夜雨聆风AI+RAG Comprehensive Application Excellent Project Packages | AI+RAG综合应用优质项目包合集
Foreword | 前言
Retrieval-Augmented Generation (RAG) technology, as a core solution to solve the problems of large language models (LLMs) such as knowledge timeliness, hallucinations, and domain knowledge deficiencies, has been widely used in intelligent question answering, document understanding, enterprise knowledge management, and other scenarios. This article collates a number of open-source, high-availability AI+RAG integrated application project packages on GitHub, covering full-link RAG frameworks, lightweight deployment tools, domain-specific applications, etc., with both Chinese and English introductions for global developers to refer to and use.
检索增强生成（RAG）技术作为解决大语言模型（LLM）知识时效性、幻觉问题、领域知识缺失等痛点的核心方案，已广泛应用于智能问答、文档理解、企业知识管理等场景。本文整理GitHub上多款开源、高可用的AI+RAG综合应用项目包，涵盖RAG全链路框架、轻量化部署工具、垂直领域专用应用等，采用中英双语介绍，供全球开发者参考使用。
Table of Contents | 目录
1. Full-link RAG Development Frameworks | 全链路RAG开发框架
2. Lightweight RAG Deployment Tools | 轻量化RAG部署工具
3. Domain-specific RAG Application Projects | 垂直领域RAG应用项目
4. RAG Enhanced AI Assistant Projects | RAG增强型AI助手项目
1. Full-link RAG Development Frameworks | 全链路RAG开发框架
1.1 LangChain
• GitHub Address | GitHub地址：https://github.com/langchain-ai/langchain
• English Introduction
The most popular open-source full-link LLM application development framework, with built-in complete RAG core modules, including document loading, text splitting, vector storage, semantic retrieval, prompt engineering, and LLM integration. It supports docking with almost all mainstream LLMs, vector databases, and data sources, provides rich extension components, and is suitable for quickly building custom RAG applications from 0 to 1. It has a complete community ecology and detailed official documents.
• 中文介绍
当下最热门的开源全链路LLM应用开发框架，内置完整RAG核心模块，包含文档加载、文本分割、向量存储、语义检索、提示工程、LLM集成等，支持对接几乎所有主流大模型、向量数据库、数据源，提供丰富拓展组件，适合从零到一快速搭建定制化RAG应用，拥有完善社区生态和详尽官方文档。
• Core Features | 核心特性
◦ Multi-source document loading (PDF, TXT, Markdown, Web pages, database, etc.) | 多源文档加载（PDF、TXT、Markdown、网页、数据库等）
◦ Flexible text segmentation strategy & vector retrieval optimization | 灵活文本分割策略+向量检索优化
◦ Seamless docking with OpenAI, Claude, domestic large models, etc. | 无缝对接OpenAI、Claude、国产大模型等
◦ Support advanced RAG technologies such as multi-retrieval and reranking | 支持多路检索、重排序等高级RAG技术
1.2 LlamaIndex
• GitHub Address | GitHub地址：https://github.com/run-llama/llama_index
• English Introduction
A professional RAG-centric development framework focused on data indexing and retrieval enhancement, optimized for complex structured/unstructured data (knowledge graphs, tables, PDFs with images). It provides a more refined data processing pipeline and retrieval enhancement strategies, and is more suitable for building high-precision RAG systems for enterprise-level knowledge bases and professional domain data.
• 中文介绍
以RAG为核心的专业开发框架，专注于数据索引与检索增强，针对复杂结构化/非结构化数据（知识图谱、表格、带图片PDF）做了深度优化，提供更精细化的数据处理流程和检索增强策略，更适合搭建企业级知识库、专业领域数据的高精度RAG系统。
• Core Features | 核心特性
◦ Professional knowledge base indexing & hybrid retrieval engine | 专业级知识库索引+混合检索引擎
◦ Optimized for complex documents such as tables and knowledge graphs | 针对表格、知识图谱等复杂文档优化
◦ Lightweight deployment & low code development support | 轻量化部署+低代码开发支持
◦ Advanced RAG solutions such as query rewriting and context compression | 查询改写、上下文压缩等高级RAG方案
2. Lightweight RAG Deployment Tools | 轻量化RAG部署工具
2.1 QAnything
• GitHub Address | GitHub地址：https://github.com/netease-youdao/QAnything
• English Introduction
An open-source, lightweight local RAG question-and-answer tool launched by NetEase Youdao, supporting local deployment without relying on cloud services, realizing private data security question-and-answer. It supports multi-format document identification, multi-round dialogue, and automatic optimization of retrieval results, with a simple Web visual interface, low deployment threshold, suitable for personal and small and medium-sized enterprises to build local knowledge Q&A.
• 中文介绍
网易有道推出的开源轻量化本地RAG问答工具，支持本地部署、不依赖云服务，实现私有数据安全问答，支持多格式文档识别、多轮对话、检索结果自动优化，拥有简洁Web可视化界面，部署门槛低，适合个人、中小企业搭建本地知识问答系统。
• Core Features | 核心特性
◦ Full local deployment & data privacy protection | 全本地部署+数据隐私保护
◦ Support multi-format documents (PDF, Word, PPT, Excel, pictures, etc.) | 支持多格式文档（PDF、Word、PPT、Excel、图片等）
◦ One-click deployment & fast start | 一键部署、快速启动
◦ Multi-round conversational RAG question and answer | 多轮对话式RAG问答
2.2 RAGFlow
• GitHub Address | GitHub地址：https://github.com/infiniflow/ragflow
• English Introduction
An open-source visual RAG orchestration tool that realizes visual configuration of the entire RAG process through drag-and-drop, without complex code development. It supports multiple vector databases, LLMs, and document parsing methods, and provides visualized retrieval result debugging and effect optimization, helping developers quickly build and iterate RAG applications.
• 中文介绍
开源可视化RAG编排工具，通过拖拽方式实现RAG全流程可视化配置，无需复杂代码开发，支持多款向量数据库、大模型、文档解析方式，提供可视化检索结果调试与效果优化，助力开发者快速搭建与迭代RAG应用。
• Core Features | 核心特性
◦ Visual drag-and-drop RAG process orchestration | 可视化拖拽式RAG流程编排
◦ Real-time retrieval effect debugging & optimization | 实时检索效果调试与优化
◦ Multi-vector database & large model flexible docking | 多向量数据库、大模型灵活对接
◦ Open source free & commercial friendly | 开源免费、商用友好
3. Domain-specific RAG Application Projects | 垂直领域RAG应用项目
3.1 MedicalGPT-RAG
• GitHub Address | GitHub地址：https://github.com/MediaBrain-SJTU/MedicalGPT-RAG
• English Introduction
A RAG-based medical domain intelligent question-and-answer project, optimized for medical professional knowledge, integrating medical literature, medical guidelines, drug manuals and other professional data, realizing accurate medical consultation, drug consultation, disease analysis and other functions, effectively reducing the hallucination problem of LLM in the medical field.
• 中文介绍
基于RAG的医疗领域智能问答项目，针对医学专业知识做专项优化，整合医学文献、医疗指南、药品说明书等专业数据，实现精准的医疗咨询、用药咨询、病症分析等功能，有效降低大模型在医疗领域的幻觉问题。
• Core Features | 核心特性
◦ Medical professional knowledge base optimization | 医学专业知识库优化
◦ Accurate retrieval of medical literature and guidelines | 医学文献、指南精准检索
◦ Reduce medical LLM hallucinations & improve answer accuracy | 降低医疗大模型幻觉、提升回答准确率
◦ Suitable for medical consultation auxiliary scenarios | 适用于医疗咨询辅助场景
3.2 LegalRAG
• GitHub Address | GitHub地址：https://github.com/ruxia-TJY/LegalRAG
• English Introduction
A RAG application project for the legal field, integrating national laws, regulations, judicial interpretations, precedent cases and other data, providing legal consultation, case analysis, legal document assistance and other services, realizing accurate legal knowledge retrieval and professional legal answer generation, to meet the needs of legal practitioners and ordinary users.
• 中文介绍
面向法律领域的RAG应用项目，整合国家法律法规、司法解释、判例等数据，提供法律咨询、案例分析、法律文书辅助等服务，实现精准的法律知识检索与专业法律回答生成，满足法律从业者、普通用户的法律问答需求。
• Core Features | 核心特性
◦ Legal database standardized sorting | 法律数据库规范化整理
◦ Legal provision & case precise retrieval | 法律条文、判例精准检索
◦ Professional legal answer generation | 专业化法律回答生成
◦ Adapt to legal consultation scenarios | 适配法律咨询场景
4. RAG Enhanced AI Assistant Projects | RAG增强型AI助手项目
4.1 Dify
• GitHub Address | GitHub地址：https://github.com/langgenius/dify
• English Introduction
An open-source AI application development platform, built-in RAG core capabilities, supporting one-click construction of RAG-based intelligent assistants, customer service systems, and knowledge base Q&A. It provides visual application configuration, dialogue management, data analysis and other functions, supports rapid deployment to online environments, and is suitable for enterprises to build commercial AI applications.
• 中文介绍
开源AI应用开发平台，内置RAG核心能力，支持一键搭建基于RAG的智能助手、客服系统、知识库问答，提供可视化应用配置、对话管理、数据分析等功能，支持快速部署至线上环境，适合企业搭建商用级AI应用。
• Core Features | 核心特性
◦ One-click construction of RAG AI assistant | 一键搭建RAG型AI助手
◦ Visual background management & data statistics | 可视化后台管理+数据统计
◦ Multi-terminal access (Web, WeChat, applet, etc.) | 多端接入（网页、微信、小程序等）
◦ Support commercial customized development | 支持商用定制化开发
4.2 FastGPT
• GitHub Address | GitHub地址：https://github.com/labring/FastGPT
• English Introduction
A fast-deployed RAG-based AI knowledge base assistant project, with a simple and efficient knowledge base construction process, supporting multi-round dialogue, intelligent retrieval, answer citation and other functions. It can quickly realize the intelligent Q&A of enterprise internal knowledge, product documents, and online customer service, with low operation and maintenance costs and high usability.
• 中文介绍
可快速部署的基于RAG的AI知识库助手项目，拥有简洁高效的知识库搭建流程，支持多轮对话、智能检索、回答溯源等功能，可快速实现企业内部知识、产品文档、在线客服的智能问答，运维成本低、实用性高。
• Core Features | 核心特性
◦ Fast knowledge base construction & indexing | 快速知识库搭建与索引
◦ Answer source traceability & retrieval result display | 回答来源溯源+检索结果展示
◦ Low-cost rapid deployment | 低成本快速部署
◦ Open source community active & continuous update | 开源社区活跃、持续更新
Summary | 总结
The above AI+RAG project packages cover full-link development, lightweight deployment, vertical domain applications and other scenarios, all of which are open-source and active projects on GitHub. Developers can select appropriate projects according to their own business needs, technical thresholds, and deployment scenarios to quickly build efficient, accurate, and secure RAG applications, and effectively improve the practical value of LLM in various fields.
以上AI+RAG项目包涵盖全链路开发、轻量化部署、垂直领域应用等多场景，均为GitHub上开源且活跃度高的项目。开发者可根据自身业务需求、技术门槛、部署场景，选择合适的项目，快速搭建高效、精准、安全的RAG应用，有效提升大模型在各领域的实用价值。
Contribution Guide | 贡献指南
Welcome to submit Issue or Pull Request to supplement more high-quality AI+RAG open-source projects, and jointly improve the comprehensiveness of this project collection!
欢迎提交Issue或Pull Request，补充更多优质AI+RAG开源项目，共同完善本项目合集的全面性！
