RAG

开源 RAG 框架

检索增强生成（RAG）让大模型根据你自己的文档回答问题：文本先被切成小块、转成向量、存进索引，提问时把最相关的几块加进提示词。这里的项目覆盖整条管线，或其中一环：文档解析、向量检索，或者完整的问答应用。

怎么选

RAG 的回答质量主要取决于检索，而不是模型。比较时看：它怎么处理 PDF、表格、扫描件这类难文档；是否支持关键词加向量的混合检索和重排序；以及能不能显示答案出自哪些来源。

Star 最多的项目

第 97–120 个，共 176 个

NVIDIA

GenerativeAIExamples

★ 4,181

Generative AI reference workflows optimized for accelerated infrastructure and microservice architecture.

Jupyter Notebook

GiovanniPasq

agentic-rag-for-dummies

★ 4,166

A modular Agentic RAG built with LangGraph — learn Retrieval-Augmented Generation Agents in minutes.

Jupyter Notebook

bragai

bRAG-langchain

★ 4,160

Everything you need to know to build your own RAG application

Jupyter Notebook

langroid

★ 4,104

Harness LLMs with Multi-Agent Programming

Python

nageoffer

ragent

★ 4,015

企业级 Agentic RAG 智能体 - 全链路覆盖文档解析、多路检索、意图识别、问题重写、会话记忆、MCP 工具调用与深度思考。面向真实业务场景，从 0 到 1 完整工程实现。

Java

yifanfeng97

Hyper-Extract

★ 3,935

Hypergraph is more powerful. Transform unstructured text into structured knowledge with LLMs. Graphs, hypergraphs, and spatio-temporal extractions — with one command.

Python

LazyAGI

LazyLLM

★ 3,885

Easiest and laziest way for building multi-agent LLMs applications.

Python

Atmosphere

Portable AI agent runtime for the JVM. One @Agent class runs on Spring AI, LangChain4j, Anthropic, or 9 more behind one SPI. Token streaming, tool calls, human approvals, and governance over WebSocket, SSE, gRPC, or WebTransport/HTTP3. Speaks MCP, A2A, and AG-UI.

Java

SciSharp

LLamaSharp

★ 3,796

A C#/.NET library to run LLM (🦙LLaMA/LLaVA) on your local device efficiently.

aiming-lab

SimpleMem

★ 3,758

SimpleMem: Efficient Lifelong Memory for LLM Agents — Text & Multimodal

Python

SamurAIGPT

llm-wiki-agent

★ 3,525

A personal knowledge base that builds and maintains itself. Drop in sources — Claude (or Codex/Gemini) reads them, extracts knowledge, and maintains a persistent interlinked wiki. Works with Claude Code, Codex, OpenCode, Gemini CLI. No API key needed.

Python

DSXiangLi

DecryptPrompt

★ 3,440

总结Prompt&LLM论文，开源数据&模型，AIGC应用

modelscope

evalscope

★ 3,423

A streamlined and customizable framework for efficient large model (LLM, VLM, AIGC) evaluation and performance benchmarking.

Python

johnbean393

Sidekick

★ 3,311

A native macOS app that allows users to chat with a local LLM that can respond with information from files, folders and websites on your Mac without installing any other software. Powered by llama.cpp.

Swift

ombharatiya

ai-system-design-guide

★ 3,302

AI system design guide for engineers building production AI systems and evals.

NVIDIA

skills

★ 3,301

适用于 NVIDIA 产品的代理技能，安装到 Claude Code、Codex 和其他编码代理中，以端到端运行物理 AI、机器人、模拟、CUDA 和 RAG 工作流程。技能目录产品|描述 |技能 |爱智商 | NVIDIA AI-Q 蓝图 - 部署本地 AI-Q 服务并作为代理技能运行浅层或深层研究工作流程。

Python