","weight":0.7},{"name":"speech-translation","source":"github","weight":0.7},{"name":"speech-synthesis","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.8,"source":"trending","url":"https://github.com/NVIDIA-NeMo/Speech","ts":"2026-08-08T00:55:52.206Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"CaviraOSS/OpenMemory","owner":"CaviraOSS","name":"OpenMemory","desc":"Local persistent memory store for LLM applications including claude desktop, github copilot, codex, antigravity, etc.","summaryCn":"给LLM装本地持久记忆，跨会话保存上下文，让AI助手不再健忘。","reasonCn":"OpenMemory是一个基于TypeScript的本地持久记忆存储系统，专为LLM应用程序设计。它使用向量数据库存储和检索对话上下文，支持Claude Desktop、GitHub Copilot等多种LLM。提供简单API集成，让AI应用记住用户偏好和历史交互，实现个性化和连续对话体验。","stars":4511,"starGrowth":0,"createdAt":"2025-10-19T16:12:49Z","silentRounds":4,"language":"TypeScript","topics":["vector-db"],"aiDims":["RAG与知识","向量数据库","AI Agent"],"zone":"AI","funScore":0,"domainTags":["向量数据库","记忆存储","本地部署","LLM应用"],"domainKey":"向量数据库","pushedAt":"2026-08-18T14:06:59Z","tags":[{"name":"RAG与知识","source":"llm","weight":1},{"name":"向量数据库","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/CaviraOSS/OpenMemory","ts":"2026-08-18T14:06:59Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"vikiboss/60s","owner":"vikiboss","name":"60s","desc":"⏰ 60s API 免费接口。每天 60 秒看世界、奥运奖牌榜、小红书/B站/微博/抖音/知乎热搜、金价、油价、天气、翻译、壁纸、Epic 游戏、二维码、猫眼票房｜一系列 高质量、开源、可靠、全球 CDN 加速 的开放 API 集合，支持 Docker / Deno / Bun / Cloudflare Workers / Node.js 部署","summaryCn":"在开发任何需要实时外部数据的产品时，常常要去寻找不同的 API，注册账号、","reasonCn":"在开发任何需要实时外部数据的产品时，常常要去寻找不同的 API，注册账号、阅读文档、处理跨域、处理限流，过程繁琐。60s 项目把这些常见的数据源统一到一个仓库，用统一的路径和参数格式，让开发者只需记住一个域名，就能快速获取所需信息。","stars":5764,"starGrowth":0,"createdAt":"2022-11-30T01:19:43Z","silentRounds":1,"language":"TypeScript","topics":["api"],"aiDims":["工具调用","非AI-实用","AI应用"],"zone":"工具","funScore":0,"domainTags":["开放API","天气查询","热搜聚合","CDN加速"],"domainKey":"开放API","pushedAt":"2026-09-10T01:41:27Z","tags":[{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/vikiboss/60s","ts":"2026-09-10T01:41:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"likeslines-maker/VectorRAG.Net","owner":"likeslines-maker","name":"VectorRAG.Net","desc":"VectorRAG.Net is a .NET-native high-performance vector database library for semantic search and RAG (Retrieval-Augmented Generation). Core search is based on Random Hyperplane LSH candidate generation with exact rerank by dot/cosine.","summaryCn":"很多开发者想在C#项目里加个大模型知识库，却发现要配置庞大的Elasticsearc","reasonCn":"它的核心巧妙之处在于使用LSH算法进行向量检索。传统数据库往往计算量巨大，而LSH通过将相似向量分入相同存储桶，从而避开了暴力穷举，极大提升了搜索速度。这种设计让它在处理中小规模的知识库时，既保持了高查询性能，又省去了搭建复杂分布式系统的烦恼。","stars":164,"starGrowth":0,"createdAt":"2026-02-08T17:25:32Z","silentRounds":20,"language":"C#","topics":["vector-db"],"aiDims":["向量数据库","RAG与知识"],"zone":"AI","funScore":0.5,"domainTags":["向量数据库","RAG检索",".NET框架"],"domainKey":"向量数据库","pushedAt":"2026-08-23T17:08:08Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.8,"source":"search","url":"https://github.com/likeslines-maker/VectorRAG.Net","ts":"2026-08-23T17:08:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sveltejs/kit","owner":"sveltejs","name":"kit","desc":"web development, streamlined","summaryCn":"让你写前端像搭积木一样爽的框架，不用懂太复杂的框架也能做超酷网页","reasonCn":"这是Svelte团队开发的现代全栈框架，采用编译时渲染技术，在构建阶段生成高效的原生JS代码，极大减少了运行时的开销。它内置了路由、服务器端渲染和端点功能，支持从数据库到前端页面的全栈开发。Kit提供了优秀的开发者体验和文档，让构建复杂的Web应用变得简单有趣。","stars":20826,"starGrowth":3,"createdAt":"2020-10-15T14:00:23Z","silentRounds":0,"language":"JavaScript","topics":["svelte"],"aiDims":["开发者工具","效率工具","创意工具"],"zone":"工具","funScore":0,"domainTags":["全栈框架","Svelte","服务端渲染","Web开发"],"domainKey":"全栈框架","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"svelte","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/sveltejs/kit","ts":"2026-08-15T00:57:24.351Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Gaurav-Gosain/tuios","owner":"Gaurav-Gosain","name":"tuios","desc":"Terminal UI OS (Terminal Multiplexer)","summaryCn":"在黑框框里玩出花样的微型操作系统，黑客的最爱","reasonCn":"这是一个基于Go语言编写的终端UI操作系统，复刻了类Unix的桌面环境。它实现了窗口管理、文件系统模拟和命令行交互，完全在终端里运行。不仅支持多任务和窗口拖拽，还能跑其他终端程序，把命令行变成了一个功能完备的微型系统。","stars":3707,"starGrowth":3,"createdAt":"2025-09-06T18:13:51Z","silentRounds":0,"language":"Go","topics":["bubbletea","charm","go","multiplexer","pty","terminal"],"aiDims":["非AI-好玩","非AI-实用","开发者工具"],"zone":"创意","funScore":1,"domainTags":["终端UI","窗口管理","微型OS","黑客美学"],"domainKey":"终端UI","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"bubbletea","source":"github","weight":0.7},{"name":"charm","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"multiplexer","source":"github","weight":0.7},{"name":"pty","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/Gaurav-Gosain/tuios","ts":"2026-08-19T01:08:47.324Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"playcanvas/supersplat","owner":"playcanvas","name":"supersplat","desc":"3D Gaussian Splat Editor","summaryCn":"这东西是个玩3D的高性能编辑器。你拍一段视频，它就能生成一个逼真的3D场景","reasonCn":"核心是Gaussian Splatting技术，把3D场景表示为很多小椭球，渲染速度快，细节保留好。它在WebGL里实现，不用安装，浏览器打开就能用。支持导入各种格式，包括Colmap、Instant Ngp等工具输出的数据。编辑功能包括移动、缩放、旋转点云，添加、删除、修改元素，还能调色、调透明度。","stars":10248,"starGrowth":4,"createdAt":"2023-10-19T14:45:07Z","silentRounds":0,"language":"TypeScript","topics":["3d-gaussian-splatting","playcanvas","webgl","webgpu","pcui","typescript","gaussian-splatting","3dgs"],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"工具","funScore":1,"domainTags":["3D渲染","Gaussian Splatting","WebGL","点云编辑"],"domainKey":"3D渲染","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"3d-gaussian-splatting","source":"github","weight":0.7},{"name":"playcanvas","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"webgpu","source":"github","weight":0.7},{"name":"pcui","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"gaussian-splatting","source":"github","weight":0.7},{"name":"3dgs","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/playcanvas/supersplat","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vearch/vearch","owner":"vearch","name":"vearch","desc":"Distributed vector search for AI-native applications","summaryCn":"专为AI应用打造的分布式向量搜索数据库，让语义检索变得简单高效","reasonCn":"vearch是一款分布式向量搜索引擎，专为AI原生应用设计，支持高维向量相似度搜索。采用分布式架构，能够水平扩展处理海量向量数据。它提供REST API和多种语言SDK，支持HNSW、IVF等多种索引算法，适合RAG系统、推荐引擎、图像检索等需要语义理解的AI应用场景。","stars":2327,"starGrowth":0,"createdAt":"2019-05-13T02:37:31Z","silentRounds":14,"language":"Python","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI搜索"],"zone":"AI","funScore":0,"domainTags":["向量数据库","分布式搜索","RAG","语义检索"],"domainKey":"向量数据库","pushedAt":"2026-07-27T05:29:53Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.82,"source":"search","url":"https://github.com/vearch/vearch","ts":"2026-07-27T05:29:53Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"rclone/rclone","owner":"rclone","name":"rclone","desc":"\"rsync for cloud storage\" - Google Drive, S3, Dropbox, Backblaze B2, One Drive, Swift, Hubic, Wasabi, Google Cloud Storage, Azure Blob, Azure Files, Yandex Files","summaryCn":"把本地文件像 rsync 一样同步到各大云盘的万能搬家工具","reasonCn":"rclone 是用 Go 编写的跨云存储同步工具，支持超过 40 种后端包括 Google Drive、S3、Dropbox、OneDrive 等，可实现类似 rsync 的增量复制、加密传输和带宽限制。它通过统一的虚拟文件系统把不同云平台抽象成同一层，用户只需一条命令即可在本地和云端、或云与云之间自由迁移数据，且支持批量删除、校验和以及并行上传，极大提升大文件和海量对象的管理效率。","stars":59900,"starGrowth":15,"createdAt":"2014-03-16T16:19:57Z","silentRounds":0,"language":"Go","topics":["golang","go","cloud-storage","s3","openstack-swift","google-drive","azure-blob","backblaze-b2","sftp","ftp","rclone","dropbox","google-cloud-storage","onedrive","sync","fuse-filesystem","encryption","webdav","azure-blob-storage","azure-files"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["云存储同步","文件传输","命令行工具","Go","备份归档"],"domainKey":"云存储同步","pushedAt":"2026-09-06T11:32:12.876Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"golang","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"cloud-storage","source":"github","weight":0.7},{"name":"s3","source":"github","weight":0.7},{"name":"openstack-swift","source":"github","weight":0.7},{"name":"google-drive","source":"github","weight":0.7},{"name":"azure-blob","source":"github","weight":0.7},{"name":"backblaze-b2","source":"github","weight":0.7},{"name":"sftp","source":"github","weight":0.7},{"name":"ftp","source":"github","weight":0.7},{"name":"rclone","source":"github","weight":0.7},{"name":"dropbox","source":"github","weight":0.7},{"name":"google-cloud-storage","source":"github","weight":0.7},{"name":"onedrive","source":"github","weight":0.7},{"name":"sync","source":"github","weight":0.7},{"name":"fuse-filesystem","source":"github","weight":0.7},{"name":"encryption","source":"github","weight":0.7},{"name":"webdav","source":"github","weight":0.7},{"name":"azure-blob-storage","source":"github","weight":0.7},{"name":"azure-files","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/rclone/rclone","ts":"2026-09-06T11:32:12.876Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04575125208681135,"copyOk":false},{"repo":"stephane-caron/pink","owner":"stephane-caron","name":"pink","desc":"Python inverse kinematics using Pinocchio and QP solvers","summaryCn":"在机器人控制领域，逆向动力学（IK）是把末端执行器的期望位置/姿态映射到关","reasonCn":"在机器人控制领域，逆向动力学（IK）是把末端执行器的期望位置/姿态映射到关节角度的核心问题。传统的解析解往往受限于机械结构，而数值解则需要高效的求解器。pink 项目正是为了解决这一痛点，基于 Pinocchio 高效的动力学库，结合多种 QP（二次规划）求解器（如 OSQP、qpOASES）","stars":845,"starGrowth":0,"createdAt":"2022-02-06T10:05:13Z","silentRounds":2,"language":"Python","topics":["robotics"],"aiDims":["AI应用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["机器人控制","逆向运动学","Pinocchio","QP求解器","URDF"],"domainKey":"机器人控制","pushedAt":"2026-09-09T21:18:48Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/stephane-caron/pink","ts":"2026-09-09T21:18:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"OpenSignLabs/OpenSign","owner":"OpenSignLabs","name":"OpenSign","desc":"🔥 The free &amp; Open Source DocuSign alternative","summaryCn":"免费开源的电子签章神器，替代DocuSign","reasonCn":"这是一个开源的电子签名平台，允许用户在浏览器中创建、编辑和签署文档。它支持多种文件格式，提供类似DocuSign的流程管理功能，如身份验证、签署历史记录和通知机制。项目使用纯JavaScript编写，完全在客户端运行，确保数据隐私。","stars":7021,"starGrowth":3,"createdAt":"2023-10-02T15:53:30Z","silentRounds":0,"language":"JavaScript","topics":["hacktoberfest","javascript","reactjs","mongodb","nodejs","open-source","digital-signature","e-signature","electronic-signature","esign","docusign-alternative","docusign-free","electronic-signatures","esignature","sign-pdf-files","document-signing","legaltech","pdf-sign","pdf-signature","self-hosted"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["电子签名","文档签署","PDF处理","在线办公"],"domainKey":"电子签名","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"reactjs","source":"github","weight":0.7},{"name":"mongodb","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"digital-signature","source":"github","weight":0.7},{"name":"e-signature","source":"github","weight":0.7},{"name":"electronic-signature","source":"github","weight":0.7},{"name":"esign","source":"github","weight":0.7},{"name":"docusign-alternative","source":"github","weight":0.7},{"name":"docusign-free","source":"github","weight":0.7},{"name":"electronic-signatures","source":"github","weight":0.7},{"name":"esignature","source":"github","weight":0.7},{"name":"sign-pdf-files","source":"github","weight":0.7},{"name":"document-signing","source":"github","weight":0.7},{"name":"legaltech","source":"github","weight":0.7},{"name":"pdf-sign","source":"github","weight":0.7},{"name":"pdf-signature","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/OpenSignLabs/OpenSign","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HuLaSpark/HuLa","owner":"HuLaSpark","name":"HuLa","desc":"🍀 A cross-platform instant messaging desktop application with exceptional performance built on Rust + Vue3, compatible with Windows, macOS, Linux, Android, and iOS（一款基于Rust+Vue3极致性能的跨平台即时通讯桌面应用，兼容Windows、MacOS、Linux、Android、IOS）","summaryCn":"一款用Rust和Vue3构建的跨平台即时通讯软件，速度快且功能全。","reasonCn":"HuLa是一款跨平台即时通讯桌面应用，基于Rust和Vue3开发，强调性能与用户体验。支持Windows、macOS、Linux、Android和iOS，提供消息发送、文件传输、群组聊天等核心功能。开源代码允许自定义和扩展，适合需要高效通讯工具的个人或团队。","stars":7731,"starGrowth":5,"createdAt":"2024-01-16T00:24:53Z","silentRounds":0,"language":"Vue","topics":["mobile","chatbot"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["跨平台即时通讯","Rust","端到端加密","隐私保护","桌面应用"],"domainKey":"跨平台即时通讯","pushedAt":"2026-09-11T01:17:37Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"chatbot","source":"github","weight":0.7},{"name":"Vue","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/HuLaSpark/HuLa","ts":"2026-09-11T01:17:37Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.01694024058983314,"copyOk":true},{"repo":"cartesapp/cartes","owner":"cartesapp","name":"cartes","desc":"⚠️ Migré sur https://codeberg.org/cartes/web. L'appli Web de cartes grand public. ","summaryCn":"开源 Web 地图应用，支持离线浏览和自定义图层。","reasonCn":"它的核心功能在于对地图数据的完全掌控。你可以轻松下载离线地图切片，然后通过自定义图层叠加自己的数据（比如 Excel 表格中的坐标点、KML 文件、或者 GeoJSON 格式的地理数据）。这对于需要在野外没有信号环境下工作的地质勘探人员、户外爱好者，或者想要做特定领域地图可视化的","stars":291,"starGrowth":0,"createdAt":"2019-02-11T21:28:07Z","silentRounds":14,"language":"TypeScript","topics":["maps"],"aiDims":["可视化"],"zone":"工具","funScore":0,"domainTags":["地图应用","离线地图","GeoJSON","Leaflet","自定义图层"],"domainKey":"地图应用","pushedAt":"2026-09-09T03:13:37Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"maps","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/cartesapp/cartes","ts":"2026-09-09T03:13:37Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"thinkwee/AgentsMeetRL","owner":"thinkwee","name":"AgentsMeetRL","desc":"Awesome List for Agentic RL","summaryCn":"专为AI Agent研究入门整理的RL与LLM结合资源清单","reasonCn":"这是一份精心整理的Awesome List，聚焦于Agentic Reinforcement Learning这一前沿交叉领域。它并非一个可直接运行的开源软件工具，而是一个资源导航库，汇集了相关的学术论文、基准测试、开源代码实现以及教程资料。对于希望深入了解大语言模型如何与强化学习结合以实现更自主智能体的研究人员和开发者来说，这是一个极具价值的起点。","stars":1850,"starGrowth":0,"createdAt":"2025-06-09T12:42:36Z","silentRounds":2,"language":"HTML","topics":["llm-model"],"aiDims":["AI Agent","学习资源","提示工程"],"zone":"资源","funScore":0,"domainTags":["Agentic RL","大语言模型","强化学习","Awesome List","资源导航"],"domainKey":"Agentic RL","pushedAt":"2026-09-15T11:58:35Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/thinkwee/AgentsMeetRL","ts":"2026-09-15T11:58:35Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NomaDamas/k-skill","owner":"NomaDamas","name":"k-skill","desc":"한국인을 위한 스킬 모음집 - 에이전트를 한국인으로","summaryCn":"对于想要进入韩国市场或与韩国人深度合作的开发者来说，语言障碍是最大的痛","reasonCn":"它的核心价值在于对韩国文化和技术生态的深度理解。无论是韩国的开发者社区、技术论坛还是求职网站，它都能提供针对性的资源。通过AI代理，用户可以实时获取韩国最新的技术趋势、职场资讯以及学习资源，大大降低了跨文化协作的门槛。","stars":7673,"starGrowth":4,"createdAt":"2026-03-24T16:39:30Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["AI Agent","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["韩国语学习","资源集合","AI代理"],"domainKey":"韩国语学习","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/NomaDamas/k-skill","ts":"2026-08-14T01:23:41.092Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HRI-EU/tulip_agent","owner":"HRI-EU","name":"tulip_agent","desc":"autonomous agent with access to a tool library","summaryCn":"欧洲人机交互实验室出品，自带丰富工具库的自主Agent框架","reasonCn":"Tulip Agent是欧洲人机交互研究团队开发的自主Agent框架，核心特色是内置丰富的工具库，让AI能主动调用计算器、浏览器、代码解释器等多种工具完成任务。设计注重人机协作和可解释性，适合研究者和开发者探索Agent的边界能力。","stars":44,"starGrowth":0,"createdAt":"2024-06-19T06:15:49Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["Agent框架","自主智能体","人机交互","工具调用"],"domainKey":"Agent框架","pushedAt":"2026-07-20T14:12:20Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.78,"source":"search","url":"https://github.com/HRI-EU/tulip_agent","ts":"2026-07-20T14:12:20Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"jewbetcha/openflight","owner":"jewbetcha","name":"openflight","desc":"","summaryCn":"让你轻松指挥飞机，像玩模拟飞行一样简单。","reasonCn":"OpenFlight是一个开源的飞行模拟器项目，旨在提供类似真实飞行体验的交互环境。它允许用户通过键盘或手柄控制飞机，体验起飞、巡航和降落的全过程。项目代码结构清晰，支持一定的自定义扩展，适合飞行爱好者和模拟器开发者研究。","stars":1111,"starGrowth":2,"createdAt":"2025-12-08T00:11:12Z","silentRounds":0,"language":"Python","topics":["diy","golf","radar","launchmonitor"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["飞行模拟","Python游戏","交互体验","模拟器"],"domainKey":"飞行模拟","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"diy","source":"github","weight":0.7},{"name":"golf","source":"github","weight":0.7},{"name":"radar","source":"github","weight":0.7},{"name":"launchmonitor","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/jewbetcha/openflight","ts":"2026-08-19T01:08:47.324Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vesoft-inc/nebula","owner":"vesoft-inc","name":"nebula","desc":"  A distributed, fast open-source graph database featuring horizontal scalability and high availability","summaryCn":"高速可扩展的开源图数据库，帮你轻松构建大规模关系数据。","reasonCn":"Nebula 是基于 C++ 实现的分布式原生图数据库，提供水平扩展、强一致性和高可用特性，支持多副本、分区存储以及内置的图查询语言 nGQL，能够在数十亿节点、万亿边的规模上保持毫秒级查询响应，适用于社交网络、推荐系统、知识图谱等复杂关系场景。","stars":12403,"starGrowth":1,"createdAt":"2018-08-28T14:25:09Z","silentRounds":1,"language":"C++","topics":["database"],"aiDims":["知识图谱","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["分布式图数据库","关系数据存储","高可用","图查询语言","Raft共识"],"domainKey":"分布式图数据库","pushedAt":"2026-09-08T06:42:15Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/vesoft-inc/nebula","ts":"2026-09-08T06:42:15Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"lakshaybhushan/vecto3d","owner":"lakshaybhushan","name":"vecto3d","desc":"A super simple tool to convert your SVG's to 3D models.","summaryCn":"做设计的同学肯定经常遇到这种需求：你在 Illustrator 里画了一个漂亮的 2D","reasonCn":"它的工作原理其实挺巧妙的。它不是用 AI 去猜测你的想法，而是基于 SVG 的路径几何结构进行算法转换。它会识别你的线条和形状，通过算法给它们添加厚度、高度和层级关系，最终生成一个 3D 模型。这个过程非常快，基本上是“秒转”。","stars":1483,"starGrowth":1,"createdAt":"2025-03-16T07:00:10Z","silentRounds":3,"language":"TypeScript","topics":["design"],"aiDims":["创意工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["SVG转3D","模型生成","几何算法","设计工具"],"domainKey":"SVG转3D","pushedAt":"2026-08-19T21:46:58Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/lakshaybhushan/vecto3d","ts":"2026-08-19T21:46:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kenn-io/msgvault","owner":"kenn-io","name":"msgvault","desc":"Archive a lifetime of email and chat. Offline search, analytics, and AI query over your full message history. Powered by SQLite and DuckDB","summaryCn":"把邮件聊天记录全归档，AI帮你一秒找到当年那句话","reasonCn":"这个项目就是来解决这个问题的。它用Go语言开发了一个本地归档工具，能把你的邮件和多种聊天平台的消息统一下载到本地，用SQLite做全文索引，用DuckDB做高性能分析查询。最核心的是它的AI查询功能——你可以直接用自然语言问\"上个月我和张三聊了什么项目\"，系统会自动把问题转成SQL去","stars":2070,"starGrowth":2,"createdAt":"2026-01-28T10:49:55Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["AI搜索","非AI-实用","AI应用"],"zone":"工具","funScore":0.6,"domainTags":["消息归档","SQLite","DuckDB","隐私"],"domainKey":"消息归档","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"AI搜索","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.78,"source":"trending","url":"https://github.com/kenn-io/msgvault","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"a9gent/mindfs","owner":"a9gent","name":"mindfs","desc":"Access your personal AI agents and workstation data anywhere, anytime through MindFS.","summaryCn":"让AI助手突破设备限制，随时随地访问你的文件和数据","reasonCn":"这是一个让AI Agent访问个人工作站数据的工具。通过MindFS，你可以把本地文件、代码仓库、开发环境都暴露给AI Agent，让它们在任意设备、任意地点工作。TypeScript编写，可轻松集成到各种Agent框架，实现跨设备数据访问，特别适合多设备协作的开发者，大幅提升工作效率。","stars":1805,"starGrowth":3,"createdAt":"2026-03-31T06:06:42Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI Agent","工具调用"],"zone":"AI","funScore":0,"domainTags":["Agent集成","本地文件访问","数据隐私","TypeScript"],"domainKey":"Agent集成","pushedAt":"2026-08-18T11:09:44Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"bigbro","url":"https://github.com/a9gent/mindfs","ts":"2026-08-18T11:09:44Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"javierbyte/img2css","owner":"javierbyte","name":"img2css","desc":"Convert any image to pure CSS. Recreates images using only box-shadows.","summaryCn":"把任意图片变成纯CSS代码，不用一根图片标签，全靠阴影堆出来","reasonCn":"技术上，它扫描图片的每个像素，把颜色信息映射到box-shadow属性上。一个像素就是一个阴影，成千上万个阴影叠在一起，就还原了原图。虽然生成的CSS代码会比较大，但优势是完全自包含——不需要任何图片服务器，加载速度反而可能更快。","stars":2495,"starGrowth":0,"createdAt":"2015-10-16T08:10:06Z","silentRounds":27,"language":"JavaScript","topics":["design","frontend","pixel-art"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["前端","CSS艺术","像素图","box-shadow"],"domainKey":"前端","pushedAt":"2026-08-23T16:32:40Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"pixel-art","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/javierbyte/img2css","ts":"2026-08-23T16:32:40Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jaywcjlove/linux-command","owner":"jaywcjlove","name":"linux-command","desc":"Linux命令大全搜索工具，内容包含Linux命令手册、详解、学习、搜集。https://git.io/linux","summaryCn":" Linux命令翻烂了不用再查，这个工具一秒帮你找到答案","reasonCn":"这个项目是个Linux命令搜索神器，支持关键词模糊匹配，自带翻译和多种用法示例。核心是一个交互式搜索界面，输入几个字母就能过滤出相关命令，还支持颜色高亮和分类标签。内置命令详解和常见场景代码，适合运维和开发者日常快速查阅。","stars":36939,"starGrowth":0,"createdAt":"2016-11-12T15:44:26Z","silentRounds":1,"language":"Markdown","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","学习资源"],"zone":"资源","funScore":0,"domainTags":["Linux","命令行","文档"],"domainKey":"Linux","pushedAt":"2026-09-06T04:01:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Markdown","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/jaywcjlove/linux-command","ts":"2026-09-06T04:01:05Z","category":"learning","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"opencx-labs/copilot","owner":"opencx-labs","name":"copilot","desc":"","summaryCn":"opencx-labs/copilot 是一个基于 TypeScript 语言开发的开源项目，虽然其在","reasonCn":"由于缺乏公开的介绍文档，项目的具体工作原理和核心卖点目前处于“黑盒”状态。不过，在 copilot 这一命名占据主导的市场环境下，同类项目通常致力于解决代码补全延迟、上下文窗口限制或特定语言支持不足的问题。opencx-labs 这个组织名称可能暗示了某种开源协作或实验室性质的探索，","stars":5109,"starGrowth":0,"createdAt":"2023-08-12T19:41:51Z","silentRounds":7,"language":"TypeScript","topics":["ai-copilot","copilot","llm","sidekick"],"aiDims":["代码助手","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["代码补全","编辑器插件","Copilot","TypeScript"],"domainKey":"代码补全","pushedAt":"2026-09-17T11:19:48.210Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-copilot","source":"github","weight":0.7},{"name":"copilot","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"sidekick","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"search","url":"https://github.com/opencx-labs/copilot","ts":"2026-09-17T11:19:48.210Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"open-metadata/OpenMetadata","owner":"open-metadata","name":"OpenMetadata","desc":"The Open Context Layer for Data and AI , OpenMetadata is the open platform for building trusted data context and business semantics for humans, AI assistants, and agents.","summaryCn":"给数据和AI都配上「说明书」，让模型不再一本正经胡说八道","reasonCn":"OpenMetadata是开源的数据上下文平台，为数据和AI构建可信的语义层。它提供数据目录、血缘追踪、数据质量监控等功能，让AI助手和人类都能理解数据的业务含义。核心能力包括元数据管理、数据血缘可视化、标签体系和数据资产发现。","stars":15303,"starGrowth":7,"createdAt":"2021-08-01T21:17:17Z","silentRounds":0,"language":"TypeScript","topics":["metadata","datadiscovery","dataquality","data-profiling","metadata-management","data-catalog","data-observability","data-discovery","data-contracts","data-governance","data-lineage","data-quality","data-collaboration","mcp","mcp-server","context","context-layer","semantics","ontologies","ontologies-api"],"aiDims":["AI基础设施","RAG与知识","知识图谱"],"zone":"工具","funScore":0,"domainTags":["元数据管理","数据治理","血缘追踪","数据目录","企业平台"],"domainKey":"元数据管理","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"metadata","source":"github","weight":0.7},{"name":"datadiscovery","source":"github","weight":0.7},{"name":"dataquality","source":"github","weight":0.7},{"name":"data-profiling","source":"github","weight":0.7},{"name":"metadata-management","source":"github","weight":0.7},{"name":"data-catalog","source":"github","weight":0.7},{"name":"data-observability","source":"github","weight":0.7},{"name":"data-discovery","source":"github","weight":0.7},{"name":"data-contracts","source":"github","weight":0.7},{"name":"data-governance","source":"github","weight":0.7},{"name":"data-lineage","source":"github","weight":0.7},{"name":"data-quality","source":"github","weight":0.7},{"name":"data-collaboration","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"mcp-server","source":"github","weight":0.7},{"name":"context","source":"github","weight":0.7},{"name":"context-layer","source":"github","weight":0.7},{"name":"semantics","source":"github","weight":0.7},{"name":"ontologies","source":"github","weight":0.7},{"name":"ontologies-api","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/open-metadata/OpenMetadata","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.022372279945108802,"copyOk":false},{"repo":"jtroo/kanata","owner":"jtroo","name":"kanata","desc":"Improve keyboard comfort and usability with advanced customization","summaryCn":"让你的键盘拥有超能力，用一行配置拯救颈椎和打字速度","reasonCn":"Kanata 是一款强大的键盘重映射软件，允许用户通过自定义配置文件重新定义键盘布局、按键功能和宏命令。它解决了机械键盘布局不适配、特殊按键难以操作以及重复性劳动繁琐的问题，支持 Linux、Mac 和 Windows，通过配置文件实现千人千面的键盘体验。","stars":7915,"starGrowth":2,"createdAt":"2022-04-18T04:43:42Z","silentRounds":0,"language":"Rust","topics":["keyboard","rust","cross-platform","linux","windows","interception-driver","keyboard-layout","mouse","mouse-emulation","macos"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["键盘重映射","宏命令","输入法","Rust"],"domainKey":"键盘重映射","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"keyboard","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"cross-platform","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"interception-driver","source":"github","weight":0.7},{"name":"keyboard-layout","source":"github","weight":0.7},{"name":"mouse","source":"github","weight":0.7},{"name":"mouse-emulation","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/jtroo/kanata","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"nCine/nCine","owner":"nCine","name":"nCine","desc":"A cross-platform 2D game engine","summaryCn":"跨平台的2D游戏引擎，适合想做复古像素风的开发者。","reasonCn":"这是一个完全开源的跨平台2D游戏引擎，基于C++编写。它支持Windows、Linux和macOS，专为2D游戏开发设计。引擎提供了物理引擎、音频合成、精灵动画管理等核心模块，以及类似Unity的编辑器界面，降低了游戏开发的门槛。","stars":1333,"starGrowth":1,"createdAt":"2019-05-28T17:31:43Z","silentRounds":12,"language":"C++","topics":["game-development"],"aiDims":["游戏","开发者工具","实用工具"],"zone":"工具","funScore":0,"domainTags":["2D游戏引擎","C++","跨平台","像素风"],"domainKey":"2D游戏引擎","pushedAt":"2026-08-24T13:51:00Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"实用工具","source":"llm","weight":1},{"name":"game-development","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/nCine/nCine","ts":"2026-08-24T13:51:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"babalae/bettergi-scripts-list","owner":"babalae","name":"bettergi-scripts-list","desc":"BetterGI 的脚本仓库，内含BetterGI 的JS脚本、路径追踪、战斗策略、七圣召唤策略。","summaryCn":"原神玩家必备！自动战斗脚本让你躺平挂机刷圣遗物","reasonCn":"BetterGI 是一个基于 JavaScript 的自动化脚本集合，专为《原神》游戏设计，提供战斗策略、七圣召唤策略及路径追踪等功能。该项目通过编写脚本模拟玩家操作，实现自动寻路、战斗技能释放和资源收集，旨在优化游戏体验并提升效率。其开源特性允许用户根据需求定制脚本，社区活跃度高，已获得 515 星标。","stars":558,"starGrowth":0,"createdAt":"2024-10-03T07:38:25Z","silentRounds":2,"language":"JavaScript","topics":[],"aiDims":["游戏","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["游戏自动化","原神","脚本集合","游戏辅助"],"domainKey":"游戏自动化","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/babalae/bettergi-scripts-list","ts":"2026-08-17T00:55:40.962Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"electerm/electerm","owner":"electerm","name":"electerm","desc":"📻Terminal/ssh/sftp/ftp/telnet/serialport/RDP/VNC/Spice client(Linux, Mac, Windows, Android, HarmonyOS)","summaryCn":"一个跨平台全能终端，连接服务器就像点外卖一样简单","reasonCn":"Electerm 是一款开源的跨平台终端模拟器和远程桌面客户端，基于 Electron 和 Node.js 构建。它完美兼容 Linux、macOS、Windows、Android 和 HarmonyOS，支持 SSH、SFTP、FTP、Telnet、串口、RDP 和 VNC 协议。项目代码量庞大，功能全面，不仅界面美观，还集成了文件管理、剪贴板同步、自定义快捷键和主题等功能，旨在替代传统命令行工具，提供一站式远程管理体验。","stars":15176,"starGrowth":1,"createdAt":"2017-10-07T09:54:11Z","silentRounds":1,"language":"JavaScript","topics":["electron","electerm","ssh","sftp","terminal","linux-app","macos-app","open-source","telnet","serialport","ftp","rdp","vnc","ai","mcp","zmodem","spice","android","harmonyos","windows-app"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["终端模拟器","SSH客户端","远程连接","文件传输"],"domainKey":"终端模拟器","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"electron","source":"github","weight":0.7},{"name":"electerm","source":"github","weight":0.7},{"name":"ssh","source":"github","weight":0.7},{"name":"sftp","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"linux-app","source":"github","weight":0.7},{"name":"macos-app","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"telnet","source":"github","weight":0.7},{"name":"serialport","source":"github","weight":0.7},{"name":"ftp","source":"github","weight":0.7},{"name":"rdp","source":"github","weight":0.7},{"name":"vnc","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"mcp","source":"github","weight":0.7},{"name":"zmodem","source":"github","weight":0.7},{"name":"spice","source":"github","weight":0.7},{"name":"android","source":"github","weight":0.7},{"name":"harmonyos","source":"github","weight":0.7},{"name":"windows-app","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/electerm/electerm","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"microsoft/vscode-copilot-chat","owner":"microsoft","name":"vscode-copilot-chat","desc":"","summaryCn":"该项目是 GitHub Copilot 聊天功能在 VS Code 中的核心实现，其最关键的状态","reasonCn":"从技术角度看，该项目展示了如何用 TypeScript 构建复杂的 AI 交互界面与后端逻辑。它支持三种主要模式：行内建议提供快速补全，行内聊天针对选中代码进行精准修改，而智能体则是处理大型任务的“自主工”。这种分层设计解决了不同复杂度编程需求，从简单的变量命名到重构整个函数模","stars":9967,"starGrowth":1,"createdAt":"2025-06-10T16:21:19Z","silentRounds":10,"language":"TypeScript","topics":[],"aiDims":["代码助手","AI应用"],"zone":"AI","funScore":0.2,"domainTags":["代码补全","VS Code","智能编程","AI聊天"],"domainKey":"代码补全","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/microsoft/vscode-copilot-chat","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"nexmoe/VidBee","owner":"nexmoe","name":"VidBee","desc":"Download videos from almost any website worldwide","summaryCn":"你是不是经常遇到这种情况：在网上看到一个超精彩的视频，想下载到本地慢慢","reasonCn":"技术实现上，VidBee 利用了现代浏览器的解析技术和开源库，能够绕过很多网站的下载限制。它不仅仅是简单的下载文件，还内置了视频格式转换功能，下载后可以直接变成 MP4 等通用格式，省去了后续处理的麻烦。它的界面采用了现代化的设计风格，操作逻辑非常直观，对于新手用户来说，","stars":10689,"starGrowth":2,"createdAt":"2025-10-22T13:43:42Z","silentRounds":0,"language":"TypeScript","topics":["downloader","facebook","tiktok","twitter","youtube"],"aiDims":["非AI-好玩","实用工具"],"zone":"工具","funScore":0,"domainTags":["视频下载","媒体工具","资源管理"],"domainKey":"视频下载","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"实用工具","source":"llm","weight":1},{"name":"downloader","source":"github","weight":0.7},{"name":"facebook","source":"github","weight":0.7},{"name":"tiktok","source":"github","weight":0.7},{"name":"twitter","source":"github","weight":0.7},{"name":"youtube","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/nexmoe/VidBee","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TanStack/router","owner":"TanStack","name":"router","desc":"🤖 A client-first, server-capable, fully type-safe router and full-stack framework for the web (React and more).","summaryCn":"React全栈框架，类型安全的路由系统，让前端开发像写后端一样丝滑。","reasonCn":"这是一款专为React生态设计的全栈路由框架，它不仅提供了类型安全的路由配置，还集成了服务端渲染和客户端路由的能力。开发者可以通过声明式的方式定义路由和数据处理逻辑，无需编写重复的样板代码，从而提升开发效率和代码可维护性。","stars":15124,"starGrowth":1,"createdAt":"2019-01-14T13:54:39Z","silentRounds":1,"language":"TypeScript","topics":["react","router","routing","javascript","searchparams","search","url","route","typesafe","typescript","state-management","fullstack","rpc","server-functions","ssr","framework"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["React路由","全栈框架","TypeScript","服务端渲染"],"domainKey":"React路由","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"react","source":"github","weight":0.7},{"name":"router","source":"github","weight":0.7},{"name":"routing","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"searchparams","source":"github","weight":0.7},{"name":"search","source":"github","weight":0.7},{"name":"url","source":"github","weight":0.7},{"name":"route","source":"github","weight":0.7},{"name":"typesafe","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"state-management","source":"github","weight":0.7},{"name":"fullstack","source":"github","weight":0.7},{"name":"rpc","source":"github","weight":0.7},{"name":"server-functions","source":"github","weight":0.7},{"name":"ssr","source":"github","weight":0.7},{"name":"framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/TanStack/router","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"kennethleungty/Finance-LLMs","owner":"kennethleungty","name":"Finance-LLMs","desc":"Comprehensive Compilation of Real-World LLM & AI Agent Use Cases in Financial Services","summaryCn":"把金融行业大模型真实落地案例一本全收罗，看完就知道钱是怎么被AI赚的","reasonCn":"该项目是一份全面的金融行业LLM与AI Agent实际应用案例汇编，覆盖了风控、投研、客服、合规等多个业务场景。作者整理了真实世界中银行、券商、保险公司如何部署大模型和智能体的完整链路，包含技术架构、落地难点和效果数据。适合想了解AI在金融垂直领域怎么落地的开发者和产品经理。","stars":140,"starGrowth":0,"createdAt":"2023-09-25T02:56:08Z","silentRounds":27,"language":"","topics":["llm-model"],"aiDims":["AI应用","AI Agent","工具调用"],"zone":"资源","funScore":0,"domainTags":["金融科技","LLM案例","行业研究"],"domainKey":"金融科技","pushedAt":"2026-08-25T16:24:07Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/kennethleungty/Finance-LLMs","ts":"2026-08-25T16:24:07Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Marksonthegamer/Mate-Engine-Linux-Port","owner":"Marksonthegamer","name":"Mate-Engine-Linux-Port","desc":"VRM desktop pet software for Linux (aka Unofficial Linux port of shinyflvre's MateEngine)","summaryCn":"简单来说，这个项目就像是在你的Linux电脑桌面上养了一只活泼可爱的动漫宠物","reasonCn":"技术上，它是作为知名项目MateEngine的非官方Linux移植版出现的。它大量利用了ShaderLab进行图形渲染，确保在各种Linux桌面环境下都能流畅呈现3D模型。通过解析VRM格式的角色文件，它支持用户自由更换喜爱的动漫形象，并且对系统资源的占用控制得相当不错，不会影响正常办公和开发","stars":310,"starGrowth":0,"createdAt":"2025-10-12T04:32:47Z","silentRounds":20,"language":"ShaderLab","topics":["anime"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0.7,"domainTags":["Linux桌面","3D桌面宠物","VRM","二次元"],"domainKey":"Linux桌面","pushedAt":"2026-08-15T12:58:17Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"anime","source":"github","weight":0.7},{"name":"ShaderLab","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/Marksonthegamer/Mate-Engine-Linux-Port","ts":"2026-08-15T12:58:17Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"theashgen/Airopath","owner":"theashgen","name":"Airopath","desc":" AiroPath calculates multiple routes between two points and ranks them by real-time AQI (Air Quality Index) — so you can choose the cleanest path, not just the fastest one. ","summaryCn":"选最干净的路线而不是最快的，雾霾天走路不呛鼻","reasonCn":"AiroPath是一个基于TypeScript开发的空气质量导航工具。它通过计算两点之间的多条路径，并依据实时AQI指数进行排序，帮助用户避开高污染区域。项目专为移动端或Web端设计，提供直观的路线选择，让通勤和户外运动更加健康。","stars":10,"starGrowth":0,"createdAt":"2026-03-09T08:03:16Z","silentRounds":13,"language":"TypeScript","topics":["fitness"],"aiDims":["非AI-实用","非AI-好玩","可视化"],"zone":"工具","funScore":0.6,"domainTags":["空气质量","路径规划","地理信息","健康"],"domainKey":"空气质量","pushedAt":"2026-09-11T12:50:18Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"fitness","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/theashgen/Airopath","ts":"2026-09-11T12:50:18Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"SLASHEM-Extended/SLASHEM-Extended","owner":"SLASHEM-Extended","name":"SLASHEM-Extended","desc":"SLASH'EM Extended (a SLASH'EM fork)","summaryCn":"简单来说，这是一款如果你不看攻略，可能连第一层楼都走不出的硬核地牢游戏","reasonCn":"技术上，它是基于 C 语言编写的，代码库规模庞大且结构极其复杂。它在继承 NetHack 复杂交互系统的基础上，进行了丧心病狂的扩展。不同于一般的游戏通过图形渲染，它利用字符界面实现了一种极其高效且信息密度极高的交互方式。","stars":147,"starGrowth":0,"createdAt":"2015-02-12T18:55:07Z","silentRounds":13,"language":"C","topics":["roguelike"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0.7,"domainTags":["Roguelike","NetHack","地牢","硬核游戏"],"domainKey":"Roguelike","pushedAt":"2026-09-11T11:57:42Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"roguelike","source":"github","weight":0.7},{"name":"C","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/SLASHEM-Extended/SLASHEM-Extended","ts":"2026-09-11T11:57:42Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"engineering-management/awesome-engineering-management","owner":"engineering-management","name":"awesome-engineering-management","desc":"Pointers and tools for learning and day-to-day practice of engineering management & leadership.","summaryCn":"工程管理大佬们整理的宝藏工具清单，职场进阶必备。","reasonCn":"这是全球最火的工程管理资源合集，汇聚了各种管理方法论、工具链和实战指南。它涵盖了团队协作、技术规划、绩效管理等核心场景，能帮助管理者系统化提升领导力。内容更新及时，社区活跃度高，是入门和进阶必读清单。","stars":2768,"starGrowth":0,"createdAt":"2021-06-11T13:37:42Z","silentRounds":13,"language":"","topics":["productivity"],"aiDims":["学习资源","效率工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["工程管理","技术管理","知识库","清单"],"domainKey":"工程管理","pushedAt":"2026-09-11T10:12:41Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"productivity","source":"github","weight":0.7}],"aiScore":0.6,"source":"search","url":"https://github.com/engineering-management/awesome-engineering-management","ts":"2026-09-11T10:12:41Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"genlayerlabs/genlayer-project-boilerplate","owner":"genlayerlabs","name":"genlayer-project-boilerplate","desc":"","summaryCn":"帮你秒搭AI应用的瑞士军刀，开源项目模板","reasonCn":"这是一个为AI应用快速开发设计的TypeScript项目脚手架，内置了现代化的开发工具链和最佳实践。它提供了一键生成的项目模板，支持多种主流AI框架和Agent框架集成，旨在解决从零搭建项目时遇到的繁琐配置问题。开发者无需重复造轮子，直接基于此模板开发，能显著缩短从构思到上线的周期，特别适合需要快速迭代和部署AI产品的团队。","stars":17836,"starGrowth":11,"createdAt":"2024-06-28T06:59:47Z","silentRounds":0,"language":"TypeScript","topics":[],"aiDims":["AI基础设施","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["AI应用","项目脚手架","TypeScript","Next.js","开发模板"],"domainKey":"AI应用","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/genlayerlabs/genlayer-project-boilerplate","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.034850190625700835,"copyOk":false},{"repo":"lusingander/serie","owner":"lusingander","name":"serie","desc":"A rich git commit graph in your terminal, like magic 📚","summaryCn":"把命令行里的Git提交记录变成一场炫酷的时间旅行视觉盛宴","reasonCn":"技术层面，它用Rust编写，性能极高，即使在拥有数万条提交的大型项目中也能快速渲染。它的独特之处在于对图形化的极致追求，通过颜色区分不同分支，用图形符号标记合并点，让复杂的代码回溯过程一目了然。相比传统的`git log --graph`，它的视觉效果更现代，阅读体验更流畅，像是给","stars":2113,"starGrowth":0,"createdAt":"2024-07-28T07:22:08Z","silentRounds":1,"language":"Rust","topics":["git"],"aiDims":["开发者工具","非AI-好玩","可视化"],"zone":"工具","funScore":0.6,"domainTags":["Git可视化","终端工具","代码历史","Rust","命令行"],"domainKey":"Git可视化","pushedAt":"2026-09-08T00:34:47Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"git","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/lusingander/serie","ts":"2026-09-08T00:34:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"oseghalep/cloud-cost-optimization-hub","owner":"oseghalep","name":"cloud-cost-optimization-hub","desc":"Cloud Cost Optimization Hub is an open-source, self-hosted platform that provides unified cost visibility and actionable recommendations across AWS, GCP, and Azure. Features include multi-cloud cost aggregation, rightsizing recommendations, idle resource detection, cost anomaly alerts, and an intuitive dashboard – all designed to help teams reduce.","summaryCn":"一键查AWS/Azure/GCP账单，省钱还能发工资","reasonCn":"这是一个自托管的多云成本管理平台，支持AWS、GCP和Azure的统一接入。它能自动扫描资源，识别闲置实例和未使用的存储，生成可执行的降本报告。通过可视化仪表盘，管理员可实时监控预算，设置告警阈值，确保云支出可控。","stars":261,"starGrowth":0,"createdAt":"2026-04-14T15:57:04Z","silentRounds":11,"language":"Go","topics":["cloud","cloud-computing","cloud-controller","cloud-infrastructure","cloud-native","cloudcost","cloudcostefficiency","cloudcostmanagement","cloudcostoptimisation","cloudformation"],"aiDims":["非AI-实用","效率工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["多云成本管理","云资源监控","DevOps","预算控制"],"domainKey":"多云成本管理","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"cloud","source":"github","weight":0.7},{"name":"cloud-computing","source":"github","weight":0.7},{"name":"cloud-controller","source":"github","weight":0.7},{"name":"cloud-infrastructure","source":"github","weight":0.7},{"name":"cloud-native","source":"github","weight":0.7},{"name":"cloudcost","source":"github","weight":0.7},{"name":"cloudcostefficiency","source":"github","weight":0.7},{"name":"cloudcostmanagement","source":"github","weight":0.7},{"name":"cloudcostoptimisation","source":"github","weight":0.7},{"name":"cloudformation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/oseghalep/cloud-cost-optimization-hub","ts":"2026-08-15T00:57:24.351Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zynaddsubfx/zynaddsubfx","owner":"zynaddsubfx","name":"zynaddsubfx","desc":"ZynAddSubFX open source synthesizer","summaryCn":"能让你用鼠标像搭积木一样合成复古合成器音效的免费软件。","reasonCn":"它的界面看起来像是一堆旋钮和滑块，但实际上是一个强大的数学模型。你不需要懂复杂的信号处理知识，只需要用鼠标拖动波形、调整包络线、修改滤波器的截止频率，就能听到声音的实时变化。它内置了超过150种振荡器，包括模拟波形、锯齿波、方波，甚至还有一些特殊的物理模型波形。","stars":1179,"starGrowth":1,"createdAt":"2014-07-05T18:33:25Z","silentRounds":13,"language":"C++","topics":["synthesizer"],"aiDims":["游戏","创意工具"],"zone":"创意","funScore":0,"domainTags":["音频合成","音乐创作","振荡器","滤波器","波形合成"],"domainKey":"音频合成","pushedAt":"2026-09-10T23:45:36Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"synthesizer","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/zynaddsubfx/zynaddsubfx","ts":"2026-09-10T23:45:36Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ScottPlot/ScottPlot","owner":"ScottPlot","name":"ScottPlot","desc":"Interactive plotting library for .NET","summaryCn":"让程序员在C#里轻松画出各种超酷图表的神器","reasonCn":"ScottPlot是一个专为.NET平台设计的交互式绘图库，支持动态动画和实时数据流。它通过灵活的API实现数据可视化，支持折线图、散点图等多种图表类型，兼容Windows、Linux和macOS。作为开源项目，它提供了丰富的自定义选项，能帮助开发者在桌面应用中快速集成专业级图表功能。","stars":6757,"starGrowth":1,"createdAt":"2018-01-04T03:34:45Z","silentRounds":1,"language":"C#","topics":["data-visualization"],"aiDims":["可视化","开发工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["数据可视化","C#",".NET","图表库"],"domainKey":"数据可视化","pushedAt":"2026-08-15T21:45:11Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/ScottPlot/ScottPlot","ts":"2026-08-15T21:45:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"eza-community/eza","owner":"eza-community","name":"eza","desc":"A modern alternative to ls","summaryCn":"比ls更好看的命令行目录工具，Rust写的现代版ls","reasonCn":"作为ls的现代化替代品，eza保留了同样的命令行接口，但大幅增加功能：自动彩色分类显示文件类型，集成Git显示文件状态，支持树状目录视图，还有各种图标让文件一目了然。它是用Rust写的，性能几乎和原生ls一样快。 安装非常简单：Mac用brew install eza，Linux用包管理器，Windows","stars":23346,"starGrowth":5,"createdAt":"2023-07-28T08:35:55Z","silentRounds":0,"language":"Rust","topics":["command-line"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["命令行工具","Rust","文件列表"],"domainKey":"命令行工具","pushedAt":"2026-08-06T19:53:22Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/eza-community/eza","ts":"2026-08-06T19:53:22Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.015642508352608584,"copyOk":false},{"repo":"palemoky/chinese-poetry-api","owner":"palemoky","name":"chinese-poetry-api","desc":"📜 诗泉：高性能中国古诗词 API 服务","summaryCn":"给古诗迷和开发者准备的宝藏API，随查随用。","reasonCn":"这是一个高性能的中国古诗词API服务，基于Go语言开发。它提供了海量的中国古诗词数据，包括唐诗宋词等，支持按作者、朝代、关键词等多种方式进行检索，能够快速返回诗词全文、注释和赏析。项目采用无状态设计，性能优异，非常适合需要集成诗词功能的Web应用或工具开发。","stars":2866,"starGrowth":1,"createdAt":"2025-12-08T23:39:04Z","silentRounds":1,"language":"Go","topics":["api","chinese","graphql","poetry"],"aiDims":["非AI-实用","学习资源","开发工具"],"zone":"资源","funScore":0,"domainTags":["古诗词","API服务","Go语言"],"domainKey":"古诗词","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"api","source":"github","weight":0.7},{"name":"chinese","source":"github","weight":0.7},{"name":"graphql","source":"github","weight":0.7},{"name":"poetry","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/palemoky/chinese-poetry-api","ts":"2026-08-15T00:57:24.351Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"sopaco/deepwiki-rs","owner":"sopaco","name":"deepwiki-rs","desc":"Turn code into clarity. Generate accurate technical docs and AI-ready context in minutes—perfectly structured for human teams and intelligent agents.","summaryCn":"把代码自动变文档，AI和人类都能看懂的神器","reasonCn":"基于Rust开发的高性能代码文档生成工具，能自动解析代码结构并生成技术文档和AI可读的上下文。支持多种编程语言，输出结构化的Markdown格式文档，帮助团队维护和对接AI Agent。项目编译速度快，内存占用低，适合大规模代码库使用。","stars":3083,"starGrowth":4,"createdAt":"2025-09-05T16:38:32Z","silentRounds":0,"language":"Rust","topics":["deepwiki","deepseek","dsh-plugin","okf-format","openwiki","grok"],"aiDims":["代码助手","开发者工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["代码文档","AST解析","Markdown","AI上下文","Rust工具"],"domainKey":"代码文档","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"deepwiki","source":"github","weight":0.7},{"name":"deepseek","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"okf-format","source":"github","weight":0.7},{"name":"openwiki","source":"github","weight":0.7},{"name":"grok","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/sopaco/deepwiki-rs","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jdx/mise","owner":"jdx","name":"mise","desc":"dev tools, env vars, task runner","summaryCn":"一个工具统一管所有开发环境，版本切换环境变量一键搞定","reasonCn":"mise 是一个跨平台开发工具管理器，支持 Node.js、Python、Go、Rust、PHP 等超百种工具的统一版本管理。进入项目目录自动切换环境，环境变量按目录自动加载，任务运行器支持自定义脚本。用 Rust 编写，安装速度快，配置简洁，解决了多语言环境切换的痛点。","stars":34214,"starGrowth":20,"createdAt":"2023-01-09T14:44:21Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["开发工具","环境管理","版本控制","Rust","DevOps"],"domainKey":"开发工具","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/jdx/mise","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.06175366808908633,"copyOk":false},{"repo":"adnanh/webhook","owner":"adnanh","name":"webhook","desc":"webhook is a lightweight incoming webhook server to run shell commands","summaryCn":"GitHub收到新提交就自动执行命令，自动化办公神器","reasonCn":"webhook是一个轻量级的Webhook服务器，用Go编写，支持接收HTTP回调请求并执行预设的Shell命令或脚本。它能与GitHub、GitLab等代码托管平台无缝集成，实现代码推送后自动触发构建、部署或测试流程。只需简单的JSON配置文件即可定义触发规则，无需编写复杂代码。项目star超过12000，是DevOps自动化领域最受欢迎的工具之一，适合追求高效工作流的开发者和运维人员。","stars":12144,"starGrowth":0,"createdAt":"2015-01-12T20:59:11Z","silentRounds":1,"language":"Go","topics":["devops","automation"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Webhook","自动化","DevOps"],"domainKey":"Webhook","pushedAt":"2026-09-04T18:02:57Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"automation","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/adnanh/webhook","ts":"2026-09-04T18:02:57Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chartdb/chartdb","owner":"chartdb","name":"chartdb","desc":"Database diagrams editor that allows you to visualize and design your DB with a single query.","summaryCn":"输入一条SQL就能自动生成精美的数据库关系图，写文档终于不用画图了","reasonCn":"这是一个基于 Web 的数据库可视化工具，支持 MySQL、PostgreSQL 等多种主流数据库。最厉害的是，你只需要输入一个 SQL 查询，它就能自动解析并生成对应的 ER 图或实体关系图。界面简洁，交互流畅，完全不需要手动拖拽连线。","stars":22947,"starGrowth":4,"createdAt":"2024-08-05T07:19:19Z","silentRounds":0,"language":"TypeScript","topics":["database","editor"],"aiDims":["可视化","非AI-实用","代码助手"],"zone":"工具","funScore":0.6,"domainTags":["数据库可视化","ER图","SQL","文档生成"],"domainKey":"数据库可视化","pushedAt":"2026-09-13T08:43:19Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"editor","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/chartdb/chartdb","ts":"2026-09-13T08:43:19Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"confeden/Antigravity","owner":"confeden","name":"Antigravity","desc":"Antigravity в России без VPN и смены региона аккаунта Google","summaryCn":"不用换区也能用谷歌服务，俄罗斯网友的“魔法”小工具","reasonCn":"这是一个基于Rust编写的工具，旨在帮助俄罗斯用户绕过复杂的网络限制，直接访问谷歌服务，无需繁琐的VPN或账号区域切换操作。该项目通过巧妙的技术手段解决了本地网络环境下的访问障碍，让谷歌地图、Gmail等服务在特定区域无缝可用。","stars":595,"starGrowth":5,"createdAt":"2026-02-08T19:51:27Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["非AI-实用","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["网络访问","Rust","谷歌服务","网络代理","本地工具"],"domainKey":"网络访问","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/confeden/Antigravity","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.04825210084033613,"copyOk":false},{"repo":"putyy/res-downloader","owner":"putyy","name":"res-downloader","desc":"视频号、小程序、抖音、快手、小红书、直播流、m3u8、酷狗、QQ音乐等常见网络资源下载!","summaryCn":"抖音快手小红书资源一键下载，免费资源统统带走","reasonCn":"这是一个专门下载各种网络媒体资源的工具，支持视频号、抖音、快手、小红书等主流平台的视频下载，还涵盖直播流、m3u8地址以及酷狗、QQ音乐等音频资源。采用Go语言编写，支持Windows、Mac、Linux多平台，界面简洁，功能强大，帮你轻松保存想看的短视频和想听的歌。","stars":19982,"starGrowth":4,"createdAt":"2023-09-01T08:03:47Z","silentRounds":0,"language":"Go","topics":["wechat-video","wechat","douyin","kuaishou","res-downloader","xiaohongshu"],"aiDims":["非AI-实用"],"zone":"工具","funScore":0,"domainTags":["视频下载","媒体资源","网络爬虫"],"domainKey":"视频下载","pushedAt":"2026-08-18T00:56:27.274Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"wechat-video","source":"github","weight":0.7},{"name":"wechat","source":"github","weight":0.7},{"name":"douyin","source":"github","weight":0.7},{"name":"kuaishou","source":"github","weight":0.7},{"name":"res-downloader","source":"github","weight":0.7},{"name":"xiaohongshu","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/putyy/res-downloader","ts":"2026-08-18T00:56:27.274Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"UseInterstellar/Interstellar","owner":"UseInterstellar","name":"Interstellar","desc":"One of the most popular modern web proxies with blazing fast speeds and a variety of games.","summaryCn":"极速网页代理+内置小游戏，翻墙又能打发时间","reasonCn":"Interstellar 是一款基于 Node.js 的现代 Web 代理，拥有极致的转发速度和自研的流量压缩算法，同时内置多款轻量小游戏，用户在科学上网的同时还能随手消遣，支持插件式扩展，可自定义规则、TLS 伪装和多协议桥接，兼容主流浏览器和系统。","stars":2147,"starGrowth":0,"createdAt":"2023-01-05T21:57:19Z","silentRounds":1,"language":"JavaScript","topics":["games","unblocked-games","discord","proxy","proxies","interstellar","unblocker","unblockers","web-proxy","youtube","interstellarnetwork","interstellarproxy","school","unblocked","interstellarunblocker"],"aiDims":["非AI-好玩","非AI-实用","游戏"],"zone":"工具","funScore":0,"domainTags":["网络代理","游戏","流量压缩","科学上网","Web代理"],"domainKey":"网络代理","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"games","source":"github","weight":0.7},{"name":"unblocked-games","source":"github","weight":0.7},{"name":"discord","source":"github","weight":0.7},{"name":"proxy","source":"github","weight":0.7},{"name":"proxies","source":"github","weight":0.7},{"name":"interstellar","source":"github","weight":0.7},{"name":"unblocker","source":"github","weight":0.7},{"name":"unblockers","source":"github","weight":0.7},{"name":"web-proxy","source":"github","weight":0.7},{"name":"youtube","source":"github","weight":0.7},{"name":"interstellarnetwork","source":"github","weight":0.7},{"name":"interstellarproxy","source":"github","weight":0.7},{"name":"school","source":"github","weight":0.7},{"name":"unblocked","source":"github","weight":0.7},{"name":"interstellarunblocker","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/UseInterstellar/Interstellar","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"evoleinik/unchatty","owner":"evoleinik","name":"unchatty","desc":"Make Claude Code write plainly. A CLAUDE.md block plus the A/B harness that proves it works.","summaryCn":"很多开发者在使用Claude编程时，都会遇到一个烦恼：AI总喜欢在给出代码前后","reasonCn":"技术上，它的思路非常硬核。它并不是简单地在Prompt里写“不准多说话”，而是利用CLAUDE.md的上下文控制能力，结合专门设计的A/B测试框架，量化并验证AI的输出行为。这种做法将AI的风格管理从玄学变成了一种可验证的工程实践，通过对比测试，让用户能直观感受到AI在不同配置下的响应","stars":17,"starGrowth":0,"createdAt":"2026-08-14T05:10:24Z","silentRounds":18,"language":"Shell","topics":["writing"],"aiDims":["AI写作","代码助手"],"zone":"工具","funScore":0,"domainTags":["提示工程","AI助手","开发工具"],"domainKey":"提示工程","pushedAt":"2026-08-14T05:10:28Z","tags":[{"name":"AI写作","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"writing","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/evoleinik/unchatty","ts":"2026-08-14T05:10:28Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Baizhige/EEGUnity","owner":"Baizhige","name":"EEGUnity","desc":"An open source tool for large-scale EEG datasets processing","summaryCn":"开源EEG数据处理神器，从原始信号到特征提取一步到位。","reasonCn":"EEGUnity是一个专为大规模脑电数据设计的开源处理工具，提供从原始数据加载到特征提取的完整流水线。它支持多种EEG设备格式，内置滤波、去伪迹、特征提取等算法，能够快速处理海量数据集。项目基于Python开发，兼容主流EEG采集系统，为脑机接口研究和神经科学分析提供了高效的数据处理解决方案。","stars":78,"starGrowth":0,"createdAt":"2024-08-14T13:27:19Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["AI基础设施","非AI-实用","学习资源"],"zone":"工具","funScore":0,"domainTags":["脑电数据","特征提取","信号处理","科研工具"],"domainKey":"脑电数据","pushedAt":"2026-08-19T06:09:23Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/Baizhige/EEGUnity","ts":"2026-08-19T06:09:23Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"anthropics/cwc-workshops","owner":"anthropics","name":"cwc-workshops","desc":"","summaryCn":"Anthropic官方出品，教你如何高效搞定大模型实战","reasonCn":"这是一套由Anthropic官方精心打磨的培训课程体系，旨在帮助开发者和产品经理从零开始掌握Claude等大模型的应用开发。课程内容涵盖了从API调用、提示工程到构建复杂AI应用的全流程，包含大量实战案例和代码示例，旨在降低AI落地的技术门槛。","stars":2124,"starGrowth":0,"createdAt":"2026-05-06T03:53:01Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["教程","AI应用","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["教程","大模型","提示工程"],"domainKey":"教程","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/anthropics/cwc-workshops","ts":"2026-08-12T01:22:42.342Z","category":"learning","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":true},{"repo":"billmei/every-chatgpt-gui","owner":"billmei","name":"every-chatgpt-gui","desc":"","summaryCn":"Every ChatGPT GUI 是一个动态维护的开源软件导航仓库，旨在帮助用户发现那","reasonCn":"该列表的分类逻辑清晰，涵盖了Web端、移动应用以及桌面软件等多个维度，并且紧跟技术趋势，同步收录支持最新LLM模型的客户端。对于开发者而言，这里不仅是工具库，更是学习如何封装AI API的丰富案例集；许多列表中的项目代码结构简单、易于二次开发，是快速搭建个人AI助手的良好起","stars":4030,"starGrowth":0,"createdAt":"2023-03-16T17:12:34Z","silentRounds":5,"language":"","topics":[],"aiDims":["学习资源","AI应用"],"zone":"资源","funScore":0.2,"domainTags":["ChatGPT客户端","GUI导航","大语言模型接口","开源应用列表"],"domainKey":"ChatGPT客户端","pushedAt":"2026-09-18T16:08:19.621Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1}],"aiScore":0.44,"source":"search","url":"https://github.com/billmei/every-chatgpt-gui","ts":"2026-09-18T16:08:19.621Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"harmonoid/harmonoid","owner":"harmonoid","name":"harmonoid","desc":"🎵 Plays & manages your music library. Looks beautiful & juicy. Available for Windows, GNU/Linux, macOS, Android & iOS.","summaryCn":"颜值超高的跨平台音乐播放器，本地曲库管理神器","reasonCn":"Harmonoid是一款界面精美的音乐播放器，支持Windows、macOS、Linux、Android和iOS全平台。采用Dart语言开发，基于Flutter框架，界面设计现代感十足。支持本地音乐文件管理，可以创建播放列表、显示歌词、调整均衡器。跨平台同步让你在不同设备上无缝切换音乐体验。","stars":4753,"starGrowth":2,"createdAt":"2020-08-04T10:28:45Z","silentRounds":0,"language":"Dart","topics":["desktop","music"],"aiDims":["非AI-好玩","创意工具","可视化"],"zone":"创意","funScore":0.7,"domainTags":["音乐播放器","跨平台","本地曲库"],"domainKey":"音乐播放器","pushedAt":"2026-09-02T19:44:49Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"desktop","source":"github","weight":0.7},{"name":"music","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/harmonoid/harmonoid","ts":"2026-09-02T19:44:49Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"SuperBruceJia/Awesome-Mixture-of-Experts","owner":"SuperBruceJia","name":"Awesome-Mixture-of-Experts","desc":"Awesome Mixture of Experts (MoE): A Curated List of Mixture of Experts (MoE) and Mixture of Multimodal Experts (MoME)","summaryCn":"MoE模型全家桶导航，从原理到论文再到代码，一站式掌握专家混合架构。","reasonCn":"这是一个精心整理的Mixture of Experts (MoE)及多模态专家混合(MoME)资源合集。涵盖了LLM领域的MoE架构论文、实现代码及最佳实践，是深入研究高效大模型架构的宝藏库，适合想了解MoE如何降低推理成本、提升训练效率的技术同学。","stars":75,"starGrowth":0,"createdAt":"2024-08-15T18:24:18Z","silentRounds":28,"language":"","topics":["llm-model"],"aiDims":["大语言模型","模型与训练","教程"],"zone":"资源","funScore":0,"domainTags":["Mixture of Experts","大模型架构","论文合集","稀疏激活","模型优化"],"domainKey":"Mixture of Experts","pushedAt":"2026-07-13T23:54:25Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.75,"source":"search","url":"https://github.com/SuperBruceJia/Awesome-Mixture-of-Experts","ts":"2026-07-13T23:54:25Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"leggedrobotics/ocs2","owner":"leggedrobotics","name":"ocs2","desc":"Optimal Control for Switched Systems","summaryCn":"提供高效最优控制库，帮助机器人和动态系统实现切换模式下的精准轨迹规划。","reasonCn":"ocs2 是面向切换系统的最优控制框架，使用 C++ 实现连续与离散时间模型的协同求解，支持非线性 MPC、时变约束和多段任务切换，提供可视化调试工具和 ROS 接口，适用于腿式机器人、无人机以及任何需要高性能轨迹生成的系统。","stars":1518,"starGrowth":0,"createdAt":"2021-07-08T07:09:10Z","silentRounds":9,"language":"C++","topics":["robotics"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["最优控制","模型预测控制","机器人轨迹规划","切换系统","ROS"],"domainKey":"最优控制","pushedAt":"2026-09-07T21:33:16Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"robotics","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.64,"source":"search","url":"https://github.com/leggedrobotics/ocs2","ts":"2026-09-07T21:33:16Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"harveyai/harvey-labs","owner":"harveyai","name":"harvey-labs","desc":"A benchmark built to evaluate and improve agent capabilities for supporting legal work.","summaryCn":"给AI律师助手出考卷，测测它法律本事有多强","reasonCn":"由Harvey AI团队开发的法律领域Agent benchmark，包含真实法律场景的测试用例，涵盖合同审查、法律研究、合规分析等任务。支持多种主流Agent框架接入，自动评分输出能力差距，帮助开发者量化优化效果，填补了法律垂直领域Agent评估工具的空白。","stars":1378,"starGrowth":4,"createdAt":"2026-03-30T18:20:43Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","AI应用","工具调用"],"zone":"工具","funScore":0.5,"domainTags":["法律AI","Agent评测","基准测试","LLM"],"domainKey":"法律AI","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/harveyai/harvey-labs","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"futrx-com/remote.futrx","owner":"futrx-com","name":"remote.futrx","desc":"","summaryCn":"用Go语言写的远程桌面工具，安全又好用。","reasonCn":"这是一个基于Go语言开发的远程桌面控制工具，专注于跨平台的高效连接。它支持Windows、Linux和macOS，通过SSH协议建立安全通道，让用户能够随时随地管理远程服务器。项目采用零配置设计，只需几行命令即可启动，适合运维人员和开发者使用。","stars":626,"starGrowth":6,"createdAt":"2026-05-28T16:41:26Z","silentRounds":0,"language":"Go","topics":[],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["远程桌面","SSH","运维工具","服务器管理","Go语言"],"domainKey":"远程桌面","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/futrx-com/remote.futrx","ts":"2026-08-15T00:57:24.351Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.05610479233226837,"copyOk":true},{"repo":"milligram/milligram","owner":"milligram","name":"milligram","desc":"A minimalist CSS framework.","summaryCn":"极简主义 CSS 框架，几行代码就能做出好看的网页。","reasonCn":"Milligram 是一个极其轻量级的 CSS 框架，它的设计哲学是“零依赖”。它只包含核心的 CSS 样式，没有任何 JavaScript 代码，文件体积非常小。开发者可以通过 CDN 引入，或者直接复制代码到项目中，快速搭建出排版整齐、视觉清爽的网页界面。","stars":10217,"starGrowth":0,"createdAt":"2015-12-10T17:03:36Z","silentRounds":28,"language":"HTML","topics":["design"],"aiDims":["设计工具","非AI-实用","创意工具"],"zone":"工具","funScore":0,"domainTags":["CSS框架","前端","极简主义","UI","样式"],"domainKey":"CSS框架","pushedAt":"2026-09-01T21:28:39Z","tags":[{"name":"设计工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/milligram/milligram","ts":"2026-09-01T21:28:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AIDASLab/Awesome-Diffusion-LLM","owner":"AIDASLab","name":"Awesome-Diffusion-LLM","desc":"A comprehensive list of papers about Large-Language-Diffusion-Models.","summaryCn":"这是一个关于扩散模型与大语言模型论文的精彩列表。","reasonCn":"该仓库是一个awesome列表，收录了扩散模型与大语言模型相关的研究论文，涵盖模型架构、训练方法与应用场景。列表按主题分类，方便研究者快速定位前沿进展，定期更新收录顶会论文，是学术跟踪的实用工具。尽管热度不高，但内容质量可靠。","stars":103,"starGrowth":0,"createdAt":"2025-06-01T15:22:30Z","silentRounds":28,"language":"","topics":["llm-model"],"aiDims":["大语言模型","模型与训练"],"zone":"资源","funScore":0.5,"domainTags":["扩散模型","大语言模型","论文列表","多模态","学术研究"],"domainKey":"扩散模型","pushedAt":"2026-09-01T12:54:57Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.7,"source":"search","url":"https://github.com/AIDASLab/Awesome-Diffusion-LLM","ts":"2026-09-01T12:54:57Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NVIDIA/personaplex","owner":"NVIDIA","name":"personaplex","desc":"PersonaPlex code.","summaryCn":"第一段：想在游戏或直播间里快速拥有一个会说话、能表情、还能走动的虚拟人","reasonCn":"第一段：想在游戏或直播间里快速拥有一个会说话、能表情、还能走动的虚拟人物，却要花费数周甚至数月调研模型、拼接音视频工具，成本高且技术门槛大。PersonaPlex 直接把这些步骤封装成一个统一框架，只要提供角色的文字描述，就能一次性生成对应的 3D 模型、语音库和表情动画，让","stars":10569,"starGrowth":2,"createdAt":"2026-01-05T19:10:35Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI Agent","多模态Agent","AI应用"],"zone":"创意","funScore":1,"domainTags":["多模态生成","虚拟人","游戏开发"],"domainKey":"多模态生成","pushedAt":"2026-09-07T22:38:09.240Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/NVIDIA/personaplex","ts":"2026-09-07T22:38:09.240Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"KoBeWi/Godot-Instance-Dock","owner":"KoBeWi","name":"Godot-Instance-Dock","desc":"An addon for Godot engine that adds a convenience dock where you can store scenes.","summaryCn":"这是专为Godot引擎打造的场景管理坞，让你拖拽场景更加高效顺手。","reasonCn":"它的核心设计非常轻量，完全用GDScript编写，没有搞那些复杂的底层魔改。它只是巧妙地扩展了Godot编辑器的UI接口，把场景引用的快捷方式管理起来。你可以在面板里对场景进行分组、重命名，支持直接拖拽生成节点。相比每次都从文件管理器里翻找并拖进场景树，这种自定义快捷坞站的方","stars":219,"starGrowth":0,"createdAt":"2020-11-21T21:57:09Z","silentRounds":14,"language":"GDScript","topics":["godot"],"aiDims":["游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Godot引擎","场景管理","游戏开发","插件"],"domainKey":"Godot引擎","pushedAt":"2026-08-23T11:19:39Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/KoBeWi/Godot-Instance-Dock","ts":"2026-08-23T11:19:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"NGT-labs/NGT","owner":"NGT-labs","name":"NGT","desc":"Nearest Neighbor Search with Neighborhood Graph and Tree for High-dimensional Data","summaryCn":"专门找相似的高维向量搜索利器，让AI记住海量特征","reasonCn":"由日本茨城大学开发的开源高维最近邻搜索库，核心算法是邻域图与动态树结构结合，专门解决向量相似度检索的效率和精度问题。支持百亿级向量索引构建，查询速度远超传统暴力搜索，提供C/C++、Python、Java等多语言绑定，可无缝接入推荐系统、图像检索、生物信息分析等场景。","stars":1374,"starGrowth":0,"createdAt":"2016-09-01T07:36:57Z","silentRounds":27,"language":"C++","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Vector Database","High-dimensional","C++"],"domainKey":"Vector Database","pushedAt":"2026-09-01T02:04:32Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/NGT-labs/NGT","ts":"2026-09-01T02:04:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"JushBJJ/Mr.-Ranedeer-AI-Tutor","owner":"JushBJJ","name":"Mr.-Ranedeer-AI-Tutor","desc":"","summaryCn":"Mr. Ranedeer AI Tutor 并不是一个可以独立运行的软件，而是一套在项目中被","reasonCn":"在使用体验上，项目提供了一套完整的提示词配置指南，允许用户开启或关闭特定工具，甚至支持多语言环境下的交互测试。从其保留的截图示例中可以看出，模型能以“第一课”、“知识点拆解”等结构化的形式进行输出，降低了直接面对大模型时的盲目性，尤其适合不知道如何向AI提问以获取高","stars":29575,"starGrowth":0,"createdAt":"2023-03-31T12:37:26Z","silentRounds":6,"language":"","topics":["ai","education","gpt-4","llm"],"aiDims":["提示工程","大语言模型","学习资源"],"zone":"资源","funScore":0.2,"domainTags":["GPT-4定制提示词","个性化学习导师","多语言支持"],"domainKey":"GPT-4定制提示词","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"ai","source":"github","weight":0.7},{"name":"education","source":"github","weight":0.7},{"name":"gpt-4","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7}],"aiScore":0.5599999999999999,"source":"search","url":"https://github.com/JushBJJ/Mr.-Ranedeer-AI-Tutor","ts":"2026-09-13T22:08:09.499Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"webhooksite/webhook.site","owner":"webhooksite","name":"webhook.site","desc":"⚓️ Easily test HTTP webhooks with this handy tool that displays requests instantly.","summaryCn":"测试Webhook的神器，无需服务器就能接收查看HTTP请求","reasonCn":"这是一个免费的在线测试平台，专门用来调试和接收HTTP请求。你不需要任何服务器，打开网页就能获得一个唯一的临时URL，用来接收来自GitHub Actions、支付网关或各类SaaS的回调。所有请求详情都会实时显示在页面上，支持复制Headers、Body和参数，还能手动重放请求进行排查，对后端开发极其友好。","stars":6762,"starGrowth":2,"createdAt":"2016-03-21T08:45:42Z","silentRounds":0,"language":"JavaScript","topics":["frontend"],"aiDims":["开发者工具","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["Webhook测试","HTTP调试","API测试","后端开发","运维工具"],"domainKey":"Webhook测试","pushedAt":"2026-07-22T10:28:58Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/webhooksite/webhook.site","ts":"2026-07-22T10:28:58Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"multimindlab/multimind-sdk","owner":"multimindlab","name":"multimind-sdk","desc":"Your SDK solves all of this. One interface. Unified logic. Local + hosted models. Fine-tuning. Agent tools. Enterprise-ready. Hybrid RAG.Star 🌟 if you like it!","summaryCn":"统一调用本地和云端AI模型，一个SDK搞定所有AI能力","reasonCn":"Multimind SDK是一个Python库，提供统一的接口调用本地和云端的大语言模型。它支持Fine-tuning、Agent工具调用、混合RAG检索等功能，让开发者无需切换多个API就能实现完整的AI应用。无论是本地部署的开源模型还是OpenAI等商业模型，都能通过同一套代码调用，大大简化了AI应用的开发流程。","stars":97,"starGrowth":0,"createdAt":"2025-05-12T13:36:47Z","silentRounds":29,"language":"Python","topics":["llm-model"],"aiDims":["AI基础设施","工具调用","模型部署"],"zone":"AI","funScore":0,"domainTags":["模型调用","Agent框架","RAG"],"domainKey":"RAG","pushedAt":"2026-08-20T12:43:59Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/multimindlab/multimind-sdk","ts":"2026-08-20T12:43:59Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"anseryuer/Local_LLM_Deployment_Guide_Chinese","owner":"anseryuer","name":"Local_LLM_Deployment_Guide_Chinese","desc":"本地部署大语言模型的中文教学","summaryCn":"手把手教你把大模型装到自家电脑上，零门槛体验AI对话乐趣","reasonCn":"这份中文教程专为普通用户设计，涵盖主流开源大模型的本地部署全流程，包括Ollama、llama.cpp、vLLM等工具的安装配置，以及模型量化、硬件加速和API调用等关键技术环节。内容详尽实用，帮助读者真正在本地跑起来大模型，是中文社区难得的零基础部署指南。","stars":54,"starGrowth":0,"createdAt":"2024-05-09T21:08:31Z","silentRounds":27,"language":"","topics":["llm-model"],"aiDims":["模型部署","教程","大语言模型"],"zone":"资源","funScore":0,"domainTags":["大语言模型","本地部署","GGUF量化","Ollama"],"domainKey":"大语言模型","pushedAt":"2026-08-31T14:32:15Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.72,"source":"search","url":"https://github.com/anseryuer/Local_LLM_Deployment_Guide_Chinese","ts":"2026-08-31T14:32:15Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aipulsedaily/f1-round2","owner":"aipulsedaily","name":"f1-round2","desc":"A 124-second 4K film of an F1 car in one unbroken take, where every model, material and sound is generated by code — no downloaded assets, no HDRIs, no photo textures, nothing AI-generated. Ships its own engineering log: ~1,300 entries recording what was measured and what turned out to be wrong.","summaryCn":"完全用代码一行AI不用画出来的震撼F1赛车动画短片","reasonCn":"在技术实现上，该项目把程序化生成（Procedural Generation）玩到了极致。通过Python脚本运用数学几何、噪声函数以及程序化音频合成技术，在代码里直接构建了几何体、计算了材质反射并合成了引擎轰鸣声。这种完全摆脱外部静态资产依赖的做法，对算力和算法逻辑的要求极高，堪称现代","stars":16,"starGrowth":0,"createdAt":"2026-08-18T15:14:45Z","silentRounds":21,"language":"Python","topics":["procedural-generation"],"aiDims":["创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["程序化生成","计算机图形学","4K短片","实时渲染","程序化音频合成"],"domainKey":"程序化生成","pushedAt":"2026-08-18T15:16:01Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/aipulsedaily/f1-round2","ts":"2026-08-18T15:16:01Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"FortAwesome/Font-Awesome","owner":"FortAwesome","name":"Font-Awesome","desc":"The iconic SVG, font, and CSS toolkit","summaryCn":"程序员必备的图标库，76000多人收藏的宝藏","reasonCn":"Font Awesome是最流行的开源图标工具包，提供超过8000个SVG图标和字体图标，涵盖社交、界面、箭头、文件等多种类型。支持通过CDN、npm、CLI等多种方式集成，兼容所有主流浏览器和框架，开发者可以灵活定制颜色、大小和样式，大幅提升UI开发效率。","stars":76947,"starGrowth":1,"createdAt":"2012-02-17T14:19:43Z","silentRounds":1,"language":"JavaScript","topics":["fontawesome","icons","webfont","font","css","svg-icons","svg-sprites"],"aiDims":["开发者工具","效率工具","可视化"],"zone":"工具","funScore":0,"domainTags":["图标库","SVG","前端开发","UI组件","开源工具"],"domainKey":"图标库","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"fontawesome","source":"github","weight":0.7},{"name":"icons","source":"github","weight":0.7},{"name":"webfont","source":"github","weight":0.7},{"name":"font","source":"github","weight":0.7},{"name":"css","source":"github","weight":0.7},{"name":"svg-icons","source":"github","weight":0.7},{"name":"svg-sprites","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/FortAwesome/Font-Awesome","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"funstory-ai/BabelDOC","owner":"funstory-ai","name":"BabelDOC","desc":"Yet Another Document Translator","summaryCn":"会理解文档结构的AI翻译神器，PDF多语言翻译保持原格式不乱","reasonCn":"BabelDOC是一个专门做文档翻译的AI工具，不只是简单翻译文字，而是能理解PDF等文档的结构布局。它用Python开发，基于大语言模型做翻译，能保持原文档的段落、表格、图片位置。适合需要处理多语言技术文档、论文的用户。","stars":9595,"starGrowth":2,"createdAt":"2024-11-13T13:57:29Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["AI应用","工具调用"],"zone":"AI","funScore":0.5,"domainTags":["文档翻译","PDF处理","AI翻译","多语言","OCR"],"domainKey":"文档翻译","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/funstory-ai/BabelDOC","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"StarlightSearch/EmbedAnything","owner":"StarlightSearch","name":"EmbedAnything","desc":"Highly Performant, Modular, Memory Safe and Production-ready Inference, Ingestion and Indexing built in Rust 🦀","summaryCn":"用Rust写的向量数据库，嵌入生成、索引和推理一条龙","reasonCn":"Rust编写的高性能向量数据库，支持嵌入生成、索引和推理全流程。零依赖设计，内存安全，原生多线程加速。相比Python方案性能提升数倍，适合对延迟和并发有严格要求的生产环境。支持批量处理，单二进制文件即可部署，无需复杂运行时环境。","stars":1306,"starGrowth":0,"createdAt":"2024-03-31T22:37:47Z","silentRounds":28,"language":"Rust","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","推理引擎"],"zone":"AI","funScore":0.8,"domainTags":["向量数据库","Rust","RAG","嵌入索引","推理引擎"],"domainKey":"向量数据库","pushedAt":"2026-08-12T08:56:59Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/StarlightSearch/EmbedAnything","ts":"2026-08-12T08:56:59Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Solido/awesome-flutter","owner":"Solido","name":"awesome-flutter","desc":"An awesome list that curates the best Flutter libraries, tools, tutorials, articles and more.","summaryCn":"Flutter开发者的终极宝典，收录了最好的库、工具和教程。","reasonCn":"它的组织结构非常清晰，通常按照功能模块进行分类。比如“UI组件”、“功能库”、“工具”、“学习资源”等。这种分类方式让你能非常直观地找到你需要的工具。例如，当你需要做一个漂亮的列表效果时，你可以直接在“UI组件”分类下找到几十个现成的组件库进行对比和选择。","stars":61283,"starGrowth":6,"createdAt":"2017-05-07T11:45:27Z","silentRounds":0,"language":"Dart","topics":["mobile","awesome"],"aiDims":["开发者工具","移动开发","创意工具"],"zone":"资源","funScore":0,"domainTags":["Flutter","移动开发","资源导航"],"domainKey":"Flutter","pushedAt":"2026-09-03T04:34:54Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"移动开发","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"awesome","source":"github","weight":0.7},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/Solido/awesome-flutter","ts":"2026-09-03T04:34:54Z","category":"learning","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.01829371930225348,"copyOk":false},{"repo":"browseros-ai/BrowserOS","owner":"browseros-ai","name":"BrowserOS","desc":"🌐 The open-source Agentic browser; alternative to ChatGPT Atlas, Perplexity Comet, Dia.","summaryCn":"开源AI浏览器，像ChatGPT一样自动搜索整理信息替你上网","reasonCn":"核心架构是典型的Agentic设计。它用LLM作为大脑，结合视觉识别能力理解网页内容，再用自动化工具操控浏览器完成操作。和传统爬虫不同，它能处理动态渲染的页面、登录状态、甚至验证码。和Perplexity的闭源方案相比，BrowserOS你可以自建部署，数据安全完全自己掌控，不会被第三方服","stars":13737,"starGrowth":0,"createdAt":"2025-05-18T16:23:54Z","silentRounds":1,"language":"TypeScript","topics":["browser","browseros","chromium","linux","macos","windows","agent","llm","lmstudio","ollama"],"aiDims":["AI Agent","多模态Agent","AI应用"],"zone":"AI","funScore":0.8,"domainTags":["AI智能体","浏览器自动化","RAG","Agent框架","信息检索"],"domainKey":"RAG","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"browser","source":"github","weight":0.7},{"name":"browseros","source":"github","weight":0.7},{"name":"chromium","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"agent","source":"github","weight":0.7},{"name":"llm","source":"github","weight":0.7},{"name":"lmstudio","source":"github","weight":0.7},{"name":"ollama","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.75,"source":"trending","url":"https://github.com/browseros-ai/BrowserOS","ts":"2026-08-10T06:58:39.267Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"itmeo/webgradients","owner":"itmeo","name":"webgradients","desc":"180 free CSS3, Figma, Sketch & PSD gradients — plus a JSON dataset for design tools, generators, and AI agents.","summaryCn":"180个免费CSS渐变组合，设计师和前端开发者的配色神器","reasonCn":"这个项目集合了180个精心设计的CSS渐变，涵盖各种风格和场景。除了网页可直接复制的CSS代码，还提供Figma、Sketch和PSD格式文件，以及JSON数据集，方便设计师和AI代理直接使用。界面美观易用，按颜色分类筛选，找渐变就像在调色盘里挑颜色一样简单。","stars":2478,"starGrowth":0,"createdAt":"2017-01-08T10:33:28Z","silentRounds":14,"language":"CSS","topics":["design","frontend"],"aiDims":["非AI-好玩","创意工具"],"zone":"资源","funScore":0,"domainTags":["CSS渐变","UI设计","配色方案","前端资源","设计素材"],"domainKey":"CSS渐变","pushedAt":"2026-08-03T21:50:58Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"frontend","source":"github","weight":0.7},{"name":"CSS","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/itmeo/webgradients","ts":"2026-08-03T21:50:58Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sinelaw/fresh","owner":"sinelaw","name":"fresh","desc":"Terminal based IDE &amp; text editor: easy, powerful and fast","summaryCn":"一个让你在黑底终端里敲代码、写文档的神器","reasonCn":"Fresh 是一个基于 Rust 编写的全功能终端编辑器，专为追求极致速度和键盘操作体验的开发者设计。它完全摒弃了传统图形界面，通过高度优化的文本渲染引擎实现丝滑流畅的交互，内置了 LSP、文件管理、任务栏等现代 IDE 的核心功能，让所有操作都能在键盘上完成。","stars":8998,"starGrowth":5,"createdAt":"2024-12-24T19:50:15Z","silentRounds":0,"language":"Rust","topics":["ide","terminal-based","text-editor"],"aiDims":["非AI-实用","开发工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["终端","IDE","Rust","编辑器"],"domainKey":"终端","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"ide","source":"github","weight":0.7},{"name":"terminal-based","source":"github","weight":0.7},{"name":"text-editor","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/sinelaw/fresh","ts":"2026-08-14T01:23:41.092Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.016667037119359858,"copyOk":false},{"repo":"gonglei007/GameDevMind","owner":"gonglei007","name":"GameDevMind","desc":"最全面的游戏开发技术图谱(Game Development Map)。帮助游戏开发者们在已知问题上节省时间，省出更多的精力投入到更有创造性的工作中去。","summaryCn":"游戏开发全领域知识图谱，从美术到程序一网打尽","reasonCn":"这是一个涵盖游戏开发全流程的技术图谱，包括程序、美术、策划、音效等各个方向的知识点。它不仅罗列了技术栈，还梳理了学习路径和推荐资源，是游戏开发新人的入门指南，也是老手的查漏补缺工具。对于想进入游戏行业或了解行业全貌的人来说，这份地图极具参考价值。","stars":6557,"starGrowth":3,"createdAt":"2022-06-22T01:22:57Z","silentRounds":0,"language":"Python","topics":["game"],"aiDims":["学习资源","教程","非AI-好玩"],"zone":"资源","funScore":0.5,"domainTags":["游戏开发","技术栈","知识图谱","游戏引擎"],"domainKey":"游戏开发","pushedAt":"2026-08-11T08:52:49Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/gonglei007/GameDevMind","ts":"2026-08-11T08:52:49Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Tonner-Zech-Group/chem-diagrams","owner":"Tonner-Zech-Group","name":"chem-diagrams","desc":"Python package for the creation of publication-quality reaction energy diagrams with Matplotlib.","summaryCn":"化学家在发表论文或者做汇报时，经常需要绘制反应路径的能量变化图。","reasonCn":"化学家在发表论文或者做汇报时，经常需要绘制反应路径的能量变化图。这种图表看起来专业，但其实很难用通用的绘图库画好，因为涉及到化学符号、能级线和能量标注的特殊规范。这个项目就是为了解决这个痛点而生的。它不是简单的封装Matplotlib，而是针对化学绘图的特定需求做了大量","stars":90,"starGrowth":0,"createdAt":"2026-01-30T13:22:41Z","silentRounds":14,"language":"Python","topics":["physics"],"aiDims":["可视化","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["化学","反应能量图","Matplotlib","科学绘图","科研工具"],"domainKey":"化学","pushedAt":"2026-09-09T14:46:11Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"physics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/Tonner-Zech-Group/chem-diagrams","ts":"2026-09-09T14:46:11Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ClarkThyLord/Voxel-Core","owner":"ClarkThyLord","name":"Voxel-Core","desc":"Voxel plugin for the Godot 3!","summaryCn":"很多独立游戏开发者想在 Godot 中实现类似《我的世界》的体素世界，却发现官","reasonCn":"很多独立游戏开发者想在 Godot 中实现类似《我的世界》的体素世界，却发现官方并没有直接的支持。Voxel-Core 正是为了解决这个痛点而诞生的，它把体素地形的生成、编辑、渲染全部封装成一个插件，让你只需几步就能在 Godot 3 项目里使用方块构建的宇宙。","stars":494,"starGrowth":0,"createdAt":"2019-06-25T00:01:55Z","silentRounds":15,"language":"GDScript","topics":["godot"],"aiDims":["游戏","非AI-好玩","创意工具"],"zone":"工具","funScore":0,"domainTags":["Godot","体素游戏","游戏开发","地形生成"],"domainKey":"Godot","pushedAt":"2026-09-07T23:16:31Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"godot","source":"github","weight":0.7},{"name":"GDScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/ClarkThyLord/Voxel-Core","ts":"2026-09-07T23:16:31Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"ultraworkers/claw-code","owner":"ultraworkers","name":"claw-code","desc":"An agent-managed museum exhibit, built in Rust with Gajae-Code / LazyCodex — developed and maintained with no human intervention.","summaryCn":"AI Agent自己管自己的博物馆，人类全程躺平不插手","reasonCn":"技术上，这个项目基于Gajae-Code/LazyCodex架构，核心是一个能自我迭代的Agent系统。Agent被赋予初始任务后，会自主读取代码、理解项目结构、分析现有功能，然后制定改进计划并执行。它能自己写代码、运行测试、修复报错、提交变更，整个过程不需要人类介入。","stars":195281,"starGrowth":0,"createdAt":"2026-03-31T08:58:08Z","silentRounds":2,"language":"Rust","topics":[],"aiDims":["AI Agent","非AI-好玩","游戏"],"zone":"AI","funScore":1,"domainTags":["AI Agent","自主运维","Rust"],"domainKey":"AI Agent","pushedAt":"2026-08-16T06:18:45Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/ultraworkers/claw-code","ts":"2026-08-16T06:18:45Z","category":"ai","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aceiii/ace-gb","owner":"aceiii","name":"ace-gb","desc":"GameBoy Emulator!","summaryCn":"把掌机经典游戏搬到电脑，随时加载 ROM，轻松玩转复古童年。","reasonCn":"ace-gb 使用 C++ 与轻量图形库 raylib 完整模拟 GameBoy 主机，包括 CPU、显卡、声音和按键，兼容大多数商业与自制 ROM，支持窗口缩放、帧率调节，可在 Windows、Linux、macOS 跨平台运行，代码结构清晰，适合二次开发与学习掌机底层原理。","stars":27,"starGrowth":0,"createdAt":"2024-04-28T04:50:07Z","silentRounds":15,"language":"C++","topics":["raylib"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":0.6,"domainTags":["GameBoy模拟器","复古游戏","游戏模拟","C++"],"domainKey":"GameBoy模拟器","pushedAt":"2026-09-07T23:07:45Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"raylib","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/aceiii/ace-gb","ts":"2026-09-07T23:07:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Battam1111/Myco","owner":"Battam1111","name":"Myco","desc":"The living armor an AI agent inhabits: eternal devouring, eternal evolution, eternal amplification.","summaryCn":"把AI智能体变成活体战甲，实现永恒吞噬与进化的知识管理系统。","reasonCn":"技术层面上，它完全用Rust编写，充分利用了这门语言在内存安全和极致性能上的先天优势。它把传统的知识库管理变成了一个动态的生态系统，AI在后台像贪吃蛇一样不断吞噬新的数据节点，通过内部复杂的演化算法自动进行关联、重构和升级。","stars":64,"starGrowth":0,"createdAt":"2026-04-08T15:52:05Z","silentRounds":19,"language":"Rust","topics":["knowledge-management"],"aiDims":["AI Agent","知识图谱"],"zone":"AI","funScore":1,"domainTags":["AI智能体","知识管理","Rust","RAG","向量数据库"],"domainKey":"RAG","pushedAt":"2026-08-12T20:28:21Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"知识图谱","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/Battam1111/Myco","ts":"2026-08-12T20:28:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Azgaar/Armoria","owner":"Azgaar","name":"Armoria","desc":"Heraldry generator and editor","summaryCn":"一键生成专属纹章， medieval风格超有范","reasonCn":"Armoria是一个纹章学生成器和编辑器，使用过程化生成技术，按照真实纹章学规则自动生成盾牌图案。支持自定义形状、颜色、金属和图案组合，渲染为高质量SVG，纯前端Svelte实现，无需服务器，直接在浏览器中编辑和导出。","stars":365,"starGrowth":1,"createdAt":"2020-11-23T21:15:06Z","silentRounds":15,"language":"Svelte","topics":["procedural-generation"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["纹章生成","过程化生成","SVG渲染"],"domainKey":"纹章生成","pushedAt":"2026-08-19T08:23:06Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"procedural-generation","source":"github","weight":0.7},{"name":"Svelte","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/Azgaar/Armoria","ts":"2026-08-19T08:23:06Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"libgdx/libgdx","owner":"libgdx","name":"libgdx","desc":"Desktop/Android/HTML5/iOS Java game development framework","summaryCn":"用Java写游戏，一套代码跑遍电脑手机网页，游戏开发者的全能工具包","reasonCn":"LibGDX是基于Java的跨平台2D游戏开发框架，支持Windows、Mac、Linux、Android、iOS和HTML5平台。采用模块化架构，内置渲染引擎、物理计算、音频处理和输入管理等核心组件，还自带Scene2D UI库可快速搭建游戏界面。Star数超过2.5万，是Java生态中最成熟的游戏开发框架之一，广泛应用于独立游戏和商业项目。","stars":25406,"starGrowth":0,"createdAt":"2012-08-10T19:34:38Z","silentRounds":1,"language":"Java","topics":["game"],"aiDims":["游戏"],"zone":"工具","funScore":0,"domainTags":["游戏开发框架","Java","跨平台渲染","2D游戏引擎"],"domainKey":"游戏开发框架","pushedAt":"2026-09-09T08:20:10Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/libgdx/libgdx","ts":"2026-09-09T08:20:10Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Mininglamp-OSS/octo-web","owner":"Mininglamp-OSS","name":"octo-web","desc":"Web &amp; desktop (Electron) client for the OCTO open workplace — one React + TypeScript codebase shipping browser and PC surfaces, with first-class AI agent UX.","summaryCn":"自带AI助手的智能办公平台，网页桌面双端都能用","reasonCn":"技术架构上，它采用单代码库同时输出Web和桌面端，React+TypeScript的前端框架保证了代码质量，Electron打包后获得原生桌面体验。AI Agent的UX设计是核心亮点，意味着它不是简单嵌入一个聊天窗口，而是从交互逻辑层面设计了AI参与的完整工作流，比如让AI理解你的任务、调用工具、返","stars":1180,"starGrowth":2,"createdAt":"2026-05-11T11:28:43Z","silentRounds":0,"language":"TypeScript","topics":["ai-agent","chat","desktop-app","electron","i18n","im","messaging","monorepo","open-source","pwa","react","turborepo","typescript","websocket"],"aiDims":["AI Agent","效率工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["办公协作","AI Agent","Electron","企业应用","前端框架"],"domainKey":"办公协作","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"ai-agent","source":"github","weight":0.7},{"name":"chat","source":"github","weight":0.7},{"name":"desktop-app","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"i18n","source":"github","weight":0.7},{"name":"im","source":"github","weight":0.7},{"name":"messaging","source":"github","weight":0.7},{"name":"monorepo","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"pwa","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"turborepo","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"websocket","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/Mininglamp-OSS/octo-web","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"explosion/spaCy","owner":"explosion","name":"spaCy","desc":"💫 Industrial-strength Natural Language Processing (NLP) in Python","summaryCn":"Python界的NLP瑞士军刀，分词命名实体提取一把梭","reasonCn":"spaCy是工业级NLP库，提供高性能文本处理流水线，涵盖分词、词性标注、句法解析、命名实体识别等核心功能。采用Cython优化，处理速度比传统NLTK快数倍。支持多语言模型，可加载预训练权重或从头训练自定义管道。API设计现代化，文档和实体提取功能在金融、医疗等行业广泛应用。","stars":33916,"starGrowth":1,"createdAt":"2014-07-03T15:15:40Z","silentRounds":1,"language":"Python","topics":["ml","ai"],"aiDims":["AI应用","文档理解","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["自然语言处理","文本分析","命名实体识别","句法解析","Python库"],"domainKey":"自然语言处理","pushedAt":"2026-08-24T08:26:10Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"文档理解","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"search","url":"https://github.com/explosion/spaCy","ts":"2026-08-24T08:26:10Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zeidalidiez/Notenotes","owner":"zeidalidiez","name":"Notenotes","desc":"Notenotes is a tiny, browser-based musical 'post it note' for capturing the idea of music with unconventional instruments and visual aids.","summaryCn":"在浏览器里贴个音乐便签，用点子音和可视化捕捉旋律灵感。","reasonCn":"Notenotes是一款基于JavaScript的轻量级网页应用，提供类似便利贴的交互界面，用户可以直接在浏览器中绘制音符、选择非传统音色并实时合成，配合颜色、形状等可视化元素帮助记谱和灵感捕捉，省去安装软件的繁琐，随时随地都能快速记录音乐创意。","stars":12,"starGrowth":0,"createdAt":"2026-05-03T03:27:18Z","silentRounds":15,"language":"JavaScript","topics":["synthesizer"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":0.7,"domainTags":["synthesizer","音乐创作","Web Audio API","浏览器应用","可视化记谱"],"domainKey":"synthesizer","pushedAt":"2026-09-07T08:16:55Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"synthesizer","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/zeidalidiez/Notenotes","ts":"2026-09-07T08:16:55Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"talkpython/mastering-pycharm-course","owner":"talkpython","name":"mastering-pycharm-course","desc":"Course demos and handouts for Talk Python's Effective PyCharm course","summaryCn":"把 PyCharm 教程拆解成实战演示，教你像高手一样写代码。","reasonCn":"这是一个专门为 Talk Python 的《Effective PyCharm》课程设计的实战仓库。它不仅包含完整的课程讲义，更重要的是提供了大量可直接运行的代码示例。内容覆盖了 PyCharm 的高级功能，包括自定义代码风格、快捷键配置、调试技巧、版本控制集成以及如何利用插件提升开发效率。通过这些精选的演示和手记，开发者可以快速掌握这款 IDE 的强大之处。","stars":1219,"starGrowth":0,"createdAt":"2018-01-04T06:56:27Z","silentRounds":14,"language":"Python","topics":["course"],"aiDims":["开发者工具","教程","学习资源"],"zone":"资源","funScore":0,"domainTags":["PyCharm教程","IDE","实战演示","Python","代码技巧"],"domainKey":"PyCharm教程","pushedAt":"2026-09-08T22:06:15Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"course","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/talkpython/mastering-pycharm-course","ts":"2026-09-08T22:06:15Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"liyupi/codefather","owner":"liyupi","name":"codefather","desc":"程序员鱼皮的编程宝典 ⭐️ 2026年最全编程学习路线图！包含Java学习路线、前端学习路线、Python学习路线、C++学习路线、算法学习路线、计算机基础学习路线、AI应用开发学习路线、AI Agent开发学习路线等。提供编程入门教程、AI大模型应用开发教程、RAG开发实战、MCP开发教程、Prompt工程指南、LLM应用开发、技术知识分享、学习资源推荐、项目实战教程、热门面试题、求职经验、简历优化、编程自学指南等内容，适用于所有零基础学编程、学习AI开发、转行程序员、计算机专业学生、求职找工作的同学 💎 编程学习，就来编程导航！","summaryCn":"程序员鱼皮整理的2026最全编程学习路线图宝典","reasonCn":"程序员鱼皮出品的编程学习路线图合集，涵盖Java、前端、Python、C++、算法、AI应用开发等全方向。内容从入门到进阶，包含项目实战、面试题、求职经验，还特别收录了AI大模型应用、RAG、MCP、Prompt工程等前沿方向的学习指南。","stars":8638,"starGrowth":4,"createdAt":"2021-04-20T14:12:13Z","silentRounds":0,"language":"TypeScript","topics":["frontend","backend"],"aiDims":["学习资源","教程","AI应用"],"zone":"资源","funScore":0,"domainTags":["编程学习","学习路线","AI应用开发","面试题"],"domainKey":"编程学习","pushedAt":"2026-07-11T03:14:33Z","tags":[{"name":"学习资源","source":"llm","weight":1},{"name":"教程","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"backend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"search","url":"https://github.com/liyupi/codefather","ts":"2026-07-11T03:14:33Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"hicccc77/WeFlow","owner":"hicccc77","name":"WeFlow","desc":"WeFlow - 一个本地的微信聊天记录导出和年度报告应用","summaryCn":"把你的微信聊天变成精美数据报告，还能自动生成年度总结。","reasonCn":"这个项目把本地的微信聊天记录解析出来，通过可视化图表展示你的聊天习惯、高频联系人和消息趋势。它能生成类似网易云音乐年度听歌报告那样的个人数据年报，让你直观看到过去一年在和谁聊天最多、什么时间段发消息最勤。适合想回顾数字生活、分析社交数据或者单纯觉得好玩的用户。","stars":14426,"starGrowth":7,"createdAt":"2026-01-07T13:40:57Z","silentRounds":0,"language":"","topics":["data-visualization"],"aiDims":["可视化","非AI-好玩"],"zone":"创意","funScore":0.8,"domainTags":["微信","数据可视化","本地导出","年度报告","SQLite解析"],"domainKey":"微信","pushedAt":"2026-08-28T13:25:42Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7}],"aiScore":0.65,"source":"search","url":"https://github.com/hicccc77/WeFlow","ts":"2026-08-28T13:25:42Z","category":"fun","momentum":["daily"],"fromOfficial":false,"heatScore":0.026946845972549566,"copyOk":false},{"repo":"jpd002/Play-","owner":"jpd002","name":"Play-","desc":"Play! - PlayStation2 Emulator","summaryCn":"用现代电脑和手机流畅运行索尼经典老游戏，重温童年回忆的模拟器。","reasonCn":"技术层面上，它是一个用C++开发的独立PS2模拟器，不需要依赖索尼的官方BIOS就能运行。它通过逆向工程还原了当年Emotion Engine处理器的架构，利用现代图形API（如Vulkan、Metal或OpenGL）进行高效的硬件加速渲染，让当年的马赛克画面在1080P甚至4K分辨率下依然清晰。","stars":2686,"starGrowth":0,"createdAt":"2014-02-23T08:45:33Z","silentRounds":9,"language":"C++","topics":["emulator"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":1,"domainTags":["PS2模拟器","复古游戏","游戏模拟","硬件逆向"],"domainKey":"PS2模拟器","pushedAt":"2026-09-03T19:10:43Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"emulator","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/jpd002/Play-","ts":"2026-09-03T19:10:43Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"TryGhost/Ghost","owner":"TryGhost","name":"Ghost","desc":"Independent technology for modern publishing, memberships, subscriptions and newsletters.","summaryCn":"一键搭建现代博客，支持会员订阅和邮件通讯，轻松发布内容。","reasonCn":"Ghost 是基于 Node.js 的开源出版平台，提供完整的内容管理、主题系统和 API，支持 Markdown 编辑、会员付费、电子邮件订阅等功能。它采用无头架构，可与前端框架无缝对接，性能优化通过缓存和静态生成实现，高并发下仍保持响应迅速。同时提供丰富的插件生态，可通过 webhooks 与第三方服务集成，实现自动化工作流。","stars":55386,"starGrowth":5,"createdAt":"2013-05-04T11:09:13Z","silentRounds":0,"language":"JavaScript","topics":["journalism","publishing","blogging","javascript","web-application","cms","nodejs","ghost"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["博客平台","内容管理","会员订阅","CMS","Headless"],"domainKey":"博客平台","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"journalism","source":"github","weight":0.7},{"name":"publishing","source":"github","weight":0.7},{"name":"blogging","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"web-application","source":"github","weight":0.7},{"name":"cms","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"ghost","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/TryGhost/Ghost","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.015270826562669267,"copyOk":false},{"repo":"endee-io/endee","owner":"endee-io","name":"endee","desc":"Endee.io – A high-performance vector database, designed to handle up to 1B vectors on a single node, delivering significant performance gains through optimized indexing and execution. Also available in cloud https://endee.io/","summaryCn":"单机就能扛住十亿向量的高性能数据库，RAG系统的强力引擎","reasonCn":"技术上，Endee的核心优势在于精心设计的索引结构和执行引擎优化。它用C++从头构建，没有多余依赖，这意味着更低的内存占用和更快的计算速度。项目支持多种向量相似度算法，包括余弦相似度、欧氏距离等常见方案，同时针对大规模数据的批量查询做了深度优化。","stars":1296,"starGrowth":0,"createdAt":"2026-01-24T15:25:20Z","silentRounds":27,"language":"C++","topics":["vector-db"],"aiDims":["向量数据库","AI基础设施","AI搜索"],"zone":"工具","funScore":0,"domainTags":["向量数据库","RAG","数据库","C++"],"domainKey":"向量数据库","pushedAt":"2026-07-29T09:56:00Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"AI搜索","source":"llm","weight":1},{"name":"vector-db","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.75,"source":"search","url":"https://github.com/endee-io/endee","ts":"2026-07-29T09:56:00Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"deepseek-ai/awesome-deepseek-agent","owner":"deepseek-ai","name":"awesome-deepseek-agent","desc":"","summaryCn":"DeepSeek智能体资源大全，一键解锁各种实用工具与教程","reasonCn":"该项目是一个由社区维护的 Awesome 列表，集中收录了基于 DeepSeek 模型构建智能体的各类开源工具、教程、API 集成示例及实战案例。内容涵盖提示词模板、多模态扩展、工具调用框架等，帮助开发者快速找到现成方案，节省调研时间。","stars":6134,"starGrowth":2,"createdAt":"2026-04-27T06:37:38Z","silentRounds":0,"language":"","topics":[],"aiDims":["AI Agent","提示工程","学习资源"],"zone":"资源","funScore":0.5,"domainTags":["DeepSeek","智能体","资源列表","工具集"],"domainKey":"DeepSeek","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"提示工程","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1}],"aiScore":0.72,"source":"trending","url":"https://github.com/deepseek-ai/awesome-deepseek-agent","ts":"2026-08-15T00:57:24.351Z","category":"learning","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"tesseract-ocr/tesseract","owner":"tesseract-ocr","name":"tesseract","desc":"Tesseract Open Source OCR Engine (main repository)","summaryCn":"老牌文字识别神器，能帮你把图片里的字抽出来","reasonCn":"它的核心技术亮点在于对多种语言的强大支持，内置训练好的数据覆盖了100多种语言。你可以用它识别中文、英文、日文、韩文等多种语言混排的文字。它还支持多种图片格式，包括PNG、JPEG、TIFF等，并且可以处理各种分辨率和质量的图片。","stars":76636,"starGrowth":12,"createdAt":"2014-08-12T18:04:59Z","silentRounds":0,"language":"C++","topics":["ml"],"aiDims":["文档理解","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["OCR","文字识别","离线","多语言","图像处理"],"domainKey":"OCR","pushedAt":"2026-09-11T05:17:27Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/tesseract-ocr/tesseract","ts":"2026-09-11T05:17:27Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.03646975311863876,"copyOk":true},{"repo":"kijai/ComfyUI-KJNodes","owner":"kijai","name":"ComfyUI-KJNodes","desc":"Various custom nodes for ComfyUI","summaryCn":"ComfyUI的万能插件库，让你的AI绘图工作流强大十倍","reasonCn":"这是ComfyUI生态中最流行的节点插件库之一，内含数百个自定义节点。覆盖了图像生成、视频处理、音频同步、提示词工程、模型加载等多个领域。它让原本复杂的节点连线工作流变得功能极其丰富，尤其是结合KJ的实用脚本，能轻松实现自动化批量生成和复杂后处理。","stars":3313,"starGrowth":2,"createdAt":"2023-09-28T10:30:20Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["创意工具","AI应用","多模态Agent"],"zone":"工具","funScore":0.6,"domainTags":["ComfyUI","AI绘图","自定义节点","工作流"],"domainKey":"ComfyUI","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/kijai/ComfyUI-KJNodes","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rgthree/rgthree-comfy","owner":"rgthree","name":"rgthree-comfy","desc":"Making ComfyUI more comfortable!","summaryCn":"让ComfyUI操作更顺手，AI绘画工作流体验升级","reasonCn":"这是一个ComfyUI的第三方扩展，主要优化了节点布局、连接线和拖拽交互，让复杂的AI绘画工作流更清晰直观。支持自定义主题、自动排列节点、快捷键操作，能大幅减少重复点击和连线混乱的问题。对于经常需要搭建复杂SD工作流的用户来说，这个扩展能显著提升使用效率，减少视觉混乱。","stars":3494,"starGrowth":0,"createdAt":"2023-08-14T00:25:25Z","silentRounds":1,"language":"JavaScript","topics":["aiart","comfyui","stable-diffusion"],"aiDims":["创意工具","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["AI绘画","ComfyUI","工作流优化","节点编辑","图像生成"],"domainKey":"图像生成","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"aiart","source":"github","weight":0.7},{"name":"comfyui","source":"github","weight":0.7},{"name":"stable-diffusion","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.72,"source":"trending","url":"https://github.com/rgthree/rgthree-comfy","ts":"2026-08-11T09:45:50.026Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"supertorpe/chessendgametraining","owner":"supertorpe","name":"chessendgametraining","desc":"Chess Endgame Training: app to practice a variety of well organized set of chess endgames. From elementary to the most complex endgames, you can practice until you learn how to face all of them.","summaryCn":"很多象棋爱好者在提升残局技巧时往往缺少系统化的练习素材和即时的反馈，导","reasonCn":"技术上项目采用 TypeScript 编写，前端使用 React 与 Redux 管理状态，棋盘渲染基于 SVG 实现高分辨率且交互流畅。每个残局都有分步提示功能，玩家可以选择查看完整走法或仅查看关键走子。练习模式支持随机抽题、指定主题、计时挑战以及错题复盘，所有数据保存在浏览器的 IndexedD","stars":155,"starGrowth":0,"createdAt":"2019-03-22T22:30:40Z","silentRounds":15,"language":"TypeScript","topics":["chess"],"aiDims":["非AI-好玩","学习资源"],"zone":"创意","funScore":0.5,"domainTags":["象棋","残局训练","Web应用"],"domainKey":"象棋","pushedAt":"2026-09-08T17:19:39Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/supertorpe/chessendgametraining","ts":"2026-09-08T17:19:39Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"rudzen/ChessLib","owner":"rudzen","name":"ChessLib","desc":"C# chess library containing a complete data structure and move generation.","summaryCn":"想在自己的 C# 项目里加入国际象棋功能，却苦于没有现成的棋盘引擎？","reasonCn":"想在自己的 C# 项目里加入国际象棋功能，却苦于没有现成的棋盘引擎？ChessLib 正是为此而生，它把国际象棋的所有核心规则和搜索算法封装成一个可直接引用的库，让开发者无需从头实现走子生成、合法性检查和局面评估，只需几行代码就能让程序下棋、分析局面或对弈。","stars":97,"starGrowth":0,"createdAt":"2018-05-21T12:36:09Z","silentRounds":14,"language":"C#","topics":["chess"],"aiDims":["AI应用","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["国际象棋","游戏引擎","C#","规则实现","算法"],"domainKey":"国际象棋","pushedAt":"2026-09-09T05:19:03Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/rudzen/ChessLib","ts":"2026-09-09T05:19:03Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"panoptes/POCS","owner":"panoptes","name":"POCS","desc":"An Observatory Control System for the PANOPTES citizen-science project designed to help find transiting exoplanets! :telescope: :stars:","summaryCn":"开源天文望远镜控制软件，帮普通人找外星人","reasonCn":"这个项目的核心在于“自动化巡天”。它不仅仅是控制望远镜转动，还包含了一套图像处理算法。当它发现某颗恒星的光度发生异常变化（这可能是系外行星经过时挡住星光造成的）时，会自动抓拍照片并分析数据，确认是否为“候选目标”。","stars":84,"starGrowth":0,"createdAt":"2013-10-02T21:19:30Z","silentRounds":12,"language":"Python","topics":["astronomy"],"aiDims":["非AI-实用","天文","开源硬件"],"zone":"工具","funScore":0.7,"domainTags":["天文望远镜控制","系外行星搜索","自动化巡天","公民科学","开源硬件"],"domainKey":"天文望远镜控制","pushedAt":"2026-09-09T10:33:51Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"天文","source":"llm","weight":1},{"name":"开源硬件","source":"llm","weight":1},{"name":"astronomy","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/panoptes/POCS","ts":"2026-09-09T10:33:51Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"IRNova/Nova-Proxy","owner":"IRNova","name":"Nova-Proxy","desc":"یک پنل گرافیکی کاربردی برای ارائه اشتراک‌های Worker با پروکسی‌های ، Trojan و Warp به همراه زنجیره پروکسی، ارائه دهنده تنظیمات کامل DNS، IP تمیز و روتینگ پیشرفته برای کاربران تمامی پلتفرم‌ها با استفاده از هسته‌های Amnezia، Wireguard، Sing-box، Clash/Mihomo و Xray.","summaryCn":"搞定Trojan和Warp代理的图形面板，全网加速一把梭","reasonCn":"这是一款基于Node.js开发的图形化管理面板，专门用于管理和分发Trojan、Warp等代理服务。它支持多级代理链路配置，能自动清洗IP地址，提供详细的DNS和路由策略设置，支持Windows、macOS、Linux和Android等多平台客户端接入。","stars":3252,"starGrowth":0,"createdAt":"2026-02-19T07:51:56Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["非AI-实用","开发工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["代理","网络加速","管理面板","Docker"],"domainKey":"代理","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/IRNova/Nova-Proxy","ts":"2026-08-17T00:55:40.962Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"jibrilsharafi/EnergyMe-Home","owner":"jibrilsharafi","name":"EnergyMe-Home","desc":"Smart energy meter capable of monitoring up to 16 channels. Supports MQTT, REST API, Modbus TCP, InfluxDB, AWS IoT Core. Friendly UI and fully open source (PCB, schematics, components, source code)","summaryCn":"给家里每一路用电装个小探头，数据想进哪就进哪","reasonCn":"这是一套面向多路电力监测的开源硬件平台，最高覆盖十六个通道，板载固件负责采集功率、电压、电流与累计电量，并对外提供消息、网页接口、工业总线及云接入能力。原理图、印制板文件和源码都开放，还能配友好前端展示历史曲线和告警阈值，方便二次开发，易于扩展。","stars":279,"starGrowth":0,"createdAt":"2024-04-19T07:04:48Z","silentRounds":14,"language":"C++","topics":["home-automation"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["智能家居","电力监测","IoT"],"domainKey":"智能家居","pushedAt":"2026-09-05T15:00:45Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"home-automation","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/jibrilsharafi/EnergyMe-Home","ts":"2026-09-05T15:00:45Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"fja05680/sp500","owner":"fja05680","name":"sp500","desc":"Current and Historical Lists of S&P 500 components since 1996","summaryCn":"1996起S&P500成分股数据，配Notebook直接分析行情","reasonCn":"想要在不花钱买专业数据平台的前提下，获取完整的美国标普500成分股历史信息？sp500 项目把自1996年起每日的成分股变动、收盘价和市值等关键数据整理成 CSV，并提供了完整的 Jupyter Notebook 示例，直接可以在本地或云端环境中加载并进行分析。","stars":946,"starGrowth":0,"createdAt":"2019-11-16T14:32:09Z","silentRounds":1,"language":"Jupyter Notebook","topics":["finance"],"aiDims":["非AI-实用","可视化"],"zone":"资源","funScore":0,"domainTags":["标普500","金融数据","股票数据"],"domainKey":"标普500","pushedAt":"2026-09-07T14:57:01Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"finance","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/fja05680/sp500","ts":"2026-09-07T14:57:01Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"3b1b/manim","owner":"3b1b","name":"manim","desc":"Animation engine for explanatory math videos","summaryCn":"用代码画数学动画的瑞士军刀，让公式自己跳舞","reasonCn":"基于Python的数学动画引擎，通过编写代码精确控制每一帧画面。支持矢量图形、路径动画、相机变换，能生成平滑的数学公式动画。创作者3Blue1Brown用它制作了百万播放的线性代数和微积分系列，代码开源免费，可自定义各种效果。","stars":94160,"starGrowth":15,"createdAt":"2015-03-22T18:50:58Z","silentRounds":0,"language":"Python","topics":["python","animation","explanatory-math-videos","3b1b-videos"],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"创意","funScore":1,"domainTags":["数学动画","代码可视化","教育视频","Python"],"domainKey":"数学动画","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"python","source":"github","weight":0.7},{"name":"animation","source":"github","weight":0.7},{"name":"explanatory-math-videos","source":"github","weight":0.7},{"name":"3b1b-videos","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/3b1b/manim","ts":"2026-08-14T01:23:41.092Z","category":"fun","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.045477909940526764,"copyOk":false},{"repo":"shiguredo/momo","owner":"shiguredo","name":"momo","desc":"WebRTC Native Client Momo","summaryCn":"给树莓派装上WebRTC触角，变身能实时看能说话的智能语音助手","reasonCn":"技术上的亮点是它的轻量化和高性能。它不是简单的Python脚本封装，而是底层用C++实现的编码器调用，优化了ARM架构下的计算效率，确保在低算力设备上也能实现低延迟传输。它还支持多种编解码格式，能够自适应网络状况，保持画面流畅。","stars":860,"starGrowth":0,"createdAt":"2018-10-11T09:09:56Z","silentRounds":15,"language":"C++","topics":["raspberry-pi"],"aiDims":["非AI-实用","开发者工具","AI应用"],"zone":"工具","funScore":0.5,"domainTags":["WebRTC","树莓派","音视频"],"domainKey":"WebRTC","pushedAt":"2026-09-08T07:57:05Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"raspberry-pi","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/shiguredo/momo","ts":"2026-09-08T07:57:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yeasy/docker_practice","owner":"yeasy","name":"docker_practice","desc":"最新Docker容器技术，从真实案例中学习最佳实践！| Learn and understand Docker&Container technologies, with real DevOps practice!","summaryCn":"中文圈最火的Docker学习指南，从入门到实战全覆盖","reasonCn":"这是一本系统讲解Docker容器技术的开源中文教程，作者yeasy整理得非常详尽，从基础概念到容器编排都有涉及。书中结合了真实的生产案例和DevOps最佳实践，适合零基础的开发者快速上手，也适合有经验的同学查漏补缺。内容紧跟Docker最新特性，社区讨论热烈，是中文环境下学习容器技术的首选资料。","stars":26265,"starGrowth":1,"createdAt":"2014-09-05T04:06:39Z","silentRounds":1,"language":"Go","topics":["devops","docker"],"aiDims":["教程","开发者工具","AI基础设施"],"zone":"资源","funScore":0,"domainTags":["Docker","容器","教程","DevOps"],"domainKey":"Docker","pushedAt":"2026-09-16T18:43:40Z","tags":[{"name":"教程","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/yeasy/docker_practice","ts":"2026-09-16T18:43:40Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kubeflow/kubeflow","owner":"kubeflow","name":"kubeflow","desc":"Machine Learning Toolkit for Kubernetes","summaryCn":"为大厂量身定制的机器学习全流程管理基建平台","reasonCn":"其架构深度绑定 Kubernetes，将机器学习的每个步骤组件化。核心组件包括 Kubeflow Pipelines 用于定义和运行工作流，Katib 用于自动化的超参数调优，以及 TFJob 等用于分布式训练的算子。这种设计使得机器学习任务可以像微服务一样被调度，支持动态扩容和故障自愈，确保了在大规模","stars":15879,"starGrowth":1,"createdAt":"2017-11-30T18:44:19Z","silentRounds":1,"language":"","topics":["ml"],"aiDims":["AI基础设施","模型部署","模型与训练"],"zone":"AI","funScore":0,"domainTags":["Kubernetes","机器学习","MLOps","分布式训练"],"domainKey":"Kubernetes","pushedAt":"2026-08-21T18:25:34Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"ml","source":"github","weight":0.7}],"aiScore":0.7,"source":"search","url":"https://github.com/kubeflow/kubeflow","ts":"2026-08-21T18:25:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"aceberg/WatchYourLAN","owner":"aceberg","name":"WatchYourLAN","desc":"Lightweight network IP scanner written in Go. With notifications, history, export to Grafana","summaryCn":"局域网设备实时管家，谁蹭你的WiFi一清二楚","reasonCn":"技术上，它用Go写成，所以跨平台且性能极好，占用资源极低。核心工作原理是定期发送ARP请求，收集网络中的IP-MAC映射表，然后解析MAC地址的前缀来识别设备厂商。它还支持ICMP ping探测来提升扫描精度。最实用的功能包括：新设备上线时发送邮件或Webhook通知、设备离线告警、连接历","stars":7634,"starGrowth":1,"createdAt":"2022-08-15T04:58:22Z","silentRounds":2,"language":"Go","topics":["arp-scan","self-hosted","selfhosted","arp-scanner","monitoring","network-security","intrusion-detection"],"aiDims":["效率工具","安全工具","开发者工具"],"zone":"工具","funScore":0,"domainTags":["网络扫描","局域网监控","设备管理","Grafana"],"domainKey":"网络扫描","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"arp-scan","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"selfhosted","source":"github","weight":0.7},{"name":"arp-scanner","source":"github","weight":0.7},{"name":"monitoring","source":"github","weight":0.7},{"name":"network-security","source":"github","weight":0.7},{"name":"intrusion-detection","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.68,"source":"trending","url":"https://github.com/aceberg/WatchYourLAN","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"GCWing/BitFun","owner":"GCWing","name":"BitFun","desc":"BitFun combines a high-performance agent runtime written in Rust with a polished desktop application. It pairs the depth of a Code Agent with open, general-purpose capabilities for work beyond software development.","summaryCn":"Rust写的超快AI工作台，代码编程和生活办公一桌全搞定","reasonCn":"它的技术底座是Rust，这意味着它在内存安全和执行效率上远超同类的Python应用。Rust的并发模型让多个Agent可以并行工作而互不干扰，桌面端用Tauri或类似框架构建，界面流畅且资源占用低。更重要的是，它支持本地运行，你的数据不用传到云端。","stars":2314,"starGrowth":2,"createdAt":"2026-02-02T09:52:10Z","silentRounds":0,"language":"Rust","topics":["ai-coding","ai-ide","vibe-coding","cowork","agent-teams","openclaw","agentic-os","agentic-runtime","agentic","ai-sdk","computer-use-agent","bitfun","dsh-plugin"],"aiDims":["AI Agent","AI应用","代码助手"],"zone":"AI","funScore":0,"domainTags":["AI Agent","桌面应用","本地运行","代码助手"],"domainKey":"AI Agent","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"ai-coding","source":"github","weight":0.7},{"name":"ai-ide","source":"github","weight":0.7},{"name":"vibe-coding","source":"github","weight":0.7},{"name":"cowork","source":"github","weight":0.7},{"name":"agent-teams","source":"github","weight":0.7},{"name":"openclaw","source":"github","weight":0.7},{"name":"agentic-os","source":"github","weight":0.7},{"name":"agentic-runtime","source":"github","weight":0.7},{"name":"agentic","source":"github","weight":0.7},{"name":"ai-sdk","source":"github","weight":0.7},{"name":"computer-use-agent","source":"github","weight":0.7},{"name":"bitfun","source":"github","weight":0.7},{"name":"dsh-plugin","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.7,"source":"trending","url":"https://github.com/GCWing/BitFun","ts":"2026-08-09T00:58:42.474Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"UnderminersTeam/UndertaleModTool","owner":"UnderminersTeam","name":"UndertaleModTool","desc":"The most complete tool for modding, decompiling and unpacking Undertale (and other GameMaker games!)","summaryCn":"像素神作Undertale改装神器，解包改图再打包，想怎么玩得你","reasonCn":"Undertale及GameMaker引擎游戏最强大的逆向工程与Mod开发工具。具备完整的解包、反编译、资源提取及重新打包功能，支持脚本层面的深度修改。内置资源浏览器可直观查看游戏内部结构，提供类Unity的可视化编辑体验。对于游戏爱好者而言，它是打开Undertale无限可能性的钥匙，允许玩家自由定制角色、剧情、音效乃至核心机制，极大拓展了游戏的生命周期与创作上限。","stars":2058,"starGrowth":0,"createdAt":"2018-09-13T21:36:55Z","silentRounds":2,"language":"C#","topics":["editor"],"aiDims":["非AI-好玩","游戏"],"zone":"创意","funScore":0.9,"domainTags":["游戏Mod","逆向工程","游戏编辑","Undertale","GameMaker"],"domainKey":"游戏Mod","pushedAt":"2026-09-07T02:25:25Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"editor","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/UnderminersTeam/UndertaleModTool","ts":"2026-09-07T02:25:25Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tryingsomestuff/Minic","owner":"tryingsomestuff","name":"Minic","desc":"A simple chess engine to learn and play with","summaryCn":"极简主义的国际象棋引擎，用C++写的超轻量级对弈玩具","reasonCn":"从技术角度看，Minic 的代码非常精简，没有引入任何复杂的第三方库，完全依靠C++标准库实现。这种设计保证了代码的可读性和移植性，你可以在任何支持C++的编译器上轻松编译运行。它虽然不像商业引擎那样拥有深不可测的搜索深度，但在业余水平上已经具备相当的挑战性。","stars":135,"starGrowth":1,"createdAt":"2018-10-14T19:28:27Z","silentRounds":22,"language":"C++","topics":["chess"],"aiDims":["游戏"],"zone":"创意","funScore":0,"domainTags":["国际象棋","游戏引擎","算法"],"domainKey":"国际象棋","pushedAt":"2026-08-29T21:35:46Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/tryingsomestuff/Minic","ts":"2026-08-29T21:35:46Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"nitrojs/nitro","owner":"nitrojs","name":"nitro","desc":"Next Generation Server Toolkit. Create web servers with everything you need and deploy them wherever you prefer.","summaryCn":"一行命令创建全功能服务器，本地开发云端部署无缝切换","reasonCn":"Nitro是下一代服务器框架，基于Vercel、Cloudflare Workers等平台，一行命令创建完整的Web服务器。支持路由、中间件、数据库ORM、缓存等开箱即用功能。从本地开发到边缘计算部署无缝衔接，TypeScript原生支持，开发体验极佳。","stars":11232,"starGrowth":1,"createdAt":"2022-01-26T12:37:12Z","silentRounds":1,"language":"TypeScript","topics":[],"aiDims":["开发者工具","效率工具","AI基础设施"],"zone":"工具","funScore":0,"domainTags":["Web框架","Serverless","Node.js","开发工具"],"domainKey":"Web框架","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"trending","url":"https://github.com/nitrojs/nitro","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"witnessmenow/ESP32-Cheap-Yellow-Display","owner":"witnessmenow","name":"ESP32-Cheap-Yellow-Display","desc":"Building a community around a cheap ESP32 Display with a touch screen","summaryCn":"几十块钱就能拥有一个带触摸屏的复古掌机，支持运行 Rust 代码","reasonCn":"这是一个基于 ESP32 芯片的超低成本开发板项目，主打触控屏和 Rust 开发体验。它利用 Rust 的内存安全特性，在资源受限的嵌入式环境中实现高效、可靠的图形界面应用开发。项目通过极简的硬件设计和开放的软件生态，降低了入门门槛，让开发者能快速构建物联网、嵌入式娱乐或实用工具等应用。","stars":4470,"starGrowth":4,"createdAt":"2023-05-19T20:45:51Z","silentRounds":0,"language":"Rust","topics":[],"aiDims":["非AI-实用","创意工具","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["嵌入式开发","Rust","ESP32","硬件开发"],"domainKey":"嵌入式开发","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/witnessmenow/ESP32-Cheap-Yellow-Display","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"deepseek-ai/DeepEP","owner":"deepseek-ai","name":"DeepEP","desc":"","summaryCn":"DeepEP（DeepEveryParallel）是一款专为现代机器学习训练和推理打造的高性能","reasonCn":"该库的核心亮点在于其极低开销的JIT（Just-In-Time）编译机制。所有GPU内核都是在运行时动态编译的，这意味着开发者在安装DeepEP时完全不需要进行耗时的CUDA编译步骤，极大地简化了部署流程。这种设计不仅让安装变得轻量级，还确保了内核针对当前具体的硬件配置和参数进行了即时优","stars":10183,"starGrowth":2,"createdAt":"2025-02-17T01:33:04Z","silentRounds":0,"language":"Cuda","topics":[],"aiDims":["AI基础设施","推理引擎"],"zone":"AI","funScore":0,"domainTags":["专家并行","分布式训练","高性能计算"],"domainKey":"专家并行","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"推理引擎","source":"llm","weight":1},{"name":"Cuda","source":"language","weight":0.5}],"aiScore":0.47000000000000003,"source":"search","url":"https://github.com/deepseek-ai/DeepEP","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"Moenupa/VTCBench","owner":"Moenupa","name":"VTCBench","desc":"[EMNLP'26] Code and data for VTCBench, a VLM benchmark for long-context understanding capabilities under vision-text compression paradigm.","summaryCn":"评估AI能否读懂超长文档和图片的综合测试题","reasonCn":"这是EMNLP 2026论文的配套项目，提供了一个叫VTCBench的基准测试，专门用来评估视觉语言模型（VLM）在长上下文场景下的理解能力。它基于视觉-文本压缩范式，模拟真实世界中信息被大幅压缩后模型仍能准确理解的情况。项目包含完整的数据集和代码，支持主流多模态大模型的评测，适合研究者和开发者检验自己模型的长文本处理能力。","stars":27,"starGrowth":0,"createdAt":"2025-11-03T02:46:34Z","silentRounds":28,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","模型与训练","多模态Agent"],"zone":"工具","funScore":0.5,"domainTags":["多模态模型","VLM评测","长上下文","学术基准"],"domainKey":"多模态模型","pushedAt":"2026-08-27T03:16:17Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.68,"source":"search","url":"https://github.com/Moenupa/VTCBench","ts":"2026-08-27T03:16:17Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"agalwood/Motrix","owner":"agalwood","name":"Motrix","desc":"A full-featured download manager.","summaryCn":"一个全能且颜值在线的下载神器，支持BT、磁力、HTTP。","reasonCn":"Motrix 是一款基于 Electron 的开源下载管理器，支持 BT、磁力链接和 HTTP 等多种协议。它界面简洁美观，功能强大，可以同时下载多个任务，断点续传，自动重试。Motrix 采用 TypeScript 开发，跨平台支持 Windows、Mac 和 Linux，无需安装即可运行，非常适合需要高效下载资源的开发者、设计师和学生。","stars":55788,"starGrowth":16,"createdAt":"2018-12-18T11:45:05Z","silentRounds":0,"language":"TypeScript","topics":["motrix","aria2","macos","windows","linux","magnet","electron","bt","mac","download","torrent"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["下载管理器","BT种子","Aria2","Electron"],"domainKey":"下载管理器","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"motrix","source":"github","weight":0.7},{"name":"aria2","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"magnet","source":"github","weight":0.7},{"name":"electron","source":"github","weight":0.7},{"name":"bt","source":"github","weight":0.7},{"name":"mac","source":"github","weight":0.7},{"name":"download","source":"github","weight":0.7},{"name":"torrent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/agalwood/Motrix","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.04886040008604001,"copyOk":true},{"repo":"LadybirdBrowser/ladybird","owner":"LadybirdBrowser","name":"ladybird","desc":"Truly independent web browser","summaryCn":"从零手写浏览器内核，没有 Chromium 代码的独立浏览器项目","reasonCn":"Ladybird 是一个完全从零开始编写的现代 Web 浏览器，不使用 Chromium 或 Gecko 的任何代码。它用 C++ 编写，目标是实现标准的 HTML/CSS/JavaScript 渲染。这个项目适合想了解浏览器内部原理的开发者，或者对 Web 引擎感兴趣的技术爱好者。目前已有数万 star，社区活跃，持续迭代中。","stars":66277,"starGrowth":4,"createdAt":"2024-05-30T09:18:10Z","silentRounds":0,"language":"C++","topics":["browser","browser-engine"],"aiDims":["非AI-好玩","开发者工具","非AI-实用"],"zone":"工具","funScore":0.5,"domainTags":["浏览器","Web引擎","C++","内核开发","浏览器开发"],"domainKey":"浏览器","pushedAt":"2026-08-11T09:45:50.026Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"browser","source":"github","weight":0.7},{"name":"browser-engine","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/LadybirdBrowser/ladybird","ts":"2026-08-11T09:45:50.026Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Acerx-AMJ/Sandbox-2D","owner":"Acerx-AMJ","name":"Sandbox-2D","desc":"Sandbox-2D is an open-source 2D sandbox survival game where you can build, destroy and explore.","summaryCn":"用C++和raylib从零搓出来的开源2D像素风沙盒生存小游戏","reasonCn":"从技术角度来看，它没有使用那些臃肿的商业游戏引擎，而是直接基于轻量级的raylib图形库开发。raylib以简单纯粹著称，这使得整个项目的架构非常清晰，没有复杂的依赖项。渲染循环、物理碰撞检测、网格地图生成等核心逻辑都用现代C++实现，运行速度极快，内存占用极小，对于想学习游","stars":12,"starGrowth":0,"createdAt":"2025-11-08T07:49:53Z","silentRounds":18,"language":"C++","topics":["raylib"],"aiDims":["游戏","非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["2D游戏","沙盒生存","raylib","C++"],"domainKey":"2D游戏","pushedAt":"2026-08-27T12:56:10Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"raylib","source":"github","weight":0.7},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/Acerx-AMJ/Sandbox-2D","ts":"2026-08-27T12:56:10Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"amirmushichge/video-states-website","owner":"amirmushichge","name":"video-states-website","desc":"Build a cinematic click-driven video website: LTX footage, image/video prompts, coding-agent build spec, workflow and seam-safe playback guide.","summaryCn":"输入电影感提示词就能自动生成的点击互动视频网站","reasonCn":"这是一个结合AI生成与前端交互的创意建站工具。它利用LTX视频模型将文字提示转化为电影级画面，并通过JavaScript实现状态驱动的点击交互。项目提供了完整的编码代理构建规范和安全播放指南，让用户能轻松搭建具有叙事感的动态网页，无需手动剪辑视频，极大降低了视频网站的制作门槛。","stars":141,"starGrowth":0,"createdAt":"2026-09-04T15:24:50Z","silentRounds":11,"language":"JavaScript","topics":["creative-coding"],"aiDims":["创意工具","AI应用"],"zone":"创意","funScore":1,"domainTags":["创意编程","交互视频","WebGL","AI生成"],"domainKey":"创意编程","pushedAt":"2026-09-04T15:28:02Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"creative-coding","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/amirmushichge/video-states-website","ts":"2026-09-04T15:28:02Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"alacritty/alacritty","owner":"alacritty","name":"alacritty","desc":"A cross-platform, OpenGL terminal emulator.","summaryCn":"让终端变快又漂亮的黑科技，程序员必备神器","reasonCn":"Alacritty 是一款基于 Rust 的跨平台终端模拟器，采用 OpenGL 加速渲染，拥有极高的性能和流畅的动画效果。它支持丰富的主题定制，兼容 iTerm2 等流行终端的配置文件，不仅解决了传统终端卡顿的问题，还提供了清爽的视觉体验。","stars":65791,"starGrowth":4,"createdAt":"2016-02-18T05:02:30Z","silentRounds":0,"language":"Rust","topics":["terminal-emulators","opengl","gpu","rust","vte","terminal","linux","macos","windows","bsd"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["终端模拟器","Rust","GPU渲染","跨平台","开发工具"],"domainKey":"终端模拟器","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"terminal-emulators","source":"github","weight":0.7},{"name":"opengl","source":"github","weight":0.7},{"name":"gpu","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"vte","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"windows","source":"github","weight":0.7},{"name":"bsd","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/alacritty/alacritty","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"terkelg/awesome-creative-coding","owner":"terkelg","name":"awesome-creative-coding","desc":"Creative Coding: Generative Art, Data visualization, Interaction Design, Resources.","summaryCn":"创意编程资源大宝库，用代码画画的入门指南","reasonCn":"收录大量创意编程资源，涵盖生成艺术、数据可视化、交互设计、声音可视化等领域。通过代码创作视觉效果、动画和交互体验，让程序员和设计师探索技术与艺术的交叉点。资源库中包括各种项目、教程、书籍、工具和在线实例，适合从入门到进阶的学习路径。社区还收录了大量创意展示作品，激发创作灵感。","stars":15357,"starGrowth":3,"createdAt":"2016-11-01T23:50:27Z","silentRounds":0,"language":"HTML","topics":["design","data-visualization"],"aiDims":["可视化","创意工具","非AI-好玩"],"zone":"资源","funScore":0,"domainTags":["创意编程","生成艺术","数据可视化","教程资源"],"domainKey":"创意编程","pushedAt":"2026-07-21T15:01:30Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"data-visualization","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/terkelg/awesome-creative-coding","ts":"2026-07-21T15:01:30Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"xwmx/nb","owner":"xwmx","name":"nb","desc":"CLI and local web plain text note‑taking, bookmarking, and archiving with linking, tagging, filtering, search, Git versioning & syncing, Pandoc conversion, + more, in a single portable script.","summaryCn":"纯文本笔记神器，本地存储零隐私泄露风险，一句话搞定想法管理","reasonCn":"xwmx/nb是一个零依赖的单文件Shell脚本，提供CLI和Web两种界面，支持纯文本笔记创建、标签分类、全文搜索、双向链接和Git版本管理。内置Pandoc转换能力，可一键导出PDF或Markdown，配合内置搜索和过滤功能，构建个人知识库得心应手。","stars":8406,"starGrowth":0,"createdAt":"2016-01-12T17:27:26Z","silentRounds":7,"language":"Shell","topics":["command-line","git","note-taking"],"aiDims":["非AI-实用","效率工具","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["笔记管理","Git","命令行","个人知识库"],"domainKey":"笔记管理","pushedAt":"2026-08-26T00:51:47Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"git","source":"github","weight":0.7},{"name":"note-taking","source":"github","weight":0.7},{"name":"Shell","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/xwmx/nb","ts":"2026-08-26T00:51:47Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"flutter/agent-plugins","owner":"flutter","name":"agent-plugins","desc":"","summaryCn":"Flutter的AI Agent插件集，让移动端也能跑Agent应用","reasonCn":"技术亮点在于它对MCP（Model Context Protocol）的支持，这是目前AI Agent工具调用的标准协议。这意味着你的Flutter应用可以调用任意兼容MCP的工具链，包括文件系统、浏览器自动化、数据库查询等等。同时它还支持多模态输入，图片、语音都能喂给模型，这在移动端场景下特别实用。","stars":2985,"starGrowth":3,"createdAt":"2026-02-25T22:22:07Z","silentRounds":0,"language":"Dart","topics":[],"aiDims":["AI Agent","开发者工具"],"zone":"AI","funScore":0.5,"domainTags":["AI Agent","Flutter","MCP协议","移动端开发"],"domainKey":"AI Agent","pushedAt":"2026-08-18T20:19:17Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"Dart","source":"language","weight":0.5}],"aiScore":0.65,"source":"bigbro","url":"https://github.com/flutter/agent-plugins","ts":"2026-08-18T20:19:17Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tree-sitter/tree-sitter","owner":"tree-sitter","name":"tree-sitter","desc":"An incremental parsing system for programming tools","summaryCn":"像给代码做手术的精密手术刀，解析语法超级快","reasonCn":"这是一个用Rust编写的增量解析器生成器，专为编程工具打造。它采用递归下降解析技术，能快速构建抽象语法树（AST）。其核心优势在于增量解析，即只重新解析修改过的代码部分，极大提高了编辑器等工具的响应速度。不同于传统工具每次都全量分析，它能精准定位语法错误，支持现代编程语言特性，是代码分析、静态检查和智能编辑器的基石。","stars":27022,"starGrowth":4,"createdAt":"2013-11-06T06:16:00Z","silentRounds":0,"language":"Rust","topics":["incremental","parsing","c","tree-sitter","rust","parser","wasm"],"aiDims":["代码助手","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["代码解析","语法分析","编辑器插件","AST"],"domainKey":"代码解析","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"incremental","source":"github","weight":0.7},{"name":"parsing","source":"github","weight":0.7},{"name":"c","source":"github","weight":0.7},{"name":"tree-sitter","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"parser","source":"github","weight":0.7},{"name":"wasm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/tree-sitter/tree-sitter","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mojs/mojs","owner":"mojs","name":"mojs","desc":"The motion graphics toolbelt for the web","summaryCn":"网页动效设计神器，轻松实现各种炫酷的动画效果，告别复杂代码","reasonCn":"这是一个专门用于创建网页动效的工具包，提供了一套简单易用的 API 来生成各种动画，如粒子效果、路径动画和形状变换。它不需要复杂的编程，通过配置即可实现流畅的视觉效果，非常适合前端设计师和想要提升网页交互体验的开发者，让网页动起来变得轻而易举。","stars":18779,"starGrowth":1,"createdAt":"2014-06-26T08:49:14Z","silentRounds":2,"language":"CoffeeScript","topics":["design"],"aiDims":["非AI-实用","创意工具","可视化"],"zone":"创意","funScore":1,"domainTags":["动效设计","JavaScript","运动图形","前端开发","动画库"],"domainKey":"动效设计","pushedAt":"2026-07-30T16:54:33Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"CoffeeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/mojs/mojs","ts":"2026-07-30T16:54:33Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chenxin-dlut/codex-anime-pets","owner":"chenxin-dlut","name":"codex-anime-pets","desc":"Tiny chibi companions for the Codex desktop app","summaryCn":"想象一下，当你熬夜改Bug或者疯狂码字时，屏幕角落里有一只像素小猫或者Q版","reasonCn":"从技术实现上看，该项目利用了Codex应用的插件扩展机制，核心逻辑在于基于像素贴图的渲染与定时触发事件。开发者巧妙地将动画帧序列嵌入到UI视图层，通过CSS动画或JS定时器控制宠物的随机行为，比如眨眼、走动或等待。","stars":13,"starGrowth":0,"createdAt":"2026-05-05T09:21:19Z","silentRounds":13,"language":"","topics":["pixel-art"],"aiDims":["非AI-好玩","创意工具"],"zone":"创意","funScore":1,"domainTags":["像素艺术","IDE插件","桌面美化","情感陪伴"],"domainKey":"像素艺术","pushedAt":"2026-08-19T05:32:32Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"pixel-art","source":"github","weight":0.7}],"aiScore":0.65,"source":"search","url":"https://github.com/chenxin-dlut/codex-anime-pets","ts":"2026-08-19T05:32:32Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"NoeFabris/opencode-antigravity-auth","owner":"NoeFabris","name":"opencode-antigravity-auth","desc":"Enable Opencode to authenticate against Antigravity (Google's IDE) via OAuth so you can use Antigravity rate limits and access models like gemini-3-pro and claude-opus-4-5-thinking with your Google credentials.","summaryCn":"让命令行工具直接调用Gemini和Claude大模型","reasonCn":"这个项目把Opencode编码工具和Google的Antigravity IDE通过OAuth打通，让你用Google账号就能调用Gemini-3-Pro、Claude-Opus-4-5-Thinking等多个顶级AI模型。不用再到处申请API Key，一个账号串起多个AI能力，编码时随时召唤最强助手，省去了繁琐的认证配置流程。","stars":10984,"starGrowth":0,"createdAt":"2025-12-09T23:46:18Z","silentRounds":29,"language":"TypeScript","topics":["claude","gemini","google","opencode"],"aiDims":["代码助手","工具调用","AI应用"],"zone":"AI","funScore":0.5,"domainTags":["AI工具","命令行","OAuth认证","Gemini","Claude"],"domainKey":"AI工具","pushedAt":"2026-06-25T18:26:55Z","tags":[{"name":"代码助手","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"claude","source":"github","weight":0.7},{"name":"gemini","source":"github","weight":0.7},{"name":"google","source":"github","weight":0.7},{"name":"opencode","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.68,"source":"bigbro","url":"https://github.com/NoeFabris/opencode-antigravity-auth","ts":"2026-06-25T18:26:55Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"carbon-design-system/carbon","owner":"carbon-design-system","name":"carbon","desc":"A design system built by IBM","summaryCn":"IBM出的官方设计系统，做网页UI的神器","reasonCn":"由IBM开源的官方设计系统，提供一套完整的设计规范和组件库。基于Google Material Design理念，拥有严谨的网格系统和原子化设计思想，支持React、Angular等主流前端框架，帮助企业快速构建专业、一致且可访问的Web应用界面。","stars":9478,"starGrowth":2,"createdAt":"2017-03-13T14:23:59Z","silentRounds":0,"language":"JavaScript","topics":["scss","design-system","react","javascript","hacktoberfest"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["设计系统","UI组件","React","前端框架"],"domainKey":"设计系统","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"scss","source":"github","weight":0.7},{"name":"design-system","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"hacktoberfest","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/carbon-design-system/carbon","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Kanaries/Rath","owner":"Kanaries","name":"Rath","desc":"Next generation of automated data exploratory analysis and visualization platform.","summaryCn":"会自己读数据的智能可视化工具，自动生成图表，告别写代码画图的痛苦","reasonCn":"Rath是开源的自动化数据分析与可视化平台，类似BI界的Copilot。它不需要写代码，直接上传数据，AI会自动分析数据结构并生成多种可视化图表。支持交互式探索，自动推荐最佳视图，还能生成Python代码供二次修改。适合非技术背景的数据分析师、需要快速出报告的业务人员，以及想验证数据的开发者。","stars":4686,"starGrowth":0,"createdAt":"2019-08-28T14:21:25Z","silentRounds":5,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用","开发者工具"],"zone":"AI","funScore":0.6,"domainTags":["数据可视化","自动化分析","BI工具","AI辅助"],"domainKey":"数据可视化","pushedAt":"2026-08-14T21:24:23Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/Kanaries/Rath","ts":"2026-08-14T21:24:23Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"paulburgess1357/nvim-mcp","owner":"paulburgess1357","name":"nvim-mcp","desc":"MCP server that connects AI agents to your running Neovim instance via msgpack-RPC — no plugins required.","summaryCn":"让Neovim成为AI Agent的外设，不用装插件就能操控你的编辑环境","reasonCn":"技术实现上，它利用msgpack-RPC协议与Neovim通信，不需要安装任何插件，降低了使用门槛。MCP是Model Context Protocol的缩写，是AI Agent连接外部工具的标准协议，这个项目让Neovim成为MCP生态的一部分。AI Agent可以通过这个服务器读取文件内容、执行shell命令、创建和修改缓冲区","stars":63,"starGrowth":0,"createdAt":"2026-04-05T01:56:42Z","silentRounds":28,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","工具调用","开发者工具"],"zone":"AI","funScore":1,"domainTags":["Neovim","MCP协议","AI Agent","编辑器"],"domainKey":"AI Agent","pushedAt":"2026-08-25T13:52:21Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/paulburgess1357/nvim-mcp","ts":"2026-08-25T13:52:21Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Eigenwise/atomic-agents","owner":"Eigenwise","name":"atomic-agents","desc":"Building AI agents, atomically","summaryCn":"原子化AI智能体框架，手把手教你搭建灵活可靠的Agent","reasonCn":"Atomic Agents是一个专注于构建模块化AI Agent的Python框架，它将复杂的智能体逻辑拆分为独立的原子组件。相比大而全的框架，它更轻量、更灵活，开发者可以自由组合搜索、计算、记忆等工具，非常适合想要深入理解Agent原理或构建定制化多步骤工作流的程序员。","stars":6254,"starGrowth":0,"createdAt":"2024-06-03T08:24:07Z","silentRounds":1,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","工具调用","AI基础设施"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","框架","LLM应用","工作流"],"domainKey":"AI Agent","pushedAt":"2026-08-24T12:20:08Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/Eigenwise/atomic-agents","ts":"2026-08-24T12:20:08Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"HearthSim/Hearthstone-Deck-Tracker","owner":"HearthSim","name":"Hearthstone-Deck-Tracker","desc":"A deck tracker and deck manager for Hearthstone on Windows","summaryCn":"炉石传说玩家的神器，自动记录卡牌打出和牌库管理的贴心助手","reasonCn":"这是炉石传说玩家必装的游戏辅助工具，能自动追踪对手打出的卡牌、记录墓地和牌库，帮助你判断对手剩余牌型和手牌。它支持多个数据统计面板，显示英雄伤害、每回合费用、回合数等关键信息。同时提供卡组管理功能，方便你整理和分享收藏的所有卡组。作为一个C#开发的Windows应用，界面美观、功能强大，是炉石玩家的实用利器。","stars":4913,"starGrowth":0,"createdAt":"2014-05-29T18:44:12Z","silentRounds":27,"language":"C#","topics":["game"],"aiDims":["游戏","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["炉石传说","卡组管理","屏幕识别","桌面应用","游戏辅助"],"domainKey":"炉石传说","pushedAt":"2026-08-14T20:25:42Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/HearthSim/Hearthstone-Deck-Tracker","ts":"2026-08-14T20:25:42Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ilya-zlobintsev/LACT","owner":"ilya-zlobintsev","name":"LACT","desc":"Linux GPU Configuration And Monitoring Tool","summaryCn":"用命令行搞定 Linux 显卡配置的神器","reasonCn":"LACT 是一款基于 Rust 编写的 Linux 显卡配置与监控工具，能够替代复杂的 nvidia-smi 和复杂的内核模块配置。它通过用户态驱动程序，实现了对显卡温度、风扇、功耗和显存频率的实时监控，支持动态调整风扇曲线和超频参数，让用户在不重启的情况下优化显卡性能。","stars":5636,"starGrowth":1,"createdAt":"2020-10-17T14:11:55Z","silentRounds":2,"language":"Rust","topics":["amdgpu","linux","rust","gpu","hactoberfest","nvidia"],"aiDims":["非AI-实用","开发工具","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["GPU监控","Linux","Rust","显卡调优","nvidia"],"domainKey":"GPU监控","pushedAt":"2026-08-15T00:57:24.351Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"amdgpu","source":"github","weight":0.7},{"name":"linux","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"gpu","source":"github","weight":0.7},{"name":"hactoberfest","source":"github","weight":0.7},{"name":"nvidia","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/ilya-zlobintsev/LACT","ts":"2026-08-15T00:57:24.351Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"antonmedv/fx","owner":"antonmedv","name":"fx","desc":"Terminal JSON viewer & processor","summaryCn":"终端里看JSON的神器，美化排版比网页还漂亮","reasonCn":"这是一个基于Go语言开发的终端JSON查看和处理工具，支持管道输入和交互式浏览。它能把杂乱的JSON数据自动美化成带颜色的高亮格式，还能通过命令行过滤器处理数据。比用浏览器插件方便得多，直接在终端就能查看API返回的原始数据，效率非常高。","stars":20639,"starGrowth":2,"createdAt":"2018-01-25T17:29:43Z","silentRounds":0,"language":"Go","topics":["command-line"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["JSON处理","终端工具","API调试","数据可视化"],"domainKey":"JSON处理","pushedAt":"2026-08-26T14:49:32Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"command-line","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/antonmedv/fx","ts":"2026-08-26T14:49:32Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"microsoft/multilspy","owner":"microsoft","name":"multilspy","desc":"multilspy is a lsp client library in Python intended to be used to build applications around language servers.","summaryCn":"微软造的桥梁，让AI读懂各种编程语言的'内部结构'","reasonCn":"这是一个Python写的LSP客户端库，LSP是微软主导的代码语言服务器协议，能精准回答代码里变量定义在哪、函数怎么调用、类型是什么。支持Python、JavaScript、Go、Java等多种语言，统一接口让AI工具读懂代码结构，而不是只看文本。安装：pip install multilspy。微软开源，GitHub 606星，持续迭代中。","stars":610,"starGrowth":0,"createdAt":"2024-08-07T12:58:24Z","silentRounds":13,"language":"Python","topics":["llm-model"],"aiDims":["开发者工具","AI应用","代码助手"],"zone":"工具","funScore":0,"domainTags":["LSP","代码分析","语言服务器","AI编程"],"domainKey":"LSP","pushedAt":"2026-08-27T13:11:56Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/microsoft/multilspy","ts":"2026-08-27T13:11:56Z","category":"tool","momentum":[],"fromOfficial":true,"heatScore":0,"copyOk":false},{"repo":"unionlabs/union","owner":"unionlabs","name":"union","desc":"The trust-minimized, zero-knowledge bridging protocol, designed for censorship resistance, extremely high security, and usage in decentralized finance.","summaryCn":"跨链桥里的安全卫士，用零知识证明保护你的资金。","reasonCn":"Union是一个基于零知识证明的跨链桥协议，旨在实现去中心化金融中的安全资产转移。它通过将跨链交易打包成证明，在不暴露具体交易细节的情况下验证交易合法性，从而防止黑客攻击和资产盗窃，是DeFi安全领域的重要创新。","stars":73793,"starGrowth":0,"createdAt":"2023-05-16T22:33:39Z","silentRounds":28,"language":"Rust","topics":["blockchain","cosmos","evm","golang","rust","solidity","ethereum","cosmwasm","nix","zero-knowledge","indexer","prover","relayer","astro","svelte","typescript","typescipt","move","interoperability"],"aiDims":["AI基础设施","安全工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["跨链桥","零知识证明","DeFi","区块链","安全协议"],"domainKey":"跨链桥","pushedAt":"2026-08-16T00:54:51.023Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"blockchain","source":"github","weight":0.7},{"name":"cosmos","source":"github","weight":0.7},{"name":"evm","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"solidity","source":"github","weight":0.7},{"name":"ethereum","source":"github","weight":0.7},{"name":"cosmwasm","source":"github","weight":0.7},{"name":"nix","source":"github","weight":0.7},{"name":"zero-knowledge","source":"github","weight":0.7},{"name":"indexer","source":"github","weight":0.7},{"name":"prover","source":"github","weight":0.7},{"name":"relayer","source":"github","weight":0.7},{"name":"astro","source":"github","weight":0.7},{"name":"svelte","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"typescipt","source":"github","weight":0.7},{"name":"move","source":"github","weight":0.7},{"name":"interoperability","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.65,"source":"trending","url":"https://github.com/unionlabs/union","ts":"2026-08-16T00:54:51.023Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"drona23/claude-token-efficient","owner":"drona23","name":"claude-token-efficient","desc":"One CLAUDE.md file. Keeps Claude responses terse. Reduces output verbosity on heavy workflows. Drop-in, no code changes.","summaryCn":"一个Claude配置秘籍，让AI废话少说，句句干货","reasonCn":"这个项目通过一个CLAUDE.md配置文件来控制Claude API的响应风格，自动注入精简指令。安装极其简单，只需把配置文档放到项目根目录，后续调用Claude时会自动生效。能显著减少Token消耗，适合频繁调用API的开发者和需要批量处理数据的场景，避免被冗长回复浪费预算。","stars":6051,"starGrowth":1,"createdAt":"2026-03-30T12:59:07Z","silentRounds":1,"language":"Python","topics":[],"aiDims":["提示工程"],"zone":"AI","funScore":0,"domainTags":["提示词工程","API调用","Token优化","Claude"],"domainKey":"提示词工程","pushedAt":"2026-06-16T03:10:06Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"bigbro","url":"https://github.com/drona23/claude-token-efficient","ts":"2026-06-16T03:10:06Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"firecrawl/pdf-inspector","owner":"firecrawl","name":"pdf-inspector","desc":"Fast Rust library for PDF inspection, classification, and text extraction. Intelligently detects scanned vs text-based PDFs to enable smart routing decisions.","summaryCn":"专门用来智能识别PDF是扫描件还是文本，方便后续处理","reasonCn":"技术上，这个库用Rust编写，执行速度非常快。它通过检测PDF中是否包含可提取的文本内容来判断类型：如果有文本层，就是文本型PDF；如果完全是图片，就是扫描版。它还支持分类功能，能识别PDF的类别，比如合同、发票、论文等。","stars":19297,"starGrowth":5,"createdAt":"2026-02-06T19:44:47Z","silentRounds":0,"language":"Rust","topics":["markdown","nodejs","pdf","pdf-extraction","pdf-parser","python","rust","text-extraction","ocr-routing","pdf-classification"],"aiDims":["文档理解","工具调用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["PDF处理","Rust","文档分类"],"domainKey":"PDF处理","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"文档理解","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"markdown","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"pdf","source":"github","weight":0.7},{"name":"pdf-extraction","source":"github","weight":0.7},{"name":"pdf-parser","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"text-extraction","source":"github","weight":0.7},{"name":"ocr-routing","source":"github","weight":0.7},{"name":"pdf-classification","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/firecrawl/pdf-inspector","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.018932787479919156,"copyOk":true},{"repo":"mudassar531/hearsay","owner":"mudassar531","name":"hearsay","desc":"crawl4ai for video & audio — turn any YouTube video, podcast, or recording into clean timestamped markdown, or a verified TTS/STT training dataset. Runs locally.","summaryCn":"Hearsay解决了处理长视频最头疼的问题：怎么快速提取关键信息？","reasonCn":"它的最大特色是“实用性”。除了普通的转写（STT），它还能生成TTS（语音合成）训练所需的对话数据。这意味着你可以用它来制作自己的语音助手训练集，或者整理播客内容。所有处理都在本地完成，不用担心视频里的隐私信息泄露给云端。","stars":19,"starGrowth":0,"createdAt":"2026-06-13T01:31:57Z","silentRounds":20,"language":"Python","topics":["podcast"],"aiDims":["AI应用","非AI-实用","工具调用"],"zone":"工具","funScore":0,"domainTags":["视频转写","TTS数据集","本地处理","音频分析"],"domainKey":"视频转写","pushedAt":"2026-09-03T05:33:05Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"podcast","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/mudassar531/hearsay","ts":"2026-09-03T05:33:05Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"android/skills","owner":"android","name":"skills","desc":"","summaryCn":"Android团队写的自动化技能库，让手机操作像写代码一样简单","reasonCn":"这是Android官方开源的一套自动化框架，用Python定义各种手机操作技能。不用写复杂的App，通过代码就能控制手机界面、点击、滑动。适合做测试自动化、游戏脚本或者 repetitive tasks，把繁琐的手动操作变成一键执行。","stars":7501,"starGrowth":8,"createdAt":"2026-03-16T12:41:46Z","silentRounds":0,"language":"Python","topics":["skills"],"aiDims":["非AI-实用","开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["Android自动化","UI测试","手机脚本","ADB"],"domainKey":"Android自动化","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"skills","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/android/skills","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.03263948806825757,"copyOk":true},{"repo":"raphamorim/rio","owner":"raphamorim","name":"rio","desc":"A hardware-accelerated GPU terminal emulator focusing to run in desktops and browsers.","summaryCn":"能在浏览器里跑的GPU加速终端，开发者远程操作新选择","reasonCn":"用Rust编写的硬件加速终端模拟器，独特之处在于支持在浏览器中运行。通过WebGPU技术实现GPU渲染，同时提供桌面客户端和Web两种访问方式，支持SSH连接远程服务器，让开发者在任何设备上都拥有一个功能完整的终端环境。","stars":7532,"starGrowth":3,"createdAt":"2022-10-05T15:18:41Z","silentRounds":0,"language":"Rust","topics":["gpu-acceleration","rust","rust-lang","terminal","terminal-emulator","terminal-ui","vte","rio","rio-terminal","terminal-emulators","wgpu","metal","vulkan"],"aiDims":["开发者工具","效率工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["终端模拟器","WebGPU","SSH","浏览器"],"domainKey":"终端模拟器","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"gpu-acceleration","source":"github","weight":0.7},{"name":"rust","source":"github","weight":0.7},{"name":"rust-lang","source":"github","weight":0.7},{"name":"terminal","source":"github","weight":0.7},{"name":"terminal-emulator","source":"github","weight":0.7},{"name":"terminal-ui","source":"github","weight":0.7},{"name":"vte","source":"github","weight":0.7},{"name":"rio","source":"github","weight":0.7},{"name":"rio-terminal","source":"github","weight":0.7},{"name":"terminal-emulators","source":"github","weight":0.7},{"name":"wgpu","source":"github","weight":0.7},{"name":"metal","source":"github","weight":0.7},{"name":"vulkan","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/raphamorim/rio","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"erincatto/box2d","owner":"erincatto","name":"box2d","desc":"Box2D is a 2D physics engine for games","summaryCn":"做2D游戏物理的绕不开的经典引擎，Flappy Bird就用它","reasonCn":"Box2D是业界最成熟的2D物理引擎之一，专为游戏开发设计。它提供刚体动力学、碰撞检测和关节约束等核心功能，性能出色且跨平台支持广泛。从独立小游戏到商业项目都在用，文档完善、社区活跃，是学习物理引擎和开发2D游戏的必备工具。","stars":10366,"starGrowth":0,"createdAt":"2015-03-14T16:52:46Z","silentRounds":4,"language":"C","topics":[],"aiDims":["游戏","非AI-好玩","非AI-实用"],"zone":"创意","funScore":0,"domainTags":["物理引擎","游戏开发","2D游戏","刚体动力学"],"domainKey":"物理引擎","pushedAt":"2026-08-18T19:31:23Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"C","source":"language","weight":0.5}],"aiScore":0.62,"source":"bigbro","url":"https://github.com/erincatto/box2d","ts":"2026-08-18T19:31:23Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"megadose/holehe","owner":"megadose","name":"holehe","desc":"holehe allows you to check if the mail is used on different sites like twitter, instagram and will retrieve information on sites with the forgotten password function.","summaryCn":"一键查询邮箱到底注册了哪些网站，还能找回丢失的账号。","reasonCn":"这是一个纯Python开发的网络工具，主要利用邮件找回密码功能来探测邮箱在各大社交平台的注册情况。它支持Twitter、Instagram、Google、Telegram等多个主流网站，并能模拟找回密码流程，提取出该邮箱在这些平台上的账户信息。项目无需API密钥，通过HTTP请求和简单的解析逻辑实现，非常适合隐私安全检测。","stars":14984,"starGrowth":6,"createdAt":"2020-06-25T23:03:02Z","silentRounds":0,"language":"Python","topics":["information-gathering","osint","osint-tools","open-source-intelligence","python","email","social-network","instagram","twitter","trio","ebay","emails","osint-python","pypi","tellonym"],"aiDims":["非AI-实用","安全工具","非AI-好玩"],"zone":"工具","funScore":0.5,"domainTags":["网络爬虫","账号探测","隐私安全"],"domainKey":"网络爬虫","pushedAt":"2026-08-17T00:55:40.962Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"information-gathering","source":"github","weight":0.7},{"name":"osint","source":"github","weight":0.7},{"name":"osint-tools","source":"github","weight":0.7},{"name":"open-source-intelligence","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"email","source":"github","weight":0.7},{"name":"social-network","source":"github","weight":0.7},{"name":"instagram","source":"github","weight":0.7},{"name":"twitter","source":"github","weight":0.7},{"name":"trio","source":"github","weight":0.7},{"name":"ebay","source":"github","weight":0.7},{"name":"emails","source":"github","weight":0.7},{"name":"osint-python","source":"github","weight":0.7},{"name":"pypi","source":"github","weight":0.7},{"name":"tellonym","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/megadose/holehe","ts":"2026-08-17T00:55:40.962Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.019201281366791244,"copyOk":true},{"repo":"objectbox/objectbox-java","owner":"objectbox","name":"objectbox-java","desc":"Database for Android and JVM - first and fast, lightweight on-device vector database","summaryCn":"给Android手机装的轻量级向量数据库，端侧也能搜向量","reasonCn":"ObjectBox是专为Android和JVM设计的关系型数据库，支持对象-关系映射，查询速度极快且内存占用低。其新增的向量数据库功能支持在设备端存储和检索向量嵌入，可配合大模型实现本地RAG检索和语义搜索，适合移动端AI应用开发，社区Star超4600，API简洁易用，支持多种相似度算法。","stars":4623,"starGrowth":0,"createdAt":"2017-01-24T10:29:48Z","silentRounds":27,"language":"Java","topics":["mobile"],"aiDims":["向量数据库","开发者工具","AI应用"],"zone":"工具","funScore":0,"domainTags":["向量数据库","移动端","语义检索","Android"],"domainKey":"向量数据库","pushedAt":"2026-08-03T10:42:55Z","tags":[{"name":"向量数据库","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"mobile","source":"github","weight":0.7},{"name":"Java","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/objectbox/objectbox-java","ts":"2026-08-03T10:42:55Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"tsinghua-fib-lab/SmartAgent","owner":"tsinghua-fib-lab","name":"SmartAgent","desc":"The official repository of \"SmartAgent: Chain-of-User-Thought for Embodied Personalized Agent in Cyber World\".","summaryCn":"清华大神级AI Agent，让智能体理解用户思维完成任务","reasonCn":"这个项目就是朝着这个方向迈出的重要一步。SmartAgent是清华大学FIB实验室的研究项目，核心创新是提出了Chain-of-User-Thought（用户思维链）方法，让AI Agent在虚拟环境中学习理解用户的意图和行为模式，从而完成任务。","stars":28,"starGrowth":0,"createdAt":"2024-12-10T09:47:33Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["AI Agent","多模态Agent","模型与训练"],"zone":"AI","funScore":0.6,"domainTags":["AI Agent","Chain-of-Thought","强化学习","虚拟环境"],"domainKey":"AI Agent","pushedAt":"2026-07-27T04:35:27Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/tsinghua-fib-lab/SmartAgent","ts":"2026-07-27T04:35:27Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Hai-chao-Zhang/ThinkJEPA","owner":"Hai-chao-Zhang","name":"ThinkJEPA","desc":"ThinkJEPA: Empowering Latent World Models with Large Vision-Language Reasoning Model","summaryCn":"用大语言模型增强多模态推理能力的研究级项目","reasonCn":"ThinkJEPA是一个研究项目，探索如何将大视觉语言模型与JEPA（联合嵌入预测架构）结合，提升多模态推理能力。项目基于Python实现，涉及潜在世界模型和视觉语言推理的最新研究方向。虽然star数只有49个，说明项目还处于早期阶段或较为小众，但从技术方向来看很有潜力，适合对多模态AI和推理模型感兴趣的开发者跟进学习。","stars":57,"starGrowth":0,"createdAt":"2026-04-01T19:57:54Z","silentRounds":7,"language":"Python","topics":["llm-model"],"aiDims":["多模态Agent","大语言模型","AI应用"],"zone":"AI","funScore":0.6,"domainTags":["JEPA","多模态推理","视觉语言模型","潜在世界模型"],"domainKey":"JEPA","pushedAt":"2026-07-19T14:35:18Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/Hai-chao-Zhang/ThinkJEPA","ts":"2026-07-19T14:35:18Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chrisliu298/awesome-llm-unlearning","owner":"chrisliu298","name":"awesome-llm-unlearning","desc":"A resource repository for machine unlearning in large language models","summaryCn":"大模型也有记忆？这个项目教你怎么让它彻底忘掉敏感隐私数据。","reasonCn":"专为大语言模型遗忘机制整理的资源库，聚焦Machine Unlearning技术。收录多篇核心论文与开源实现，帮研究者快速掌握如何移除模型中的特定数据或敏感信息。虽然偏学术，但对关注AI隐私合规的团队极具参考价值，是入门该领域的优质导航站。","stars":627,"starGrowth":0,"createdAt":"2024-03-27T15:44:23Z","silentRounds":27,"language":"","topics":["llm-model"],"aiDims":["大语言模型","微调","学习资源"],"zone":"资源","funScore":0,"domainTags":["大模型遗忘","Machine Unlearning","论文合集","AI安全"],"domainKey":"大模型遗忘","pushedAt":"2026-08-06T09:51:49Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"微调","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.65,"source":"search","url":"https://github.com/chrisliu298/awesome-llm-unlearning","ts":"2026-08-06T09:51:49Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Abhisheklearn12/chess-engine","owner":"Abhisheklearn12","name":"chess-engine","desc":"Chess Engine","summaryCn":"用 Rust 打造的开源国际象棋引擎，快如闪电，随时对弈练棋。","reasonCn":"该项目用 Rust 实现了完整的国际象棋引擎，支持标准走法生成、合法性校验和局面评估，可作为学习棋类算法的教材或嵌入自研棋盘 UI。实现上采用位棋盘表示法，运算高效且内存占用低，兼容 Linux、macOS 与 Windows，适合跨平台使用。","stars":47,"starGrowth":0,"createdAt":"2025-11-11T17:32:11Z","silentRounds":17,"language":"Rust","topics":["chess"],"aiDims":["非AI-好玩","游戏"],"zone":"工具","funScore":0.5,"domainTags":["国际象棋","位棋盘","alpha-beta剪枝","Rust"],"domainKey":"国际象棋","pushedAt":"2026-08-13T20:12:48Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/Abhisheklearn12/chess-engine","ts":"2026-08-13T20:12:48Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"AnInsomniacy/motrix-next","owner":"AnInsomniacy","name":"motrix-next","desc":"A full-featured download manager — rebuilt from the ground up","summaryCn":"一款全能且开源的下载工具，支持协议多且界面清爽。","reasonCn":"Motrix 是一款基于 Electron 构建的全功能下载管理器，旨在提供极致的下载体验。它支持 HTTP、FTP、BitTorrent 等多种主流下载协议，并采用类 Safari 的清爽 UI 设计，让用户在下载文件时不再感到繁琐。项目从零开始重写，不仅优化了性能，还修复了旧版 Motrix 的诸多问题，是一款非常实用且美观的桌面级下载工具。","stars":10413,"starGrowth":7,"createdAt":"2026-03-04T00:46:49Z","silentRounds":0,"language":"JavaScript","topics":[],"aiDims":["非AI-实用","开发工具"],"zone":"工具","funScore":0,"domainTags":["下载管理","BitTorrent","跨平台","Electron"],"domainKey":"下载管理","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发工具","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/AnInsomniacy/motrix-next","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.027620051858254103,"copyOk":false},{"repo":"repomonkey/PhotoScribe","owner":"repomonkey","name":"PhotoScribe","desc":"AI-powered photo metadata generator using local models. Analyses your photographs and writes title, caption, and keywords directly to IPTC/XMP metadata","summaryCn":"用本地 AI 给照片自动生成标题和关键词，隐私安全无需联网","reasonCn":"PhotoScribe 是一个利用本地运行的人工智能模型为照片生成元数据的工具。它分析照片内容后自动写入标题、描述和关键词到 IPTC/XMP 标准格式，完全在本地运行，不上传照片到任何云端服务。支持多种本地模型后端，适合重视隐私的摄影师和内容创作者，一键批量处理整个相册。","stars":24,"starGrowth":0,"createdAt":"2026-04-27T01:55:03Z","silentRounds":11,"language":"Python","topics":["photography"],"aiDims":["AI应用","多模态Agent"],"zone":"AI","funScore":0,"domainTags":["本地AI","元数据生成","照片处理","隐私安全"],"domainKey":"本地AI","pushedAt":"2026-08-18T01:52:36Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"photography","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/repomonkey/PhotoScribe","ts":"2026-08-18T01:52:36Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"HatScripts/circle-flags","owner":"HatScripts","name":"circle-flags","desc":"A collection of 400+ minimal circular SVG country, state and language flags","summaryCn":"400多个圆滚滚的国旗小图标，做UI设计超实用","reasonCn":"汇集全球400多个国家的圆形SVG旗帜图标，每个图标都是极简路径设计，配色精准还原国旗原貌。支持国家和语言代码命名，方便开发者直接引用。相比庞大的通用图标库，这个项目专攻旗帜领域，体积小巧下载快，设计师和前端开发者都能轻松上手使用。","stars":1593,"starGrowth":0,"createdAt":"2018-12-05T12:22:56Z","silentRounds":14,"language":"HTML","topics":["design"],"aiDims":["创意工具","可视化","非AI-好玩"],"zone":"资源","funScore":0,"domainTags":["国旗图标","SVG矢量图","UI设计素材"],"domainKey":"国旗图标","pushedAt":"2026-07-26T05:09:00Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"可视化","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"design","source":"github","weight":0.7},{"name":"HTML","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/HatScripts/circle-flags","ts":"2026-07-26T05:09:00Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"zarzouram/image_captioning_with_transformers","owner":"zarzouram","name":"image_captioning_with_transformers","desc":"Pytorch implementation of image captioning using transformer-based model.","summaryCn":"如果你经常为给相册或者社交媒体配文而发愁，这个项目能帮你直接把照片变文","reasonCn":"核心架构采用了Encoder-Decoder的Transformer设计模式。Encoder部分通常使用预训练的视觉模型（如ViT或ResNet的变体）来捕捉图像的关键特征，Decoder则利用自注意力机制（Self-Attention）生成符合语法逻辑的描述序列。","stars":68,"starGrowth":0,"createdAt":"2021-09-26T08:01:45Z","silentRounds":17,"language":"Jupyter Notebook","topics":["transformers"],"aiDims":["多模态Agent","模型与训练"],"zone":"资源","funScore":0,"domainTags":["图像描述","Transformer","多模态","Jupyter"],"domainKey":"图像描述","pushedAt":"2026-08-17T12:03:20Z","tags":[{"name":"多模态Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"transformers","source":"github","weight":0.7},{"name":"Jupyter Notebook","source":"language","weight":0.5}],"aiScore":0.65,"source":"search","url":"https://github.com/zarzouram/image_captioning_with_transformers","ts":"2026-08-17T12:03:20Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"etcd-io/etcd","owner":"etcd-io","name":"etcd","desc":"Distributed reliable key-value store for the most critical data of a distributed system","summaryCn":"分布式系统里的超级记忆体，Kubernetes全靠它存配置","reasonCn":"etcd是分布式系统的基础设施级组件，基于Raft一致性算法实现高可用键值存储。它被广泛用于Kubernetes、CockroachDB、Cloudflare等核心系统，存储集群配置、服务发现信息、分布式锁等关键数据。支持事务性写操作、watch机制和Leader选举，保证了即使在节点故障时也能维持数据的强一致性。虽然偏底层基础设施，但它是支撑大量AI和现代应用跑起来的关键拼图。","stars":52306,"starGrowth":8,"createdAt":"2013-07-06T21:57:21Z","silentRounds":0,"language":"Go","topics":["database","kubernetes"],"aiDims":["AI基础设施"],"zone":"工具","funScore":0,"domainTags":["分布式存储","etcd","Raft","Kubernetes"],"domainKey":"分布式存储","pushedAt":"2026-09-23T12:55:34Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/etcd-io/etcd","ts":"2026-09-23T12:55:34Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.024458838374182695,"copyOk":false},{"repo":"davyxu/cellnet","owner":"davyxu","name":"cellnet","desc":"High performance, simple, extensible golang open source network library","summaryCn":"用Go语言快速搭建高并发游戏服务端的简易网络开发框架。","reasonCn":"cellnet 是一款基于 Go 语言打造的高性能事件驱动网络库，专为游戏服务器和长连接系统设计。它采用高扩展性的队列管道架构，解耦了会话管理、编码解码与消息路由机制。原生支持 Protobuf、JSON 以及自定义二进制协议，配合多线程事件队列调度，能够轻松应对海量高并发长连接通信需求。","stars":4055,"starGrowth":0,"createdAt":"2015-08-18T09:54:01Z","silentRounds":15,"language":"Go","topics":["game"],"aiDims":["开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["网络库","游戏服务端","高并发","Go语言"],"domainKey":"网络库","pushedAt":"2026-08-24T21:06:27Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"game","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/davyxu/cellnet","ts":"2026-08-24T21:06:27Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"cloud-hypervisor/cloud-hypervisor","owner":"cloud-hypervisor","name":"cloud-hypervisor","desc":"A Virtual Machine Monitor for modern Cloud workloads. Features include CPU, memory and device hotplug, support for running Windows and Linux guests, device offload with vhost-user and a minimal compact footprint. Written in Rust with a strong focus on security.","summaryCn":"Rust写的虚拟机神器，让云服务器跑得更飞快更省资源","reasonCn":"cloud-hypervisor是用Rust编写的轻量级虚拟机监控器，专为现代云工作负载设计。支持CPU、内存和设备热插拔，可同时运行Windows和Linux虚拟机。通过vhost-user实现设备卸载，显著降低主机开销。相比传统QEMU，启动更快、资源占用更少，适合容器化场景和AI训练集群的底层虚拟化需求。","stars":6251,"starGrowth":1,"createdAt":"2019-04-30T22:49:03Z","silentRounds":1,"language":"Rust","topics":["rust-vmm","cloud-workloads","virtualization","kvm"],"aiDims":["AI基础设施","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["虚拟化","Rust","云原生"],"domainKey":"虚拟化","pushedAt":"2026-08-09T00:58:42.474Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"rust-vmm","source":"github","weight":0.7},{"name":"cloud-workloads","source":"github","weight":0.7},{"name":"virtualization","source":"github","weight":0.7},{"name":"kvm","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.62,"source":"trending","url":"https://github.com/cloud-hypervisor/cloud-hypervisor","ts":"2026-08-09T00:58:42.474Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"ethereum/go-ethereum","owner":"ethereum","name":"go-ethereum","desc":"Go implementation of the Ethereum protocol","summaryCn":"以太坊的官方“心脏”，全球最火的区块链开源实现","reasonCn":"这是以太坊协议的官方 Go 语言实现，也是目前 GitHub 上最流行的区块链项目之一。它不仅实现了完整的以太坊虚拟机（EVM），还包含完整的客户端功能，如节点运行、挖矿、交易处理和智能合约执行。作为区块链领域的核心基础设施，它支持主网和各种测试网，被广泛应用于去中心化应用的开发、节点部署和学术研究中。","stars":51365,"starGrowth":2,"createdAt":"2013-12-26T13:05:46Z","silentRounds":0,"language":"Go","topics":["go","blockchain","ethereum","p2p","geth"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["以太坊","EVM","区块链","Go语言"],"domainKey":"以太坊","pushedAt":"2026-08-19T01:08:47.324Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"blockchain","source":"github","weight":0.7},{"name":"ethereum","source":"github","weight":0.7},{"name":"p2p","source":"github","weight":0.7},{"name":"geth","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/ethereum/go-ethereum","ts":"2026-08-19T01:08:47.324Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"mui/material-ui","owner":"mui","name":"material-ui","desc":"Material UI: Comprehensive React component library that implements Google's Material Design. Free forever.","summaryCn":"React界最流行的UI组件库，让网页开发像搭积木一样快","reasonCn":"这个项目解决了React开发中UI重复造轮子的痛点，直接提供一套工业级标准组件，从按钮到复杂的表格和图表一应俱全。开发者只需要关注业务逻辑，界面部分通过配置就能快速搭建，极大地提升了开发效率。它的核心优势在于高度可定制，基于CSS-in-JS技术，样式完全可控，同时保持了代码","stars":99085,"starGrowth":2,"createdAt":"2014-08-18T19:11:54Z","silentRounds":0,"language":"JavaScript","topics":["react-components","react","material-design","material-ui","design-system"],"aiDims":["开发者工具","效率工具"],"zone":"工具","funScore":0,"domainTags":["React","UI组件库","Material Design"],"domainKey":"React","pushedAt":"2026-08-14T01:23:41.092Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"react-components","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"material-design","source":"github","weight":0.7},{"name":"material-ui","source":"github","weight":0.7},{"name":"design-system","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/mui/material-ui","ts":"2026-08-14T01:23:41.092Z","category":"tool","momentum":["hot"],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"polius/FileSync","owner":"polius","name":"FileSync","desc":"Send files from one device to many in real-time.","summaryCn":"一个让文件在多台设备间实时同步的神器，随时随地都能传文件。","reasonCn":"这是一个基于WebSocket的实时文件同步工具，支持多设备同时在线监听并接收文件更新。项目使用JavaScript编写，无依赖后端，通过P2P直连实现高速传输。核心功能包括实时文件传输、自动重连机制和跨平台支持，解决了传统FTP和云盘同步慢的痛点。","stars":1774,"starGrowth":2,"createdAt":"2023-07-12T22:05:42Z","silentRounds":0,"language":"JavaScript","topics":["fastapi","fileshare","javascript","self-hosted","webrtc","docker"],"aiDims":["非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["文件同步","WebSocket","P2P传输"],"domainKey":"文件同步","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"fastapi","source":"github","weight":0.7},{"name":"fileshare","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"webrtc","source":"github","weight":0.7},{"name":"docker","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/polius/FileSync","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"esengine/estella","owner":"esengine","name":"estella","desc":"A fast 2D game engine — TypeScript SDK, C++/WebAssembly core, visual editor. Ship one project to web, desktop, WeChat MiniGames, playable ads, and native Android / iOS.","summaryCn":"用代码和可视化编辑器打造2D游戏，一键发布到网页手机小游戏","reasonCn":"TypeScript SDK加上C++与WebAssembly核心，让开发效率与运行性能兼得。内置可视化编辑器降低2D游戏开发门槛。支持一次开发多端发布，涵盖Web、桌面、微信小游戏、可玩广告及原生iOS与安卓。生态对独立开发者友好，跨平台能力突出，适合快速验证创意并推向市场。","stars":679,"starGrowth":1,"createdAt":"2026-01-25T07:34:59Z","silentRounds":3,"language":"TypeScript","topics":["2d-game-engine","android","cpp","ecs","game-editor","game-engine","gamedev","ios","typescript","webassembly","webgl","webgpu","wechat-minigame"],"aiDims":["游戏","创意工具","非AI-好玩"],"zone":"工具","funScore":0,"domainTags":["2D游戏引擎","TypeScript","跨平台发布"],"domainKey":"2D游戏引擎","pushedAt":"2026-08-19T00:57:46Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"创意工具","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"2d-game-engine","source":"github","weight":0.7},{"name":"android","source":"github","weight":0.7},{"name":"cpp","source":"github","weight":0.7},{"name":"ecs","source":"github","weight":0.7},{"name":"game-editor","source":"github","weight":0.7},{"name":"game-engine","source":"github","weight":0.7},{"name":"gamedev","source":"github","weight":0.7},{"name":"ios","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"webassembly","source":"github","weight":0.7},{"name":"webgl","source":"github","weight":0.7},{"name":"webgpu","source":"github","weight":0.7},{"name":"wechat-minigame","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"bigbro","url":"https://github.com/esengine/estella","ts":"2026-08-19T00:57:46Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"a2fpga/a2fpga_core","owner":"a2fpga","name":"a2fpga_core","desc":"Apple II FPGA Co-Processor","summaryCn":"给老古董苹果II加装FPGA芯片，瞬间变身神速超级计算机。","reasonCn":"这是给经典苹果II电脑加装FPGA加速协处理器的开源项目，通过Verilog硬核逻辑实现，让这台老机焕发新生。项目提供详细的原理图和工程文件，支持在FPGA板上运行多种模拟器，甚至可以直接在苹果II上运行现代操作系统。对于复古极客和硬件爱好者来说，这是连接过去与未来的硬核实验。","stars":76,"starGrowth":0,"createdAt":"2024-02-25T18:51:09Z","silentRounds":16,"language":"Verilog","topics":["retrocomputing"],"aiDims":["非AI-好玩","retrocomputing"],"zone":"创意","funScore":1,"domainTags":["苹果II","FPGA","硬件加速","复古计算","Verilog"],"domainKey":"苹果II","pushedAt":"2026-08-28T04:27:54Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"retrocomputing","source":"llm","weight":1},{"name":"Verilog","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/a2fpga/a2fpga_core","ts":"2026-08-28T04:27:54Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"cocos/cocos-cli","owner":"cocos","name":"cocos-cli","desc":"COCOS CLI","summaryCn":"无需GUI，命令行即可快速初始化、构建、发布Cocos游戏","reasonCn":"Cocos CLI是Cocos游戏引擎的命令行工具，支持项目创建、代码编译、包体发布等全流程操作。无需图形界面，适合自动化构建和CI/CD集成。提供cocos new、cocos build等丰富子命令，能处理资源打包、代码混淆等细节，提升开发效率。","stars":434,"starGrowth":0,"createdAt":"2025-09-04T06:59:42Z","silentRounds":2,"language":"TypeScript","topics":[],"aiDims":["游戏","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Cocos引擎","游戏开发工具","命令行工具"],"domainKey":"Cocos引擎","pushedAt":"2026-08-18T03:33:50Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"bigbro","url":"https://github.com/cocos/cocos-cli","ts":"2026-08-18T03:33:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"actions/actions-runner-controller","owner":"actions","name":"actions-runner-controller","desc":"Kubernetes controller for GitHub Actions self-hosted runners","summaryCn":"在K8s上托管GitHub Actions跑者的控制器","reasonCn":"技术上，它是一个Kubernetes Operator，监控runner资源，自动创建、删除pod。支持配置runner标签、容器镜像、环境变量、资源限制等。可以和GitHub仓库集成，实现安全访问。适合需要运行大型工作流、自定义环境、或者数据敏感的场景。","stars":6512,"starGrowth":3,"createdAt":"2020-01-28T12:59:43Z","silentRounds":0,"language":"Go","topics":["github-actions","kubernetes","operator"],"aiDims":["开发者工具","AI基础设施","效率工具"],"zone":"工具","funScore":0.5,"domainTags":["CI/CD","Kubernetes","GitHub Actions","DevOps"],"domainKey":"CI/CD","pushedAt":"2026-09-05T02:38:58.868Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"github-actions","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"operator","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/actions/actions-runner-controller","ts":"2026-09-05T02:38:58.868Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"frappe/hrms","owner":"frappe","name":"hrms","desc":"Open Source HR and Payroll Software","summaryCn":"开源版的HR管理系统，从招聘到发工资全搞定，中小企业福音","reasonCn":"Frappe HRMS是一套功能完整的开源人力资源管理系统，覆盖招聘、考勤、薪资、绩效等核心模块。基于Python的Frappe框架构建，可高度定制。虽然不直接涉及AI算法，但作为实用的企业管理工具，对于需要搭建内部系统或优化HR流程的团队极具价值。","stars":8819,"starGrowth":1,"createdAt":"2022-06-08T14:45:49Z","silentRounds":1,"language":"Python","topics":["hr","hrms","payroll","erpnext","hcm","hris","employee","performance-management","recruitment","attendance","leave-management","shift-management","exits","frappe","frappe-framework","javascript","onboarding","open-source","python","pwa"],"aiDims":["效率工具"],"zone":"工具","funScore":0,"domainTags":["人力资源管理","薪资系统","考勤管理","ERP"],"domainKey":"人力资源管理","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"hr","source":"github","weight":0.7},{"name":"hrms","source":"github","weight":0.7},{"name":"payroll","source":"github","weight":0.7},{"name":"erpnext","source":"github","weight":0.7},{"name":"hcm","source":"github","weight":0.7},{"name":"hris","source":"github","weight":0.7},{"name":"employee","source":"github","weight":0.7},{"name":"performance-management","source":"github","weight":0.7},{"name":"recruitment","source":"github","weight":0.7},{"name":"attendance","source":"github","weight":0.7},{"name":"leave-management","source":"github","weight":0.7},{"name":"shift-management","source":"github","weight":0.7},{"name":"exits","source":"github","weight":0.7},{"name":"frappe","source":"github","weight":0.7},{"name":"frappe-framework","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"onboarding","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"python","source":"github","weight":0.7},{"name":"pwa","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/frappe/hrms","ts":"2026-08-07T04:03:56.389Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Unity-Technologies/DesktopSamples","owner":"Unity-Technologies","name":"DesktopSamples","desc":"This repository contains Unity samples for desktop platforms","summaryCn":"Unity官方桌面开发示例合集，直接抄代码","reasonCn":"技术上，示例覆盖UI系统、图形渲染、输入控制、音频播放等核心模块。每个项目都经过官方测试，代码规范，可直接导入Unity Editor运行学习。相比第三方教程，示例更贴近实际开发流程，且随Unity版本更新。它展示了跨平台开发的注意事项，比如分辨率适配和输入映射。","stars":244,"starGrowth":0,"createdAt":"2019-10-29T19:17:52Z","silentRounds":27,"language":"C++","topics":[],"aiDims":["游戏","开发者工具"],"zone":"资源","funScore":0,"domainTags":["Unity","桌面开发","示例代码"],"domainKey":"Unity","pushedAt":"2026-08-12T17:41:00Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"C++","source":"language","weight":0.5}],"aiScore":0.6,"source":"bigbro","url":"https://github.com/Unity-Technologies/DesktopSamples","ts":"2026-08-12T17:41:00Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"sleeepeer/PISanitizer","owner":"sleeepeer","name":"PISanitizer","desc":"[EMNLP 2026] PISanitizer: Preventing Prompt Injection to Long-Context LLMs via Prompt Sanitization","summaryCn":"防提示注入的AI安检员，保护大模型不被套路偷信息","reasonCn":"这是个针对大模型提示注入攻击的防御工具，通过prompt sanitization技术检测并过滤恶意注入内容。针对长上下文场景做了专门优化，能识别各种prompt injection攻击模式。支持主流LLM框架，可以直接集成到现有AI应用中保护用户数据安全，防止敏感信息泄露和越权操作，保障AI应用安全运行。","stars":20,"starGrowth":0,"createdAt":"2025-11-12T04:16:58Z","silentRounds":1,"language":"Python","topics":["llm-model"],"aiDims":["提示工程","大语言模型","安全工具"],"zone":"AI","funScore":0.5,"domainTags":["提示注入","Prompt Security","LLM安全","EMNLP"],"domainKey":"提示注入","pushedAt":"2026-08-21T18:57:44Z","tags":[{"name":"提示工程","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.62,"source":"search","url":"https://github.com/sleeepeer/PISanitizer","ts":"2026-08-21T18:57:44Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"EleutherAI/gpt-neo","owner":"EleutherAI","name":"gpt-neo","desc":"","summaryCn":"EleutherAI 团队曾通过这个仓库探索 GPT-Neo 系列模型的训练，其核心亮点在","reasonCn":"对于想要体验预训练模型能力的用户，官方在文档中强烈建议转向 HuggingFace 的 Transformer 集成，因为那里有更完善、更易用的接口和模型支持。这个仓库的主要价值在于其历史文献意义，它展示了在 GPU 普及之前，研究人员如何利用 TPU 集群进行大规模语言模型的并行计算，以及当时","stars":8260,"starGrowth":0,"createdAt":"2020-07-05T10:23:46Z","silentRounds":10,"language":"Python","topics":["language-model","transformers","gpt","gpt-2","gpt-3"],"aiDims":["大语言模型","模型与训练"],"zone":"AI","funScore":0.5,"domainTags":["语言模型","Transformer","TPU","分布式训练","mesh-tensorflow"],"domainKey":"语言模型","pushedAt":"2026-09-14T23:04:28.868Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"language-model","source":"github","weight":0.7},{"name":"transformers","source":"github","weight":0.7},{"name":"gpt","source":"github","weight":0.7},{"name":"gpt-2","source":"github","weight":0.7},{"name":"gpt-3","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.44,"source":"search","url":"https://github.com/EleutherAI/gpt-neo","ts":"2026-09-14T23:04:28.868Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"man-group/dtale","owner":"man-group","name":"dtale","desc":"Visualizer for pandas data structures","summaryCn":"pandas数据可视化神器，一行代码生成交互仪表板，数据分析从此有颜","reasonCn":"技术上，DTale用Flask做后端API，React做前端界面，两者通过WebSocket实时通信。当你加载一个DataFrame，前端会立即显示表格视图，还可以切换到各种图表模式：条形图、折线图、散点图等。它支持过滤、排序、分页等交互操作，还能生成统计摘要，包括均值、标准差、分位数等。","stars":5223,"starGrowth":0,"createdAt":"2019-07-15T09:34:48Z","silentRounds":9,"language":"TypeScript","topics":["data-visualization"],"aiDims":["可视化","非AI-实用","效率工具"],"zone":"工具","funScore":0,"domainTags":["pandas可视化","数据探索","交互式仪表板","Python"],"domainKey":"pandas可视化","pushedAt":"2026-07-24T05:18:50Z","tags":[{"name":"可视化","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"效率工具","source":"llm","weight":1},{"name":"data-visualization","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/man-group/dtale","ts":"2026-07-24T05:18:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"yusing/godoxy","owner":"yusing","name":"godoxy","desc":"High-performance reverse proxy and container orchestrator for self-hosters","summaryCn":"用Go编写的高性能反向代理与容器编排利器","reasonCn":"这是一个用Go语言开发的高性能反向代理和容器编排工具，专为自托管用户设计。整合反向代理和容器管理能力，让个人和小团队轻松管理多服务部署，统一入口转发，还能自动启停容器。代码简洁文档友好，不需要复杂配置就能跑起来，适合自托管玩家和小型运维场景。","stars":4158,"starGrowth":0,"createdAt":"2024-02-29T03:52:35Z","silentRounds":1,"language":"Go","topics":["docker","go","golang","letsencrypt","microservice","reverse-proxy","homepage","self-hosted","startpage"],"aiDims":["非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["反向代理","容器编排","自托管","Go"],"domainKey":"反向代理","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"docker","source":"github","weight":0.7},{"name":"go","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"letsencrypt","source":"github","weight":0.7},{"name":"microservice","source":"github","weight":0.7},{"name":"reverse-proxy","source":"github","weight":0.7},{"name":"homepage","source":"github","weight":0.7},{"name":"self-hosted","source":"github","weight":0.7},{"name":"startpage","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.6,"source":"trending","url":"https://github.com/yusing/godoxy","ts":"2026-08-10T06:58:39.267Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Cammin/LDtkToUnity","owner":"Cammin","name":"LDtkToUnity","desc":"Unity importer for the LDtk level editor","summaryCn":"游戏地图编辑器LDtk在Unity里的完美适配器","reasonCn":"技术上，这个项目是一个Unity Package，通过OpenUPM发布，安装方式很规范。它实现了LDtk文件格式到Unity场景的自动映射，支持层、实体、网格等不同数据的解析和导入。相比手动写脚本解析JSON或Tiled地图文件，这个工具的优势在于开箱即用，不需要你理解底层数据结构，导入器会自动","stars":425,"starGrowth":0,"createdAt":"2020-11-08T11:59:39Z","silentRounds":27,"language":"C#","topics":["2d","2d-game-development","importer","ldtk","ldtk-api","level-editor","openupm","unity","unity-editor","unity-package-manager","unity2d","unity2d-game"],"aiDims":["游戏"],"zone":"工具","funScore":0,"domainTags":["Unity插件","2D游戏开发","关卡编辑器","LDtk"],"domainKey":"Unity插件","pushedAt":"2026-04-11T09:49:49Z","tags":[{"name":"游戏","source":"llm","weight":1},{"name":"2d","source":"github","weight":0.7},{"name":"2d-game-development","source":"github","weight":0.7},{"name":"importer","source":"github","weight":0.7},{"name":"ldtk","source":"github","weight":0.7},{"name":"ldtk-api","source":"github","weight":0.7},{"name":"level-editor","source":"github","weight":0.7},{"name":"openupm","source":"github","weight":0.7},{"name":"unity","source":"github","weight":0.7},{"name":"unity-editor","source":"github","weight":0.7},{"name":"unity-package-manager","source":"github","weight":0.7},{"name":"unity2d","source":"github","weight":0.7},{"name":"unity2d-game","source":"github","weight":0.7},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.6,"source":"bigbro","url":"https://github.com/Cammin/LDtkToUnity","ts":"2026-04-11T09:49:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"40rbidd3n/Hydro0x01","owner":"40rbidd3n","name":"Hydro0x01","desc":"Secure, production-grade IoT hydroponic automation system with ESP32, MQTT telemetry, and real-time dashboard","summaryCn":"如果你想在阳台上搞个家庭农场，这个项目就是现成的“种菜机器人”。","reasonCn":"它的亮点在于“生产级”的定位。这意味着它不仅是个玩具，而是考虑到了稳定性、错误处理和功耗管理。通过内置的实时仪表盘，你可以像看股票走势一样，随时查看土壤湿度、PH值和营养液浓度。即使你人不在家，也能通过手机远程查看大棚状态，甚至远程启动灌溉。","stars":543,"starGrowth":0,"createdAt":"2026-04-14T21:51:02Z","silentRounds":16,"language":"TypeScript","topics":["arduino"],"aiDims":["非AI-实用","物联网"],"zone":"创意","funScore":1,"domainTags":["物联网","水培","自动化"],"domainKey":"物联网","pushedAt":"2026-08-18T19:33:18Z","tags":[{"name":"非AI-实用","source":"llm","weight":1},{"name":"物联网","source":"llm","weight":1},{"name":"arduino","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.6,"source":"search","url":"https://github.com/40rbidd3n/Hydro0x01","ts":"2026-08-18T19:33:18Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"scverse/pertpy","owner":"scverse","name":"pertpy","desc":"Single-cell perturbation analysis","summaryCn":"生物学家用来解析细胞对药物反应的神秘指南针工具包。","reasonCn":"技术上是基于scanpy生态系统构建的Python库。它的核心亮点是提供了模块化的分析管道，包括标准化、差异表达分析和轨迹推断等步骤。与通用的生物信息学工具相比，Pertpy更专注于“扰动”这一场景，内置了针对药物敏感性、基因敲除等实验设计的专用算法，能直接关联表型标签，减少了研","stars":346,"starGrowth":0,"createdAt":"2021-06-23T14:15:02Z","silentRounds":12,"language":"Python","topics":["bioinformatics"],"aiDims":["AI应用"],"zone":"工具","funScore":0,"domainTags":["单细胞","药物反应","生物信息学"],"domainKey":"单细胞","pushedAt":"2026-09-23T13:55:56Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"bioinformatics","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.25,"source":"search","url":"https://github.com/scverse/pertpy","ts":"2026-09-23T13:55:56Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"marswaveai/TypeNo","owner":"marswaveai","name":"TypeNo","desc":"A free, open source, privacy-first voice input app for macOS.","summaryCn":"macOS本地语音输入神器，隐私优先完全离线识别","reasonCn":"TypeNo是一款专为macOS设计的开源语音输入应用，强调隐私保护和本地优先。所有语音处理都在本地完成，不上传云端，适合对数据敏感的用户。支持菜单栏快捷调用、多语言识别，是自由软件爱好者和注重隐私的Mac用户的实用工具，值得一试。","stars":900,"starGrowth":0,"createdAt":"2026-03-24T09:15:02Z","silentRounds":24,"language":"Swift","topics":["accessibility","dictation","local-first","macos","menu-bar-app","open-source","privacy","productivity","speech-to-text","swift","voice-input"],"aiDims":["效率工具","AI应用","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["macOS","语音识别","本地优先","隐私保护","菜单栏应用"],"domainKey":"macOS","pushedAt":"2026-08-11T06:08:26Z","tags":[{"name":"效率工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"accessibility","source":"github","weight":0.7},{"name":"dictation","source":"github","weight":0.7},{"name":"local-first","source":"github","weight":0.7},{"name":"macos","source":"github","weight":0.7},{"name":"menu-bar-app","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"privacy","source":"github","weight":0.7},{"name":"productivity","source":"github","weight":0.7},{"name":"speech-to-text","source":"github","weight":0.7},{"name":"swift","source":"github","weight":0.7},{"name":"voice-input","source":"github","weight":0.7},{"name":"Swift","source":"language","weight":0.5}],"aiScore":0.6,"source":"bigbro","url":"https://github.com/marswaveai/TypeNo","ts":"2026-08-11T06:08:26Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"payloadcms/payload","owner":"payloadcms","name":"payload","desc":"Payload is the open-source, fullstack Next.js framework, giving you instant backend superpowers. Get a full TypeScript backend and admin panel instantly. Use Payload as a headless CMS or for building powerful applications.","summaryCn":"Next.js全栈框架，自动生成后台管理面板，写代码的同时后台也有了","reasonCn":"Payload是一个基于Next.js的开源全栈框架，能自动生成TypeScript类型安全的后端和可视化Admin管理面板。支持自定义字段类型、权限控制、文件上传、本地化等完整CMS功能。开发者只需定义数据模型，剩下的API和后台界面全部自动生成，是快速构建内容驱动应用的高效方案。","stars":44899,"starGrowth":11,"createdAt":"2021-01-05T18:49:45Z","silentRounds":0,"language":"TypeScript","topics":["headless-cms","cms","express","graphql","react","payload","nodejs","typescript","javascript","mongodb","headless","content-management-system","content-management","mit-license","open-source","payloadcms","jamstack","postgres","nextjs"],"aiDims":["AI应用","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["Next.js","CMS","TypeScript","全栈框架"],"domainKey":"Next.js","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"headless-cms","source":"github","weight":0.7},{"name":"cms","source":"github","weight":0.7},{"name":"express","source":"github","weight":0.7},{"name":"graphql","source":"github","weight":0.7},{"name":"react","source":"github","weight":0.7},{"name":"payload","source":"github","weight":0.7},{"name":"nodejs","source":"github","weight":0.7},{"name":"typescript","source":"github","weight":0.7},{"name":"javascript","source":"github","weight":0.7},{"name":"mongodb","source":"github","weight":0.7},{"name":"headless","source":"github","weight":0.7},{"name":"content-management-system","source":"github","weight":0.7},{"name":"content-management","source":"github","weight":0.7},{"name":"mit-license","source":"github","weight":0.7},{"name":"open-source","source":"github","weight":0.7},{"name":"payloadcms","source":"github","weight":0.7},{"name":"jamstack","source":"github","weight":0.7},{"name":"postgres","source":"github","weight":0.7},{"name":"nextjs","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/payloadcms/payload","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":["hot","daily"],"fromOfficial":false,"heatScore":0.033734982961758614,"copyOk":true},{"repo":"DLS5-Omics/GEMGen","owner":"DLS5-Omics","name":"GEMGen","desc":"Inverse design of chemical perturbations for desired cellular state transitions with large language models","summaryCn":"用大模型给细胞'下指令'，让它变成你想要的样子","reasonCn":"GEMGen将大语言模型引入生物化学逆向设计领域，通过学习已知扰动与细胞状态转换之间的关联规律，预测能实现特定目标细胞状态的最优化学方案。它可以分析大规模组学数据、理解基因调控网络，辅助药物发现和精准医疗研究。","stars":11,"starGrowth":0,"createdAt":"2025-12-17T04:34:10Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["AI应用","模型与训练"],"zone":"AI","funScore":0,"domainTags":["生物化学","LLM","细胞状态","药物发现","逆向设计"],"domainKey":"生物化学","pushedAt":"2026-09-01T03:37:15Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/DLS5-Omics/GEMGen","ts":"2026-09-01T03:37:15Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"pallets/quart","owner":"pallets","name":"quart","desc":"An async Python micro framework for building web applications. ","summaryCn":"Python写的异步微框架，开发轻量Web服务超好用","reasonCn":"Quart是Flask的异步双胞胎，基于ASGI标准构建。它保留了你熟悉的Flask API风格，但原生支持async/await，完美契合现代Python生态。适合想快速搭建高性能API或WebSocket服务的开发者，无需大量重构即可体验异步优势。","stars":3673,"starGrowth":0,"createdAt":"2017-11-10T20:10:33Z","silentRounds":2,"language":"Python","topics":["web-framework"],"aiDims":["开发者工具","非AI-实用","代码助手"],"zone":"工具","funScore":0,"domainTags":["Python","异步框架","Web开发"],"domainKey":"Python","pushedAt":"2026-09-12T09:07:36Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/pallets/quart","ts":"2026-09-12T09:07:36Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"volcano-sh/volcano","owner":"volcano-sh","name":"volcano","desc":"A Cloud Native Batch System (Project under CNCF)","summaryCn":"K8s上批处理调度神器，让大规模任务跑得稳又快","reasonCn":"Volcano 作为 CNCF 项目，为 Kubernetes 提供专业的批量任务调度能力，支持作业依赖、资源配额、容错重试和多租户隔离，可在大数据清洗、机器学习离线训练等场景中实现高并发、低延迟的任务执行，且通过声明式 API 与原生调度器深度融合，降低运维复杂度。","stars":5977,"starGrowth":1,"createdAt":"2019-03-14T09:47:29Z","silentRounds":1,"language":"Go","topics":["batch-systems","kubernetes","golang","hpc","bigdata","machine-learning","gene","ai","serving","training"],"aiDims":["AI基础设施","非AI-实用","开发者工具"],"zone":"工具","funScore":0,"domainTags":["Kubernetes","Batch Processing","Scheduler"],"domainKey":"Kubernetes","pushedAt":"2026-09-09T22:18:37.171Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"batch-systems","source":"github","weight":0.7},{"name":"kubernetes","source":"github","weight":0.7},{"name":"golang","source":"github","weight":0.7},{"name":"hpc","source":"github","weight":0.7},{"name":"bigdata","source":"github","weight":0.7},{"name":"machine-learning","source":"github","weight":0.7},{"name":"gene","source":"github","weight":0.7},{"name":"ai","source":"github","weight":0.7},{"name":"serving","source":"github","weight":0.7},{"name":"training","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.48,"source":"trending","url":"https://github.com/volcano-sh/volcano","ts":"2026-09-09T22:18:37.171Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"glitternetwork/pinme","owner":"glitternetwork","name":"pinme","desc":"Deploy Your Frontend in a Single Command. Claude Code Skills supported.","summaryCn":"一条命令搞定前端部署，连Claude Code都为你准备好了专属技能包","reasonCn":"这是一个极简的前端部署工具，主打单命令上线。有趣的是它深度集成了Claude Code Skills，开发者可以直接通过自然语言指令调用部署能力，无需手写复杂脚本。对于追求效率的全栈工程师来说，这是降低部署摩擦的神器。","stars":3747,"starGrowth":0,"createdAt":"2025-04-13T01:41:33Z","silentRounds":6,"language":"TypeScript","topics":["frontend"],"aiDims":["模型部署","开发者工具","代码助手"],"zone":"工具","funScore":0,"domainTags":["前端部署","Claude Code","DevOps","部署自动化"],"domainKey":"前端部署","pushedAt":"2026-09-12T05:23:49Z","tags":[{"name":"模型部署","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"frontend","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/glitternetwork/pinme","ts":"2026-09-12T05:23:49Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"openbao/openbao","owner":"openbao","name":"openbao","desc":"OpenBao is a software solution to manage, store, and distribute sensitive data including secrets, certificates, and keys.","summaryCn":"保管密钥证书的加密保险箱，敏感信息再也不怕泄露","reasonCn":"OpenBao 是 Vault 的开源分支，专管敏感数据的安全存储与分发。支持密钥管理、证书颁发、动态凭证生成，提供统一 API 和细粒度访问控制。企业用它保护数据库密码、API 密钥、TLS 证书等，防止泄露。还内置审计日志和实时监控，满足合规要求。","stars":7461,"starGrowth":6,"createdAt":"2023-11-09T18:17:33Z","silentRounds":0,"language":"Go","topics":["go","secret-management","security"],"aiDims":["安全工具","AI基础设施","开发者工具"],"zone":"工具","funScore":0,"domainTags":["密钥管理","Vault","DevOps","安全"],"domainKey":"密钥管理","pushedAt":"2026-08-08T00:55:52.206Z","tags":[{"name":"安全工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"go","source":"github","weight":0.7},{"name":"secret-management","source":"github","weight":0.7},{"name":"security","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.55,"source":"trending","url":"https://github.com/openbao/openbao","ts":"2026-08-08T00:55:52.206Z","category":"tool","momentum":["daily"],"fromOfficial":false,"heatScore":0.020412545235223158,"copyOk":false},{"repo":"dg/ai-access","owner":"dg","name":"ai-access","desc":"A flexible PHP library providing access to various AI models (OpenAI, Claude, Gemini, DeepSeek and Grok) via a consistent interface.","summaryCn":"一个PHP库统一调用各大AI模型，OpenAI Claude Gemini一次对接全搞定","reasonCn":"技术上，它定义了一组标准化的请求和响应接口，为每个支持的AI模型提供具体的适配器实现。支持常见的文本生成、流式输出和嵌入向量功能。设计比较轻量，没有过多抽象，代码量不大但覆盖了主流模型。缺点是作为PHP库，适用场景相对有限，主要面向PHP技术栈的开发者。","stars":63,"starGrowth":0,"createdAt":"2025-04-11T15:23:01Z","silentRounds":27,"language":"PHP","topics":["llm-model"],"aiDims":["大语言模型","工具调用"],"zone":"工具","funScore":0,"domainTags":["PHP","AI接口","模型适配","API封装","开发库"],"domainKey":"PHP","pushedAt":"2026-08-26T03:47:50Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"工具调用","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"PHP","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/dg/ai-access","ts":"2026-08-26T03:47:50Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"galilai-group/stable-pretraining","owner":"galilai-group","name":"stable-pretraining","desc":"Reliable, minimal and scalable library for pretraining foundation and world models","summaryCn":"给想自己训练大模型的你，准备的一套极简、抗造又好用的标准化配方","reasonCn":"这不是现成的聊天机器人，而是教AI学会说话的基础设施。它主打最小化代码和极高稳定性，专门用于从头训练或继续预训练基础模型。对于想深入研究AI底层原理、或者需要稳定复现论文效果的工程师来说，这是非常实用的开源工具。","stars":319,"starGrowth":0,"createdAt":"2024-07-23T15:20:57Z","silentRounds":6,"language":"Python","topics":["llm-model"],"aiDims":["模型与训练","大语言模型","AI基础设施"],"zone":"AI","funScore":0,"domainTags":["大模型训练","预训练","Transformer","PyTorch"],"domainKey":"大模型训练","pushedAt":"2026-07-16T01:34:01Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.55,"source":"search","url":"https://github.com/galilai-group/stable-pretraining","ts":"2026-07-16T01:34:01Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"konamgil/mandu","owner":"konamgil","name":"mandu","desc":"An open-source framework that prevents AI agents from breaking your codebase.","summaryCn":"它把智能体改码当作工程流程问题来治理，思路偏流程约束。","reasonCn":"它把契约、架构约束和受控工作流放进 Bun 与 React 全栈运行环境里。当自动编程代理反复修改项目时，它试图阻止项目结构越改越乱、接口契约失配或构建状态失效，给人和代理保留可继续修改的安全边界。版本偏早期，适合关注自动编程协作与工程安全边界的人做初步评估。","stars":47,"starGrowth":0,"createdAt":"2026-01-21T15:48:41Z","silentRounds":12,"language":"TypeScript","topics":["web-framework"],"aiDims":["AI Agent","开发者工具"],"zone":"AI","funScore":0,"domainTags":["AI代理","代码安全","自动编程","工程治理"],"domainKey":"AI代理","pushedAt":"2026-08-27T06:01:34Z","tags":[{"name":"AI Agent","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"web-framework","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.53,"source":"search","url":"https://github.com/konamgil/mandu","ts":"2026-08-27T06:01:34Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Bouwles/rl-chess-snake-lab","owner":"Bouwles","name":"rl-chess-snake-lab","desc":"Reinforcement learning experiments for Snake and Chess using Python, PyTorch, Gymnasium, and python-chess.","summaryCn":"Python跑蛇棋强化学习实验，探索智能决策","reasonCn":"本项目利用PyTorch与Gymnasium搭建Snake与Chess的强化学习环境，支持DQN、Actor-Critic等多种算法，代码结构清晰易扩展；内置可视化训练日志与模型保存，帮助快速实验与调优；适合对RL感兴趣的开发者和研究者，且可直接运行无额外配置。","stars":23,"starGrowth":0,"createdAt":"2026-08-23T08:07:26Z","silentRounds":15,"language":"Python","topics":["chess"],"aiDims":["模型与训练","游戏","AI基础设施"],"zone":"创意","funScore":0.7,"domainTags":["强化学习","Python","Snake游戏","国际象棋"],"domainKey":"强化学习","pushedAt":"2026-09-09T10:45:41Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"游戏","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"chess","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/Bouwles/rl-chess-snake-lab","ts":"2026-09-09T10:45:41Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"Comfy-Org/workflow_templates","owner":"Comfy-Org","name":"workflow_templates","desc":"ComfyUI template workflows","summaryCn":"ComfyUI工作流模板库，拿来即用不踩坑","reasonCn":"这是ComfyUI官方维护的节点流工作流收藏库。里面全是现成的流程图文件，涵盖文生图、图生图、视频生成等主流任务。下载JSON后拖进ComfyUI即可直接运行，省去了繁琐的参数调试和节点连接过程，非常适合想快速验证效果或学习复杂工作流的AI创作者。","stars":974,"starGrowth":2,"createdAt":"2025-02-28T21:39:51Z","silentRounds":0,"language":"Python","topics":[],"aiDims":["创意工具","AI应用","AI基础设施"],"zone":"资源","funScore":0,"domainTags":["ComfyUI","AI绘图","工作流模板","Stable Diffusion"],"domainKey":"ComfyUI","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.52,"source":"trending","url":"https://github.com/Comfy-Org/workflow_templates","ts":"2026-08-07T04:03:56.389Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"amark/gun","owner":"amark","name":"gun","desc":"An open source cybersecurity protocol for syncing decentralized graph data.","summaryCn":"去中心化数据库黑科技，让数据在设备间自动同步且无需中心服务器","reasonCn":"开源的去中心化图数据库协议，基于JavaScript实现，支持设备间P2P数据同步、端到端加密存储。无需云端服务器，数据主权完全归用户所有，适合构建抗审查、高可用的分布式应用架构，是Web3和边缘计算时代的数据库新范式。","stars":19135,"starGrowth":1,"createdAt":"2014-07-31T15:52:10Z","silentRounds":6,"language":"JavaScript","topics":["database","ml"],"aiDims":["AI基础设施","安全工具","开发者工具"],"zone":"工具","funScore":0.6,"domainTags":["P2P","数据同步","Web3"],"domainKey":"P2P","pushedAt":"2026-09-16T01:19:08Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"安全工具","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"database","source":"github","weight":0.7},{"name":"ml","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/amark/gun","ts":"2026-09-16T01:19:08Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Unity-Technologies/barracuda-release","owner":"Unity-Technologies","name":"barracuda-release","desc":"","summaryCn":"Unity官方机器学习推理引擎，让游戏里跑AI模型变得像搭积木一样简单","reasonCn":"Barracuda是Unity官方推出的机器学习推理引擎，支持TensorFlow、PyTorch等框架导入的模型，可在CPU和GPU上运行推理。提供模型量化和格式转换能力，让AI能力轻松嵌入Unity游戏项目，支持图像分类、手势识别等常见推理任务。","stars":606,"starGrowth":0,"createdAt":"2019-07-22T12:34:26Z","silentRounds":27,"language":"C#","topics":[],"aiDims":["AI基础设施","模型部署","非AI-游戏"],"zone":"AI","funScore":0,"domainTags":["Unity","机器学习","ONNX","游戏AI"],"domainKey":"Unity","pushedAt":"2025-02-13T14:07:24Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型部署","source":"llm","weight":1},{"name":"非AI-游戏","source":"llm","weight":1},{"name":"C#","source":"language","weight":0.5}],"aiScore":0.55,"source":"bigbro","url":"https://github.com/Unity-Technologies/barracuda-release","ts":"2025-02-13T14:07:24Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"stenciljs/core","owner":"stenciljs","name":"core","desc":"","summaryCn":"Stencil 是一个专门为生成 Web Components 设计的编译器，由知名开源团队 I","reasonCn":"Stencil 的核心优势在于其强大的跨框架兼容性。生成的组件不仅能在 React、Vue 等主流框架中无缝使用，甚至可以在没有框架的纯 HTML 页面中直接调用，完全不需要引入额外的库或进行复杂的适配。这种特性使得开发者能够构建出独立、轻量且可复用的组件，极大地提升了开发效率和组件","stars":13129,"starGrowth":0,"createdAt":"2017-02-15T18:57:07Z","silentRounds":3,"language":"TypeScript","topics":["typescript","stencil","stenciljs","ionic","ssr","ssg","static-site-generator","custom-element","design-system","webcomponent"],"aiDims":["开发者工具","代码助手"],"zone":"工具","funScore":0,"domainTags":["Web Components","TypeScript","前端框架"],"domainKey":"Web Components","pushedAt":"2026-09-13T22:08:09.499Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"代码助手","source":"llm","weight":1},{"name":"typescript","source":"github","weight":0.7},{"name":"stencil","source":"github","weight":0.7},{"name":"stenciljs","source":"github","weight":0.7},{"name":"ionic","source":"github","weight":0.7},{"name":"ssr","source":"github","weight":0.7},{"name":"ssg","source":"github","weight":0.7},{"name":"static-site-generator","source":"github","weight":0.7},{"name":"custom-element","source":"github","weight":0.7},{"name":"design-system","source":"github","weight":0.7},{"name":"webcomponent","source":"github","weight":0.7},{"name":"TypeScript","source":"language","weight":0.5}],"aiScore":0.38,"source":"tier","url":"https://github.com/stenciljs/core","ts":"2026-09-13T22:08:09.499Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"open-telemetry/opentelemetry.io","owner":"open-telemetry","name":"opentelemetry.io","desc":"The OpenTelemetry website and documentation","summaryCn":"AI系统可观测性标准配置，监控追踪大模型应用","reasonCn":"OpenTelemetry是云原生计算基金会旗下的开源可观测性框架，提供日志、指标、追踪的统一采集方案。官网包含完整文档、SDK指南和最佳实践。对AI应用来说，用它监控大模型推理延迟、追踪请求链路、收集性能指标，是构建生产级AI系统的必要基础设施之一。","stars":971,"starGrowth":0,"createdAt":"2019-04-29T18:56:56Z","silentRounds":2,"language":"JavaScript","topics":["opentelemetry","documentation"],"aiDims":["AI基础设施","开发者工具","非AI-实用"],"zone":"资源","funScore":0,"domainTags":["可观测性","日志追踪","CNCF","APM"],"domainKey":"可观测性","pushedAt":"2026-08-10T06:58:39.267Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"opentelemetry","source":"github","weight":0.7},{"name":"documentation","source":"github","weight":0.7},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.5,"source":"trending","url":"https://github.com/open-telemetry/opentelemetry.io","ts":"2026-08-10T06:58:39.267Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"vrgamegirl19/comfyui-vrgamedevgirl","owner":"vrgamegirl19","name":"comfyui-vrgamedevgirl","desc":"Custom ComfyUI nodes for film grain, color matching, and video enhancement.","summaryCn":"给ComfyUI加料，视频调色增强特效一键搞定","reasonCn":"这是一个针对ComfyUI工作流的扩展项目，主要提供电影颗粒感、色彩匹配和视频增强等自定义节点。虽然底层基于AI绘图，但更偏向后期处理工具链，无需复杂训练。对于喜欢做视频的创作者来说，它能批量处理素材，省去手动调色的麻烦，提升成片质感。","stars":731,"starGrowth":1,"createdAt":"2025-07-03T23:19:24Z","silentRounds":1,"language":"JavaScript","topics":[],"aiDims":["创意工具","AI应用","非AI-好玩"],"zone":"创意","funScore":0.6,"domainTags":["ComfyUI","视频增强","调色特效"],"domainKey":"ComfyUI","pushedAt":"2026-08-07T04:03:56.389Z","tags":[{"name":"创意工具","source":"llm","weight":1},{"name":"AI应用","source":"llm","weight":1},{"name":"非AI-好玩","source":"llm","weight":1},{"name":"JavaScript","source":"language","weight":0.5}],"aiScore":0.45,"source":"trending","url":"https://github.com/vrgamegirl19/comfyui-vrgamedevgirl","ts":"2026-08-07T04:03:56.389Z","category":"fun","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"liguge/Awesome-large-language-model-for-Prognostics-and-health-management","owner":"liguge","name":"Awesome-large-language-model-for-Prognostics-and-health-management","desc":"用于预测性维护与健康管理的大型语言模型（故障诊断；寿命预测）","summaryCn":"用大模型做设备故障诊断和寿命预测的工业AI指南","reasonCn":"这是一个专注于预测性维护与健康管理领域的LLM资源合集，针对故障诊断和寿命预测等工业场景提供大模型应用方案。收录了相关论文、开源代码和工具链，帮助工程师快速落地AI能力。项目由国内开发者维护，star数130，属于相对冷门但垂直度很高的专业资源库。","stars":136,"starGrowth":0,"createdAt":"2025-04-26T03:08:13Z","silentRounds":2,"language":"","topics":["llm-model"],"aiDims":["AI应用","大语言模型","学习资源"],"zone":"资源","funScore":0,"domainTags":["预测性维护","故障诊断","工业AI"],"domainKey":"预测性维护","pushedAt":"2026-08-05T12:10:28Z","tags":[{"name":"AI应用","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"学习资源","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7}],"aiScore":0.45,"source":"search","url":"https://github.com/liguge/Awesome-large-language-model-for-Prognostics-and-health-management","ts":"2026-08-05T12:10:28Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":true},{"repo":"tomomasayamasaki/RBFleX-NAS","owner":"tomomasayamasaki","name":"RBFleX-NAS","desc":"Training-free Neural Architecture Search (NAS) Using Radial Basis Function (RBF) Kernel ","summaryCn":"用数学魔法自动设计神经网络架构的AI搜索工具","reasonCn":"这是一个不用训练就能跑的网络架构搜索(NAS)工具。它巧妙利用径向基函数核技术，跳过传统NAS耗时的训练评估过程，直接通过数学近似来评估架构优劣。对于研究AI基础设施和自动建模的开发者来说，这种“零成本”筛选方法在学术探索上非常有趣。","stars":11,"starGrowth":0,"createdAt":"2025-03-11T04:55:12Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["非AI-好玩","AI基础设施","模型与训练"],"zone":"AI","funScore":0.6,"domainTags":["神经架构搜索","模型压缩","深度学习"],"domainKey":"神经架构搜索","pushedAt":"2026-07-29T09:55:41Z","tags":[{"name":"非AI-好玩","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.45,"source":"search","url":"https://github.com/tomomasayamasaki/RBFleX-NAS","ts":"2026-07-29T09:55:41Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"kubesphere/kubesphere","owner":"kubesphere","name":"kubesphere","desc":"The container platform tailored for Kubernetes multi-cloud, datacenter, and edge management ⎈ 🖥 ☁️","summaryCn":"K8s界的瑞士军刀，可视化界面让运维不再头秃，多云管理一把梭","reasonCn":"这是基于Kubernetes构建的分布式操作系统，提供全栈IT自动化运维能力。它通过可视化界面大幅降低了K8s的使用门槛，支持多集群管理、微服务治理、应用商店等功能。相比原生K8s，它内置了日志监控、CI/CD流水线等实用组件，让企业能更专注于业务创新而非基础设施维护，适合中大型团队进行数字化转型。","stars":17057,"starGrowth":0,"createdAt":"2018-04-21T02:03:04Z","silentRounds":1,"language":"Go","topics":["devops"],"aiDims":["开发者工具","AI基础设施","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["容器编排","DevOps","可视化监控"],"domainKey":"容器编排","pushedAt":"2026-07-15T01:14:39Z","tags":[{"name":"开发者工具","source":"llm","weight":1},{"name":"AI基础设施","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"devops","source":"github","weight":0.7},{"name":"Go","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/kubesphere/kubesphere","ts":"2026-07-15T01:14:39Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"Amirhosein-gh98/Gnosis","owner":"Amirhosein-gh98","name":"Gnosis","desc":"Can LLMs Predict Their Own Failures? Self-Awareness via Internal Circuits","summaryCn":"研究AI大模型能否自我察觉并预测自己的错误","reasonCn":"这是一项前沿AI研究，探索LLM是否能在内部层面识别自身可能犯错的时刻。通过分析模型内部电路和激活模式，研究者试图让模型具备自我意识的苗头，从而提升AI的可靠性和可解释性。star较少属于学术研究项目，但选题非常新颖有趣，代表了AI安全研究的前沿方向。","stars":46,"starGrowth":0,"createdAt":"2025-12-19T16:50:07Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["大语言模型","AI Agent","模型与训练"],"zone":"AI","funScore":0.6,"domainTags":["LLM","可解释性","AI安全"],"domainKey":"LLM","pushedAt":"2026-07-15T21:03:50Z","tags":[{"name":"大语言模型","source":"llm","weight":1},{"name":"AI Agent","source":"llm","weight":1},{"name":"模型与训练","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.4,"source":"search","url":"https://github.com/Amirhosein-gh98/Gnosis","ts":"2026-07-15T21:03:50Z","category":"ai","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"paritytech/polkadot-sdk","owner":"paritytech","name":"polkadot-sdk","desc":"The Parity Polkadot Blockchain SDK","summaryCn":"基于Rust的区块链开发套件，想玩Web3开发别错过。","reasonCn":"Polkadot SDK是Polkadot网络的核心底层开发框架，完全用Rust语言编写。它提供了一套模块化、可复用的开发工具集，让开发者能够快速构建、部署和治理区块链应用。该SDK集成了Substrate框架的核心功能，包括状态管理、共识机制、交易处理以及跨链互操作性等，极大地降低了区块链开发的门槛。","stars":2812,"starGrowth":1,"createdAt":"2023-08-04T08:33:14Z","silentRounds":2,"language":"Rust","topics":["blockchain","cumulus","polkadot","substrate"],"aiDims":["AI基础设施","开发者工具","非AI-实用"],"zone":"工具","funScore":0,"domainTags":["区块链","Substrate","Web3开发"],"domainKey":"区块链","pushedAt":"2026-08-12T01:22:42.342Z","tags":[{"name":"AI基础设施","source":"llm","weight":1},{"name":"开发者工具","source":"llm","weight":1},{"name":"非AI-实用","source":"llm","weight":1},{"name":"blockchain","source":"github","weight":0.7},{"name":"cumulus","source":"github","weight":0.7},{"name":"polkadot","source":"github","weight":0.7},{"name":"substrate","source":"github","weight":0.7},{"name":"Rust","source":"language","weight":0.5}],"aiScore":0.35,"source":"trending","url":"https://github.com/paritytech/polkadot-sdk","ts":"2026-08-12T01:22:42.342Z","category":"tool","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"chenshuang-zhang/imagenet_d","owner":"chenshuang-zhang","name":"imagenet_d","desc":"[CVPR 2024 Highlight] ImageNet-D","summaryCn":"这个仓库来自一篇CVPR 2024的Highlight论文，作者专门构建了一个叫ImageNet","reasonCn":"它的核心技术亮点在于系统性地收集了13种不同类型的图像退化场景，包括高斯模糊、运动模糊、雪灾、冻结、亮度干扰、噪声、脉冲噪声、像素化、喷雾、对比度干扰、弹性变换、椒盐噪声和数字压缩失真。每种退化都有多个强度等级，模型在这些退化图像上的性能下降幅度，直接反映了它的","stars":47,"starGrowth":0,"createdAt":"2024-03-14T07:39:22Z","silentRounds":27,"language":"Python","topics":["llm-model"],"aiDims":["模型与训练","大语言模型","多模态Agent"],"zone":"资源","funScore":0.6,"domainTags":["图像分类","模型评估","数据集"],"domainKey":"图像分类","pushedAt":"2026-07-24T01:53:11Z","tags":[{"name":"模型与训练","source":"llm","weight":1},{"name":"大语言模型","source":"llm","weight":1},{"name":"多模态Agent","source":"llm","weight":1},{"name":"llm-model","source":"github","weight":0.7},{"name":"Python","source":"language","weight":0.5}],"aiScore":0.35,"source":"search","url":"https://github.com/chenshuang-zhang/imagenet_d","ts":"2026-07-24T01:53:11Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false},{"repo":"branch-danya-dev/ssad-methodology","owner":"branch-danya-dev","name":"ssad-methodology","desc":"Practical system-analysis methodology for structuring, validating, and evolving knowledge around real system boundaries, responsibilities, ownership, and delivery workflows.","summaryCn":"教工程师怎么把复杂系统拆明白的知识管理手册","reasonCn":"它的核心是SSAD方法论，全称系统结构化分析与设计。简单说就是给你一套框架，帮你回答\"这个系统到底是什么、边界在哪、谁负责什么、怎么交付\"这些让人头疼的问题。项目中包含模板、示例和实操指南，不是空洞的理论，而是可以直接套用的工具集。","stars":64,"starGrowth":0,"createdAt":"2026-08-29T01:15:33Z","silentRounds":13,"language":"","topics":["knowledge-management"],"aiDims":["知识图谱","RAG与知识"],"zone":"资源","funScore":0,"domainTags":["系统设计","工程方法论","架构文档"],"domainKey":"系统设计","pushedAt":"2026-08-30T16:51:06Z","tags":[{"name":"知识图谱","source":"llm","weight":1},{"name":"RAG与知识","source":"llm","weight":1},{"name":"knowledge-management","source":"github","weight":0.7}],"aiScore":0.25,"source":"search","url":"https://github.com/branch-danya-dev/ssad-methodology","ts":"2026-08-30T16:51:06Z","category":"learning","momentum":[],"fromOfficial":false,"heatScore":0,"copyOk":false}]