YB AI

A free AI in your browser. Sign in with your email, and chat straight away.

Open YB AI — sign in with your email, free

What it is

YB AI in your browser: chat, translation, goals, memory, Gmail and settings in one app. You sign in with your email — a 6-digit code is sent to your inbox — and chat answers instantly from the server brain. The brain is one intelligence, one identity: Younesse.

Your chat goes to the server brain (Cloudflare Workers AI), which picks one of two model families per task: Llama for general chat, Qwen for code. Your words are sent to that server to be answered — they do not stay only on your device. Chats, goals, memory and translation history are stored in this browser, on your device.

The on-device brain is the upgrade, not the gate: three Qwen models (Qwen3 4B recommended, Qwen3.5 2B, Qwen3 0.6B) can be downloaded into your browser — about 0.5–2.4 GB, free, once — and then run on your device with no server involved. There is no router in this browser app; you pick one of the three models.

Status — 17 November 2026

Every claim on this page traces to a row in the public capability registry (audit/capability-registry.md, 20 capabilities). A claim with no row is not made here. What is not yet proven is listed as not yet proven.

How it is built

Spec v3.19 (442 pages) is the contract. Every phase is tested before it ships, and a capability is only claimed after it survives a live test against the machine as it is. The code lives in a private repository; the Founder governs releases.

Last updated 17 November 2026. This page is generated from the project's own record, so it changes when the record changes.