OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

LLM Graph Builder

An open-source application that uses selectable LLMs and LangChain to extract structured knowledge graphs from unstructured sources and store them in Neo4j.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

LLM Graph Builder — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a project needs to turn mixed unstructured inputs into a Neo4j knowledge graph that can support graph search, visualization or conversational retrieval.

Who it is for

  • Teams building GraphRAG or graph-search applications
  • Researchers organizing mixed document and web collections
  • Developers creating domain-specific knowledge graph pipelines
  • Organizations working with Neo4j-backed conversational retrieval

Documented uses

  • Ingest documents, videos and web pages from several source types
  • Extract nodes, relationships and properties with an LLM
  • Store and visualize the resulting graph in Neo4j
  • Query graph-backed data conversationally
  • Configure schemas, embedding models and chat modes
Before you choose

Requirements & limits

The repository is public, active and licensed under Apache-2.0. Documented deployment options include web, cloud and self-hosted environments; local setup requirements include Python 3.12 or later and Neo4j 5.23 or later with APOC.

  • Neo4j is required, with version 5.23 or later and APOC for the documented local setup.
  • Input coverage includes local files, Google Cloud Storage, Amazon S3, YouTube and web sources, with available sources depending on configuration.
  • Embedding models can be selected from the frontend when user tracking is enabled; otherwise they are configured through backend environment variables.
  • Neo4j Bloom is used for graph visualization, while conversational querying is available through the database and a documented chat-only route.
  • Adding a domain-specific extraction schema and validation layer is Likely compatible, but the supplied material does not verify an implementation.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up the backend for a local deployment

Prepare the documented backend environment for LLM Graph Builder.

Before you start

  • Python 3.12 or later
  • Neo4j 5.23 or later with APOC installed
  • Neo4j connection details

Steps

  1. Copy backend/example.env to create a .env file in the backend folder.

  2. Add the Neo4j URI, username, password and database name to the environment file.

  3. Change to the backend directory and create a Python 3.12 virtual environment.

  4. Activate the environment and install the requirements from requirements.txt.

  5. Start the FastAPI application with the documented Uvicorn command.

Expected outcome

The backend is prepared and started for a local or separate deployment.

Guide limitations

  • The documented setup does not by itself describe frontend setup or a complete ingestion run.
  • Neo4j Desktop requires separate frontend and backend deployment; docker-compose is not supported for that setup.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
1 October 2026
Original website title
neo4j-labs/llm-graph-builder · Official homepage title · retained in its source language
How we review sourcesSuggest a correction