Intelligence, on your terms.

Meet an open, on-device AI workspace. Run Liquid AI's 1.7B parameter model with your own hardware. Your conversations become optional, local retrieval memory—not a server-side inference bill.

✦ LFM3 1.7B◈ WebGPU + WASM⌘ Local vector memory

Responses are generated on your computer · Local memory is not model fine-tuning · v0.1.0