[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-ThousandBirdsInc--chidori":3,"tool-ThousandBirdsInc--chidori":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",138956,2,"2026-04-05T11:33:21",[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":80,"owner_twitter":79,"owner_website":81,"owner_url":82,"languages":83,"stars":115,"forks":116,"last_commit_at":117,"license":118,"difficulty_score":119,"env_os":120,"env_gpu":121,"env_ram":121,"env_deps":122,"category_tags":126,"github_topics":127,"view_count":10,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":135,"updated_at":136,"faqs":137,"releases":168},1097,"ThousandBirdsInc\u002Fchidori","chidori","A reactive runtime for building durable AI agents","Chidori 是一个专注于构建持久化AI代理的开源工具，通过反应式运行时实现对复杂AI流程的高效管理。它解决了传统AI代理在状态追踪困难、执行流程复杂、难以与人类交互等问题，提供可视化调试、时间回溯和分支控制等功能，帮助开发者更直观地理解和优化AI代理的行为。工具采用Rust语言构建，支持Python和JavaScript代码执行，内置缓存机制和状态图可视化，可快速恢复部分执行流程。适合需要开发和调试长期运行AI代理的开发者与研究人员，尤其适用于涉及多步骤推理、状态空间探索的场景。其独特的分支控制和时间旅行调试功能，让复杂流程的测试与优化更加灵活高效。","\n\n[\u002F\u002F]: # ([![Demo Video](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FThousandBirdsInc_chidori_readme_39062a46e671.png)](https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fassets\u002F515757\u002F6b088f7d-d8f7-4c7e-9006-4360ae40d1de))\n[![Demo Video](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FThousandBirdsInc_chidori_readme_39062a46e671.png)]()\n\n\u003Cdiv align=\"center\">\n\n#   Chidori (v2)  \n\n**A reactive runtime for building durable AI agents**\n\n\u003Cp>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fcommits\">\u003Cimg alt=\"Current Build Status\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FThousandBirdsInc\u002Fchidori\u002Fpush.yml\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fcommits\">\u003Cimg alt=\"GitHub Last Commit\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002FThousandBirdsInc\u002Fchidori\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fcrates.io\u002Fcrates\u002Fchidori-debugger\">\u003Cimg alt=\"Cargo.io download\" src=\"https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fv\u002Fchidori-debugger\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fblob\u002Fmain\u002FLICENSE\">\u003Cimg alt=\"GitHub License\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-green.svg\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\u003Cbr \u002F>\n\u003C\u002Fdiv>\n\nStar us on GitHub! Join us on [Discord](https:\u002F\u002Fdiscord.gg\u002FCJwKsPSgew).\n\n## Contents\n- [📖 Chidori V2](#-chidori-v2)\n- [⚡️ Getting Started](#️-getting-started)\n  - [Installation](#installation)\n  - [Environment Variables](#environment-variables)\n  - [Example](#example)\n- [🤔 About](#-about)\n  - [Reactive Runtime](#reactive-runtime)\n  - [Monitoring and Observability](#monitoring-and-observability)\n  - [Branching and Time-Travel](#branching-and-time-travel)\n  - [Code Interpreter Environments](#code-interpreter-environments)\n- [🛣️ Roadmap](#️-roadmap)\n  - [Short term](#short-term)\n  - [Medium term](#med-term)\n- [Contributing](#contributing)\n- [Inspiration](#inspiration)\n- [License](#license)\n- [Help us out!](#help-us-out)\n\n\n## 📖 Chidori V2\nChidori is an open-source orchestrator, runtime, and IDE for building software in symbiosis with modern AI tools.\nIt is especially catered towards building AI agents by providing solutions to the following problems:\n\n- How do we understand what an agent is doing and how it got into a given state?\n- How can we pause execution and then resume after interaction with a human?\n- How do we handle the accidental complexity of state-space exploration, evaluating and reverting execution throughout our software?\n\nWhen using Chidori, you author code with python or javascript, we provide a layer for interfacing\nwith the complexities of AI models in long-running workflows. We have avoided the need for declaring a new language \nor SDK in order to provide these capabilities so that you can leverage software patterns that you are already familiar with.\n\nFeatures:\n\n- Runtime written in Rust, supporting Python and JavaScript code execution\n- The ability to cache behaviors and resume from partially executed agents\n- Time travel debugging, execution of the program can be reverted to prior states\n- Visual debugging environment, visualize and manipulate the graph of states your code has executed through.\n- Create and navigate tree-searching code execution workflows\n\n## ⚡️ Getting Started\n\n### Installation\nChidori is available on [crates.io](https:\u002F\u002Fcrates.io\u002Fcrates\u002Fchidori) and can be installed using cargo. Our expected entrypoint for\nprototype development is `chidori-debugger` which wraps our runtime in a useful visual interface.\n\n```bash\n# Install the rust toolchain and the nightly channel\ncurl --proto '=https' --tlsv1.2 -sSf https:\u002F\u002Fsh.rustup.rs | sh\nrustup toolchain install nightly\n\n# Required for building dependencies\nxcode-select --install\n\n# These dependencies are necessary for a successful build\nbrew install cmake  \n\n# We are investigating if this is necessary or can be removed\nbrew install python@3.12\n\n# Chidori uses uv for handling python dependencies \nbrew install uv\n\n# We depend on features only supported by nightly at the moment\ncargo +nightly install chidori-debugger --locked\n```\n\nIf you prefer to use a different python interpreter you can set PYO3_PYTHON=python3.12 (or whichever version > 3.7) during\nyour installation to change which is linked against.\n\n\n### Setting Up The Runtime Environment\nChidori's interactions with LLMs default to http:\u002F\u002Flocalhost:4000 to hook into LiteLLM's proxy.\nIf you'd like to leverage gpt-3.5-turbo the included config file will support that.\nYou will need to install `pip install litellm[proxy]` in order to run the below:\n```bash\nexport OPENAI_API_KEY=...\nuv pip install \"litellm[proxy]\"\nuv run litellm --config .\u002Flitellm_config.yaml\n```\n\n## Examples\n\nThe following example shows how to build a simple agent that fetches the top stories from Hacker News and call the OpenAI API to filter to AI related launches and then format that data into markdown.\n\n------\n\n### Beginning here is an example executable Chidori agent:\n\nChidori agents can be a single file, or a collection of files structured as a typical Typescript or Python project. \nThe following example is a single file agent. Consider this similar to something like a jupyter\u002FiPython notebook \nrepresented as a markdown file.\n\n\u003Cpre>\n\n```javascript (load_hacker_news)\nconst axios = require('https:\u002F\u002Fdeno.land\u002Fx\u002Faxiod\u002Fmod.ts');\n\nconst HN_URL_TOP_STORIES = \"https:\u002F\u002Fhacker-news.firebaseio.com\u002Fv0\u002Ftopstories.json\";\n\nfunction fetchStory(id) {\n    return axios.get(`https:\u002F\u002Fhacker-news.firebaseio.com\u002Fv0\u002Fitem\u002F${id}.json?print=pretty`)\n        .then(response => response.data);\n}\n\nasync function fetchHN() {\n    const stories = await axios.get(HN_URL_TOP_STORIES);\n    const storyIds = stories.data;\n    \u002F\u002F only the first 30 \n    const tasks = storyIds.slice(0, 30).map(id => fetchStory(id));\n    return Promise.all(tasks)\n      .then(stories => {\n        return stories.map(story => {\n          const { title, url, score } = story;\n          return {title, url, score};\n        });\n      });\n}\n```\n\nPrompt \"interpret_the_group\"\n```prompt (interpret_the_group)\n  Based on the following list of HackerNews threads,\n  filter this list to only launches of \n  new AI projects: {{fetched_articles}}\n```\n\nPrompt \"format_and_rank\"\n```prompt (format_and_rank)\nFormat this list of new AI projects in markdown, ranking the most \ninteresting projects from most interesting to least. \n{{interpret_the_group}}\n```\n\nUsing a python cell as our entrypoint, demonstrating inter-language execution:\n```python\narticles = await fetchHN()\nformat_and_rank(articles=articles)\n```\n\u003C\u002Fpre>\n------\n\n\n## About\n\n### Reactive Runtime\nAt its core, Chidori brings a reactive runtime that orchestrates interactions between different agents and their components. \nChidori accepts arbitrary Python or JavaScript code, taking over brokering and execution of it to allow for interruptions and reactivity.\nThis allows you to get the benefits of these runtime behaviors while leveraging the patterns you're already familiar with.\n\n### Monitoring and Observability\nChidori ensures comprehensive monitoring and observability of your agents. We record all the inputs and outputs emitted by functions throughout the execution of your agent, enabling us to explain precisely what led to what, enhancing your debugging experience and understanding of the system’s production behavior.\n\n### Branching and Time-Travel\nWith Chidori, you can take snapshots of your system and explore different possible outcomes from that point (branching), or rewind the system to a previous state (time-travel). This functionality improves error handling, debugging, and system robustness by offering alternative pathways and do-overs.\n\n### Code Interpreter Environments\nChidori comes with first-class support for code interpretation for both Python and JavaScript. You can execute code directly within your system, providing quick startup, ease of use, and secure execution. We're continually working on additional safeguards against running untrusted code, with containerized environment support coming soon.\n\n### Code Generation During Evaluation\nWith our execution graph, preservation of state, and tools for debugging - Chidori is an exceptional environment for generating code during the evaluation of your agent.\nYou can use this to leverage LLMs to achieve more generalized behavior and to evolve your agents over time.\n\n\n\n\n## 🛣️ Roadmap\n\n### Short term\n* [x] Reactive subscriptions between nodes\n* [x] Branching and time travel debugging, reverting execution of a graph\n* [x] Node.js, Python, and Rust support for building and executing graphs\n* [x] Simple local vector db for development\n* [ ] Adding support for containerized nodes\n\n### Medium term\n* [x] Analysis tools for comparing executions\n* [x] Adding support for more vector databases\n* [x] Adding support for other LLM sources\n* [x] Adding support for more code interpreter environments\n* [ ] Agent re-evaluation with feedback\n* [ ] Definitive patterns for human in the loop agents\n\n\n## Contributing\nThis is an early open source release and we're looking for collaborators from the community. \nA good place to start would be to join our [discord](https:\u002F\u002Fdiscord.gg\u002FCJwKsPSgew)!\n\n## Inspiration\nOur framework is inspired by the work of many others, including:\n* [Temporal.io](https:\u002F\u002Ftemporal.io) - providing reliability and durability to workflows\n* [Eve](http:\u002F\u002Fwitheve.com) - developing patterns for building reactive systems and reducing accidental complexity\n* [Timely Dataflow](https:\u002F\u002Ftimelydataflow.github.io\u002Ftimely-dataflow) - efficiently streaming changes\n* [Langchain](https:\u002F\u002Fwww.langchain.com) - developing tools and patterns for building with LLMs\n\n## License\nChidori is under the MIT license. See the [LICENSE](LICENSE) for more information.\n\n## Help us out!\nPlease star the GitHub repo and join our [discord](https:\u002F\u002Fdiscord.gg\u002FCJwKsPSgew)!","[\u002F\u002F]: # ([![Demo Video](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FThousandBirdsInc_chidori_readme_39062a46e671.png)](https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fassets\u002F515757\u002F6b088f7d-d8f7-4c7e-9006-4360ae40d1de))\n[![Demo Video](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FThousandBirdsInc_chidori_readme_39062a46e671.png)]()\n\n\u003Cdiv align=\"center\">\n\n#   Chidori (v2)  \n\n**用于构建持久化 AI 代理的反应式运行时（reactive runtime）**\n\n\u003Cp>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fcommits\">\u003Cimg alt=\"当前构建状态\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FThousandBirdsInc\u002Fchidori\u002Fpush.yml\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fcommits\">\u003Cimg alt=\"GitHub 最后提交\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002FThousandBirdsInc\u002Fchidori\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fcrates.io\u002Fcrates\u002Fchidori-debugger\">\u003Cimg alt=\"Cargo.io 下载量\" src=\"https:\u002F\u002Fimg.shields.io\u002Fcrates\u002Fv\u002Fchidori-debugger\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fblob\u002Fmain\u002FLICENSE\">\u003Cimg alt=\"GitHub 协议\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-green.svg\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\u003Cbr \u002F>\n\u003C\u002Fdiv>\n\n在 GitHub 上为我们点赞！加入我们的 [Discord](https:\u002F\u002Fdiscord.gg\u002FCJwKsPSgew) 社区。\n\n## 目录\n- [📖 Chidori V2](#-chidori-v2)\n- [⚡️ 快速开始](#️-快速开始)\n  - [安装](#安装)\n  - [环境变量](#环境变量)\n  - [示例](#示例)\n- [🤔 关于](#-关于)\n  - [反应式运行时](#反应式运行时)\n  - [监控与可观测性](#监控与可观测性)\n  - [分支与时间旅行](#分支与时间旅行)\n  - [代码解释器环境](#代码解释器环境)\n- [🛣️ 路线图](#️-路线图)\n  - [短期目标](#短期目标)\n  - [中期目标](#中期目标)\n- [贡献指南](#贡献指南)\n- [灵感来源](#灵感来源)\n- [开源协议](#开源协议)\n- [帮助我们！](#帮助我们)\n\n## 📖 Chidori V2\nChidori 是一个开源的编排器、运行时和集成开发环境（IDE），旨在与现代 AI 工具协同构建软件。它特别针对以下问题提供了解决方案：\n\n- 如何理解代理的行为及其状态演变过程？\n- 如何暂停执行并在与人类交互后继续执行？\n- 如何处理状态空间探索的复杂性，在软件中评估和回滚执行？\n\n使用 Chidori 时，您可以使用 Python 或 JavaScript 编写代码，我们提供了一层接口来处理 AI 模型在长期运行工作流中的复杂性。我们避免了声明新语言或 SDK 的需求，使您能够利用已熟悉的软件模式。\n\n核心特性：\n\n- Rust 编写的运行时，支持 Python 和 JavaScript 代码执行\n- 支持缓存行为并从部分执行的代理中恢复\n- 时间旅行调试功能，可将程序执行回滚到先前状态\n- 可视化调试环境，可视化并操作代码执行状态图\n- 创建和导航树状搜索代码执行工作流\n\n## ⚡️ 快速开始\n\n### 安装\nChidori 发布在 [crates.io](https:\u002F\u002Fcrates.io\u002Fcrates\u002Fchidori)，可通过 cargo 安装。我们推荐使用 `chidori-debugger` 作为原型开发入口，它将运行时封装在可视化界面中。\n\n```bash\n# 安装 Rust 工具链和 nightly 通道\ncurl --proto '=https' --tlsv1.2 -sSf https:\u002F\u002Fsh.rustup.rs | sh\nrustup toolchain install nightly\n\n# 构建依赖所需工具\nxcode-select --install\n\n# 必要构建依赖\nbrew install cmake  \n\n# 正在验证是否可移除该依赖\nbrew install python@3.12\n\n# 使用 uv 管理 Python 依赖 \nbrew install uv\n\n# 当前依赖 nightly 特性\ncargo +nightly install chidori-debugger --locked\n```\n\n如需使用其他 Python 解释器，可在安装时设置 PYO3_PYTHON=python3.12（或 3.7+ 版本）指定链接版本。\n\n### 运行时环境配置\nChidori 默认通过 http:\u002F\u002Flocalhost:4000 与 LiteLLM 代理交互。\n如需使用 gpt-3.5-turbo，配置文件已支持该功能。需先安装 `pip install litellm[proxy]`：\n```bash\nexport OPENAI_API_KEY=...\nuv pip install \"litellm[proxy]\"\nuv run litellm --config .\u002Flitellm_config.yaml\n```\n\n## 示例\n\n以下示例展示如何构建一个简单代理：获取 Hacker News 热门故事，调用 OpenAI API 过滤 AI 相关项目，并将结果格式化为 Markdown。\n\n------\n### 以下是一个可执行的 Chidori 代理示例：\n\nChidori 代理可以是单个文件，也可以是典型的 Typescript\u002FPython 项目结构。以下为单文件示例，类似于 jupyter\u002FiPython 笔记本的 Markdown 表示。\n\n\u003Cpre>\n\n```javascript (load_hacker_news)\nconst axios = require('https:\u002F\u002Fdeno.land\u002Fx\u002Faxiod\u002Fmod.ts');\n\nconst HN_URL_TOP_STORIES = \"https:\u002F\u002Fhacker-news.firebaseio.com\u002Fv0\u002Ftopstories.json\";\n\nfunction fetchStory(id) {\n    return axios.get(`https:\u002F\u002Fhacker-news.firebaseio.com\u002Fv0\u002Fitem\u002F${id}.json?print=pretty`)\n        .then(response => response.data);\n}\n\nasync function fetchHN() {\n    const stories = await axios.get(HN_URL_TOP_STORIES);\n    const storyIds = stories.data;\n    \u002F\u002F 仅取前30个\n    const tasks = storyIds.slice(0, 30).map(id => fetchStory(id));\n    return Promise.all(tasks)\n      .then(stories => {\n        return stories.map(story => {\n          const { title, url, score } = story;\n          return {title, url, score};\n        });\n      });\n}\n```\n\n提示词 \"interpret_the_group\"\n```prompt (interpret_the_group)\n  根据以下 HackerNews 主题列表，\n  仅保留新 AI 项目的发布信息：{{fetched_articles}}\n```\n\n提示词 \"format_and_rank\"\n```prompt (format_and_rank)\n将以下新 AI 项目列表用 Markdown 格式呈现，按最有趣到最无趣排序：\n{{interpret_the_group}}\n```\n\n使用 Python 单元作为入口点，演示跨语言执行：\n```python\narticles = await fetchHN()\nformat_and_rank(articles=articles)\n```\n\u003C\u002Fpre>\n------\n\n## 关于\n\n### 反应式运行时\nChidori 的核心是一个反应式运行时（reactive runtime），协调不同代理及其组件之间的交互。它接受任意 Python 或 JavaScript 代码，接管其执行和调度，支持中断和响应式行为。这使您能在保持熟悉开发模式的同时，获得这些运行时特性带来的优势。\n\n### 监控与可观察性（Monitoring and Observability）\nChidori 确保对您的智能体（agent）实现全面监控与可观测性。我们会记录执行过程中所有函数产生的输入输出，这种机制能够精确解释系统行为的因果关系，提升调试体验并加深对生产环境行为的理解。\n\n### 分支与时间旅行（Branching and Time-Travel）\n通过 Chidori，您可以对系统状态进行快照，并探索该节点后的不同可能结果（分支），或回溯系统到先前状态（时间旅行）。该功能通过提供替代路径和重试机制，显著增强错误处理、调试能力和系统健壮性。\n\n### 代码解释器环境（Code Interpreter Environments）\nChidori 原生支持 Python 和 JavaScript 的代码解释功能。您可以在系统内直接执行代码，实现快速启动、便捷使用和安全执行。我们正在持续开发针对不可信代码的防护措施，即将支持容器化环境执行。\n\n### 评估过程中的代码生成（Code Generation During Evaluation）\n依托执行图、状态持久化和调试工具，Chidori 为在智能体评估过程中生成代码提供了理想环境。您可以利用此功能结合大语言模型（LLM）实现更通用的行为模式，并随时间推移动态优化智能体。\n\n## 🛣️ 路线图\n\n### 短期\n* [x] 节点间响应式订阅（Reactive subscriptions between nodes）\n* [x] 分支与时间旅行调试，支持图执行回滚\n* [x] 支持 Node.js、Python 和 Rust 构建执行图\n* [x] 开发用本地向量数据库\n* [ ] 添加容器化节点支持\n\n### 中期\n* [x] 执行对比分析工具\n* [x] 扩展更多向量数据库支持\n* [x] 集成更多 LLM 数据源\n* [x] 支持更多代码解释器环境\n* [ ] 基于反馈的智能体重评估\n* [ ] 定义人机协同智能体模式\n\n## 贡献\n这是早期开源版本，我们期待社区开发者加入。  \n建议从加入我们的 [discord](https:\u002F\u002Fdiscord.gg\u002FCJwKsPSgew) 开始！\n\n## 灵感来源\n本框架受以下项目启发：\n* [Temporal.io](https:\u002F\u002Ftemporal.io) - 提供工作流可靠性与持久性\n* [Eve](http:\u002F\u002Fwitheve.com) - 开发响应式系统模式并降低意外复杂度\n* [Timely Dataflow](https:\u002F\u002Ftimelydataflow.github.io\u002Ftimely-dataflow) - 高效流式数据变更处理\n* [Langchain](https:\u002F\u002Fwww.langchain.com) - 构建 LLM 工具与模式开发\n\n## 许可证\nChidori 遵循 MIT 许可证。详见 [LICENSE](LICENSE) 文件。\n\n## 支持我们！\n请为 GitHub 仓库加星标，并加入我们的 [discord](https:\u002F\u002Fdiscord.gg\u002FCJwKsPSgew)！","# Chidori 快速上手指南\n\n## 环境准备\n### 系统要求\n- macOS（需要xcode-select）或Linux\n- 至少8GB内存\n- Python 3.7+ 环境\n\n### 前置依赖\n```bash\n# macOS安装xcode命令行工具\nxcode-select --install\n\n# 安装Homebrew（若未安装）\n\u002Fbin\u002Fbash -c \"$(curl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002FHomebrew\u002Finstall\u002FHEAD\u002Finstall.sh)\"\n\n# 国内用户可选：替换brew源（中科大镜像）\ncd $(brew --repo)\ngit remote set-url origin https:\u002F\u002Fmirrors.ustc.edu.cn\u002Fbrew.git\ncd $(brew --repo)\u002FLibrary\u002FTaps\u002Fhomebrew\u002Fhomebrew-core\ngit remote set-url origin https:\u002F\u002Fmirrors.ustc.edu.cn\u002Fhomebrew-core.git\n```\n\n## 安装步骤\n```bash\n# 安装Rust工具链（国内加速版）\ncurl --proto '=https' --tlsv1.2 -sSf https:\u002F\u002Fsh.rustup.rs | sh -s -- --default-toolchain nightly\n\n# 安装构建依赖\nbrew install cmake python@3.12 uv\n\n# 安装Chidori调试器（使用国内镜像加速）\ncargo +nightly install chidori-debugger --locked --git https:\u002F\u002Fmirrors.ustc.edu.cn\u002Fgitee.com\u002FThousandBirdsInc\u002Fchidori.git\n```\n\n## 基本使用\n### 启动LLM代理服务\n```bash\n# 安装litellm代理（国内镜像加速）\nuv pip install \"litellm[proxy]\" -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n\n# 创建litellm配置文件（示例）\necho \"model: gpt-3.5-turbo\" > litellm_config.yaml\n\n# 启动代理服务\nexport OPENAI_API_KEY=your_api_key_here\nuv run litellm --config .\u002Flitellm_config.yaml\n```\n\n### 运行示例Agent\n```python\n# 创建示例文件 hn_ai_filter.py\narticles = await fetchHN()\nformat_and_rank(articles=articles)\n```\n\n```bash\n# 通过调试器运行示例\nchidori-debugger run hn_ai_filter.py\n```\n\n### 示例代码说明\n该示例包含：\n1. JavaScript编写的Hacker News抓取函数\n2. 两个Prompt模板（AI项目过滤\u002FMarkdown格式化）\n3. Python入口点调用流程\n\n完整代码请参考README中的示例部分，可直接复制到`.py`文件中运行。\n\n> 注意：首次运行需要配置OPENAI_API_KEY环境变量，并确保代理服务正常运行。","数据分析师团队正在开发一个自动化金融报告生成系统，需要AI代理持续从多个数据源抓取信息并动态调整分析逻辑。由于数据源频繁变更且分析逻辑复杂，团队面临调试和维护难题。\n\n### 没有 chidori 时\n- 状态追踪困难：代理在多步骤分析中出现异常时，需要手动在分散的日志文件中定位问题节点，平均耗时2小时以上\n- 调试过程不可逆：修复错误后无法回溯到出错前的状态继续验证，必须重新运行整个分析流程\n- 人工干预低效：当需要人工审核异常数据时，必须终止整个流程并重新启动，导致30%的重复计算\n- 协作障碍：多个开发者同时调试不同分支逻辑时，状态冲突导致频繁覆盖彼此的临时修改\n\n### 使用 chidori 后\n- 可视化状态追踪：通过时间轴面板实时查看每个数据处理节点的状态快照，定位问题时间缩短至15分钟内\n- 精确回溯验证：利用时间旅行调试功能直接跳转到异常发生前的状态，修改修复后仅需重放受影响的分支流程\n- 动态暂停恢复：在数据审核环节自动暂停流程并弹出待办事项，审核完成后无缝续传计算状态，减少85%重复工作\n- 分支协同开发：每个开发者在独立的状态分支上调试，通过可视化合并工具解决冲突，代码集成效率提升3倍\n\n核心价值：chidori 通过状态可追溯性和流程可操控性，使复杂AI代理的开发调试效率提升一个数量级，让团队能聚焦核心业务逻辑创新。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FThousandBirdsInc_chidori_39062a46.png","ThousandBirdsInc","Thousand Birds","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FThousandBirdsInc_0b9ef961.png","",null,"team@thousandbirds.ai","https:\u002F\u002Fdocs.thousandbirds.ai\u002F","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc",[84,88,92,96,100,104,108,111],{"name":85,"color":86,"percentage":87},"Rust","#dea584",95.6,{"name":89,"color":90,"percentage":91},"Python","#3572A5",3.2,{"name":93,"color":94,"percentage":95},"WGSL","#1a5e9a",0.5,{"name":97,"color":98,"percentage":99},"TypeScript","#3178c6",0.3,{"name":101,"color":102,"percentage":103},"Shell","#89e051",0.2,{"name":105,"color":106,"percentage":107},"Nix","#7e7eff",0.1,{"name":109,"color":110,"percentage":107},"Dockerfile","#384d54",{"name":112,"color":113,"percentage":114},"JavaScript","#f1e05a",0,1340,56,"2026-04-01T13:21:53","MIT",4,"macOS","未说明",{"notes":123,"python":124,"dependencies":125},"需要安装Rust nightly工具链，xcode-select，cmake，python@3.12及uv工具。需设置PYO3_PYTHON环境变量指定Python版本，并配置OPENAI_API_KEY。使用brew安装依赖库如cmake和python。","3.7+",[121],[26,13,15,14],[128,129,130,131,132,133,134],"agents","ai","debugging","framework","llmops","llms","orchestration","2026-03-27T02:49:30.150509","2026-04-06T07:15:06.868945",[138,143,148,153,158,163],{"id":139,"question_zh":140,"answer_zh":141,"source_url":142},4934,"在M1 Mac上安装Chidori时出现404错误，如何解决？","这是已知问题，目前尚未提供ARM架构的二进制文件。维护者计划在近期发布支持ARM的版本，建议关注项目更新。","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fissues\u002F4",{"id":144,"question_zh":145,"answer_zh":146,"source_url":147},4935,"运行Rust示例时出现未解析的导入错误（如create_change_value），如何处理？","项目代码已发生重大变更，当前示例可能与最新版本不兼容。建议查看最新文档或加入Discord讨论获取适配方案（https:\u002F\u002Fdiscord.com\u002Fchannels\u002F1132365768986742927\u002F1132907110557024256\u002F1163529198611222579）。","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fissues\u002F24",{"id":149,"question_zh":150,"answer_zh":151,"source_url":152},4936,"运行JS示例时提示invalid ELF header错误，如何解决？","该问题与本地构建的native模块兼容性有关。维护者正在优化发布流程，后续将通过容器化示例解决此问题。临时可尝试使用Docker运行示例。","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fissues\u002F7",{"id":154,"question_zh":155,"answer_zh":156,"source_url":157},4937,"运行Python示例时提示无法导入GraphBuilder，原因是什么？","项目结构已发生重大调整，当前版本可能移除了相关模块。建议检查安装版本与示例代码的兼容性，或参考最新文档更新代码。","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fissues\u002F5",{"id":159,"question_zh":160,"answer_zh":161,"source_url":162},4938,"Rust HN示例在Docker中运行时卡住，如何排查？","需要确保正确设置环境变量OPEN_AI_KEY。未配置该密钥会导致程序在执行时陷入无限循环。","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fissues\u002F26",{"id":164,"question_zh":165,"answer_zh":166,"source_url":167},4939,"能否通过Docker-compose简化本地开发环境搭建？","维护者已认可该需求，并计划在项目重构后提供容器化方案。当前建议参考项目文档中的Docker配置手动部署。","https:\u002F\u002Fgithub.com\u002FThousandBirdsInc\u002Fchidori\u002Fissues\u002F22",[169,173,177,181,186,191,195,199,203,207,212,216,220,224,228],{"id":170,"version":171,"summary_zh":79,"released_at":172},104434,"v0.1.28","2023-09-26T05:24:35",{"id":174,"version":175,"summary_zh":79,"released_at":176},104435,"v0.1.27","2023-09-26T04:49:13",{"id":178,"version":179,"summary_zh":79,"released_at":180},104436,"v0.1.26","2023-07-30T23:33:26",{"id":182,"version":183,"summary_zh":184,"released_at":185},104437,"v0.1.25","Test release\n","2023-07-30T04:07:35",{"id":187,"version":188,"summary_zh":189,"released_at":190},104438,"prompt-graph-exec-v0.1.24","### Fixed\n\n- Fixing releases\n\n","2023-07-28T07:36:16",{"id":192,"version":193,"summary_zh":189,"released_at":194},104439,"prompt-graph-exec-v0.1.23","2023-07-28T07:20:52",{"id":196,"version":197,"summary_zh":189,"released_at":198},104440,"prompt-graph-core-v0.1.24","2023-07-28T07:36:15",{"id":200,"version":201,"summary_zh":189,"released_at":202},104441,"prompt-graph-core-v0.1.23","2023-07-28T07:20:51",{"id":204,"version":205,"summary_zh":189,"released_at":206},104442,"chidori-v0.1.24","2023-07-28T07:36:17",{"id":208,"version":209,"summary_zh":210,"released_at":211},104443,"v0.1.23","Node.js SDK is now in line with the Python and Rust SDKs\r\n","2023-07-27T17:58:19",{"id":213,"version":214,"summary_zh":79,"released_at":215},104444,"v0.1.22","2023-07-25T21:47:29",{"id":217,"version":218,"summary_zh":79,"released_at":219},104445,"v0.1.21","2023-07-22T20:08:19",{"id":221,"version":222,"summary_zh":79,"released_at":223},104446,"v0.1.20","2023-07-22T17:20:58",{"id":225,"version":226,"summary_zh":79,"released_at":227},104447,"v0.1.19","2023-07-22T17:12:53",{"id":229,"version":230,"summary_zh":79,"released_at":231},104448,"v0.1.18","2023-07-22T17:02:10"]