We speak about three times faster than we type. That is the whole point of WhisperDictation, the voice dictation tool we make available: you speak, and the text appears live, in any application.
What it does, in practice
You press a shortcut (a quick double tap on Ctrl), speak naturally, and the text appears wherever your cursor is — messaging app, document, form, ticketing tool. A second shortcut switches to French. Punctuation, accents and proper names are handled by the Whisper large-v3 engine, one of the best speech recognition models available today. A nice touch: your music pauses automatically while you dictate, then starts again.
Three modes, depending on your priorities
- Cloud — the simplest option, free with Groq, very fast. Ideal to get started.
- Local GPU — everything stays on your computer, nothing leaves your premises. Perfect for sensitive content.
- Remote GPU — a powerful machine on your network does the processing; your lightweight computer just listens and types.
In our tests, the same audio clip goes from about 3.8 seconds of processing on a laptop processor to under 0.3 seconds on a recent graphics card: dictation becomes almost instant.
Designed for non-technical users too
Installation is done through a guided wizard that explains each step: which mode to choose, how to get a free key, how the shortcuts work. Built-in tutorials help beginners along.
Windows, macOS and Linux
The tool runs on all three systems. You can use it as is with Python or build a standalone executable. A native version is available for Linux (KDE/Wayland).
Privacy
In local or remote GPU mode, your audio never leaves your hardware or your network. In cloud mode, the audio is sent to the provider for the duration of the transcription — a mode to avoid for confidential information, and one you are free not to enable.
Who is it for?
For anyone who writes a lot and wants to save time: freelancers, digital craftspeople, support teams, writers. It is a tool we use ourselves every day. Download it from our Free tools section and dictate your first message in under five minutes.