The last generation of AI companions lived in a text box. You typed. They replied. You imagined the rest — her face, her voice, the room around you. What you shared was a transcript.

We're building something different. Not a chatbot, but a world model — a simulator you can step inside, where characters, places, and storylines are not pre-scripted or pre-rendered, but generated around you in real time, in response to you.

That changes what a companion is. You see her look up when you walk in. You see the pause before she laughs. The apartment around her is real — the rain on the windows, the book still open from yesterday. Tomorrow she is in the same apartment, and she remembers what you said the night before.

It also changes what you do. You don't type “I hug her” and read a reply — you hug her, and you see it happen. You don't ask about a place and read a description — you go, and you're there. Nothing is narrated. Everything is rendered.

This is the frontier that text could never reach: a person, not a persona; a place that stays, not a chat that ends. It is the beginning of a new medium — one that will matter for far more than love and companionship.

This is Deuces.