graphrag
by microsoft
GraphRAG is a Python toolkit from Microsoft Research that uses a large language model to turn unstructured text into a knowledge graph. That graph is then used to answer questions about private documents more accurately than simple search.
- Stars
- 36.5k
- Licence
- MIT
- Runs on
- Your computer
- Setup
- Technical
- Language
- Python
- Age
- 943 days
What you can do with it
- Index a collection of documents into a knowledge graph using an LLM
- Ask questions that need an overview of a whole dataset, not just single passages
- Tune the prompts to suit your own subject area and data
Make money with it
A freelancer could set up document question-answering systems for small firms and charge for the build and ongoing tuning.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- Build an internal assistant that answers questions across company documents, reports or support archives
- Offer clients a searchable knowledge base built from their contracts, research or meeting notes
Can you use it commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
Install the graphrag package with pip, then use the command line (graphrag init, then indexing and querying) as described in the quickstart.
Follow the official instructions in the project README.
Things to check
- Project is in maintenance mode: no new features or pull requests accepted
- Not an officially supported Microsoft product; described as a demonstration
- Indexing can be expensive because it makes many calls to a paid LLM API
- Results usually need prompt tuning to work well on your own data
- Config format changes between versions and may need re-indexing
Who's talking about it
- Rost (rosgluk): Trending Open Source AI Projects on GitHub, October 2026 Featured in a roundup
- Rost Glukhov: Trending Open Source AI Projects on GitHub, October 2026 Featured in a roundup
Checked 27 Oct 2026. Summary drafted by AI from the project's README.