Live lecture translator — English → 中文. Everything runs locally in this browser; no audio leaves your machine.

1) Pick an ASR model and source above. 2) Click Start and allow the microphone. 3) Speak English — Chinese captions appear below. Use Full screen to project.

First run downloads the models once (≈150 MB for base.en), then they are cached. Use Chrome or Edge for WebGPU speed. Speak into a good microphone for best accuracy.