neo4j-labs/llm-graph-builder

What is it?

Key features

• 01Builds knowledge graphs with nodes, relationships and properties from unstructured files.
• 02Imports data from local files, S3, Google Cloud Storage, YouTube and web pages.
• 03Supports multiple LLM providers including OpenAI, Gemini, Diffbot, Anthropic and Ollama.
• 04Allows custom schemas for entity and relationship labels.
• 05Provides a chat interface that answers questions over the graph with source metadata.
• 06Offers graph visualization in Neo4j Bloom.

Use cases

• Turning a collection of PDFs or reports into a queryable graph of entities and relationships.
• Building a GraphRAG chatbot over your own documents stored in Neo4j.
• Extracting structured knowledge from YouTube transcripts or web articles.
• Running extraction with a local Ollama model when data must stay on your own machine.

Quick start

1. 1

   Prerequisite: Python 3.12 and a Neo4j 5.23+ database with APOC

2. 2

   Create backend/.env from backend/example.env and set NEO4J_URI, NEO4J_USERNAME, NEO4J_PASSWORD, NEO4J_DATABASE

3. 3

   Create a virtual environment and install backend requirements

   cd backend && python3.12 -m venv venv && source venv/bin/activate && pip install -r requirements.txt
4. 4

   Start the backend server

   uvicorn score:app --reload

Things to know

• !Requires a Neo4j 5.23+ database with APOC installed; Neo4j Desktop needs separate backend and frontend deployment.
• !Needs an LLM API key (for example OpenAI or Diffbot) to build graphs.

Topics

FAQ

What is llm-graph-builder?

Web app that uses LLMs and LangChain to turn PDFs, docs, videos and web pages into a knowledge graph stored in Neo4j. Knowledge Graph Builder converts unstructured content such as PDFs, DOCs, TXT files, YouTube videos and web pages into a structured knowledge graph in Neo4j. It uses large language models through the LangChain framework to extract nodes, relationships and their properties, and lets you define a custom schema for the extraction.

Is llm-graph-builder free to use? What license does it have?

llm-graph-builder is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.

Is llm-graph-builder still maintained?

Actively maintained. The last commit to llm-graph-builder was on Oct 21, 2026.

What can I use llm-graph-builder for?

Turning a collection of PDFs or reports into a queryable graph of entities and relationships. Building a GraphRAG chatbot over your own documents stored in Neo4j. Extracting structured knowledge from YouTube transcripts or web articles. Running extraction with a local Ollama model when data must stay on your own machine.

How do I install llm-graph-builder?

Prerequisite: Python 3.12 and a Neo4j 5.23+ database with APOC Create backend/.env from backend/example.env and set NEO4J_URI, NEO4J_USERNAME, NEO4J_PASSWORD, NEO4J_DATABASE Create a virtual environment and install backend requirements: cd backend && python3.12 -m venv venv && source venv/bin/activate && pip install -r requirements.txt Start the backend server: uvicorn score:app --reload

Where can I get llm-graph-builder?

The source code and installation instructions are on GitHub: https://github.com/neo4j-labs/llm-graph-builder