Why context, not faces.
Contexted is an experiment in what happens when the first thing two people share isn't a photograph — it's the memory their AI has built with them.
What we're making
Contexted matches people through shared context. You bring a slice of the memory your assistant has accumulated about you — a ChatGPT or Claude export, an excerpt, a paragraph that sounds like you. We read it for recurring themes, tone, the texture of how you think, and then we discard the raw text. What's left is a quiet portrait that other people can be matched against.
Matches arrive in scheduled drops. Each one surfaces a few shared synergy points and a mutual prompt — a small, slightly vulnerable question — that both sides answer before any chat opens. The chat itself is anonymous until you decide otherwise.
Why this is different
Dating apps optimize for selection. You scan, you swipe, you decline. The interface trains you to be quick and harsh, and to value people for the parts of them that fit on a phone screen.
Contexted is closer to recognition than selection. Two people end up matched because something in how they think overlaps — a shared preoccupation, a similar way of holding a question, a kind of humor that survives translation. There is no scrolling. There is no leaderboard.
Words over faces. Warmth over performance. Context over coincidence.
How AI memory fits in
If you've used ChatGPT or Claude for a while, the assistant has built up something like a sketch of you — what you reread, what you wrestle with, the kinds of metaphors you reach for. That sketch is more honest than a profile, because you weren't writing it to be seen.
We treat that material with care. The raw export never leaves processing. What we keep are the abstractions — the shape of you, not the transcript. You can read what we kept, edit it, and delete it.
What it isn't
Contexted is not a dating app, even when matches turn into something. It's not a friendship app either. It's a small bet that the AI-native generation already lives in language, and that language is a better starting point for connection than appearance.
It's also not for everyone. If you want speed, optionality, and a stack of faces, this won't feel like home. If you've been quietly tired of all of that, it might.
The shape of the experiment
Contexted is invite-only and runs in batched drops. New invites are earned by people already in — a quiet referral loop rather than a viral one. We'd rather grow slowly and stay coherent than grow fast and become a feed.
If that's the kind of thing you want to be early to, you can join the waitlist or read more about how we think.