VibeSuperTonic

Select text in any application and press a key: VibeSuperTonic reads it
aloud in a neural voice. Speech is generated on your own machine, so
nothing you read is sent anywhere.

It is also a Speech Dispatcher voice, so Orca and other programs that
speak on Linux can use the same voices.

- Two engines: Supertonic, and Piper with 65 voices in 35 languages
- Global hotkeys: Ctrl+` reads the selection, Ctrl+~ stops
- A window that follows along as it reads, with speed, voice and pronunciation controls
- Works on X11, and on Wayland with KDE Plasma. On GNOME's Wayland session the
hotkey cannot read the selection, because GNOME does not offer the protocol
that allows it; the screen reader voices work everywhere
- Optional GPU acceleration, falling back to the CPU on its own

The voice models are downloaded on first run, from Hugging Face, after you
accept their licence. Connecting the hotkeys and the screen reader is one
command in a terminal, which the application shows you.