neo4j/neo4j-graphrag-python

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

1275 stars · 237 forks observed · 2026-09-03

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

The official Neo4j first-party Python library for building graph retrieval-augmented generation (GraphRAG) applications. It provides retrievers (vector, hybrid, Text2Cypher), LLM and embedding provider integrations, and an experimental pipeline for building knowledge graphs from unstructured data.

Use cases

• build a graphrag application on neo4j
• retrieve context from a knowledge graph with llm
• convert natural language questions to cypher queries
• extract entities and relations from pdfs into a knowledge graph
• combine vector search with graph traversal for retrieval
• create vector indexes and run similarity search in neo4j
• build a rag pipeline with openai or ollama over a graph database

When to choose

• you use Neo4j as the graph store and want first-party, long-term-supported GraphRAG tooling
• you need hybrid retrieval, graph traversal enrichment, or Text2Cypher retrieval in Python
• you want an end-to-end pipeline to build knowledge graphs from unstructured documents
• you need pluggable LLM/embedding providers (OpenAI, Anthropic, Ollama, Bedrock, etc.) with a LangChain-compatible interface

When to avoid

• your RAG stack does not involve Neo4j or graph databases
• you need a general-purpose agent framework or non-graph RAG library like LangChain or LlamaIndex
• you require a stable, non-experimental knowledge graph construction pipeline (that feature is still experimental)
• you are not using Python 3.10+ or Neo4j >= 5.18.1

Facets

library · maturity active

rag llm-inference agent-framework search-engine vector-database nlp database-driver prompt-engineering large-language-models databases artificial-intelligence python cross-platform graphrag neo4j knowledge-graph cypher text2cypher retrievers kg-builder embeddings genai retrieval-augmented-generation natural-language-processing search data-engineering

Member repositories

For agents

markdown · JSON · MCP: product_card(name="neo4j/neo4j-graphrag-python")