Manifesto
Your data should work for you.
For years we've poured ourselves into other people's platforms. Every connection, message, post, hire, and deal taught those companies who we are. They built a model of us — and kept it.
You can get it back. Not as a favour — the law compels it. Portability rules oblige every platform that matters to hand you a machine-readable copy, free, on request. But an export is a tombstone: thousands of rows you'll never read, in a zip you'll never open twice.
That's backwards. The most valuable dataset about your life — or your company — should belong to you and be usable by you. Not locked in a feed, but structured, connected, and handed to an AI that works on your behalf.
Here is what changed. The models became a commodity. Everyone rents the same ones, they get cheaper every quarter, and every competitor gets the same upgrade on the same morning. Renting intelligence is now the easy part. The part nobody else can buy is the record of what you did, who you know, and what you decided.
So we built the layer that turns that record into something an agent can use: a local-first ingestion and identity engine that reads your exports and returns a connected brain — for a person or a company — on your machine, in plain Markdown, with the rules enforced in source you can read.
We think this ends somewhere specific. Not one system that knows everyone. Millions of private brains — one per person, one per company — interoperable because they are plain files, and owned because they were built that way from the first line. Ownership can't be added later. Once your history lives in someone else's database, you are renting access to your own past.
Models depreciate. Your graph compounds. Every release from every lab makes the brain you already own worth more, and you rebuild nothing to collect it.
Your data already exists. It's already yours. It's time it started working for you.