llame
Private AI chats in your browser
On-device ONNX via WebGPU
Private AI chats in your browser
Run local AI models in your browser. No installs, no API keys, and your chats stay on your device.
Ready state
Default model
Qwen3.5 0.8B
Open the full chat workspace only when you need it. This keeps the landing page fast on mobile while leaving the full local inference stack one tap away.
Private by default
Prompts and responses stay on your device.
WebGPU required
GPU-accelerated inference in supported browsers.
Zero backend
No Python. No CUDA. No server roundtrip.