HKUDS/MiniRAG
A lightweight RAG framework designed to help small language models perform retrieval-augmented generation efficiently.
Not currently ranked — collecting fresh signals.
star history
MiniRAG is an open-source retrieval-augmented generation framework optimized for small language models. It employs heterogeneous graph indexing and lightweight topology-enhanced retrieval to simplify RAG implementation while maintaining performance. The project originates from Hong Kong University (HKUDS) research group and supports both API and Docker deployment.
Frequently asked
- What is HKUDS/MiniRAG?
- A lightweight RAG framework designed to help small language models perform retrieval-augmented generation efficiently.
- Is MiniRAG open source?
- Yes — HKUDS/MiniRAG is open source, released under the MIT license.
- What language is MiniRAG written in?
- HKUDS/MiniRAG is primarily written in Python.
- How popular is MiniRAG?
- HKUDS/MiniRAG has 2k stars on GitHub.
- Where can I find MiniRAG?
- HKUDS/MiniRAG is on GitHub at https://github.com/HKUDS/MiniRAG.