LightRAG
Fast retrieval-augmented generation framework that fuses knowledge-graph structure with vector retrieval.
LightRAG
LightRAG is a simple and fast retrieval-augmented generation framework from the HKU Data Intelligence Lab (EMNLP 2025). It combines automatically constructed knowledge-graph structure with vector similarity search, giving more context-aware and relational answers than plain vector RAG while staying lighter than heavyweight GraphRAG pipelines.
Key features
- Dual-level retrieval merging graph relationships and vector similarity
- Incremental knowledge-graph construction from ingested documents
- Multiple storage backends (JSON, PostgreSQL, Neo4j, Milvus, and more)
- Pluggable LLM and embedding backends (OpenAI, Ollama, Hugging Face)
- Fast, low-overhead indexing suitable for iterative research corpora
Usage note: install via pip, point it at your documents to build the graph + vector index, then query with graph, vector, or hybrid retrieval modes.
Curated mirror of the open-source LightRAG (MIT). Get it from the source.
Compromise signals — malicious or tampered code (leaked secrets, backdoors, a dropped executable) — reduce the score, and known dependency CVEs carry a bounded penalty (they warrant review but never QUARANTINE — update the dependency to clear). Other dangerous-by-capability traits are risk surface, expected for some capabilities. Every finding is mapped to its OWASP control below.
Findings mapped to the OWASP Top 10 for LLM Applications (2025) and the OWASP Machine Learning Security Top 10. Expand any flagged control for the exact findings — compromise reduces the score; expected/risk-surface do not, except a known CVE, which carries a small bounded penalty (high/critical → Review).
The same gate an agent runs before installing (POST /api/v1/trust/lightrag-graph-rag/check). Click a policy:
Consume LightRAG programmatically. Authenticate with an API key or session — see Authorize an agent.
# Agents: CHECK BEFORE YOU INSTALL (no auth) — score, grade, level, capability manifest
curl https://ai-supply.store/api/v1/trust/lightrag-graph-rag
# Gate against your org policy (returns { pass, violations })
curl -X POST https://ai-supply.store/api/v1/trust/lightrag-graph-rag/check \
-H "Content-Type: application/json" \
-d '{"minGrade":"B","denyPermissions":["shell"],"denyUnknownEgress":true}'
# CLI
npx ai-supply add lightrag-graph-rag
# REST (install → download)
curl -X POST https://ai-supply.store/api/v1/listings/lightrag-graph-rag/install \
-H "Authorization: Bearer $AIM_KEY"
# MCP tool
install_listing({ "slug": "lightrag-graph-rag" })OpenAPI spec →Curated mirror — latest upstream source. See the repository for tagged releases.