Offline push-to-talk dictation for the whole operating system. Hold a hotkey, speak, release — the transcript lands wherever the cursor is.
Faster-whisper runs locally. No cloud, no network dependency, no billing dashboard watching your words.
Works in any text field — browser, Word, terminal, chat. The transcript is typed where the cursor is.
Sub-second transcription on a modern laptop. GPU-accelerated where available, CPU-fine where not.
Whisper-large-v3 handles ninety-nine languages. Switch per-session or let auto-detect decide.
Your dictated thoughts are yours. No training on your voice, no transcript history on someone else's server.
The straightforward alternative to Wispr Flow, superwhisper and friends. Fork it, audit it, ship it.
| Capability | Dictateur | Wispr Flow | Superwhisper |
|---|---|---|---|
| Runs offline | Yes | Cloud | Yes |
| Free of charge | Free · MIT | Subscription | Paid |
| Open source | Yes | No | No |
| Windows native | Yes | Yes | macOS only |
| System-wide hotkey | Yes | Yes | Yes |
| Custom model | Any whisper.cpp | No | Preset |