graphrag

A modular Retrieval-Augmented Generation system from Microsoft that uses graphs to build context for question answering. The README describes it as a research project.

Share on XLicense: MIT

Overview

GraphRAG is a modular graph-based Retrieval-Augmented Generation system from Microsoft. It is a data pipeline and transformation suite that uses LLMs to pull structured data out of unstructured text, then uses graphs to build a focused context for question answering. The README calls it a research project. It is largely in maintenance mode, with no new features or pull requests, only bug fixes and dependency updates.

Key features

  • Extracts structured data from unstructured text using LLMs
  • Builds graphs to form targeted context for question answering
  • Modular pipeline and transformation suite
  • Documentation and an arXiv paper are available

Best for

Researchers and developers who want to study how graphs can improve LLM answers over private text. Expect no new features, since the project is in maintenance mode.

Upstream
microsoft/graphrag
Fork on GitHub
Guo-astro/graphrag
Upstream stars
36k
Category
AI agents and LLM tools
License
MIT
Forked
2025-07-15
Sync status
Behind upstream