LightRAG

Haystack 是企业级 RAG 与搜索应用框架，支持文档处理、检索、生成与评估全链路。

A MemAgent framework that can extrapolate to 3.5M context tokens, along with a training framework for RL training of any agent workflow.

KAG 是基于 OpenSPG 引擎和 LLM 的逻辑形式引导推理与检索框架，用于构建专业领域知识库的逻辑推理和事实问答解决方案，有效克服传统 RAG 向量相似度计算模型的不足。

Opinionated RAG framework for integrating GenAI into your apps. Works with any LLM, any vectorstore, any files — so you can focus on your product instead of building RAG pipelines.