Matteo

Every "second brain" tool promises AI that actually knows your notes, then ships something that can't be trusted to read them. Vague "related to your other notes" suggestions and silent processing nobody trusts are the norm. The real need: a companion that only speaks up when it has something grounded to say, and never touches what you wrote without asking first.

Every capture becomes a living note that keeps working long after it's filed. Every night, while I sleep, Matteo quietly re-reads what I kept, follows the connections between ideas, and surfaces the ones I'd never think to look for myself, always as a suggestion I approve, never a silent edit. A multi-agent AI system with its own private, searchable knowledge graph underneath.

In daily personal use as my own second brain. Every capture - a voice memo, a clipped article, a half-formed idea - becomes a note; every night the Dream cycle finds what connects them. The chat surface, the connection graph, and the accept/dismiss review queue are all live and in active iteration.