Input
Local Settings
Model
Voice
Device
Precision
Speed
The first load downloads the model and voice assets (large). After that, the browser cache will be used.
This is fully local speech synthesis. It does not call any external API.