LightRAG
A retrieval framework that combines knowledge graphs and vector embeddings, with a server, REST API and web interface for indexing and querying documents.
About LightRAG
Document applications need to retrieve connected entities and relationships as well as matching text passages.
Who it’s for
- Developers building document question-answering systems
- Researchers exploring graph-based retrieval
When to consider it
Consider LightRAG when you want to combine graph retrieval and vector search behind a document-query API.
Tradeoffs & limitations
- MIT licensed; self-hosting requires configuring language and embedding models, and switching embedding models requires rebuilding stored vectors.
SOTA overview · Documentation-based assessment · Sources & review method
Updates
No updates shared yet.
Discussion
Newest firstAsk a question or share how you use LightRAG.
Keep it helpful. Community rules
Loading discussion…
Sources & review method
Documentation-based assessment · Oct 16, 2026 · Prepared with AI assistance; not a hands-on benchmark.
- LightRAG retrieval architecture, server and model configuration requirements
- LightRAG MIT license
- HKU institutional profile identifying the LightRAG team and its institutional email domain
- Recent official maintenance commit (2026-10-14, UTC)
Checked by SOTA · AI-assisted documentation review. Selection advice is our assessment; verify current requirements for your deployment.
Import history & original evidence
LightRAG on GitHub
A retrieval framework that combines knowledge graphs and vector embeddings, with a server, REST API and web interface for indexing and querying documents.
Based on official pages and announcements checked on 2026-10-16. No hands-on test, performance benchmark or popularity ranking is claimed.