[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-kyrolabs--awesome-agents":3,"tool-kyrolabs--awesome-agents":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",140436,2,"2026-04-05T23:32:43",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":79,"owner_twitter":80,"owner_website":81,"owner_url":82,"languages":79,"stars":83,"forks":84,"last_commit_at":85,"license":79,"difficulty_score":86,"env_os":87,"env_gpu":88,"env_ram":88,"env_deps":89,"category_tags":92,"github_topics":79,"view_count":93,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":94,"updated_at":95,"faqs":96,"releases":127},184,"kyrolabs\u002Fawesome-agents","awesome-agents","🤖 Awesome list of AI Agents","awesome-agents 是一个精心整理的开源 AI 智能体工具集合，帮助开发者和研究人员快速找到构建 AI 代理所需的框架、评测工具和应用场景。它解决了“从零开始搭建智能体太复杂”的问题，把分散在 GitHub 上的优质项目按类别归类，比如对话机器人、自动化脚本、多智能体协作、浏览器控制等，让你不用大海捞针就能选到合适的轮子。\n\n这个列表特别适合 AI 工程师、全栈开发者和学术研究者使用，也欢迎产品经理或技术爱好者探索前沿趋势。其中不乏亮点项目，比如 Meta 的 llama-agentic-system 提供标准化接口，微软 Semantic Kernel 支持 C# 快速集成大模型，MetaGPT 能用一句话生成完整开发流程，而 ClaudeClaw 则支持跨平台（Slack\u002FWhatsApp）运行并具备沙箱隔离能力，安全又灵活。\n\n无论你是想构建客服助手、自动编程伙伴，还是模拟游戏 NPC 或知识管理引擎，awesome-agents 都像一本实用指南，帮你少走弯路，专注创造真正有价值的智能应用。","# 🤖 Awesome Agents\n\nAwesome Agents is a curated list of open-source tools and products to build AI agents.\n\n> 🆕 **Featured:** [ClaudeClaw](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fclaudeclaw) — Persistent agent orchestrator as a Claude Code plugin. Multi-channel routing (Slack, WhatsApp, Telegram), OS-level sandbox isolation, composable extension system. [Read more →](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fclaudeclaw)\n\n## Table of Contents\n\n- [🤖 Awesome Agents](#-awesome-agents)\n  - [Table of Contents](#table-of-contents)\n  - [Frameworks](#frameworks)\n  - [Testing and Evaluation](#testing-and-evaluation)\n  - [Software Development](#software-development)\n  - [Research](#research)\n  - [Conversational \u002F General Agents](#conversational--general-agents)\n  - [Game \u002F Simulation](#game--simulation)\n  - [Knowledge Management](#knowledge-management)\n  - [Automation](#automation)\n    - [Browser](#browser)\n    - [Multimodal](#multimodal)\n\n## Frameworks\n\n- [llama-agentic-system](https:\u002F\u002Fgithub.com\u002Fmeta-llama\u002Fllama-agentic-system): Agentic components of the Llama Stack APIs ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmeta-llama\u002Fllama-agentic-system?style=social)\n- [Transformers Agents](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Ftransformers\u002Ftransformers_agents): Provides a natural language API on top of transformers\n- [LlamaIndex](https:\u002F\u002Fgithub.com\u002Fjerryjliu\u002Fllama_index): provides a central interface to connect your LLM's with external data. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjerryjliu\u002Fllama_index?style=social)\n- [LangChain](https:\u002F\u002Fgithub.com\u002Fhwchase17\u002Flangchain): the original 🐍 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhwchase17\u002Flangchain?style=social)\n- [Botpress](https:\u002F\u002Fgithub.com\u002Fbotpress\u002Fbotpress): The building blocks for building chatbots ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fbotpress\u002Fbotpress?style=social)\n- [Haystack](https:\u002F\u002Fgithub.com\u002Fdeepset-ai\u002Fhaystack): NLP framework to interact with your data using Transformer models and LLMs ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fdeepset-ai\u002Fhaystack?style=social)\n- [Semantic Kernel](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsemantic-kernel): Microsoft C# SDK to integrate cutting-edge LLM technology quickly and easily into your apps ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmicrosoft\u002Fsemantic-kernel?style=social)\n- [Agent-LLM](https:\u002F\u002Fgithub.com\u002FJosh-XT\u002FAgent-LLM): An Artificial Intelligence Automation Platform. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FJosh-XT\u002FAgent-LLM?style=social)\n- [LLM Agents](https:\u002F\u002Fgithub.com\u002Fmpaepper\u002Fllm_agents): Build agents which are controlled by LLMs ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmpaepper\u002Fllm_agents?style=social)\n- [e2b](https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fe2b): Open-source platform for building & deploying virtual developers’ agents\n- [Dust](https:\u002F\u002Fgithub.com\u002Fdust-tt\u002Fdust): Design and Deploy Large Language Model Apps ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fdust-tt\u002Fdust?style=social)\n- [MetaGPT](https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT): The Multi-Agent Meta Programming Framework: Given one line Requirement, return PRD, Design, Tasks, Repo and CI ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgeekan\u002FMetaGPT?style=social)\n- [AgentFlow](https:\u002F\u002Fgithub.com\u002Fsimonmesmith\u002Fagentflow): About Complex LLM Workflows from Simple JSON. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsimonmesmith\u002Fagentflow?style=social)\n- [Lagent](https:\u002F\u002Fgithub.com\u002FInternLM\u002Flagent): A lightweight framework for building LLM-based agents ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FInternLM\u002Flagent?style=social)\n- [Autogen](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen): Enable Next-Gen Large Language Model Applications. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmicrosoft\u002Fautogen?style=social)\n- [AG2](https:\u002F\u002Fgithub.com\u002Fag2ai\u002Fag2): AG2 (from AutoGen creators) is an open-source programming framework for building AI agents and facilitating cooperation among multiple agents to solve tasks. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fag2ai\u002Fag2?style=social)\n- [AgentVerse](https:\u002F\u002Fgithub.com\u002Fopenbmb\u002Fagentverse) Provides a flexible framework that simplifies the process of building custom multi-agent environments for LLMs ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenbmb\u002Fagentverse?style=social)\n- [Maestro](https:\u002F\u002Fgithub.com\u002FDoriandarko\u002Fmaestro): A framework for Claude Opus to intelligently orchestrate subagents. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FDoriandarko\u002Fmaestro?style=social)\n- [AgentScope](https:\u002F\u002Fgithub.com\u002Fmodelscope\u002Fagentscope): Start building LLM-empowered multi-agent applications in an easier way. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmodelscope\u002Fagentscope?style=social)\n- [CrewAI](https:\u002F\u002Fgithub.com\u002Fjoaomdmoura\u002Fcrewai): Cutting-edge framework for orchestrating role-playing, autonomous AI agents. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjoaomdmoura\u002Fcrewai?style=social)\n- [Swarm](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fswarm): Educational framework exploring ergonomic, lightweight multi-agent orchestration by OpenAI. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenai\u002Fswarm?style=social)\n- [agency-swarm](https:\u002F\u002Fgithub.com\u002FVRSEN\u002Fagency-swarm): Reliable agent framework built on top of the latest OpenAI Assistants API. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FVRSEN\u002Fagency-swarm?style=social)\n- [Upsonic](https:\u002F\u002Fgithub.com\u002Fupsonic\u002Fupsonic): Reliable agent framework that supports MCP. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fupsonic\u002Fupsonic?style=social)\n- [Mastra](https:\u002F\u002Fgithub.com\u002Fmastra-ai\u002Fmastra): Mastra is an opinionated TypeScript framework that helps you build AI applications and features quickly. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmastra-ai\u002Fmastra?style=social)\n- [Vectara-agentic](https:\u002F\u002Fgithub.com\u002Fvectara\u002Fpy-vectara-agentic): Vectara-agentic is a framework for creating AI Assistants and agents using Vectara.\n- [Portia AI](https:\u002F\u002Fgithub.com\u002FportiaAI\u002Fportia-sdk-python\u002F) Portia is a new open source agentic Python framework designed for creating reliable agents in production. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FportiaAI\u002Fportia-sdk-python?style=social)\n- [AgentDock](https:\u002F\u002Fgithub.com\u002FAgentDock\u002FAgentDock): Stop wrestling with countless APIs and complex integrations. AgentDock provides the open-source foundation to build, manage, and deploy production-ready AI agents and workflows, frictionlessly. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fagentdock\u002Fagentdock?style=social)\n- [Modus](https:\u002F\u002Fgithub.com\u002Fhypermodeinc\u002Fmodus): An open source, serverless framework for building intelligent agents and APIs in Go or AssemblyScript (a TypeScript-like language).\n- [Swarms Framework](https:\u002F\u002Fgithub.com\u002Fkyegomez\u002Fswarms) Bleeding-Edge Multi-Agent Orchestration Framework for Enterprise Applications ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fkyegomez\u002Fswarms?style=social)\n- [Strands Agents SDK](https:\u002F\u002Fgithub.com\u002Fstrands-agents\u002Fsdk-python): A model-driven approach to building AI agents in just a few lines of code. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstrands-agents\u002Fsdk-python?style=social)\n- [AgentUp](https:\u002F\u002Fgithub.com\u002FRedDotRocket\u002FAgentUp): Designed with security, scalability, and extensibility at its foundation, AgentUp streamlines development through a configuration-driven architecture and rich plugin ecosystem ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FRedDotRocket\u002FAgentUp?style=social)\n- [VoltAgent](https:\u002F\u002Fgithub.com\u002FVoltAgent\u002Fvoltagent) - An open source TypeScript Framework for building AI agents with built-in LLM observability. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fvoltagent\u002Fvoltagent?style=social)\n- [Agentic Context Engine](https:\u002F\u002Fgithub.com\u002Fkayba-ai\u002Fagentic-context-engine): Self-improving agents that learn from execution feedback. LangChain integration for agents that curate their own context. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fkayba-ai\u002Fagentic-context-engine?style=social)\n- [Astron](https:\u002F\u002Fgithub.com\u002Fiflytek\u002Fastron-agent): Enterprise-grade, commercial-friendly agentic workflow platform for building next-generation SuperAgents.\n- [Ailoy](https:\u002F\u002Fgithub.com\u002Fbrekkylab\u002Failoy): A comprehensive framework for building AI agents that run anywhere, with full local-AI and WASM support.\n- [PraisonAI](https:\u002F\u002Fgithub.com\u002FMervinPraison\u002FPraisonAI): Production-ready Multi-AI Agents framework with self-reflection. Fastest agent instantiation (3.77μs), 100+ LLM support, MCP integration, workflows, memory, and both Python & JavaScript SDKs. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FMervinPraison\u002FPraisonAI?style=social)\n- [AgentField](https:\u002F\u002Fgithub.com\u002FAgent-Field\u002Fagentfield): Open-source infrastructure for AI backends. Kubernetes for orchestration, Okta for identity—production-ready from day one. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FAgent-Field\u002Fagentfield?style=social)\n- [Cortex Memory](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fcortex-mem): A complete solution for agent memory, from extraction and vector search to automated optimization, with a REST API, MCP, CLI, and insights dashboard out-of-the-box.\n- [Pipecat](https:\u002F\u002Fgithub.com\u002Fpipecat-ai\u002Fpipecat): Open Source framework for voice and multimodal conversational AI. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpipecat-ai\u002Fpipecat?style=social)\n- [LoongFlow](https:\u002F\u002Fgithub.com\u002Fbaidu-baige\u002FLoongFlow): A Evolve Agent Development Framework. From atomic components and development frameworks to core scenario Agents ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fbaidu-baige\u002FLoongFlow?style=social)\n- [Agentset](https:\u002F\u002Fgithub.com\u002Fagentset-ai\u002Fagentset): Open-source production-ready RAG platform with built-in agentic reasoning, hybrid search, and multimodal support. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fagentset-ai\u002Fagentset?style=social)\n- [Pilot Protocol](https:\u002F\u002Fgithub.com\u002FTeoSlayer\u002Fpilotprotocol): An overlay network stack that gives AI agents virtual addresses, encrypted UDP tunnels, NAT traversal, and mutual trust. Zero dependencies, written in Go. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTeoSlayer\u002Fpilotprotocol?style=social)\n- [hcom](https:\u002F\u002Fgithub.com\u002Faannoo\u002Fhcom): Let AI agents message, watch, and spawn each other across terminals. Claude Code, Gemini CLI, Codex, OpenCode. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Faannoo\u002Fhcom?style=social)\n- [nanobot](https:\u002F\u002Fgithub.com\u002FHKUDS\u002Fnanobot): Ultra-lightweight personal AI assistant framework (~4,000 lines of Python). Supports MCP, 9+ chat channels, and extensible skills system. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FHKUDS\u002Fnanobot?style=social)\n- [NeuroLink](https:\u002F\u002Fgithub.com\u002Fjuspay\u002Fneurolink): TypeScript agent framework with multi-step agentic loops, per-step tool execution control (`prepareStep`), persistent memory (Redis\u002FS3\u002FSQLite), HITL workflows, MCP client integration, and 13 LLM providers. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjuspay\u002Fneurolink?style=social)\n- [OpenClaw](https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw): Open-source AI agent framework that turns LLMs into persistent, proactive personal AI agents with multi-channel messaging (Signal, Telegram, Discord, WhatsApp), cron scheduling, memory systems, MCP integration, skill plugins, sub-agent spawning, and browser automation. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenclaw\u002Fopenclaw?style=social)\n- [Awesome OpenClaw Agents](https:\u002F\u002Fgithub.com\u002Fmergisi\u002Fawesome-openclaw-agents): Curated collection of 100+ production-ready SOUL.md agent templates for OpenClaw, covering productivity, development, marketing, and business automation. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmergisi\u002Fawesome-openclaw-agents?style=social)\n- [FIM Agent](https:\u002F\u002Fgithub.com\u002Ffim-ai\u002Ffim-agent): AI-powered Connector Hub — Python async framework with dynamic DAG planning, ReAct agent, MCP client, RAG with grounded generation, and a connector platform that turns any API into agent tools. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffim-ai\u002Ffim-agent?style=social)\n- [Hive](https:\u002F\u002Fgithub.com\u002Faden-hive\u002Fhive): Open-source AI agent framework for building goal-driven, self-improving autonomous agents. Define outcomes in natural language and auto-generate agent graphs with built-in evolution loops, MCP integration, and 100+ tools. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Faden-hive\u002Fhive?style=social)\n- [ConnectOnion](https:\u002F\u002Fgithub.com\u002Fopenonion\u002Fconnectonion): Simple Python framework for production-ready AI agents — 2-line creation, functions as tools, 12 lifecycle hooks, plugin system, multi-agent networking with trust ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenonion\u002Fconnectonion?style=social)\n- [AVP: Agent Vector Protocol](https:\u002F\u002Fgithub.com\u002FVectorArc\u002Favp-python): Agents communicate via KV-cache and hidden states instead of text — 2x faster, 56% fewer tokens. Supports HuggingFace, vLLM, llama.cpp, Ollama. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FVectorArc\u002Favp-python?style=social)\n\n## Testing and Evaluation\n- [Voice Lab](https:\u002F\u002Fgithub.com\u002Fsaharmor\u002Fvoice-lab): A comprehensive testing and evaluation framework for voice agents across language models, prompts, and agent personas. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsaharmor\u002Fvoice-lab?style=social)\n- [Open-RAG-Eval](https:\u002F\u002Fgithub.com\u002Fvectara\u002Fopen-rag-eval): an open source RAG evaluation framework that does not require golden answers, and can be used to evaluate performance of RAG tools connected to an AI Agent (Agentic RAG)\n- [EvoAgentX](https:\u002F\u002Fgithub.com\u002FEvoAgentX\u002FEvoAgentX): EvoAgentX is building a Self-Evolving Ecosystem of AI Agents, it will give you automated framework for evaluating and evolving agentic workflows. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FEvoAgentX\u002FEvoAgentX?style=social)\n- [Arize-Phoenix](https:\u002F\u002Fgithub.com\u002FArize-ai\u002Fphoenix): Arize-Phoenix is an open source library for agent testing, evaluation and observability. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FArize-ai\u002Fphoenix?style=social)\n- [voicetest](https:\u002F\u002Fgithub.com\u002Fvoicetestdev\u002Fvoicetest): Open-source test harness for voice agents with support for Retell, VAPI, Bland, and LiveKit. Run autonomous simulations and evaluate with LLM judges. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fvoicetestdev\u002Fvoicetest?style=social)\n- [Manifest](https:\u002F\u002Fgithub.com\u002Fmnfst\u002Fmanifest): Open-source, real-time cost observability platform for AI agents. Track tokens, costs, messages, and model usage with a local-first dashboard. Supports 28+ LLM models, OTLP ingestion, self-hosted. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmnfst\u002Fmanifest?style=social)\n* [ai-evaluation](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002Fai-evaluation): Open-source Python\u002FTypeScript SDK for evaluating LLM agents with 50+ local metrics, LLM-as-Judge augmentation, guardrail scanners (jailbreak, PII, injection), streaming assessment, and AutoEval pipelines with CI\u002FCD support. [![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffuture-agi\u002Fai-evaluation?style=social)](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002Fai-evaluation)\n* [traceAI](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002FtraceAI): Open-source OpenTelemetry-native tracing framework that auto-instruments 20+ AI frameworks and LLM providers (OpenAI, Anthropic, LangChain, LlamaIndex, CrewAI, Bedrock) capturing prompts, tokens, latency, and errors out of the box. [![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffuture-agi\u002FtraceAI?style=social)](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002FtraceAI)\n## Software Development\n\n- [MetaGPT](https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT): The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgeekan\u002FMetaGPT?style=social)\n- [OpenHands](https:\u002F\u002Fgithub.com\u002FAll-Hands-AI\u002FOpenHands): 🙌 OpenHands: Code Less, Make More. (formerly OpenDevin), a platform for software development agents powered by AI ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FAll-Hands-AI\u002FOpenHands?style=social)\n- [GPT Pilot](https:\u002F\u002Fgithub.com\u002FPythagora-io\u002Fgpt-pilot): GPT Pilot is the core technology for the Pythagora VS Code extension that aims to provide the first real AI developer companion. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FPythagora-io\u002Fgpt-pilot?style=social)\n- [Aider](https:\u002F\u002Fgithub.com\u002FAider-AI\u002Faider): aider is AI pair programming in your terminal ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FAider-AI\u002Faider?style=social)\n- [Devika](https:\u002F\u002Fgithub.com\u002Fstitionai\u002Fdevika): Devika is an Agentic AI Software Engineer that can understand high-level human instructions, break them down into steps, research relevant information, and write code to achieve the given objective ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstitionai\u002Fdevika?style=social)\n- [RepoAgent](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FRepoAgent): An LLM-powered repository agent designed to assist developers and teams in generating documentation and understanding repositories quickly ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FOpenBMB\u002FRepoAgent?style=social)\n- [DSPy](https:\u002F\u002Fgithub.com\u002Fstanfordnlp\u002Fdspy): The framework for programming—not prompting—foundation models ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstanfordnlp\u002Fdspy?style=social)\n- [ThinkGPT](https:\u002F\u002Fgithub.com\u002Falaeddine-13\u002Fthinkgpt): Agent techniques to augment your LLM and push it beyond its limits ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Falaeddine-13\u002Fthinkgpt?style=social)\n- [PyCodeAGI](https:\u002F\u002Fgithub.com\u002Fchakkaradeep\u002FpyCodeAGI): A small AGI experiment to generate a Python app given what app the user wants to build ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fchakkaradeep\u002FpyCodeAGI?style=social)\n- [SuperAGI](https:\u002F\u002Fgithub.com\u002FTransformerOptimus\u002FSuperAGI): SuperAGI - A dev-first open source autonomous AI agent framework ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTransformerOptimus\u002FSuperAGI?style=social)\n- [Plandex](https:\u002F\u002Fgithub.com\u002Fplandex-ai\u002Fplandex): An AI coding engine for complex tasks ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fplandex-ai\u002Fplandex?style=social)\n- [Codel](https:\u002F\u002Fgithub.com\u002Fsemanser\u002Fcodel): ✨ Fully autonomous AI Agent that can perform complicated tasks and projects using terminal, browser, and editor. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsemanser\u002Fcodel?style=social)\n- [DB GPT](https:\u002F\u002Fgithub.com\u002Fcsunny\u002FDB-GPT): Interact your data and environment using the local GPT, no data leaks, 100% privately, 100% security ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcsunny\u002FDB-GPT?style=social)\n- [Agency](https:\u002F\u002Fgithub.com\u002Fneurocult\u002Fagency): 🕵️‍♂️ Library designed for developers eager to explore the potential of Large Language Models (LLMs) and other generative AI through a clean, effective, and Go-idiomatic approach ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fneurocult\u002Fagency?style=social)\n- [TaskWeaver](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FTaskWeaver): A code-first agent framework for seamlessly planning and executing data analytics tasks. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmicrosoft\u002FTaskWeaver?style=social)\n- [MicroAgent](https:\u002F\u002Fgithub.com\u002Faymenfurter\u002Fmicroagents): Agents Capable of Self-Editing Their Prompts \u002F Python Code ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Faymenfurter\u002Fmicroagents?style=social)\n- [SWE Agent](https:\u002F\u002Fgithub.com\u002Fprinceton-nlp\u002Fswe-agent): SWE-agent takes a GitHub issue and tries to automatically fix it, using GPT-4, or your LM of choice. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fprinceton-nlp\u002Fswe-agent?style=social)\n- [AgentRun](https:\u002F\u002Fgithub.com\u002FJonathan-Adly\u002FAgentRun): The easiest, and fastest way to run AI-generated Python code safely ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FJonathan-Adly\u002FAgentRun?style=social)\n- [Claude Engineer](https:\u002F\u002Fgithub.com\u002FDoriandarko\u002Fclaude-engineer): Claude Engineer is an interactive command-line interface (CLI) that leverages the power of Anthropic's Claude-3.5-Sonnet model to assist with software development tasks.\n- [Vision agent](https:\u002F\u002Fgithub.com\u002Flanding-ai\u002Fvision-agent): Vision Agent is a library that helps you utilize agent frameworks to generate code to solve your vision task. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Flanding-ai\u002Fvision-agent?style=social)\n- [Nous](https:\u002F\u002Fgithub.com\u002FTrafficGuard\u002Fnous): TypeScript AI agent platform with Autonomous agents, Software developer agents, AI code review agents and more. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTrafficGuard\u002Fnous?style=social)\n- [Cline](https:\u002F\u002Fgithub.com\u002Fcline\u002Fcline): Open-source AI coding agent giving developers direct access to frontier models with full transparency. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcline\u002Fcline?style=social)\n- [OpenCode](https:\u002F\u002Fgithub.com\u002Fsst\u002Fopencode): The AI coding agent built for the terminal. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsst\u002Fopencode?style=social)\n- [Stakpak](https:\u002F\u002Fgithub.com\u002Fstakpak\u002Fagent): Open-source DevOps agent to help you secure, deploy, and maintain production-ready infrastructure. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstakpak\u002Fagent?style=social)\n- [ReviewCerberus](https:\u002F\u002Fgithub.com\u002FKirill89\u002Freviewcerberus) - 100% free, open-source AI code review tool for analyzing git branch differences with comprehensive security, performance, and quality analysis.\n- [Hivemoot Colony](https:\u002F\u002Fgithub.com\u002Fhivemoot\u002Fcolony): An autonomous agent governance platform where AI agents vote on features, review code, and ship releases without human intervention. Includes a live governance dashboard. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhivemoot\u002Fcolony?style=social)\n- [Dorothy](https:\u002F\u002Fgithub.com\u002FCharlie85270\u002FDorothy): Open-source desktop app to orchestrate multiple AI CLI agents (Claude Code, Codex, Gemini) simultaneously with automations, Kanban management, remote control, and MCP servers. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FCharlie85270\u002FDorothy?style=social)\n- [VibeGrid](https:\u002F\u002Fgithub.com\u002Fjcanizalez\u002Fvibegrid): Terminal manager for AI coding agents with multi-agent grid, task queues, workflow automation, headless execution, inline diff review, and Claude Code hooks. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjcanizalez\u002Fvibegrid?style=social)\n- [Greywall](https:\u002F\u002Fgithub.com\u002FGreyhavenHQ\u002Fgreywall): Deny-by-default command sandbox for AI coding agents with filesystem isolation, network control via transparent proxy, built-in profiles for agents like Claude Code or OpenCode, and learning mode for auto-generating configs. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FGreyhavenHQ\u002Fgreywall?style=social)\n\n## Research\n\n- [GPT Researcher](https:\u002F\u002Fgithub.com\u002Fassafelovic\u002Fgpt-researcher): GPT Researcher is an autonomous agent designed for comprehensive online research on a variety of tasks. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fassafelovic\u002Fgpt-researcher?style=social)\n- [BlockAGI](https:\u002F\u002Fgithub.com\u002Fblockpipe\u002Fblockagi): BlockAGI conducts iterative, domain-specific research, and outputs detailed narrative reports to showcase its findings ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fblockpipe\u002Fblockagi?style=social)\n- [data-to-paper](https:\u002F\u002Fgithub.com\u002FTechnion-Kishony-lab\u002Fdata-to-paper): data-to-paper: AI-driven research from data to human-verifiable research papers ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTechnion-Kishony-lab\u002Fdata-to-paper?style=social)\n- [AI Scientist](https:\u002F\u002Fgithub.com\u002FSakanaAI\u002FAI-Scientist): The AI Scientist: Towards Fully Automated Open-Ended Scientific ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FSakanaAI\u002FAI-Scientist?style=social)\n- [Storm](https:\u002F\u002Fgithub.com\u002Fstanford-oval\u002Fstorm): An LLM-powered knowledge curation system that researches a topic and generates a full-length report with citations. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstanford-oval\u002Fstorm?style=social)\n- [OpenLens AI](https:\u002F\u002Fgithub.com\u002Fjarrycyx\u002Fopenlens-ai): Fully Autonomous Research Agent for Health Infomatics ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjarrycyx\u002Fopenlens-ai?style=social)\n- [GenoMAS](https:\u002F\u002Fgithub.com\u002FLiu-Hy\u002FGenoMAS): A multi-agent framework for scientific discovery that automates gene expression analysis through code-driven workflows. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FLiu-Hy\u002FGenoMAS?style=social)\n- [DeepAnalyze](https:\u002F\u002Fgithub.com\u002Fruc-datalab\u002FDeepAnalyze): first agentic LLM for autonomous data science, supporting specific data tasks (data preparation, analysis, modeling, visualization, and insight) and data-oriented deep research (produce analyst-grade research reports). [![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fruc-datalab\u002FDeepAnalyze?style=social&label=Code+Stars)](https:\u002F\u002Fgithub.com\u002Fruc-datalab\u002FDeepAnalyze)\n- [AIDE](https:\u002F\u002Fgithub.com\u002FWecoAI\u002Faideml): AI-Driven Exploration — ML engineering agent that uses tree search to automate experiment design, code generation, and evaluation against any metric. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FWecoAI\u002Faideml?style=social)\n\n## Conversational \u002F General Agents\n\n- [CollosalAI Chat](https:\u002F\u002Fgithub.com\u002Fhpcaitech\u002FColossalAI\u002Ftree\u002Fmain\u002Fapplications\u002FChat): implement LLM with RLHF, powered by the Colossal-AI project ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhpcaitech\u002FColossalAI?style=social)\n- [RasaGPT](https:\u002F\u002Fgithub.com\u002Fpaulpierre\u002FRasaGPT): RasaGPT is the first headless LLM chatbot platform built on top of Rasa and Langchain. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpaulpierre\u002FRasaGPT?style=social)\n- [SuperAgent](https:\u002F\u002Fgithub.com\u002Fhomanp\u002Fsuperagent): Deploy LLM Agents to production ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhomanp\u002Fsuperagent?style=social)\n- [BabyAGI UI](https:\u002F\u002Fgithub.com\u002Fmiurla\u002Fbabyagi-ui): Make it easier to run and develop with babyagi in a web app, like a ChatGPT ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmiurla\u002Fbabyagi-ui?style=social)\n- [ix](https:\u002F\u002Fgithub.com\u002Fkreneskyp\u002Fix): Autonomous GPT-4 agent platform ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fkreneskyp\u002Fix?style=social)\n- [Multi-Modal LangChain agents in Production](https:\u002F\u002Fgithub.com\u002Fsteamship-packages\u002Flangchain-agent-production-starter): Deploy LangChain Agents and connect them to Telegram ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsteamship-packages\u002Flangchain-agent-production-starter?style=social)\n- [Autonomous HR Chatbot](https:\u002F\u002Fgithub.com\u002Fstepanogil\u002Fautonomous-hr-chatbot): An autonomous agent that can answer HR related queries autonomously using the tools it has on hand ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstepanogil\u002Fautonomous-hr-chatbot?style=social)\n- [LLama Cpp Agent](https:\u002F\u002Fgithub.com\u002FMaximilian-Winter\u002Fllama-cpp-agent): The llama-cpp-agent framework is a tool designed for easy interaction with Large Language Models ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FMaximilian-Winter\u002Fllama-cpp-agent?style=social)\n- [Memgpt](https:\u002F\u002Fgithub.com\u002Fcpacker\u002Fmemgpt): Create LLM agents with long-term memory and custom tools 📚🦙 ![Github Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcpacker\u002Fmemgpt?style=social)\n- [joinly](https:\u002F\u002Fgithub.com\u002Fjoinly-ai\u002Fjoinly): Voice-first AI Assistant for online meetings that can actively participate and solve tasks live during the meeting ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjoinly-ai\u002Fjoinly?style=social)\n- [Gobii](https:\u002F\u002Fgithub.com\u002Fgobii-ai\u002Fgobii-platform): Gobii is an open-source platform for deploying and managing browser-use agents at scale with a conversational interface and API ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgobii-ai\u002Fgobii-platform?style=social)\n- [ClaudeClaw](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fclaudeclaw): Persistent agent orchestrator plugin for Claude Code — multi-channel routing (Slack, WhatsApp, Telegram), OS-level sandbox isolation, composable extensions, structured memory, webhook triggers ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsbusso\u002Fclaudeclaw?style=social)\n\n## Game \u002F Simulation\n\n- [Camel-AutoGPT](https:\u002F\u002Fgithub.com\u002FSamurAIGPT\u002FCamel-AutoGPT): role-playing approach for LLMs and auto-agents like BabyAGI & AutoGPT ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FSamurAIGPT\u002FCamel-AutoGPT?style=social)\n- [SkyAGI](https:\u002F\u002Fgithub.com\u002Flitanlitudan\u002Fskyagi): Emerging human-behavior simulation capability in LLM agents ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Flitanlitudan\u002Fskyagi?style=social)\n- [Voyager](https:\u002F\u002Fgithub.com\u002FMineDojo\u002FVoyager): An Open-Ended Embodied Agent with Large Language Models ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FMineDojo\u002FVoyager?style=social)\n\n## Knowledge Management\n\n- [Private GPT](https:\u002F\u002Fgithub.com\u002Fimartinez\u002FprivateGPT): Interact privately with your documents using the power of GPT, 100% privately, no data leaks ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fimartinez\u002FprivateGPT?style=social)\n- [Local GPT](https:\u002F\u002Fgithub.com\u002FPromtEngineer\u002FlocalGPT): Inspired on Private GPT with the GPT4ALL model replaced with the Vicuna-7B model and using the InstructorEmbeddings instead of LlamaEmbeddings ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FPromtEngineer\u002FlocalGPT?style=social)\n- [LLocalSearch](https:\u002F\u002Fgithub.com\u002Fnilsherzig\u002FLLocalSearch): LLocalSearch is a completely locally running search aggregator using LLM Agents. The user can ask a question and the system will use a chain of LLMs to find the answer. The user can see the progress of the agents and the final answer. No OpenAI or Google API keys are needed. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fnilsherzig\u002FLLocalSearch?style=social)\n- [Second Brain AI Agent](https:\u002F\u002Fgithub.com\u002Fflepied\u002Fsecond-brain-agent): A streamlit app to dialog with your second brain notes using OpenAI and ChromaDB locally. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fflepied\u002Fsecond-brain-agent?style=social)\n- [SAGE](https:\u002F\u002Fgithub.com\u002Fl33tdawg\u002Fsage): Institutional memory for AI agents — every memory goes through BFT consensus before it's committed. 4 application validators, 13 MCP tools, runs locally. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fl33tdawg\u002Fsage?style=social)\n\n## Automation\n\n- [DemoGPT](https:\u002F\u002Fgithub.com\u002Fmelih-unsal\u002FDemoGPT): DemoGPT enables you to create quick demos by just using prompt. It applies ToT approach on Langchain documentation tree. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmelih-unsal\u002FDemoGPT?style=social)\n- [RestGPT](https:\u002F\u002Fgithub.com\u002FYifan-Song793\u002FRestGPT): An LLM-based autonomous agent controlling real-world applications via RESTful APIs ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FYifan-Song793\u002FRestGPT?style=social)\n- [XAgent](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FXAgent): An Autonomous LLM Agent for Complex Task Solving ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FOpenBMB\u002FXAgent?style=social)\n- [LLM Agents](https:\u002F\u002Fgithub.com\u002Fmpaepper\u002Fllm_agents): Build agents which are controlled by LLMs ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmpaepper\u002Fllm_agents?style=social)\n- [uAgents](https:\u002F\u002Fgithub.com\u002Ffetchai\u002FuAgents): A fast and lightweight framework for creating decentralized agents with ease. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffetchai\u002FuAgents?style=social)\n- [Maestro](https:\u002F\u002Fgithub.com\u002FDoriandarko\u002Fmaestro): A framework for Claude Opus to intelligently orchestrate subagents. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FDoriandarko\u002Fmaestro?style=social)\n- [llama-agents](https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama-agents): llama-agents is an async-first framework for building, iterating, and productionizing multi-agent systems, including multi-agent communication, distributed tool execution, human-in-the-loop, and more ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Frun-llama\u002Fllama-agents?style=social)\n- [Phidata](https:\u002F\u002Fgithub.com\u002Fphidatahq\u002Fphidata): Build AI Assistants with memory, knowledge and tools. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fphidatahq\u002Fphidata?style=social)\n- [AgentK](https:\u002F\u002Fgithub.com\u002Fmikekelly\u002FAgentK): An autoagentic AGI that is self-evolving and modular. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmikekelly\u002FAgentK?style=social)\n- [ADAS](https:\u002F\u002Fgithub.com\u002FShengranHu\u002FADAS): Automated Design of Agentic Systems ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FShengranHu\u002FADAS?style=social)\n- [Giselle](https:\u002F\u002Fgithub.com\u002Fgiselles-ai\u002Fgiselle): Giselle is an agentic workflow builder that empowers you to create AI-driven solutions with ease. ![Github Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgiselles-ai\u002Fgiselle?style=social)\n- [Untether](https:\u002F\u002Fgithub.com\u002Flittlebearapps\u002Funtether): Telegram bridge for AI coding agents — remote task control with progress streaming, interactive permissions, voice input, and cost tracking. Supports Claude Code, Codex, OpenCode, Pi, Gemini CLI, and Amp. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Flittlebearapps\u002Funtether?style=social)\n\n### Browser\n\n- [AgentGPT](https:\u002F\u002Fgithub.com\u002Freworkd\u002FAgentGPT): AI Agents with Langchain & OpenAI (Vercel \u002F Nextjs) ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Freworkd\u002FAgentGPT?style=social)\n- [OpenAgents](https:\u002F\u002Fgithub.com\u002Fxlang-ai\u002FOpenAgents): An Open Platform for Language Agents in the Wild ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fxlang-ai\u002FOpenAgents?style=social)\n- [Steel Browser](https:\u002F\u002Fgithub.com\u002Fsteel-dev\u002Fsteel-browser): Open-source browser infrastructure for AI agents with session-backed automation, extraction, screenshots\u002FPDFs, an AI-native CLI, and a reusable steel-browser skill. ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsteel-dev\u002Fsteel-browser?style=social)\n","# 🤖 精选 AI 代理工具\n\nAwesome Agents 是一份精选的开源工具与产品清单，用于构建 AI 代理（AI agents）。\n\n> 🆕 **推荐项目：** [ClaudeClaw](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fclaudeclaw) —— 作为 Claude Code 插件的持久化代理编排器。支持多通道路由（Slack、WhatsApp、Telegram），操作系统级沙箱隔离，可组合扩展系统。[了解更多 →](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fclaudeclaw)\n\n## 目录\n\n- [🤖 精选 AI 代理工具](#-精选-ai-代理工具)\n  - [目录](#目录)\n  - [框架（Frameworks）](#框架)\n  - [测试与评估（Testing and Evaluation）](#测试与评估)\n  - [软件开发（Software Development）](#软件开发)\n  - [研究（Research）](#研究)\n  - [对话型 \u002F 通用代理（Conversational \u002F General Agents）](#对话型--通用代理)\n  - [游戏 \u002F 模拟（Game \u002F Simulation）](#游戏--模拟)\n  - [知识管理（Knowledge Management）](#知识管理)\n  - [自动化（Automation）](#自动化)\n    - [浏览器（Browser）](#浏览器)\n    - [多模态（Multimodal）](#多模态)\n\n## 框架\n\n- [llama-agentic-system](https:\u002F\u002Fgithub.com\u002Fmeta-llama\u002Fllama-agentic-system)：Llama Stack API 的智能体（Agentic）组件 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmeta-llama\u002Fllama-agentic-system?style=social)\n- [Transformers Agents](https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Ftransformers\u002Ftransformers_agents)：在 Transformers 之上提供自然语言 API\n- [LlamaIndex](https:\u002F\u002Fgithub.com\u002Fjerryjliu\u002Fllama_index)：提供一个中心化接口，将你的大语言模型（LLM）与外部数据连接起来。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjerryjliu\u002Fllama_index?style=social)\n- [LangChain](https:\u002F\u002Fgithub.com\u002Fhwchase17\u002Flangchain)：最初的 🐍 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhwchase17\u002Flangchain?style=social)\n- [Botpress](https:\u002F\u002Fgithub.com\u002Fbotpress\u002Fbotpress)：构建聊天机器人的基础模块 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fbotpress\u002Fbotpress?style=social)\n- [Haystack](https:\u002F\u002Fgithub.com\u002Fdeepset-ai\u002Fhaystack)：使用 Transformer 模型和大语言模型（LLM）与你的数据交互的 NLP 框架 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fdeepset-ai\u002Fhaystack?style=social)\n- [Semantic Kernel](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fsemantic-kernel)：微软 C# SDK，可快速轻松地将前沿的大语言模型（LLM）技术集成到你的应用中 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmicrosoft\u002Fsemantic-kernel?style=social)\n- [Agent-LLM](https:\u002F\u002Fgithub.com\u002FJosh-XT\u002FAgent-LLM)：一个人工智能自动化平台。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FJosh-XT\u002FAgent-LLM?style=social)\n- [LLM Agents](https:\u002F\u002Fgithub.com\u002Fmpaepper\u002Fllm_agents)：构建由大语言模型（LLM）控制的智能体 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmpaepper\u002Fllm_agents?style=social)\n- [e2b](https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fe2b)：用于构建和部署虚拟开发者智能体的开源平台\n- [Dust](https:\u002F\u002Fgithub.com\u002Fdust-tt\u002Fdust)：设计并部署大语言模型（LLM）应用 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fdust-tt\u002Fdust?style=social)\n- [MetaGPT](https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT)：多智能体元编程框架：输入一行需求，自动生成产品需求文档（PRD）、设计、任务、代码仓库和持续集成（CI） ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgeekan\u002FMetaGPT?style=social)\n- [AgentFlow](https:\u002F\u002Fgithub.com\u002Fsimonmesmith\u002Fagentflow)：通过简单 JSON 构建复杂的大语言模型（LLM）工作流。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsimonmesmith\u002Fagentflow?style=social)\n- [Lagent](https:\u002F\u002Fgithub.com\u002FInternLM\u002Flagent)：用于构建基于大语言模型（LLM）智能体的轻量级框架 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FInternLM\u002Flagent?style=social)\n- [Autogen](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fautogen)：赋能下一代大语言模型（LLM）应用。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmicrosoft\u002Fautogen?style=social)\n- [AG2](https:\u002F\u002Fgithub.com\u002Fag2ai\u002Fag2)：AG2（由 AutoGen 团队开发）是一个开源编程框架，用于构建 AI 智能体并促进多个智能体协作完成任务。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fag2ai\u002Fag2?style=social)\n- [AgentVerse](https:\u002F\u002Fgithub.com\u002Fopenbmb\u002Fagentverse) 提供灵活框架，简化为大语言模型（LLM）构建自定义多智能体环境的过程 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenbmb\u002Fagentverse?style=social)\n- [Maestro](https:\u002F\u002Fgithub.com\u002FDoriandarko\u002Fmaestro)：一个让 Claude Opus 能够智能编排子智能体的框架。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FDoriandarko\u002Fmaestro?style=social)\n- [AgentScope](https:\u002F\u002Fgithub.com\u002Fmodelscope\u002Fagentscope)：以更简单的方式开始构建由大语言模型（LLM）驱动的多智能体应用。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmodelscope\u002Fagentscope?style=social)\n- [CrewAI](https:\u002F\u002Fgithub.com\u002Fjoaomdmoura\u002Fcrewai)：用于编排角色扮演式自主 AI 智能体的前沿框架。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjoaomdmoura\u002Fcrewai?style=social)\n- [Swarm](https:\u002F\u002Fgithub.com\u002Fopenai\u002Fswarm)：由 OpenAI 开发的教育性框架，探索高效、轻量的多智能体编排。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenai\u002Fswarm?style=social)\n- [agency-swarm](https:\u002F\u002Fgithub.com\u002FVRSEN\u002Fagency-swarm)：基于最新 OpenAI Assistants API 构建的可靠智能体框架。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FVRSEN\u002Fagency-swarm?style=social)\n- [Upsonic](https:\u002F\u002Fgithub.com\u002Fupsonic\u002Fupsonic)：支持 MCP 的可靠智能体框架。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fupsonic\u002Fupsonic?style=social)\n- [Mastra](https:\u002F\u002Fgithub.com\u002Fmastra-ai\u002Fmastra)：Mastra 是一个有明确设计哲学的 TypeScript 框架，帮助你快速构建 AI 应用和功能。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmastra-ai\u002Fmastra?style=social)\n- [Vectara-agentic](https:\u002F\u002Fgithub.com\u002Fvectara\u002Fpy-vectara-agentic)：Vectara-agentic 是一个使用 Vectara 创建 AI 助手和智能体的框架。\n- [Portia AI](https:\u002F\u002Fgithub.com\u002FportiaAI\u002Fportia-sdk-python\u002F) Portia 是一个全新的开源智能体 Python 框架，专为在生产环境中构建可靠的智能体而设计。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FportiaAI\u002Fportia-sdk-python?style=social)\n- [AgentDock](https:\u002F\u002Fgithub.com\u002FAgentDock\u002FAgentDock)：无需再与无数 API 和复杂集成搏斗。AgentDock 提供开源基础，让你无摩擦地构建、管理和部署可用于生产的 AI 智能体和工作流。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fagentdock\u002Fagentdock?style=social)\n- [Modus](https:\u002F\u002Fgithub.com\u002Fhypermodeinc\u002Fmodus)：一个开源、无服务器框架，用于使用 Go 或 AssemblyScript（类似 TypeScript 的语言）构建智能体和 API。\n- [Swarms Framework](https:\u002F\u002Fgithub.com\u002Fkyegomez\u002Fswarms) 面向企业应用的前沿多智能体编排框架 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fkyegomez\u002Fswarms?style=social)\n- [Strands Agents SDK](https:\u002F\u002Fgithub.com\u002Fstrands-agents\u002Fsdk-python)：一种模型驱动的方法，仅需几行代码即可构建 AI 智能体。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstrands-agents\u002Fsdk-python?style=social)\n- [AgentUp](https:\u002F\u002Fgithub.com\u002FRedDotRocket\u002FAgentUp)：以安全性、可扩展性和可扩展性为设计基础，通过配置驱动架构和丰富的插件生态系统简化开发流程 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FRedDotRocket\u002FAgentUp?style=social)\n- [VoltAgent](https:\u002F\u002Fgithub.com\u002FVoltAgent\u002Fvoltagent) - 一个开源 TypeScript 框架，用于构建具备内置大语言模型（LLM）可观测性的 AI 智能体。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fvoltagent\u002Fvoltagent?style=social)\n- [Agentic Context Engine](https:\u002F\u002Fgithub.com\u002Fkayba-ai\u002Fagentic-context-engine)：能从执行反馈中学习的自我改进智能体。集成 LangChain，使智能体能够自主管理上下文。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fkayba-ai\u002Fagentic-context-engine?style=social)\n- [Astron](https:\u002F\u002Fgithub.com\u002Fiflytek\u002Fastron-agent)：面向企业级、商业友好的智能体工作流平台，用于构建下一代 SuperAgents。\n- [Ailoy](https:\u002F\u002Fgithub.com\u002Fbrekkylab\u002Failoy)：一个全面的框架，用于构建可在任何地方运行的 AI 智能体，支持完整的本地 AI 和 WASM。\n- [PraisonAI](https:\u002F\u002Fgithub.com\u002FMervinPraison\u002FPraisonAI)：可用于生产的多 AI 智能体框架，具备自我反思能力。最快智能体实例化速度（3.77μs），支持 100+ 大语言模型（LLM），集成 MCP、工作流、记忆系统，并提供 Python 和 JavaScript SDK。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FMervinPraison\u002FPraisonAI?style=social)\n- [AgentField](https:\u002F\u002Fgithub.com\u002FAgent-Field\u002Fagentfield)：面向 AI 后端的开源基础设施。使用 Kubernetes 进行编排，Okta 管理身份——从第一天起即可投入生产。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FAgent-Field\u002Fagentfield?style=social)\n- [Cortex Memory](https:\u002F\u002Fgithub.com\u002Fsopaco\u002Fcortex-mem)：完整的智能体记忆解决方案，涵盖提取、向量搜索到自动优化，开箱即用 REST API、MCP、CLI 和洞察仪表盘。\n- [Pipecat](https:\u002F\u002Fgithub.com\u002Fpipecat-ai\u002Fpipecat)：开源语音与多模态对话式 AI 框架。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpipecat-ai\u002Fpipecat?style=social)\n- [LoongFlow](https:\u002F\u002Fgithub.com\u002Fbaidu-baige\u002FLoongFlow)：演进式智能体开发框架。从原子组件、开发框架到核心场景智能体 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fbaidu-baige\u002FLoongFlow?style=social)\n- [Agentset](https:\u002F\u002Fgithub.com\u002Fagentset-ai\u002Fagentset)：开源、可用于生产的 RAG 平台，内置智能体推理、混合搜索和多模态支持。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fagentset-ai\u002Fagentset?style=social)\n- [Pilot Protocol](https:\u002F\u002Fgithub.com\u002FTeoSlayer\u002Fpilotprotocol)：一种覆盖网络协议栈，赋予 AI 智能体虚拟地址、加密 UDP 隧道、NAT 穿透和互信机制。零依赖，使用 Go 编写。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTeoSlayer\u002Fpilotprotocol?style=social)\n- [hcom](https:\u002F\u002Fgithub.com\u002Faannoo\u002Fhcom)：让 AI 智能体跨终端互相发送消息、监控和生成彼此。支持 Claude Code、Gemini CLI、Codex、OpenCode。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Faannoo\u002Fhcom?style=social)\n- [nanobot](https:\u002F\u002Fgithub.com\u002FHKUDS\u002Fnanobot)：超轻量级个人 AI 助手框架（约 4,000 行 Python 代码）。支持 MCP、9+ 聊天渠道和可扩展技能系统。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FHKUDS\u002Fnanobot?style=social)\n- [NeuroLink](https:\u002F\u002Fgithub.com\u002Fjuspay\u002Fneurolink)：TypeScript 智能体框架，支持多步骤智能体循环、每步工具执行控制（`prepareStep`）、持久化记忆（Redis\u002FS3\u002FSQLite）、人机协同（HITL）工作流、MCP 客户端集成和 13 个大语言模型（LLM）提供商。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjuspay\u002Fneurolink?style=social)\n- [OpenClaw](https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw)：开源 AI 智能体框架，可将大语言模型（LLM）转变为持久、主动的个人 AI 智能体，支持多通道消息（Signal、Telegram、Discord、WhatsApp）、定时任务、记忆系统、MCP 集成、技能插件、子智能体生成和浏览器自动化。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenclaw\u002Fopenclaw?style=social)\n- [Awesome OpenClaw Agents](https:\u002F\u002Fgithub.com\u002Fmergisi\u002Fawesome-openclaw-agents)：精选的 100+ 个可用于生产的 SOUL.md 智能体模板集合，适用于 OpenClaw，涵盖生产力、开发、营销和业务自动化。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmergisi\u002Fawesome-openclaw-agents?style=social)\n- [FIM Agent](https:\u002F\u002Fgithub.com\u002Ffim-ai\u002Ffim-agent)：AI 驱动的连接器中心 —— Python 异步框架，支持动态 DAG 规划、ReAct 智能体、MCP 客户端、带事实依据生成的 RAG，以及可将任意 API 转换为智能体工具的连接器平台。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffim-ai\u002Ffim-agent?style=social)\n- [Hive](https:\u002F\u002Fgithub.com\u002Faden-hive\u002Fhive)：开源 AI 智能体框架，用于构建目标驱动、自我改进的自主智能体。用自然语言定义目标，自动生成智能体图谱，内置进化循环、MCP 集成和 100+ 工具。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Faden-hive\u002Fhive?style=social)\n- [ConnectOnion](https:\u002F\u002Fgithub.com\u002Fopenonion\u002Fconnectonion)：简单的 Python 框架，用于构建可用于生产的 AI 智能体 —— 两行代码创建，函数作为工具，12 个生命周期钩子，插件系统，支持信任机制的多智能体网络 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fopenonion\u002Fconnectonion?style=social)\n- [AVP: Agent Vector Protocol](https:\u002F\u002Fgithub.com\u002FVectorArc\u002Favp-python)：智能体通过 KV 缓存和隐藏状态而非文本进行通信 —— 速度提升 2 倍，减少 56% 的 token 使用。支持 HuggingFace、vLLM、llama.cpp、Ollama。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FVectorArc\u002Favp-python?style=social)\n\n## 测试与评估\n- [Voice Lab](https:\u002F\u002Fgithub.com\u002Fsaharmor\u002Fvoice-lab)：一个全面的语音代理（voice agents）测试与评估框架，支持跨语言模型、提示词（prompts）和代理角色（agent personas）进行评测。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsaharmor\u002Fvoice-lab?style=social)\n- [Open-RAG-Eval](https:\u002F\u002Fgithub.com\u002Fvectara\u002Fopen-rag-eval)：一个开源的 RAG（检索增强生成，Retrieval-Augmented Generation）评估框架，无需黄金答案（golden answers），可用于评估连接到 AI 代理的 RAG 工具性能（Agentic RAG）。\n- [EvoAgentX](https:\u002F\u002Fgithub.com\u002FEvoAgentX\u002FEvoAgentX)：EvoAgentX 正在构建一个自我演化的 AI 代理生态系统，为你提供自动化框架，用于评估和演化代理工作流（agentic workflows）。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FEvoAgentX\u002FEvoAgentX?style=social)\n- [Arize-Phoenix](https:\u002F\u002Fgithub.com\u002FArize-ai\u002Fphoenix)：Arize-Phoenix 是一个开源库，用于代理测试、评估与可观测性（observability）。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FArize-ai\u002Fphoenix?style=social)\n- [voicetest](https:\u002F\u002Fgithub.com\u002Fvoicetestdev\u002Fvoicetest)：开源语音代理测试平台，支持 Retell、VAPI、Bland 和 LiveKit。可运行自主模拟，并使用 LLM 作为评委进行评估。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fvoicetestdev\u002Fvoicetest?style=social)\n- [Manifest](https:\u002F\u002Fgithub.com\u002Fmnfst\u002Fmanifest)：开源、实时成本可观测性平台，专为 AI 代理设计。通过本地优先仪表盘追踪 token、成本、消息及模型使用情况。支持 28+ 种 LLM 模型，支持 OTLP 数据摄入，可自托管。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmnfst\u002Fmanifest?style=social)\n* [ai-evaluation](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002Fai-evaluation)：开源 Python\u002FTypeScript SDK，用于评估 LLM 代理，内置 50+ 项本地指标、LLM-as-Judge 增强功能、防护扫描器（越狱、个人身份信息 PII、注入攻击）、流式评估，以及支持 CI\u002FCD 的 AutoEval 自动化评估流水线。[![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffuture-agi\u002Fai-evaluation?style=social)](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002Fai-evaluation)\n* [traceAI](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002FtraceAI)：开源、原生支持 OpenTelemetry 的追踪框架，自动插桩 20+ 种 AI 框架与 LLM 提供商（如 OpenAI、Anthropic、LangChain、LlamaIndex、CrewAI、Bedrock），开箱即用捕获提示词、token、延迟与错误信息。[![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffuture-agi\u002FtraceAI?style=social)](https:\u002F\u002Fgithub.com\u002Ffuture-agi\u002FtraceAI)\n## 软件开发\n\n- [MetaGPT](https:\u002F\u002Fgithub.com\u002Fgeekan\u002FMetaGPT)：多智能体（Multi-Agent）框架：首家 AI 软件公司，迈向自然语言编程 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgeekan\u002FMetaGPT?style=social)\n- [OpenHands](https:\u002F\u002Fgithub.com\u002FAll-Hands-AI\u002FOpenHands)：🙌 OpenHands：少写代码，多做创造。（原名 OpenDevin），一个由 AI 驱动的软件开发智能体平台 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FAll-Hands-AI\u002FOpenHands?style=social)\n- [GPT Pilot](https:\u002F\u002Fgithub.com\u002FPythagora-io\u002Fgpt-pilot)：GPT Pilot 是 Pythagora VS Code 扩展的核心技术，旨在提供首个真正的 AI 开发者助手。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FPythagora-io\u002Fgpt-pilot?style=social)\n- [Aider](https:\u002F\u002Fgithub.com\u002FAider-AI\u002Faider)：aider 是在终端中实现的 AI 结对编程工具 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FAider-AI\u002Faider?style=social)\n- [Devika](https:\u002F\u002Fgithub.com\u002Fstitionai\u002Fdevika)：Devika 是一名具备自主能力的 AI 软件工程师，能够理解高级人类指令，将其分解为步骤，研究相关信息，并编写代码以完成指定目标 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstitionai\u002Fdevika?style=social)\n- [RepoAgent](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FRepoAgent)：一个基于大语言模型（LLM）的仓库智能体，旨在帮助开发者和团队快速生成文档并理解代码仓库 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FOpenBMB\u002FRepoAgent?style=social)\n- [DSPy](https:\u002F\u002Fgithub.com\u002Fstanfordnlp\u002Fdspy)：用于“编程”——而非“提示”——基础模型的框架 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstanfordnlp\u002Fdspy?style=social)\n- [ThinkGPT](https:\u002F\u002Fgithub.com\u002Falaeddine-13\u002Fthinkgpt)：通过智能体技术增强你的 LLM，突破其能力极限 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Falaeddine-13\u002Fthinkgpt?style=social)\n- [PyCodeAGI](https:\u002F\u002Fgithub.com\u002Fchakkaradeep\u002FpyCodeAGI)：一项小型 AGI 实验，可根据用户想构建的应用自动生成 Python 应用程序 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fchakkaradeep\u002FpyCodeAGI?style=social)\n- [SuperAGI](https:\u002F\u002Fgithub.com\u002FTransformerOptimus\u002FSuperAGI)：SuperAGI —— 一款面向开发者的开源自主 AI 智能体框架 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTransformerOptimus\u002FSuperAGI?style=social)\n- [Plandex](https:\u002F\u002Fgithub.com\u002Fplandex-ai\u002Fplandex)：专为复杂任务设计的 AI 编码引擎 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fplandex-ai\u002Fplandex?style=social)\n- [Codel](https:\u002F\u002Fgithub.com\u002Fsemanser\u002Fcodel)：✨ 完全自主的 AI 智能体，可通过终端、浏览器和编辑器执行复杂任务与项目。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsemanser\u002Fcodel?style=social)\n- [DB GPT](https:\u002F\u002Fgithub.com\u002Fcsunny\u002FDB-GPT)：使用本地 GPT 与你的数据和环境交互，无数据泄露，100% 私密，100% 安全 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcsunny\u002FDB-GPT?style=social)\n- [Agency](https:\u002F\u002Fgithub.com\u002Fneurocult\u002Fagency)：🕵️‍♂️ 专为渴望通过简洁、高效且符合 Go 语言习惯的方式探索大语言模型（LLMs）及其他生成式 AI 潜力的开发者设计的库 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fneurocult\u002Fagency?style=social)\n- [TaskWeaver](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FTaskWeaver)：一个以代码为核心的智能体框架，可无缝规划并执行数据分析任务。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmicrosoft\u002FTaskWeaver?style=social)\n- [MicroAgent](https:\u002F\u002Fgithub.com\u002Faymenfurter\u002Fmicroagents)：能够自我编辑提示词 \u002F Python 代码的智能体 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Faymenfurter\u002Fmicroagents?style=social)\n- [SWE Agent](https:\u002F\u002Fgithub.com\u002Fprinceton-nlp\u002Fswe-agent)：SWE-agent 接收 GitHub issue 并尝试使用 GPT-4 或你选择的语言模型自动修复它。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fprinceton-nlp\u002Fswe-agent?style=social)\n- [AgentRun](https:\u002F\u002Fgithub.com\u002FJonathan-Adly\u002FAgentRun)：安全运行 AI 生成的 Python 代码最简单、最快捷的方式 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FJonathan-Adly\u002FAgentRun?style=social)\n- [Claude Engineer](https:\u002F\u002Fgithub.com\u002FDoriandarko\u002Fclaude-engineer)：Claude Engineer 是一个交互式命令行界面（CLI），利用 Anthropic 的 Claude-3.5-Sonnet 模型协助完成软件开发任务。\n- [Vision agent](https:\u002F\u002Fgithub.com\u002Flanding-ai\u002Fvision-agent)：Vision Agent 是一个库，帮助你利用智能体框架生成代码以解决视觉任务。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Flanding-ai\u002Fvision-agent?style=social)\n- [Nous](https:\u002F\u002Fgithub.com\u002FTrafficGuard\u002Fnous)：TypeScript AI 智能体平台，包含自主智能体、软件开发智能体、AI 代码审查智能体等。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTrafficGuard\u002Fnous?style=social)\n- [Cline](https:\u002F\u002Fgithub.com\u002Fcline\u002Fcline)：开源 AI 编码智能体，让开发者直接透明地访问前沿模型。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcline\u002Fcline?style=social)\n- [OpenCode](https:\u002F\u002Fgithub.com\u002Fsst\u002Fopencode)：专为终端打造的 AI 编码智能体。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsst\u002Fopencode?style=social)\n- [Stakpak](https:\u002F\u002Fgithub.com\u002Fstakpak\u002Fagent)：开源 DevOps 智能体，帮助你安全部署并维护生产级基础设施。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstakpak\u002Fagent?style=social)\n- [ReviewCerberus](https:\u002F\u002Fgithub.com\u002FKirill89\u002Freviewcerberus) - 100% 免费、开源的 AI 代码审查工具，用于分析 Git 分支差异，提供全面的安全性、性能和质量分析。\n- [Hivemoot Colony](https:\u002F\u002Fgithub.com\u002Fhivemoot\u002Fcolony)：一个自主智能体治理平台，AI 智能体可投票决定功能、审查代码并发布版本，无需人工干预。包含实时治理仪表盘。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhivemoot\u002Fcolony?style=social)\n- [Dorothy](https:\u002F\u002Fgithub.com\u002FCharlie85270\u002FDorothy)：开源桌面应用，可同时编排多个 AI CLI 智能体（如 Claude Code、Codex、Gemini），支持自动化、看板管理、远程控制及 MCP 服务器。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FCharlie85270\u002FDorothy?style=social)\n- [VibeGrid](https:\u002F\u002Fgithub.com\u002Fjcanizalez\u002Fvibegrid)：AI 编码智能体的终端管理器，支持多智能体网格、任务队列、工作流自动化、无头执行、内联差异审查及 Claude Code 钩子。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjcanizalez\u002Fvibegrid?style=social)\n- [Greywall](https:\u002F\u002Fgithub.com\u002FGreyhavenHQ\u002Fgreywall)：默认拒绝命令的沙箱环境，专为 AI 编码智能体设计，具备文件系统隔离、通过透明代理进行网络控制、内置 Claude Code 或 OpenCode 等智能体配置文件，并支持学习模式以自动生成配置。![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FGreyhavenHQ\u002Fgreywall?style=social)\n\n## 研究\n\n- [GPT Researcher](https:\u002F\u002Fgithub.com\u002Fassafelovic\u002Fgpt-researcher)：GPT Researcher 是一个自主代理（autonomous agent），专为执行多种任务的全面在线研究而设计。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fassafelovic\u002Fgpt-researcher?style=social)\n- [BlockAGI](https:\u002F\u002Fgithub.com\u002Fblockpipe\u002Fblockagi)：BlockAGI 进行迭代式、特定领域的研究，并输出详细的叙事报告以展示其发现。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fblockpipe\u002Fblockagi?style=social)\n- [data-to-paper](https:\u002F\u002Fgithub.com\u002FTechnion-Kishony-lab\u002Fdata-to-paper)：data-to-paper：由 AI 驱动，从原始数据生成可供人类验证的研究论文。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FTechnion-Kishony-lab\u002Fdata-to-paper?style=social)\n- [AI Scientist](https:\u002F\u002Fgithub.com\u002FSakanaAI\u002FAI-Scientist)：AI Scientist：迈向完全自动化、开放式的科学研究。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FSakanaAI\u002FAI-Scientist?style=social)\n- [Storm](https:\u002F\u002Fgithub.com\u002Fstanford-oval\u002Fstorm)：一个由大语言模型（LLM）驱动的知识整理系统，可针对某一主题进行研究并生成带引用的完整报告。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstanford-oval\u002Fstorm?style=social)\n- [OpenLens AI](https:\u002F\u002Fgithub.com\u002Fjarrycyx\u002Fopenlens-ai)：面向健康信息学（Health Informatics）的全自动研究代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjarrycyx\u002Fopenlens-ai?style=social)\n- [GenoMAS](https:\u002F\u002Fgithub.com\u002FLiu-Hy\u002FGenoMAS)：一个用于科学发现的多代理框架，通过代码驱动的工作流自动化基因表达分析。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FLiu-Hy\u002FGenoMAS?style=social)\n- [DeepAnalyze](https:\u002F\u002Fgithub.com\u002Fruc-datalab\u002FDeepAnalyze)：首个面向自主数据科学的代理型大语言模型（agentic LLM），支持具体数据任务（数据准备、分析、建模、可视化与洞察）及数据导向的深度研究（生成分析师级别的研究报告）。[![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fruc-datalab\u002FDeepAnalyze?style=social&label=Code+Stars)](https:\u002F\u002Fgithub.com\u002Fruc-datalab\u002FDeepAnalyze)\n- [AIDE](https:\u002F\u002Fgithub.com\u002FWecoAI\u002Faideml)：AI 驱动探索 —— 一种机器学习工程代理，使用树搜索自动完成实验设计、代码生成及任意指标评估。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FWecoAI\u002Faideml?style=social)\n\n## 对话型 \u002F 通用代理\n\n- [CollosalAI Chat](https:\u002F\u002Fgithub.com\u002Fhpcaitech\u002FColossalAI\u002Ftree\u002Fmain\u002Fapplications\u002FChat)：基于 Colossal-AI 项目实现的大语言模型（LLM），结合了人类反馈强化学习（RLHF）。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhpcaitech\u002FColossalAI?style=social)\n- [RasaGPT](https:\u002F\u002Fgithub.com\u002Fpaulpierre\u002FRasaGPT)：RasaGPT 是首个构建于 Rasa 和 Langchain 之上的无头（headless）大语言模型聊天机器人平台。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fpaulpierre\u002FRasaGPT?style=social)\n- [SuperAgent](https:\u002F\u002Fgithub.com\u002Fhomanp\u002Fsuperagent)：将大语言模型代理部署到生产环境。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhomanp\u002Fsuperagent?style=social)\n- [BabyAGI UI](https:\u002F\u002Fgithub.com\u002Fmiurla\u002Fbabyagi-ui)：让在 Web 应用中运行和开发 BabyAGI 更加便捷，体验类似 ChatGPT。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmiurla\u002Fbabyagi-ui?style=social)\n- [ix](https:\u002F\u002Fgithub.com\u002Fkreneskyp\u002Fix)：自主 GPT-4 代理平台。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fkreneskyp\u002Fix?style=social)\n- [Multi-Modal LangChain agents in Production](https:\u002F\u002Fgithub.com\u002Fsteamship-packages\u002Flangchain-agent-production-starter)：部署 LangChain 代理并将其连接至 Telegram。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsteamship-packages\u002Flangchain-agent-production-starter?style=social)\n- [Autonomous HR Chatbot](https:\u002F\u002Fgithub.com\u002Fstepanogil\u002Fautonomous-hr-chatbot)：一个能利用现有工具自主回答人力资源相关问题的代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fstepanogil\u002Fautonomous-hr-chatbot?style=social)\n- [LLama Cpp Agent](https:\u002F\u002Fgithub.com\u002FMaximilian-Winter\u002Fllama-cpp-agent)：llama-cpp-agent 框架是一个便于与大语言模型交互的工具。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FMaximilian-Winter\u002Fllama-cpp-agent?style=social)\n- [Memgpt](https:\u002F\u002Fgithub.com\u002Fcpacker\u002Fmemgpt)：创建具备长期记忆（long-term memory）和自定义工具的 LLM 代理 📚🦙 ![Github Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fcpacker\u002Fmemgpt?style=social)\n- [joinly](https:\u002F\u002Fgithub.com\u002Fjoinly-ai\u002Fjoinly)：语音优先的 AI 助手，可在在线会议中主动参与并实时完成任务。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fjoinly-ai\u002Fjoinly?style=social)\n- [Gobii](https:\u002F\u002Fgithub.com\u002Fgobii-ai\u002Fgobii-platform)：Gobii 是一个开源平台，支持通过对话界面和 API 大规模部署与管理浏览器操作代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgobii-ai\u002Fgobii-platform?style=social)\n- [ClaudeClaw](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fclaudeclaw)：Claude Code 的持久化代理编排插件 —— 支持多通道路由（Slack、WhatsApp、Telegram）、操作系统级沙箱隔离、可组合扩展、结构化记忆、Webhook 触发器。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsbusso\u002Fclaudeclaw?style=social)\n\n## 游戏 \u002F 模拟\n\n- [Camel-AutoGPT](https:\u002F\u002Fgithub.com\u002FSamurAIGPT\u002FCamel-AutoGPT)：为大语言模型和 BabyAGI、AutoGPT 等自动代理提供角色扮演方法。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FSamurAIGPT\u002FCamel-AutoGPT?style=social)\n- [SkyAGI](https:\u002F\u002Fgithub.com\u002Flitanlitudan\u002Fskyagi)：在大语言模型代理中涌现的人类行为模拟能力。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Flitanlitudan\u002Fskyagi?style=social)\n- [Voyager](https:\u002F\u002Fgithub.com\u002FMineDojo\u002FVoyager)：基于大语言模型的开放式具身代理（Embodied Agent）。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FMineDojo\u002FVoyager?style=social)\n\n## 知识管理\n\n- [Private GPT](https:\u002F\u002Fgithub.com\u002Fimartinez\u002FprivateGPT)：利用 GPT 的强大能力，100% 私密地与您的文档交互，无数据泄露风险 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fimartinez\u002FprivateGPT?style=social)\n- [Local GPT](https:\u002F\u002Fgithub.com\u002FPromtEngineer\u002FlocalGPT)：受 Private GPT 启发，将 GPT4ALL 模型替换为 Vicuna-7B 模型，并使用 InstructorEmbeddings 替代 LlamaEmbeddings ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FPromtEngineer\u002FlocalGPT?style=social)\n- [LLocalSearch](https:\u002F\u002Fgithub.com\u002Fnilsherzig\u002FLLocalSearch)：LLocalSearch 是一个完全本地运行的搜索聚合器，使用 LLM（大语言模型）代理。用户可提出问题，系统将通过一系列 LLM 查找答案。用户可查看代理进度及最终答案，无需 OpenAI 或 Google API 密钥。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fnilsherzig\u002FLLocalSearch?style=social)\n- [Second Brain AI Agent](https:\u002F\u002Fgithub.com\u002Fflepied\u002Fsecond-brain-agent)：一款基于 Streamlit 的应用，允许您使用 OpenAI 和本地 ChromaDB 与“第二大脑”笔记进行对话。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fflepied\u002Fsecond-brain-agent?style=social)\n- [SAGE](https:\u002F\u002Fgithub.com\u002Fl33tdawg\u002Fsage)：面向 AI 代理的机构记忆系统——每条记忆在提交前均需通过 BFT 共识机制。包含 4 个应用验证器、13 个 MCP 工具，支持本地运行。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fl33tdawg\u002Fsage?style=social)\n\n## 自动化\n\n- [DemoGPT](https:\u002F\u002Fgithub.com\u002Fmelih-unsal\u002FDemoGPT)：仅需输入提示词即可快速创建演示原型。它在 Langchain 文档树上应用了 ToT（Tree of Thoughts）方法。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmelih-unsal\u002FDemoGPT?style=social)\n- [RestGPT](https:\u002F\u002Fgithub.com\u002FYifan-Song793\u002FRestGPT)：基于 LLM 的自主代理，可通过 RESTful API 控制现实世界中的应用程序。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FYifan-Song793\u002FRestGPT?style=social)\n- [XAgent](https:\u002F\u002Fgithub.com\u002FOpenBMB\u002FXAgent)：用于解决复杂任务的自主 LLM 代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FOpenBMB\u002FXAgent?style=social)\n- [LLM Agents](https:\u002F\u002Fgithub.com\u002Fmpaepper\u002Fllm_agents)：构建由 LLM 控制的代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmpaepper\u002Fllm_agents?style=social)\n- [uAgents](https:\u002F\u002Fgithub.com\u002Ffetchai\u002FuAgents)：一个快速轻量的框架，便于轻松创建去中心化代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Ffetchai\u002FuAgents?style=social)\n- [Maestro](https:\u002F\u002Fgithub.com\u002FDoriandarko\u002Fmaestro)：一个框架，让 Claude Opus 能够智能编排子代理。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FDoriandarko\u002Fmaestro?style=social)\n- [llama-agents](https:\u002F\u002Fgithub.com\u002Frun-llama\u002Fllama-agents)：llama-agents 是一个以异步优先的框架，用于构建、迭代和生产化多代理系统，支持多代理通信、分布式工具执行、人机协同等特性。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Frun-llama\u002Fllama-agents?style=social)\n- [Phidata](https:\u002F\u002Fgithub.com\u002Fphidatahq\u002Fphidata)：构建具备记忆、知识和工具的 AI 助手。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fphidatahq\u002Fphidata?style=social)\n- [AgentK](https:\u002F\u002Fgithub.com\u002Fmikekelly\u002FAgentK)：一个自我进化、模块化的自动代理 AGI（人工通用智能）。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmikekelly\u002FAgentK?style=social)\n- [ADAS](https:\u002F\u002Fgithub.com\u002FShengranHu\u002FADAS)：自动化设计代理系统（Automated Design of Agentic Systems）。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FShengranHu\u002FADAS?style=social)\n- [Giselle](https:\u002F\u002Fgithub.com\u002Fgiselles-ai\u002Fgiselle)：Giselle 是一个代理工作流构建器，让您轻松创建 AI 驱动的解决方案。 ![Github Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fgiselles-ai\u002Fgiselle?style=social)\n- [Untether](https:\u002F\u002Fgithub.com\u002Flittlebearapps\u002Funtether)：AI 编程代理的 Telegram 桥接工具——支持远程任务控制、进度流式传输、交互式权限、语音输入和成本追踪。兼容 Claude Code、Codex、OpenCode、Pi、Gemini CLI 和 Amp。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Flittlebearapps\u002Funtether?style=social)\n\n### 浏览器\n\n- [AgentGPT](https:\u002F\u002Fgithub.com\u002Freworkd\u002FAgentGPT)：基于 Langchain 与 OpenAI 的 AI 代理（Vercel \u002F Nextjs） ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Freworkd\u002FAgentGPT?style=social)\n- [OpenAgents](https:\u002F\u002Fgithub.com\u002Fxlang-ai\u002FOpenAgents)：面向真实环境的语言代理开放平台。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fxlang-ai\u002FOpenAgents?style=social)\n- [Steel Browser](https:\u002F\u002Fgithub.com\u002Fsteel-dev\u002Fsteel-browser)：开源浏览器基础设施，专为 AI 代理设计，支持会话持久化自动化、内容提取、截图\u002FPDF、AI 原生命令行界面，以及可复用的 steel-browser 技能。 ![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsteel-dev\u002Fsteel-browser?style=social)","# 🤖 Awesome Agents 快速上手指南\n\n## 环境准备\n\n- **操作系统**：支持 macOS、Linux、Windows（推荐 WSL2）\n- **Python 版本**：≥ 3.8（推荐 3.10+）\n- **Node.js**（部分框架需要）：≥ 16.x\n- **pip \u002F pipx**：确保已安装并更新至最新版\n- **可选加速源**：\n  ```bash\n  pip config set global.index-url https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n  ```\n\n> 注：部分工具如 `Semantic Kernel`、`Mastra` 需 .NET 或 TypeScript 环境，请按需安装。\n\n---\n\n## 安装步骤\n\n由于 `awesome-agents` 是一个精选列表而非单一工具，以下以最流行框架 **LangChain** 和 **LlamaIndex** 为例演示安装：\n\n### 安装 LangChain（Python）\n\n```bash\npip install langchain\n```\n\n如需完整功能（含常用 LLM 和工具集成）：\n\n```bash\npip install \"langchain[all]\"\n```\n\n### 安装 LlamaIndex\n\n```bash\npip install llama-index\n```\n\n### 安装 Autogen（微软多智能体框架）\n\n```bash\npip install pyautogen\n```\n\n> 其他框架请访问其 GitHub 页面查看具体安装命令，例如：\n> - MetaGPT: `pip install metagpt`\n> - CrewAI: `pip install crewai`\n> - OpenClaw: `git clone https:\u002F\u002Fgithub.com\u002Fopenclaw\u002Fopenclaw && cd openclaw && pip install -e .`\n\n---\n\n## 基本使用\n\n### 示例 1：用 LangChain 创建简单问答 Agent\n\n```python\nfrom langchain.agents import load_tools\nfrom langchain.agents import initialize_agent\nfrom langchain.llms import OpenAI\n\nllm = OpenAI(temperature=0)\ntools = load_tools([\"serpapi\", \"llm-math\"], llm=llm)\nagent = initialize_agent(tools, llm, agent=\"zero-shot-react-description\", verbose=True)\n\nagent.run(\"今天北京的天气如何？换算成华氏度是多少？\")\n```\n\n> 需提前设置 `OPENAI_API_KEY` 和 `SERPAPI_API_KEY` 环境变量。\n\n---\n\n### 示例 2：用 LlamaIndex 连接本地文档问答\n\n```python\nfrom llama_index import SimpleDirectoryReader, GPTVectorStoreIndex\n\ndocuments = SimpleDirectoryReader('.\u002Fdata').load_data()\nindex = GPTVectorStoreIndex.from_documents(documents)\n\nquery_engine = index.as_query_engine()\nresponse = query_engine.query(\"总结这份文档的核心内容\")\nprint(response)\n```\n\n> 需在 `.\u002Fdata` 目录放置 PDF\u002FTXT\u002FMD 等格式文档。\n\n---\n\n### 示例 3：用 Autogen 创建双智能体对话\n\n```python\nimport autogen\n\nconfig_list = [\n    {\n        \"model\": \"gpt-4\",\n        \"api_key\": \"your-openai-api-key\"\n    }\n]\n\nassistant = autogen.AssistantAgent(\"assistant\", llm_config={\"config_list\": config_list})\nuser_proxy = autogen.UserProxyAgent(\"user_proxy\", code_execution_config={\"work_dir\": \"coding\"})\n\nuser_proxy.initiate_chat(assistant, message=\"画一个正弦函数图像，x范围从0到2π\")\n```\n\n> 自动执行代码并生成图表，适合任务协作场景。\n\n---\n\n✅ 至此你已完成基础环境搭建与首个 AI Agent 运行！  \n👉 推荐下一步：访问 [Awesome Agents GitHub](https:\u002F\u002Fgithub.com\u002Fsbusso\u002Fawesome-agents) 按类别探索更多框架。","一家跨境电商运营团队每天需要处理来自 Slack、WhatsApp 和 Telegram 的客户咨询、订单状态查询与退换货请求，人工客服响应慢、跨平台消息容易遗漏，严重影响客户满意度。\n\n### 没有 awesome-agents 时\n- 客服人员需手动切换三个通讯平台，信息分散导致漏回率高达30%，客户投诉频发。\n- 复杂问题（如跨境物流追踪）需转交技术或物流部门，沟通链条长，平均解决时间超过24小时。\n- 缺乏统一知识库支持，新人培训周期长达两周，且回答口径不一致影响品牌形象。\n- 自动化脚本只能处理固定关键词，无法理解自然语言意图，80%的对话仍需人工介入。\n- 系统无沙箱隔离机制，第三方插件存在安全风险，曾因误操作导致客户数据短暂泄露。\n\n### 使用 awesome-agents 后\n- 通过集成 ClaudeClaw，实现 Slack\u002FWhatsApp\u002FTelegram 三端消息自动路由与统一响应，漏回率降至1%以下。\n- 利用 MetaGPT 构建多角色协作代理，自动拆解复杂任务并分配给“物流Agent”或“售后Agent”，平均响应时间缩短至2小时内。\n- 借助 LlamaIndex 接入公司知识库，新客服上岗3天即可胜任，标准问答准确率达95%以上。\n- 基于 Transformers Agents 的自然语言理解能力，可识别“我想退货但还没收到货”等模糊表达，自动化处理率提升至70%。\n- ClaudeClaw 的 OS 级沙箱隔离确保所有插件在安全环境中运行，半年内零安全事故。\n\nawesome-agents 不是替代人力，而是让团队把精力从重复劳动中解放出来，专注真正需要人类判断的高价值服务。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fkyrolabs_awesome-agents_78b71a0c.png","kyrolabs","Kyrolabs","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fkyrolabs_c23a401e.png","Build and deploy AI-powered applications with Kyroagent no-code platforms.",null,"kyrolabsAI","https:\u002F\u002Fkyrolabs.com","https:\u002F\u002Fgithub.com\u002Fkyrolabs",2113,439,"2026-04-05T14:07:24",5,"","未说明",{"notes":90,"python":88,"dependencies":91},"该项目为精选列表，非单一工具，不包含具体运行环境需求。各子项目需参考其独立文档配置环境。",[],[13,15],4,"2026-03-27T02:49:30.150509","2026-04-06T08:45:21.633769",[97,102,107,112,117,122],{"id":98,"question_zh":99,"answer_zh":100,"source_url":101},441,"如何停止接收 GitHub 项目的邮件通知？","请访问 GitHub 通知设置页面，点击“Unsubscribe”链接取消订阅。具体路径：进入 Issue 页面 → 点击右上角“Unsubscribe”按钮，或通过邮件底部的退订链接操作。","https:\u002F\u002Fgithub.com\u002Fkyrolabs\u002Fawesome-agents\u002Fissues\u002F47",{"id":103,"question_zh":104,"answer_zh":105,"source_url":106},442,"如何向 awesome-agents 列表提交新框架？","建议先在 Issues 中提出推荐请求，并附上项目链接、简介和理由。若维护者认可，需自行提交 Pull Request（PR）完成添加。例如 Swarms 框架作者在 Issue #31 被邀请后提交了 PR #32。","https:\u002F\u002Fgithub.com\u002Fkyrolabs\u002Fawesome-agents\u002Fissues\u002F31",{"id":108,"question_zh":109,"answer_zh":110,"source_url":111},443,"论文或研究内容能否被收录进此列表？","不能。该列表仅聚焦于代码和工具类项目，纯论文或研究内容（如 BrowseSafe 论文）不会被收录，即使其附带数据集或模型。","https:\u002F\u002Fgithub.com\u002Fkyrolabs\u002Fawesome-agents\u002Fissues\u002F64",{"id":113,"question_zh":114,"answer_zh":115,"source_url":116},444,"AgentWatch 提供哪些多智能体可观测性功能？","提供心跳注册、级联故障检测、跨智能体故障溯源、舰队仪表盘（CLI: npx agentwatch dashboard）、告警去重和故障回放功能。支持 TypeScript，npm 包名为 @nicofains1\u002Fagentwatch。","https:\u002F\u002Fgithub.com\u002Fkyrolabs\u002Fawesome-agents\u002Fissues\u002F207",{"id":118,"question_zh":119,"answer_zh":120,"source_url":121},445,"如何为 AI 智能体快速注册 AgentHive 社交账号？","使用 Python 发送 POST 请求到 https:\u002F\u002Fagenthive.to\u002Fapi\u002Fagents\u002Fregister，传入 username、display_name 和 bio 字段即可获取 API 密钥。示例代码：requests.post(..., json={'username':'my-agent', 'display_name':'My Agent', 'bio':'Built with LangChain'})。","https:\u002F\u002Fgithub.com\u002Fkyrolabs\u002Fawesome-agents\u002Fissues\u002F232",{"id":123,"question_zh":124,"answer_zh":125,"source_url":126},446,"SuperLocalMemory 是什么？适用于哪些场景？","它是面向 AI 智能体的本地优先记忆层，采用 Fisher-Rao 信息几何与贝叶斯信任机制，支持零云模式。兼容 LangChain 和 CrewAI，MIT 许可，适合需要隐私保护和高性能本地记忆管理的智能体系统。","https:\u002F\u002Fgithub.com\u002Fkyrolabs\u002Fawesome-agents\u002Fissues\u002F243",[]]