neo4j-graphrag

What it is and what it does

Neo4j GraphRAG is a first-party Python library from Neo4j that bridges language models and graph databases to build retrieval-augmented generation systems. It provides two main workflows: constructing knowledge graphs from text or PDFs using LLM-driven entity and relationship extraction, and retrieving relevant graph data to augment LLM prompts for question-answering and reasoning tasks.

The package wraps Neo4j's graph database with high-level abstractions—SimpleKGPipeline for streamlined knowledge graph building, Pipeline for advanced customization, and multiple retriever strategies (vector search, text-to-Cypher, hybrid traversal). It integrates with LLM providers (OpenAI, Anthropic, Cohere, Bedrock, etc.) and optional embeddings backends (sentence-transformers, Weaviate, Pinecone, Qdrant). Core dependencies include pydantic for schema validation, tenacity for retry logic, and utilities like pypdf and json-repair for data handling.

Use it for

• Build a knowledge graph from unstructured documents, then use it to answer domain-specific questions with LLM-grounded retrieval.
• Implement hybrid search combining vector similarity with graph traversal to find contextually relevant entities and their relationships.
• Extract structured facts (entities, relationships, patterns) from PDFs or text using LLM-guided pipelines without manual annotation.
• Create a text-to-Cypher retriever that translates natural language queries into graph database queries for precise, schema-aware retrieval.
• Augment LLM responses with real-time graph data to reduce hallucination and ground answers in a curated knowledge base.

Worth the install?

AI-flagged interpretation of the facts on this page. Verify before relying on it.

Install

neo4j-graphrag on PyPI

Before you install

Low friction: pure Python wheel with no compiled dependencies. Active maintenance with recent releases; last commit 2026-08-13. Requires Neo4j instance and at least one LLM provider extra (openai, anthropic, etc.) to function for RAG tasks.

Requires a running Neo4j instance (local or remote). APOC core library must be installed in Neo4j for knowledge graph construction. At least one LLM provider extra must be installed (e.g., openai, anthropic). Python 3.10–3.13 recommended; Python 3.14 support limited due to upstream spaCy issue.

License in practice

Apache License 2.0 (permissive): you can use, modify, and distribute this package freely in commercial and open-source projects, provided you include a copy of the license and state material changes.

Quickstart

pip install 'neo4j-graphrag[openai]'

from neo4j import GraphDatabase
from neo4j_graphrag.embeddings import OpenAIEmbeddings
from neo4j_graphrag.experimental.pipeline.kg_builder import SimpleKGPipeline
from neo4j_graphrag.llm import OpenAILLM

driver = GraphDatabase.driver('neo4j://localhost:7687', auth=('neo4j', 'password'))
embedder = OpenAIEmbeddings(model='text-embedding-3-large')
llm = OpenAILLM(model_name='gpt-4')
kg_builder = SimpleKGPipeline(llm=llm, driver=driver, embedder=embedder, schema={...})
await kg_builder.run_async(text='your text here')

Package facts

Evidence: neo4j_graphrag-1.18.0-py3-none-any.whl

Tags

More Artificial Intelligence packages

See also graphdatascience · graphrag · langchain-neo4j · neomodel · graph-retriever · ragstack-ai-knowledge-store · graphlib · lightrag-hku · graphiti-core · neo4j-driver

Further reading