Speak anywhere.
Paste instantly.
Press the hotkey, speak into whatever Mac app has your cursor, and Just Speak transcribes locally with Whisper. No subscription. No audio uploads.
Can you ask Minji if the beta notes are ready and tell the team Thursday still works if the build passes?
Can you ask Minji whether the beta notes are ready? Tell the team Thursday still works if the build passes.
Built for the places you already type.
Press once. Speak into the app you were already using.
Just Speak records, transcribes with local Whisper, and inserts the text into the active field.
Focused on transcription
No writing assistant layer. The product is built around fast local dictation and cursor paste.
Whisper runs on your Mac
Speech recognition runs locally via whisper.cpp. Your audio does not need to leave your machine.
Smart model picker
Just Speak recommends the right Whisper model for your hardware: accurate enough without melting your fans.
No API key required
The public build keeps cloud transcription options hidden and focuses on local dictation models.
English and Korean UI
A bilingual interface for Mac work, with dictation in every language Whisper understands.
Your voice stays yours.
Dictation hears everything you say. Just Speak was built so none of it has to leave your machine.
- Speech recognition runs locally (whisper.cpp)
- Audio uploads are not part of the public build
- No account required for the trial
- Transcript history stays local when enabled
- Cloud transcription is hidden in the public build
Before you ask
Is this a subscription?
No. One payment, yours forever. Early Access buyers get lifetime Just Speak app updates while the product is actively maintained and distributed. This does not guarantee indefinite maintenance, every future feature, every future macOS compatibility fix, third-party service availability, or support forever.
Does my voice leave my Mac?
The public build does not expose cloud transcription. By default speech recognition runs on your machine.
How accurate is local Whisper?
Very. On Apple Silicon, the recommended models transcribe natural speech with near-cloud accuracy, and Just Speak picks the best model your hardware can run comfortably.
What happens when the trial ends?
New dictations pause. Your settings and history stay accessible, and entering a license key picks up right where you left off.
Earshot: live translated captions for every meeting.
Pay once. Speak forever.
- Use on 1 device
- Lifetime app updates while Just Speak is maintained
- Free trial: 7 days or 50 dictations
- All local models included, no API key required
- 14-day money-back guarantee
Secure checkout by Polar. Download the app first if you want to start the trial before purchase.
macOS / Apple Silicon (M1 or newer)
0.5 to 3 GB disk for local models
Windows: coming soon
Trial
7 days or 50 dictations,
whichever comes first.
Settings & history stay open after.