# What is Graph RAG?

> Graph Retrieval-Augmented Generation

It is a method where artificial intelligence works by looking not only at documents but also at the networks of connections between data when generating answers.

## Overview
While standard RAG methods usually perform searches on text chunks, Graph RAG stores data as a knowledge graph. This allows artificial intelligence to bridge information that may seem distant but is actually related. It enables the generation of deeper and more contextual answers.

*Analogy: It is like mapping the entire historical development of a subject by following references between books instead of reading them randomly in a library.*

## How it works
Data is first converted into a knowledge graph in the form of nodes and relationships. When a query arrives, the system retrieves the most relevant information by looking not just at the text, but at the connection paths in this graph.

## Where it is used
It is used in the analysis of complex corporate data, medical diagnostic systems, and academic research.

## Commonly confused with
It can be confused with classic RAG; the classic one focuses on text similarity, while this one focuses on structural relationships.

## Frequently asked questions
**When should it be preferred?**
If you have complex hierarchies or multiple connections between your data, Graph RAG provides much better results.


## Related terms
- [RAG](/en/dictionary/rag/)
- [Knowledge Graph](/en/dictionary/knowledge-graph/)
- [Vector Database](/en/dictionary/vector-database/)

## Related tools
- [GitNexus](/en/discover/gitnexus/)
- [Code Graph RAG](/en/discover/code-graph-rag/)

---
Source: TreScout Dictionary · https://trescout.com/en/dictionary/graph-rag/