Private AI Chat

A fully capable AI that runs inside your browser. No servers, no accounts, no data collection — ever.

🔒 100% On-Device 🚫 Zero Data Sent 📶 Works Offline 🆓 Always Free

⚠️ WebGPU not available. Your browser doesn't support WebGPU, which is required to run AI models locally. Please use Chrome 113+ or Edge 113+ on desktop for the best experience.

Choose a Model

SmolLM 1.7B ~1 GB · Fast Llama 3.2 1B ~0.9 GB · Fastest Llama 3.2 3B ~2 GB · Smarter Phi-3.5 Mini ~2.4 GB · Best quality

Initializing…

🤖

Your private AI is ready. Ask me anything — all processing happens on your device.

How it works

The AI model downloads to your browser cache (~1–2 GB, one time only) — subsequent visits are instant.
All inference runs on your GPU or CPU using WebGPU — nothing is transmitted to any server.
Closing the tab clears all conversation history from memory.
No account, no API key, no tracking, no subscription — completely free forever.