On this browser

Voice engine

Vozub picks the best local voice model for this device. You can override that choice below. The engine is separate from your books, reading position, and generated audio.

Currently using

HD Voice

Not loaded yet · 326 MB model

Needs download

Your setting

Automatic

Vozub picked HD Voice because this device can run it smoothly.

Choose a setting

Standard Voice engine

Standard Voice is ~401 MB on WebGPU. HD Voice uses a separate ~326 MB model.

Stage timings: /debug/supertonic-bench

Downloaded on this device
~326 MB
Engine status
Needs download
Persistent storage
Unknown
Last successful load
Not yet loaded

Removing downloaded models frees space in this browser. Vozub will download the selected model again the next time you create narration.