围绕 GPT-6 Astra 的 7 个 RAG 开源仓库清单
把公司文档变成可召回知识库的七层工具链：解析、切块、向量库、图检索、记忆、评测、追踪，模型可随时替换。
7 GitHub repos around GPT 6 Astra turn company docs into a knowledge system that actually recalls 01 Docling ▸ https://github.com/docling-project/docling → parses PDFs and Office into structured model input 02 Chonkie ▸ https://github.com/chonkie-inc/chonkie → chunks and ships documents for retrieval 03 Qdrant ▸ https://github.com/qdrant/qdrant → stores embeddings for fast similarity search 04 LightRAG ▸ https://github.com/HKUDS/LightRAG → mixes graph and vector retrieval in one pipeline 05 Zep ▸ https://github.com/getzep/zep → keeps agent memory and context across sessions 06 Ragas ▸ https://github.com/explodinggradients/ragas → scores whether the retrieval answers are any good 07 Opik ▸ https://github.com/comet-ml/opik → traces the full run when answers go wrong Astra should spend tokens on judgment, not rediscovering the same PDF Parsing, chunking, search, memory and evals stay in dedicated layers Replace the model tomorrow and the knowledge stack still holds
这条是引用着另一条写的：x.com/promixyz/status/2102424036416348494