VOCO

Dictation is the first step. VOCO is building the voice layer for Linux: private, local, and under your control. Today it's private dictation: press a shortcut, speak, and VOCO pastes your words into the app you are typing in, while you speak. Speech is recognized on your computer, with no account, subscription or cloud service. VOCO listens only when you start it and never presses Enter.

VOCO stays in the system tray, or in the GNOME top bar with live microphone bars. Pasting depends on your desktop session and the app you type in.

- Local English speech recognition with NVIDIA Nemotron
- Words appear while you speak, in whichever app has keyboard focus
- Words VOCO cannot paste are copied to the clipboard when you stop, and Review recovers dictation after a crash
- The Snap Store keeps VOCO up to date

The snap types in GNOME on Wayland, through the VOCO panel from extensions.gnome.org, and in X11 sessions. On other Wayland desktops, use VOCO's .deb or RPM from its GitHub releases.

VOCO is MIT-licensed, and its speech runtime Apache-2.0. The speech model, NVIDIA's Nemotron, comes under the NVIDIA Open Model License, and the snap carries every license and notice.