LightRAG
HKU Data Science LabLightweight graph-based RAG framework from HKU Data Science Lab (EMNLP 2025) combining knowledge graphs with vector embeddings via dual-layer retrieval. Drops GraphRAG's expensive community-report step and supports incremental updates (add data without rebuilding the global index).
MIT. ~37k stars, ~251k monthly PyPI downloads (lightrag-hku).
Openness
5 high confidence5.0
- license
- MIT(OSI)
- source
- public
- core-gated
- ungated
Fully MIT.
- https://github.com/HKUDS/LightRAG recorded 2026-06-17
MIT, ~36.7k stars, EMNLP 2025
Adoption
4 medium confidence4.0
~251k PyPI downloads/month (lightrag-hku) corroborating ~37k stars.
- https://pepy.tech/projects/lightrag-hku recorded 2026-06-17
~251k downloads/month
Capability
4 medium confidence4.0
Deeper and more efficient than GraphRAG for graph-RAG; narrower than general orchestrators.
- https://github.com/HKUDS/LightRAG recorded 2026-06-17
feature set: dual-level retrieval, incremental updates
Unchanged since 2026-07-30 (last edited, not re-checked)