Knowledge Graph

As Trace indexes the pages you visit, it also extracts entities (people, companies, products, concepts, whatever the page is actually about) and the relationships between them. Over time, this builds a graph you can browse rather than a flat list of pages.

The Knowledge Graph Builder background agent handles this extraction automatically as you browse, so there's nothing to trigger manually.

Why it's useful

A search finds you a page. The knowledge graph shows you how that page's subject connects to everything else you've read: the same company mentioned across five articles, a concept that keeps recurring in your research, a person referenced from multiple angles. It's a way to see the shape of a topic, not just retrieve one document about it.

This is a separate graph from the personal memory graph described in Personal memory, which is specifically about facts extracted about *you*. The Knowledge Graph is about the content you've read.

Next steps

• Memory & search for searching pages directly by meaning.
• Personal memory for the graph of facts about you specifically.
• Notes & highlights for saving a specific passage as its own graph node.