OnVoice keeps dictation local
OnVoice is a macOS dictation app. Audio and transcripts are processed on your Mac by default. OnVoice does not operate a cloud transcription service and does not sell, rent, or use your content for advertising or tracking.
Audio and transcripts
Microphone audio is captured only while you use dictation. The app uses it to produce the transcript you requested. Transcripts are kept on your Mac for delivery, history, or pending results, according to your settings. OnVoice does not upload them to an OnVoice server.
Optional AI providers
AI text cleanup is optional. Built-in Local uses a model running on your Mac. If you choose an Ollama or other external API provider, the selected text is sent to that provider and handled under its privacy policy. You can turn this feature off and keep the raw transcript.
Models and permissions
Downloaded speech and language models are stored in OnVoice's app-managed support folder. Microphone permission enables recording. Accessibility permission enables safe insertion into the focused text field. These permissions can be revoked in macOS System Settings.
Changes and contact
This policy may be updated as OnVoice changes. The date above identifies the latest version. For questions, please open an issue in the OnVoice project repository or use the support contact published with the app.