Skip to content

GraphRAG

Enrich text datasets with graphs and retrieval augmented generation

Visit official site

What GraphRAG is

The technique is described as Graphs plus Retrieval Augmented Generation, combining text extraction, network analysis and LLM prompting and summarisation to enrich understanding of text datasets. It is part of a Microsoft research project and an end-to-end system from data input to result output.

What you can do with it

  • Extract useful information from large amounts of data
  • Analyse data structurally and build a knowledge graph
  • Use LLM prompts to strengthen text understanding and generation
  • Run a complete flow from data input to result output

Who it is for

  • Researchers analyse complex text data
  • Data scientists explore links between data through knowledge graphs
  • Developers integrate the system into their own applications

Getting started

  • Visit the GitHub page to learn the project background
  • Read the preprint paper for the technical details
  • Follow the documentation to download and install the open source implementation
  • Prepare the text dataset to be processed

Pros & cons

✓ What we like

  • The preprint of the research paper is available to read
  • A GitHub open source implementation is stated as coming soon

! What to watch out for

  • The open source release is described as upcoming rather than available
  • The technique is tied to LLM prompting and summarisation

FAQ

What does GraphRAG combine?

The material names text extraction, network analysis, and LLM prompting and summarisation.

How can it be tried?

The documented steps cover the GitHub page, the preprint, installing the open source implementation and preparing a text dataset.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review