[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-wrtnlabs--agentica":3,"tool-wrtnlabs--agentica":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":70,"readme_en":71,"readme_zh":72,"quickstart_zh":73,"use_case_zh":74,"hero_image_url":75,"owner_login":76,"owner_name":77,"owner_avatar_url":78,"owner_bio":79,"owner_company":80,"owner_location":80,"owner_email":81,"owner_twitter":80,"owner_website":82,"owner_url":83,"languages":84,"stars":109,"forks":110,"last_commit_at":111,"license":112,"difficulty_score":23,"env_os":113,"env_gpu":113,"env_ram":113,"env_deps":114,"category_tags":123,"github_topics":124,"view_count":23,"oss_zip_url":80,"oss_zip_packed_at":80,"status":16,"created_at":141,"updated_at":142,"faqs":143,"releases":173},2630,"wrtnlabs\u002Fagentica","agentica","TypeScript AI AI Function Calling Framework enhanced by compiler skills.","Agentica 是一个专为 TypeScript 开发者打造的 AI 智能体（Agent）开发框架，核心专注于简化\"AI 函数调用”的实现过程。它旨在解决传统 AI 代理开发中复杂的集成难题，让开发者无需深入钻研大模型底层原理，只需列出已有的业务函数，即可快速构建具备执行能力的智能体。\n\n无论是后端工程师还是前端开发者，只要熟悉 TypeScript，就能轻松上手。Agentica 支持三种灵活的函数接入方式：直接复用现有的 TypeScript 类、导入标准的 Swagger\u002FOpenAPI 文档，或连接 MCP（模型上下文协议）服务器。这意味着你可以像搭积木一样，将电商交易、新闻检索或文件管理等现有功能迅速转化为 AI 能力。\n\n其独特的技术亮点在于“编译器增强”技能，利用静态类型检查确保 AI 调用的函数参数准确无误，大幅减少了运行时错误。配合内置的项目初始化向导，用户可一键生成包含常见服务（如 GitHub、Google Calendar 等）的 Node.js 或 NestJS 项目模板。如果你希望用熟悉的后端开发经验来构建可靠的 AI 应用，Agentica 能让这一过程","Agentica 是一个专为 TypeScript 开发者打造的 AI 智能体（Agent）开发框架，核心专注于简化\"AI 函数调用”的实现过程。它旨在解决传统 AI 代理开发中复杂的集成难题，让开发者无需深入钻研大模型底层原理，只需列出已有的业务函数，即可快速构建具备执行能力的智能体。\n\n无论是后端工程师还是前端开发者，只要熟悉 TypeScript，就能轻松上手。Agentica 支持三种灵活的函数接入方式：直接复用现有的 TypeScript 类、导入标准的 Swagger\u002FOpenAPI 文档，或连接 MCP（模型上下文协议）服务器。这意味着你可以像搭积木一样，将电商交易、新闻检索或文件管理等现有功能迅速转化为 AI 能力。\n\n其独特的技术亮点在于“编译器增强”技能，利用静态类型检查确保 AI 调用的函数参数准确无误，大幅减少了运行时错误。配合内置的项目初始化向导，用户可一键生成包含常见服务（如 GitHub、Google Calendar 等）的 Node.js 或 NestJS 项目模板。如果你希望用熟悉的后端开发经验来构建可靠的 AI 应用，Agentica 能让这一过程变得简单而高效。","# Agentica, AI Function Calling Framework\n\n![Agentica - ReadMe Diagram](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_readme_815bb1e88d05.png)\n\n\u003C!-- Github\u002FNPM Badges -->\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fblob\u002Fmaster\u002FLICENSE\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-blue.svg\" alt=\"MIT License\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@agentica\u002Fcore\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002F@agentica\u002Fcore.svg\" alt=\"NPM Version\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@agentica\u002Fcore\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fdm\u002F@agentica\u002Fcore.svg\" alt=\"NPM Downloads\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdormoshe.io\u002Fnewsletters\u002F373\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDorMoshe%20Newsletter-Top%20%236%20of%201K-orange?style=flat&logo=rss\" alt=\"Newsletter Top #6\"\u002F>\u003C\u002Fa>\n \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Factions?query=workflow%3Abuild\">\u003Cimg src=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fworkflows\u002Fbuild\u002Fbadge.svg\" alt=\"Build Status\"\u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003C!-- Youtube + Discord -->\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fwww.youtube.com\u002F@wrtnlabs\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FYouTube%20Tutorial-0d1117?style=social&logo=youtube\" alt=\"YouTube\"\u002F>\n  \u003C\u002Fa>\n   \n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FaMhRmzkqCx\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-0d1117?style=social&logo=discord\" alt=\"Discord\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fwww.bloomberg.com\u002Fnews\u002Fvideos\u002F2025-03-31\u002Fwtrn-on-series-b-funding-growth-strategy-video\">\n    \u003Cimg src=\"https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fimages\u002Fbadges\u002Ffund-raising-news-202503.svg\" \u002F>\n  \u003C\u002Fa>\n    \n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\">\n    \u003Cimg src=\"https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fimages\u002Fbadges\u002Fopen-source-mission.svg\" \u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\nAgentic AI framework specialized in AI Function Calling.\n\nDon't be afraid of AI agent development. Just list functions from three protocols below. This is everything you should do for AI agent development.\n\n- TypeScript Class\n- Swagger\u002FOpenAPI Document\n- MCP (Model Context Protocol) Server\n\nWanna make an e-commerce agent? Bring in e-commerce functions. Need a newspaper agent? Get API functions from the newspaper company. Just prepare any functions that you need, then it becomes an AI agent.\n\nAre you a TypeScript developer? Then you're already an AI developer. Familiar with backend development? You're already well-versed in AI development. Anyone who can make functions can make AI agents.\n\n\u003C!-- eslint-skip -->\n\n```typescript\n\nimport { Agentica, assertHttpController } from \"@agentica\u002Fcore\";\nimport OpenAI from \"openai\";\nimport typia from \"typia\";\n\nimport { MobileFileSystem } from \".\u002Fservices\u002FMobileFileSystem\";\n\nconst agent = new Agentica({\n  vendor: {\n    api: new OpenAI({ apiKey: \"********\" }),\n    model: \"gpt-4o-mini\",\n  },\n  controllers: [\n    \u002F\u002F functions from TypeScript class\n    typia.llm.controller\u003CMobileFileSystem>(\n      \"filesystem\",\n      new MobileFileSystem(),\n    ),\n    \u002F\u002F functions from Swagger\u002FOpenAPI\n    assertHttpController({\n      name: \"shopping\",\n      model: \"chatgpt\",\n      document: await fetch(\n        \"https:\u002F\u002Fshopping-be.wrtn.ai\u002Feditor\u002Fswagger.json\",\n      ).then(r => r.json()),\n      connection: {\n        host: \"https:\u002F\u002Fshopping-be.wrtn.ai\",\n        headers: { Authorization: \"Bearer ********\" },\n      },\n    }),\n  ],\n});\nawait agent.conversate(\"I wanna buy MacBook Pro\");\n\n```\n\n## 📦 Setup\n\n```bash\n$ npx agentica start \u003Cdirectory>\n\n----------------------------------------\n Agentica Setup Wizard\n----------------------------------------\n? Package Manager (use arrow keys)\n  > npm\n    pnpm\n    yarn (berry is not supported)\n? Project Type\n    NodeJS Agent Server\n  > NestJS Agent Server\n    React Client Application\n    Standalone Application\n? Embedded Controllers (multi-selectable)\n    (none)\n    Google Calendar\n    Google News\n  > Github\n    Reddit\n    Slack\n    ...\n```\n\nThe setup wizard helps you create a new project tailored to your needs.\n\nFor reference, when selecting a project type, any option other than \"Standalone Application\" will implement the [WebSocket Protocol](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fwebsocket\u002F) for client-server communication.\n\nFor comprehensive setup instructions, visit our [Getting Started](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002F) guide.\n\n## 💻 Playground\n\nExperience Agentica firsthand through our [interactive playground](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground) before installing.\n\nOur demonstrations showcase the power and simplicity of Agentica's function calling capabilities across different integration methods.\n\n- [TypeScript Class](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground\u002Fbbs)\n- [Swagger\u002FOpenAPI Document](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground\u002Fuploader)\n- [Enterprise E-commerce Agent](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground\u002Fshopping)\n\n![E-commerce Agent Demo](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_readme_f35042e17404.png)\n\n\u003C!--\n@todo this section would be changed after making tutorial playground\n-->\n\n## 📚 Documentation Resources\n\nFind comprehensive resources at our [official website](https:\u002F\u002Fwrtnlabs.io\u002Fagentica).\n\n- [Home](https:\u002F\u002Fwrtnlabs.io\u002Fagentica)\n- [Guide Documents](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs)\n  - [Setup](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fsetup\u002Fcli\u002F)\n  - [Concepts](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Ffunction-calling\u002F)\n  - [Core Library](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fcore\u002F)\n  - [WebSocket Protocol](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fwebsocket\u002F)\n  - [Plugin Modules](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fplugins\u002Fbenchmark\u002F)\n- [Tutorial](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial)\n  - [Productivity](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Fproductivity\u002Farxiv\u002F)\n  - [Coding](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Fcoding\u002Ffile-system\u002F)\n  - [React Native](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Freact-native\u002Fsms\u002F)\n  - [Enterprise](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Fenterprise\u002Fshopping\u002F)\n- [API Documents](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fapi)\n- [Youtube](https:\u002F\u002Fwww.youtube.com\u002F@wrtnlabs)\n- [Paper](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fpaper)\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F2f2a4cdc-6cf1-4304-b82d-04a8ed0be0dd\n\n> Tutorial Videos: https:\u002F\u002Fwww.youtube.com\u002F@wrtnlabs\n\n## 🌟 Why Agentica?\n\n```mermaid\nflowchart\n  subgraph \"JSON Schema Specification\"\n    schemav4(\"JSON Schema v4 ~ v7\") --upgrades--> emended[[\"OpenAPI v3.1 (emended)\"]]\n    schema2910(\"JSON Schema 2019-03\") --upgrades--> emended\n    schema2020(\"JSON Schema 2020-12\") --emends--> emended\n  end\n  subgraph \"Agentica\"\n    emended --\"Artificial Intelligence\"--> fc{{\"AI Function Calling\"}}\n    fc --\"OpenAI\"--> chatgpt(\"ChatGPT\")\n    fc --\"Google\"--> gemini(\"Gemini\")\n    fc --\"Anthropic\"--> claude(\"Claude\")\n    fc --\"High-Flyer\"--> deepseek(\"DeepSeek\")\n    fc --\"Meta\"--> llama(\"Llama\")\n    chatgpt --\"3.1\"--> custom([\"Custom JSON Schema\"])\n    gemini --\"3.0\"--> custom([\"Custom JSON Schema\"])\n    claude --\"3.1\"--> standard([\"Standard JSON Schema\"])\n    deepseek --\"3.1\"--> standard\n    llama --\"3.1\"--> standard\n  end\n```\n\nAgentica enhances AI function calling by the following strategies:\n\n- [**Compiler Driven Development**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Fcompiler-driven-development): constructs function calling schema automatically by compiler skills without hand-writing.\n- [**JSON Schema Conversion**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fcore\u002Fvendor\u002F#schema-specification): automatically handles specification differences between LLM vendors, ensuring seamless integration regardless of your chosen AI model.\n- [**Validation Feedback**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Ffunction-calling#validation-feedback): detects and corrects AI mistakes in argument composition, dramatically reducing errors and improving reliability.\n- [**Selector Agent**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Ffunction-calling#orchestration-strategy): filtering candidate functions to minimize context usage, optimize performance, and reduce token consumption.\n\nThanks to these innovations, Agentica makes AI function calling easier, safer, and more accurate than before. Development becomes more intuitive since you only need to prepare functions relevant to your specific use case, and scaling your agent's capabilities is as simple as adding or removing functions.\n\nIn 2023, when OpenAI announced function calling, many predicted that function calling-driven AI development would become the mainstream. However, in reality, due to the difficulty and instability of function calling, the trend in AI development became agent workflow. Agent workflow, which is inflexible and must be created for specific purposes, has conquered the AI agent ecosystem.\n\nBy the way, as Agentica has resolved the difficulty and instability problems of function calling, the time has come to embrace function-driven AI development once again.\n\n| Type        | Workflow      | Vanilla Function Calling | Agentica Function Calling |\n| ----------- | ------------- | ------------------------ | ------------------------- |\n| Purpose     | ❌ Specific   | 🟢 General               | 🟢 General                |\n| Difficulty  | ❌ Difficult  | ❌ Difficult             | 🟢 Easy                   |\n| Stability   | 🟢 Stable     | ❌ Unstable              | 🟢 Stable                 |\n| Flexibility | ❌ Inflexible | 🟢 Flexible              | 🟢 Flexible               |\n\n## 💬 Community & Support\n\nFor support, questions, or to provide feedback, join our Discord community:\n\n[![Discord](https:\u002F\u002Fdcbadge.limes.pink\u002Fapi\u002Fserver\u002Fhttps:\u002F\u002Fdiscord.gg\u002FaMhRmzkqCx)](https:\u002F\u002Fdiscord.gg\u002FaMhRmzkqCx)\n\n## ⚖️ License\n\nAgentica is open-source and available under the [MIT License](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fblob\u002Fmaster\u002FLICENSE).\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_readme_1d0f03e72e5d.png\" alt=\"Wrtn Labs Logo\" \u002F>\n\u003C\u002Fp>\n\u003Cdiv align=\"center\">\n  Agentica is maintained by \u003Ca href=\"https:\u002F\u002Fwrtnlabs.io\">Wrtn Technologies\u003C\u002Fa> — Empowering developers to transform TypeScript functions and OpenAPI specs into powerful AI agents.\n\u003C\u002Fdiv>\n","# Agentica，AI 函数调用框架\n\n![Agentica - ReadMe 图表](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_readme_815bb1e88d05.png)\n\n\u003C!-- GitHub\u002FNPM 徽章 -->\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fblob\u002Fmaster\u002FLICENSE\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-blue.svg\" alt=\"MIT 许可证\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@agentica\u002Fcore\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002F@agentica\u002Fcore.svg\" alt=\"NPM 版本\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@agentica\u002Fcore\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fdm\u002F@agentica\u002Fcore.svg\" alt=\"NPM 下载量\"\u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdormoshe.io\u002Fnewsletters\u002F373\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDorMoshe%20Newsletter-Top%20%236%20of%201K-orange?style=flat&logo=rss\" alt=\"时事通讯第6名\"\u002F>\u003C\u002Fa>\n \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Factions?query=workflow%3Abuild\">\u003Cimg src=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fworkflows\u002Fbuild\u002Fbadge.svg\" alt=\"构建状态\"\u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003C!-- YouTube + Discord -->\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fwww.youtube.com\u002F@wrtnlabs\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FYouTube%20Tutorial-0d1117?style=social&logo=youtube\" alt=\"YouTube\"\u002F>\n  \u003C\u002Fa>\n   \n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FaMhRmzkqCx\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-0d1117?style=social&logo=discord\" alt=\"Discord\"\u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fwww.bloomberg.com\u002Fnews\u002Fvideos\u002F2025-03-31\u002Fwtrn-on-series-b-funding-growth-strategy-video\">\n    \u003Cimg src=\"https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fimages\u002Fbadges\u002Ffund-raising-news-202503.svg\" \u002F>\n  \u003C\u002Fa>\n    \n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fwrtnlabs\">\n    \u003Cimg src=\"https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fimages\u002Fbadges\u002Fopen-source-mission.svg\" \u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n\n专注于 AI 函数调用的智能体 AI 框架。\n\n别再害怕开发 AI 智能体了。只需列出以下三种协议中的函数即可。这就是你进行 AI 智能体开发需要做的全部工作。\n\n- TypeScript 类\n- Swagger\u002FOpenAPI 文档\n- MCP（模型上下文协议）服务器\n\n想打造一个电商智能体吗？引入电商相关功能即可。需要一个新闻智能体吗？从报社获取 API 功能就行了。只要准备好你需要的任何功能，它就能成为一个 AI 智能体。\n\n你是 TypeScript 开发者吗？那你就已经是 AI 开发者了。熟悉后端开发吗？那你已经对 AI 开发非常熟练了。任何能够编写函数的人，都能创建 AI 智能体。\n\n\u003C!-- eslint-skip -->\n\n```typescript\n\nimport { Agentica, assertHttpController } from \"@agentica\u002Fcore\";\nimport OpenAI from \"openai\";\nimport typia from \"typia\";\n\nimport { MobileFileSystem } from \".\u002Fservices\u002FMobileFileSystem\";\n\nconst agent = new Agentica({\n  vendor: {\n    api: new OpenAI({ apiKey: \"********\" }),\n    model: \"gpt-4o-mini\",\n  },\n  controllers: [\n    \u002F\u002F 来自 TypeScript 类的函数\n    typia.llm.controller\u003CMobileFileSystem>(\n      \"filesystem\",\n      new MobileFileSystem(),\n    ),\n    \u002F\u002F 来自 Swagger\u002FOpenAPI 的函数\n    assertHttpController({\n      name: \"shopping\",\n      model: \"chatgpt\",\n      document: await fetch(\n        \"https:\u002F\u002Fshopping-be.wrtn.ai\u002Feditor\u002Fswagger.json\",\n      ).then(r => r.json()),\n      connection: {\n        host: \"https:\u002F\u002Fshopping-be.wrtn.ai\",\n        headers: { Authorization: \"Bearer ********\" },\n      },\n    }),\n  ],\n});\nawait agent.conversate(\"我想买 MacBook Pro\");\n\n```\n\n## 📦 设置\n\n```bash\n$ npx agentica start \u003C目录>\n\n----------------------------------------\n Agentica 设置向导\n----------------------------------------\n? 包管理器（使用方向键）\n  > npm\n    pnpm\n    yarn（berry 不支持）\n? 项目类型\n    NodeJS 智能体服务器\n  > NestJS 智能体服务器\n    React 客户端应用\n    独立应用\n? 内置控制器（可多选）\n    （无）\n    Google 日历\n    Google 新闻\n  > Github\n    Reddit\n    Slack\n    ...\n```\n\n设置向导可以帮助你根据需求创建一个新项目。\n\n作为参考，在选择项目类型时，除了“独立应用”之外的任何选项都会实现 [WebSocket 协议](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fwebsocket\u002F) 用于客户端与服务器之间的通信。\n\n有关完整的设置说明，请访问我们的 [入门指南](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002F)。\n\n## 💻 体验区\n\n在安装之前，可以通过我们的 [交互式体验区](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground) 亲身体验 Agentica。\n\n我们的演示展示了 Agentica 在不同集成方式下函数调用功能的强大与简单性。\n\n- [TypeScript 类](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground\u002Fbbs)\n- [Swagger\u002FOpenAPI 文档](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground\u002Fuploader)\n- [企业级电商智能体](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground\u002Fshopping)\n\n![电商智能体演示](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_readme_f35042e17404.png)\n\n\u003C!--\n@todo 这部分内容将在制作教程体验区后进行更新\n-->\n\n## 📚 文档资源\n\n您可以在我们的 [官方网站](https:\u002F\u002Fwrtnlabs.io\u002Fagentica) 找到全面的资源。\n\n- [首页](https:\u002F\u002Fwrtnlabs.io\u002Fagentica)\n- [指南文档](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs)\n  - [设置](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fsetup\u002Fcli\u002F)\n  - [概念](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Ffunction-calling\u002F)\n  - [核心库](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fcore\u002F)\n  - [WebSocket 协议](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fwebsocket\u002F)\n  - [插件模块](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fplugins\u002Fbenchmark\u002F)\n- [教程](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial)\n  - [生产力](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Fproductivity\u002Farxiv\u002F)\n  - [编码](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Fcoding\u002Ffile-system\u002F)\n  - [React Native](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Freact-native\u002Fsms\u002F)\n  - [企业](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Ftutorial\u002Fenterprise\u002Fshopping\u002F)\n- [API 文档](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fapi)\n- [YouTube](https:\u002F\u002Fwww.youtube.com\u002F@wrtnlabs)\n- [论文](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fpaper)\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F2f2a4cdc-6cf1-4304-b82d-04a8ed0be0dd\n\n> 教程视频：https:\u002F\u002Fwww.youtube.com\u002F@wrtnlabs\n\n## 🌟 为什么选择 Agentica？\n\n```mermaid\nflowchart\n  subgraph \"JSON Schema 规范\"\n    schemav4(\"JSON Schema v4 ~ v7\") --升级--> emended[[\"OpenAPI v3.1（修订版）\"]]\n    schema2910(\"JSON Schema 2019-03\") --升级--> emended\n    schema2020(\"JSON Schema 2020-12\") --修订--> emended\n  end\n  subgraph \"Agentica\"\n    emended --\"人工智能\"--> fc{{\"AI 函数调用\"}}\n    fc --\"OpenAI\"--> chatgpt(\"ChatGPT\")\n    fc --\"Google\"--> gemini(\"Gemini\")\n    fc --\"Anthropic\"--> claude(\"Claude\")\n    fc --\"High-Flyer\"--> deepseek(\"DeepSeek\")\n    fc --\"Meta\"--> llama(\"Llama\")\n    chatgpt --\"3.1\"--> custom([\"自定义 JSON Schema\"])\n    gemini --\"3.0\"--> custom([\"自定义 JSON Schema\"])\n    claude --\"3.1\"--> standard([\"标准 JSON Schema\"])\n    deepseek --\"3.1\"--> standard\n    llama --\"3.1\"--> standard\n  end\n```\n\nAgentica 通过以下策略增强了 AI 函数调用的能力：\n\n- [**编译器驱动开发**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Fcompiler-driven-development)：无需手动编写，直接利用编译器技术自动构建函数调用模式。\n- [**JSON Schema 转换**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fcore\u002Fvendor\u002F#schema-specification)：自动处理不同大模型供应商之间的规范差异，确保无论你选择哪种 AI 模型都能无缝集成。\n- [**验证反馈**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Ffunction-calling#validation-feedback)：检测并纠正 AI 在参数构造中的错误，大幅减少失误、提升可靠性。\n- [**选择器智能体**](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs\u002Fconcepts\u002Ffunction-calling#orchestration-strategy)：筛选候选函数，以最小化上下文使用量、优化性能并降低 token 消耗。\n\n得益于这些创新，Agentica 使 AI 函数调用比以往更加简单、安全和精准。开发者只需准备与特定用例相关的函数，开发过程便更为直观；而扩展智能体的能力也只需增删函数即可，十分简便。\n\n2023 年，当 OpenAI 宣布推出函数调用功能时，许多人预测基于函数调用的 AI 开发将成为主流。然而，实际上由于函数调用的复杂性和不稳定性，AI 开发的趋势却转向了工作流式智能体。这种工作流式智能体虽然灵活度较低，且需针对特定场景定制，却迅速占据了 AI 智能体生态系统的主导地位。\n\n值得一提的是，随着 Agentica 解决了函数调用的复杂性与不稳定性问题，重新拥抱函数驱动型 AI 开发的时机已经到来。\n\n| 类型        | 工作流式      | 原生函数调用 | Agentica 函数调用 |\n| ----------- | ------------- | ------------------------ | ------------------------- |\n| 目的        | ❌ 特定       | 🟢 通用               | 🟢 通用                |\n| 难度        | ❌ 困难        | ❌ 困难             | 🟢 简单                   |\n| 稳定性      | 🟢 稳定        | ❌ 不稳定          | 🟢 稳定                 |\n| 灵活性      | ❌ 不灵活     | 🟢 灵活              | 🟢 灵活               |\n\n## 💬 社区与支持\n\n如需支持、提问或提供反馈，请加入我们的 Discord 社区：\n\n[![Discord](https:\u002F\u002Fdcbadge.limes.pink\u002Fapi\u002Fserver\u002Fhttps:\u002F\u002Fdiscord.gg\u002FaMhRmzkqCx)](https:\u002F\u002Fdiscord.gg\u002FaMhRmzkqCx)\n\n## ⚖️ 许可证\n\nAgentica 是开源项目，采用 [MIT 许可证](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fblob\u002Fmaster\u002FLICENSE) 授权。\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_readme_1d0f03e72e5d.png\" alt=\"Wrtn Labs Logo\" \u002F>\n\u003C\u002Fp>\n\u003Cdiv align=\"center\">\n  Agentica 由 \u003Ca href=\"https:\u002F\u002Fwrtnlabs.io\">Wrtn Technologies\u003C\u002Fa> 维护——助力开发者将 TypeScript 函数和 OpenAPI 规范转化为强大的 AI 智能体。\n\u003C\u002Fdiv>","# Agentica 快速上手指南\n\nAgentica 是一个专注于 **AI 函数调用 (Function Calling)** 的智能体框架。它允许开发者直接将 TypeScript 类、Swagger\u002FOpenAPI 文档或 MCP 服务器转换为 AI 智能体，无需手动编写复杂的提示词或工作流。\n\n## 1. 环境准备\n\n在开始之前，请确保您的开发环境满足以下要求：\n\n*   **操作系统**: Windows, macOS, 或 Linux\n*   **Node.js**: 建议安装 LTS 版本 (v18 或更高)\n*   **包管理器**: npm, pnpm, 或 yarn\n*   **API Key**: 需要准备一个大模型服务商的 API Key (如 OpenAI, Google Gemini, Anthropic Claude 等)\n\n> **国内开发者提示**：如果访问 npm 官方源较慢，建议使用国内镜像源加速安装。\n> ```bash\n> # 临时使用淘宝镜像源\n> npm config set registry https:\u002F\u002Fregistry.npmmirror.com\n> ```\n\n## 2. 安装步骤\n\nAgentica 提供了交互式向导来快速初始化项目。\n\n### 步骤一：运行初始化命令\n\n在终端中执行以下命令（将 `\u003Cyour-project-name>` 替换为你的项目目录名）：\n\n```bash\nnpx agentica start \u003Cyour-project-name>\n```\n\n### 步骤二：配置项目选项\n\n运行命令后，向导会引导你完成以下配置（使用方向键选择，回车确认）：\n\n1.  **Package Manager**: 选择包管理器 (推荐 `npm` 或 `pnpm`)。\n2.  **Project Type**: 选择项目类型。\n    *   `NodeJS Agent Server`: 纯后端服务。\n    *   `NestJS Agent Server`: 基于 NestJS 的后端服务 (推荐)。\n    *   `React Client Application`: 带前端界面的应用。\n    *   `Standalone Application`: 独立应用程序。\n3.  **Embedded Controllers**: 选择预置的功能模块 (可多选)，例如：\n    *   `Github`: GitHub 操作功能\n    *   `Google Calendar`: 日历管理\n    *   `Slack`: 消息通知\n    *   `(none)`: 不选择任何预置模块，完全自定义。\n\n### 步骤三：安装依赖\n\n向导完成后，进入项目目录并安装依赖：\n\n```bash\ncd \u003Cyour-project-name>\nnpm install\n# 或者如果你选择了 pnpm\u002Fyarn\n# pnpm install\n# yarn install\n```\n\n## 3. 基本使用\n\nAgentica 的核心逻辑非常简单：**定义函数 -> 注册控制器 -> 对话**。\n\n以下是一个最简化的代码示例，展示如何结合 **TypeScript 类** 和 **OpenAPI 文档** 创建一个购物助手智能体。\n\n### 代码示例 (`index.ts`)\n\n```typescript\nimport { Agentica, assertHttpController } from \"@agentica\u002Fcore\";\nimport OpenAI from \"openai\";\nimport typia from \"typia\";\n\n\u002F\u002F 1. 定义你的业务逻辑类 (TypeScript Class)\nclass MobileFileSystem {\n  async writeFile(path: string, content: string): Promise\u003Cvoid> {\n    console.log(`Writing to ${path}: ${content}`);\n  }\n  \n  async readFile(path: string): Promise\u003Cstring> {\n    return `Content of ${path}`;\n  }\n}\n\nconst agent = new Agentica({\n  \u002F\u002F 2. 配置大模型供应商\n  vendor: {\n    api: new OpenAI({ apiKey: process.env.OPENAI_API_KEY }),\n    model: \"gpt-4o-mini\",\n  },\n  controllers: [\n    \u002F\u002F 方式 A: 直接注册 TypeScript 类\n    \u002F\u002F typia 会自动分析类型生成 Schema\n    typia.llm.controller\u003CMobileFileSystem>(\n      \"filesystem\",\n      new MobileFileSystem(),\n    ),\n    \n    \u002F\u002F 方式 B: 注册 Swagger\u002FOpenAPI 文档 (例如电商接口)\n    assertHttpController({\n      name: \"shopping\",\n      model: \"chatgpt\",\n      \u002F\u002F 加载远程或本地的 Swagger JSON\n      document: await fetch(\n        \"https:\u002F\u002Fshopping-be.wrtn.ai\u002Feditor\u002Fswagger.json\",\n      ).then(r => r.json()),\n      connection: {\n        host: \"https:\u002F\u002Fshopping-be.wrtn.ai\",\n        headers: { Authorization: \"Bearer YOUR_SHOPPING_API_KEY\" },\n      },\n    }),\n  ],\n});\n\n\u002F\u002F 3. 开始对话\nasync function main() {\n  const response = await agent.conversate(\"我想买一台 MacBook Pro，并请帮我记录到文件中。\");\n  console.log(response);\n}\n\nmain().catch(console.error);\n```\n\n### 核心特性说明\n\n*   **自动 Schema 生成**: 无需手写 JSON Schema，Agentica 利用编译器技术从 TypeScript 类型或 OpenAPI 文档自动生成。\n*   **多模型兼容**: 自动处理不同大模型厂商（OpenAI, Google, Anthropic 等）之间的 Schema 格式差异。\n*   **错误自愈**: 内置验证反馈机制，如果 AI 生成的参数有误，框架会自动检测并引导 AI 修正，提高稳定性。\n\n### 下一步\n\n*   **在线体验**: 访问 [Agentica Playground](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fplayground) 直接在浏览器中测试不同场景。\n*   **详细文档**: 查阅 [官方文档](https:\u002F\u002Fwrtnlabs.io\u002Fagentica\u002Fdocs) 了解 WebSocket 协议、插件开发和高级配置。","某电商初创团队的后端工程师需要快速构建一个能处理用户自然语言指令（如“帮我买台 MacBook\"）的智能客服 Agent，并对接内部现有的订单系统与第三方物流 API。\n\n### 没有 agentica 时\n- **开发门槛高**：工程师需深入研究 Prompt Engineering 和复杂的 LLM 函数调用协议，将后端业务逻辑手动转换为 AI 可理解的 Schema，耗时且易错。\n- **维护成本大**：一旦后端 API 接口变更（如参数调整），必须同步手动更新 AI 层的描述文件，极易出现文档与代码不一致导致的调用失败。\n- **集成碎片化**：难以统一纳管不同类型的接口，内部 TypeScript 类、外部 Swagger 文档和 MCP 服务需要编写多套适配代码，架构臃肿。\n- **调试困难**：缺乏类型安全保护，AI 生成的参数错误往往在运行时才暴露，排查问题如同大海捞针。\n\n### 使用 agentica 后\n- **零样本开发**：工程师只需列出原有的 TypeScript 类或 Swagger 文档，agentica 利用编译器技能自动将其转化为高精度的 AI 函数定义，无需重写任何业务逻辑。\n- **实时同步**：后端代码或 API 文档更新后，agentica 自动感知并刷新 AI 能力，彻底消除了人工维护描述文件的滞后性与错误率。\n- **统一协议栈**：通过简单的配置即可同时挂载内部 TS 服务、外部 OpenAPI 及 MCP 服务器，将异构数据源无缝整合为统一的 Agent 大脑。\n- **编译级保障**：依托 Typia 技术，在编译阶段即可校验 AI 输出参数的类型安全性，将运行时错误拦截在开发期，大幅提升系统稳定性。\n\nagentica 让传统后端开发者无需学习复杂的 AI 原理，仅凭现有的函数代码即可瞬间构建出高可靠、易维护的智能 Agent 应用。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fwrtnlabs_agentica_815bb1e8.png","wrtnlabs","Wrtn Technologies","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fwrtnlabs_96370e4b.png","",null,"wrtnlabs@wrtn.io","https:\u002F\u002Fwrtnlabs.io","https:\u002F\u002Fgithub.com\u002Fwrtnlabs",[85,89,93,97,101,105],{"name":86,"color":87,"percentage":88},"TypeScript","#3178c6",64.2,{"name":90,"color":91,"percentage":92},"MDX","#fcb32c",33.4,{"name":94,"color":95,"percentage":96},"JavaScript","#f1e05a",1.2,{"name":98,"color":99,"percentage":100},"HTML","#e34c26",0.8,{"name":102,"color":103,"percentage":104},"CSS","#663399",0.4,{"name":106,"color":107,"percentage":108},"Shell","#89e051",0,1012,61,"2026-04-02T17:00:42","MIT","未说明",{"notes":115,"python":116,"dependencies":117},"该工具是基于 TypeScript\u002FNode.js 的 AI 函数调用框架，非 Python 项目。运行需安装 Node.js 环境及包管理器（npm、pnpm 或 yarn，不支持 yarn berry）。支持通过 TypeScript 类、Swagger\u002FOpenAPI 文档或 MCP 服务器定义功能。部分项目类型（非独立应用）需实现 WebSocket 协议进行通信。建议使用官方提供的 CLI 向导 (`npx agentica start`) 初始化项目。","不适用 (基于 Node.js\u002FTypeScript)",[118,119,120,121,122],"@agentica\u002Fcore","openai","typia","Node.js","npm\u002Fpnpm\u002Fyarn",[15,53,54,26,13,14],[125,126,127,128,129,119,130,131,132,133,134,135,136,137,138,139,140],"agent","ai","chatbot","llm-function-calling","multi-agent-system","openapi","swagger","claude","llama","rag","retrieval-augmented-generation","agentic","agentic-ai","agentic-framework","typescript","function-calling","2026-03-27T02:49:30.150509","2026-04-06T06:46:10.250896",[144,149,154,159,164,169],{"id":145,"question_zh":146,"answer_zh":147,"source_url":148},12163,"如何在基于 NodeNext ESM 的 TypeScript 项目中正确配置和使用 @agentica\u002Fvector-selector？","确保将相关包更新到最新版本（例如 v0.29.1 或更高），以解决兼容性问题。需要更新的包包括：@agentica\u002Frpc、@agentica\u002Fvector-selector 和 @agentica\u002Fcore。可以通过运行以下命令进行更新：\n\nnpm install @agentica\u002Frpc@v0.29.1 @agentica\u002Fvector-selector@v0.29.1 @agentica\u002Fcore@v0.29.1\n\n更新后，原有的导入方式（如 BootAgenticaVectorSelector）和配置代码应能正常工作。","https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F466",{"id":150,"question_zh":151,"answer_zh":152,"source_url":153},12164,"在使用 Swagger 文档初始化 Agentica 时遇到类型不匹配（TS2345: Argument of type 'unknown'...）导致构建失败，如何解决？","该错误通常是因为 fetch 返回的 JSON 数据被推断为 'unknown' 类型，而 OpenApi.convert 需要特定的文档类型。虽然官方指南中使用了简化的 fetch 写法，但在严格的 TypeScript 配置下需要显式断言或转换类型。\n\n建议的解决方法是确保获取的数据符合 SwaggerV2.IDocument 或 OpenApiV3.IDocument 类型。如果问题依然存在，请参考项目最新的修复版本（相关修复已在 PR #269 中完成），或者避免直接在参数中使用异步 fetch 链式调用，改为先定义变量并断言类型后再传入。","https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F253",{"id":155,"question_zh":156,"answer_zh":157,"source_url":158},12165,"在 Swagger Playground 中粘贴文本时 JsonInput 组件失效，或者无法写入 Basic Auth 头信息怎么办？","维护者已修复了 Swagger Playground 的相关 Bug。如果遇到 JsonInput 组件粘贴失效的问题，现在可以使用已修复的版本。对于无法写入 Basic Auth 头信息的问题，请确保使用的是最新版本的 playground。如果问题涉及函数调用（function calling）而非单纯的进入聊天机器人，建议提供具体的提示词示例、截图或录屏以便进一步排查。目前可以通过 Swagger uploader playground 进行测试。","https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F89",{"id":160,"question_zh":161,"answer_zh":162,"source_url":163},12166,"CLI 中的 inquirer 版本过旧，是否有计划升级以减少打包体积或优化性能？","项目维护者正在对 CLI 目录进行大量重构，并计划将交互库从 inquirer 迁移到 @clack\u002Fprompts。@clack\u002Fprompts 不仅体积小巧（参考 bundlephobia 数据），而且能够提供更现代的用户体验。如果您没有特殊阻碍，可以期待在后续版本中看到这一变更。","https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F172",{"id":165,"question_zh":166,"answer_zh":167,"source_url":168},12167,"提交 Pull Request (PR) 时，如何自动关联并关闭对应的 GitHub Issue？","在 Pull Request 的描述（Description）或评论中使用特定的关键字语法即可。常用的格式是 `fixes #\u003CIssue 编号>` 或 `closes #\u003CIssue 编号>`。\n\n例如，如果要在合并 PR 时关闭 Issue #304，请在 PR 描述中写入：\nfixes #304\n\n这样当 PR 被合并到默认分支时，对应的 Issue 会自动关闭。","https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F304",{"id":170,"question_zh":171,"answer_zh":172,"source_url":158},12168,"Agentica 与 MCP (Model Context Protocol) 或其他库（如 vercel-ai）相比有什么优势？是否支持本地 LLM？","Agentica 主要是一个结合了 Typia 的库，旨在通过 TypeScript 类型系统平滑协议差异并提供更强的类型安全，而 MCP 是一个通用协议。关于成本，使用 Agentica 配合本地 LLM 可以实现比使用云端模型（如 Claude via MCP）更低的成本。虽然其他工具组合（如 goose + MCP）也能实现本地部署，但 Agentica 通过 Typia 自动生成逻辑和 Swagger 文档，在错误检测和安全性方面提供了独特的优势，特别是在处理复杂类型结构时能显著减少运行时错误。",[174,179,184,189,194,199,204,209,214,219,224,229,234,239,244,249,254,259,264,269],{"id":175,"version":176,"summary_zh":177,"released_at":178},62553,"v0.44.1","*无重大变更*\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.44.0...v0.44.1)","2026-03-15T00:47:52",{"id":180,"version":181,"summary_zh":182,"released_at":183},62554,"v0.44.0","###    🚀 功能特性\n\n- **station**: 下一步发布配置。 -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F535 中提出 [\u003Csamp>(1fff0)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F1fff00a07)\n\n###    🐞 错误修复\n\n- **cli**:\n  - 恢复损坏的构建配置。 -  由 @samchon 和 **Copilot 自动修复（AI 驱动）** 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F536 中完成 [\u003Csamp>(76a93)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F76a934675)\n- **core**:\n  - 引入新的 `typia`，用于宽松的 JSON 解析与类型转换。 -  由 @samchon 和 **sunrabbit123** 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F534 中实现 [\u003Csamp>(ef92f)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fef92f3113)\n  - LLM 函数的返回值类型必须为对象。 -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F537 中修复 [\u003Csamp>(c60fe)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fc60fed074)\n\n#####     [在 GitHub 上查看变更](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.43.3...v0.44.0)","2026-03-13T11:33:43",{"id":185,"version":186,"summary_zh":187,"released_at":188},62555,"v0.43.3","###    🐞 错误修复\n\n- **core**: 将 JSON 解析器拆分为针对每个属性的解析。 -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F533 中完成 [\u003Csamp>(e14d4)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fe14d47563)\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.43.2...v0.43.3)","2026-03-04T16:56:12",{"id":190,"version":191,"summary_zh":192,"released_at":193},62556,"v0.43.2","*无重大变更*\n\n#####     [在GitHub上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.43.1...v0.43.2)","2026-03-04T15:55:29",{"id":195,"version":196,"summary_zh":197,"released_at":198},62557,"v0.43.1","###    🚀 功能特性\n\n- **core**: 对 `JsonUtil.parse` 的字符串化属性进行分解  -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F531 中提出 [\u003Csamp>(4ed4b)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F4ed4b9b65)\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.43.0...v0.43.1)","2026-03-04T15:25:06",{"id":200,"version":201,"summary_zh":202,"released_at":203},62558,"v0.43.0","###    🐞 错误修复\n\n- **core**: 将通用系统提示放置到最后一位。 -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F530 中提出 [\u003Csamp>(1f01c)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F1f01cc01e)\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.42.0...v0.43.0)","2026-02-25T04:51:37",{"id":205,"version":206,"summary_zh":207,"released_at":208},62559,"v0.42.0","*无重大变更*\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.41.4...v0.42.0)","2026-02-25T04:51:33",{"id":210,"version":211,"summary_zh":212,"released_at":213},62560,"v0.41.4","###    🐞 Bug 修复\n\n- **core**: 即使实际未调用，也重试函数调用反馈。 -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F529 中提出 [\u003Csamp>(5ced3)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F5ced32ac7)\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.41.3...v0.41.4)","2026-02-13T16:06:31",{"id":215,"version":216,"summary_zh":217,"released_at":218},62561,"v0.41.3","###    🐞 错误修复\n\n- **核心**: `stringifyValidationFailure` 现在会考虑 `array[]` 路径。 -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F528 中提出 [\u003Csamp>(8c14c)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F8c14c2588)\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.41.2...v0.41.3)","2026-02-11T18:25:10",{"id":220,"version":221,"summary_zh":222,"released_at":223},62562,"v0.41.2","###    🚀 功能\n\n- **agent**: 修复 `json_parse_error.md` 中的占位符语法  -  由 @samchon 在 https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F527 中完成 [\u003Csamp>(889a8)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F889a801b6)\n\n#####     [在 GitHub 上查看更改](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.41.1...v0.41.2)","2026-02-09T13:22:04",{"id":225,"version":226,"summary_zh":227,"released_at":228},62563,"v0.41.1","###    🚀 Features\n\n- **core**: Consider `choices` and undefindable.  -  by @samchon and **Copilot** in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F526 [\u003Csamp>(cde84)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fcde848486)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.41.0...v0.41.1)","2026-01-30T04:56:50",{"id":230,"version":231,"summary_zh":232,"released_at":233},62564,"v0.41.0","###    🐞 Bug Fixes\n\n- **core**:\n  - Token usage and response event type for none stream mode.  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F524 [\u003Csamp>(2f8ae)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F2f8aecebe)\n  - Request\u002Fresponse events' `stream` discrimination.  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F525 [\u003Csamp>(41b95)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F41b953d27)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.40.0...v0.41.0)","2026-01-28T07:42:43",{"id":235,"version":236,"summary_zh":237,"released_at":238},62565,"v0.40.0","###    🔨 Code Refactoring\n\n- **agent**: Unify chat completion request handling and introduce new context request result type  -  by @sunrabbit123 in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F522 [\u003Csamp>(bb4ab)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fbb4ab4951)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.39.0...v0.40.0)","2026-01-26T09:27:20",{"id":240,"version":241,"summary_zh":242,"released_at":243},62566,"v0.39.0","###    🐞 Bug Fixes\n\n- **agent**: Error instance for validation feedback.  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F520 [\u003Csamp>(ac805)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fac8058748)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.38.0...v0.39.0)","2026-01-11T17:00:30",{"id":245,"version":246,"summary_zh":247,"released_at":248},62567,"v0.38.0","###    🚀 Features\n\n- **agent**: MicroAgentica defaultly throws  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F518 [\u003Csamp>(053e9)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F053e9a802)\n\n###    🐞 Bug Fixes\n\n- **agent**: `JsonUtil.stringifyValidationFailure` with unmappable.  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F519 [\u003Csamp>(f61f3)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Ff61f3c087)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.37.0...v0.38.0)","2026-01-08T14:42:21",{"id":250,"version":251,"summary_zh":252,"released_at":253},62568,"v0.37.0","###    🚀 Features\n\n- **agent**: `IMicroAgenticaConfig.throw`  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F517 [\u003Csamp>(09655)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F09655380c)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.36.4...v0.37.0)","2026-01-07T05:26:33",{"id":255,"version":256,"summary_zh":257,"released_at":258},62569,"v0.36.4","###    🐞 Bug Fixes\n\n- **agent**: Revive `es-jsonkit` for `qwen3` models.  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F516 [\u003Csamp>(92a00)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F92a0042fb)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.36.3...v0.36.4)","2026-01-01T05:51:26",{"id":260,"version":261,"summary_zh":262,"released_at":263},62570,"v0.36.3","###    🚀 Features\n\n- **agent**: `es-jsonkit` + `jsonrepair`  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F514 [\u003Csamp>(b0d62)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fb0d623b90)\n\n###    🐞 Bug Fixes\n\n- **agent**: Remake `JsonUtil`  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F515 [\u003Csamp>(8d9fc)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F8d9fc1ee6)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.36.2...v0.36.3)","2026-01-01T02:10:29",{"id":265,"version":266,"summary_zh":267,"released_at":268},62571,"v0.36.2","###    🚀 Features\n\n- **core**: Previous validation errors also customized.  -  by @samchon and **Copilot** in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F513 [\u003Csamp>(df75c)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002Fdf75cdae2)\n\n###    🐞 Bug Fixes\n\n- **website**: Revise contents about universal `ILlmSchema`  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F512 [\u003Csamp>(85e56)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F85e5606e8)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.36.1...v0.36.2)","2025-12-31T16:53:34",{"id":270,"version":271,"summary_zh":272,"released_at":273},62572,"v0.36.1","###    🚀 Features\n\n- **agent**: New `stringifyValidateFailure()` function.  -  by @samchon in https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fissues\u002F511 [\u003Csamp>(48939)\u003C\u002Fsamp>](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcommit\u002F489395ec2)\n\n#####     [View changes on GitHub](https:\u002F\u002Fgithub.com\u002Fwrtnlabs\u002Fagentica\u002Fcompare\u002Fv0.36.0...v0.36.1)","2025-12-29T16:57:46"]