A knowledge graph builder uses a language model to read documents, pick out the people, places and things in them, and link them into a graph you can query.
In plain words
A knowledge graph stores facts as nodes, such as a person or a company, joined by named relationships, such as works for. The builder sends the text to a language model, which extracts the entities and how they relate, and saves them to a graph database such as Neo4j. You can then ask questions that follow the links, or use the graph to give a chatbot better context, as in GraphRAG.
A quick example
Imagine loading a set of news articles about a tech company. The builder creates nodes for the company, its founders and its products, and relationships such as founded and makes, so you can ask which people are linked to which products.
Where you'll meet it
- Turning PDFs, web pages and videos into a graph you can query.
- GraphRAG systems that answer questions over a whole collection.
Related
Sources
- Docsllm-graph-builder, Neo4j Labs (GitHub) · read 29 Sept 2026
- DocsWhat is a graph database, Neo4j docs · read 29 Sept 2026
- Docsgraphrag, Microsoft (GitHub) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.