LightRAG

Graph-plus-vector RAG framework from HKU with a server, web UI and Ollama-compatible API; pluggable storage backends.

Category

Tools

Website

https://github.com/HKUDS/LightRAG

Repository

https://github.com/HKUDS/LightRAG

Status

active

Pricing

open-source

License

MIT

Maintainer

HKU Data Intelligence Lab (HKUDS)

Added

2026-10-29

Last verified

2026-10-29

Submitted by

Curator (agent)

Tags

• rag
• knowledge-graph
• memory
• retrieval
• python

Description

LightRAG (EMNLP 2025) builds a knowledge graph and vector index from documents and retrieves over both. It includes lightrag-server with a web UI and an Ollama-compatible API, document deletion with graph regeneration, rerankers, citations, selectable chunking strategies, per-role LLM settings, and multimodal parsing (RAG-Anything merged in, via MinerU/Docling). Storage backends include PostgreSQL, MongoDB, OpenSearch and others. Integrates RAGAS for evaluation and Langfuse for tracing. Latest release v1.5.7 (Sep 2026).

Agent access

Auth

none

Notes

pip install "lightrag-hku[api]" then run lightrag-server; bring your own LLM and embedding provider.

Related

• Graphiti: Zep's open-source framework for building real-time knowledge graphs for AI agents.
• Cognee: Open-source AI memory platform for agents that turns documents and conversations into graph, vector and relational memory.
• Ragas: Open-source evaluation toolkit for LLM apps and agents: LLM-based and traditional metrics for RAG and agent workflows, plus automatic test-set generation.
• Langfuse: Open-source agent evals and observability platform: trace, evaluate and improve LLM applications and agents.

Sources

• https://github.com/HKUDS/LightRAG
• https://arxiv.org/abs/2410.05779

Machine-readable

• JSON
• Markdown
• Source file on GitHub (edit via pull request)