Opus Vela Voice · Early access
Speak anywhere. Keep it private.
Private system-wide dictation for Windows. Hold a shortcut, speak, and return the words to the application you were already using.
What works today
The current development build uses local Whisper recognition, system-wide hotkeys, selectable microphones, local history, language settings, styles, and transform recipes. Recognition stays on the computer by default.
Release boundary
Voice remains early access. The currently evidenced prerelease is unsigned, automatic updates are not yet built, and clean-machine upgrade testing under the Opus Vela name is still pending.
Privacy boundary
Normal dictation audio is not silently written to disk or uploaded. Transcript content is excluded from analytics, and secure or elevated targets are refused when the application cannot establish a safe insertion path.
Licence
The Voice codebase remains MIT-licensed. Historical Toolshape releases and repository links stay available while the display-name migration is completed safely.