It knows your world by name.

Documents are where things get written down. The knowledge graph is the layer underneath that knows the things.

Each person, each project, each decision, each event is held as its own entry — and linked to the others it touches. Not a pile of pages to search through, but a map of who and what is in your life, and how it all joins up.

It is built quietly as you talk. Mention that Maya leads the Meridian project and a connection forms between them. Decide something on Friday and it is filed against the project, dated, ready to recall. You never maintain it — you just keep talking, and the map keeps drawing itself.

Because every person and project is its own entry, asking about one goes straight to it. "What did we decide about pricing?" "What do I know about Maya?" "Where did Meridian get to?" — each lands on the right entry rather than the nearest-looking page.

And it works hand in hand with the way memory finds things by meaning. The map gives the precise answer when you name a thing exactly; meaning-based search catches it when you only half-remember the name. Between them, a fuzzy memory still finds its way home.

The map is not a flat list — it is connected, and it is dated. The same decision links the project it belongs to, the people who made it, and the day it happened. So a question in one corner can pull in what matters from another.

Ask what you were working on in March and it walks the timeline. Ask about a person and their projects, your last exchanges and the threads left open come with them. This is what turns recall into understanding — the connections are the point, not the individual facts.

Every entry is traceable to where it came from, and editable the moment it is wrong. Tell it "that's out of date" and the change shows up in every future answer — not buried somewhere you cannot reach.

And the whole map is yours to take. It exports as a single structured file — your world, in a form you own, trapped inside nothing. The graph is the connective tissue of your memory; cross-tool synthesis is where it pays off most — joining what you know with what your tools know.