OiPer Desktop
Voice-to-text that runs on your machine and types into any application.
OiPer Desktop turns speech into text without leaving the application you are working in. You hold a global hotkey, speak, and release. The recording is transcribed and the result is inserted at your cursor.
How a dictation works
- Hold your hotkey. Recording starts and an overlay shows the current state.
- Speak. Audio is captured from your selected input device.
- Release the hotkey. Recording stops and transcription begins.
- Transcription runs. Depending on the active profile this happens locally on your machine, through a cloud provider you configured, or through OiPer's own service.
- Optional formatting. A language model can clean up, rewrite, or translate the transcript before it is inserted.
- The text is inserted into the application that had focus, and the run is saved to your local history.
Local by default
A profile in local mode does everything on your machine using a Whisper model you
downloaded. Audio, transcripts, history, and configuration stay in OiPer's application
data directory. Nothing is sent anywhere.
Cloud and OiPer modes exist for when you want a hosted model instead, and they only send data while a request is running. See Security & Data Handling for exactly what is stored and what leaves the device.
Where to go next
- Installation — download the app and grant permissions
- Quickstart — get through first-run setup
- Profiles — hotkeys, languages, and transcription modes
- Models — local Whisper models, cloud providers, OiPer models