Appearance under control
Window appearance, shortcut, prompt, API — make the app your own.
Save hours of typing every week. Think - speak - combine.
Dictation on the left, keyboard on the right. We speak about 3× faster than we type.
To: team@company.com
Hi there, I'm sending over the materials for tomorrow's meeting.
To: team@company.com
Hi there, I'm sending over the materials for tomorrow's meeting.
No window switching, no copy-pasting. Three steps — and you're typing with your voice.
Set your own global shortcut — just hold it.
The equalizer shows you it's listening. Just speak naturally.
Let go and your text lands instantly, right where your cursor was.
A powerful tool you don't see until you need it.
Window appearance, shortcut, prompt, API — make the app your own.
While you speak, an indicator with an equalizer pops up — and disappears once your text is inserted.
from speech to finished text — depending on length
The system is always ready in the background.
Lives in the tray icon — no windows, no visual noise. And it uses just ~90 MB of RAM.
You enter your own API key and it's stored only locally with you.
Works in every app
Dictation runs on your own Groq API key — free and no strings attached. No credit card or phone number needed.
Set up in a minute
Even with heavy dictation. Two Groq models run under the hood:
On by default — for more accurate, higher-quality results.
Go over and there are no charges or commitments — the limit simply resets.
The current version, fresh features and what's in the works right now.
It works in every Windows application — browser, email, Slack, VS Code, Word, notes. Wherever you place the cursor, the text will appear there.
Slovak, Ukrainian, and English — all processed with the same accuracy and natural quality. No compromises, no favorites.
Because it is a very effective alternative to traditional STT solutions and LLM-based postprocessing. We also take into account that not every user has a powerful enough device to run large LLM models locally. Nevertheless, we are actively working on local processing and optimization for different types of devices. The current solution provides significantly higher quality and faster processing without unnecessarily using your device's resources. Modern LLM models for text processing are very powerful, and thanks to the optimized infrastructure on Groq servers, they can deliver fast and high-quality results.
Exclusively on your device. It never passes through our servers — we cannot see it and we never send it anywhere. Control remains entirely with you.
Not at all. The text is inserted exactly where your cursor was. If something goes wrong or you forget to select a field, the text remains in your clipboard — just press Ctrl+V.
No. The application uses around 90 MB of RAM, which has no noticeable impact on performance or FPS. It runs quietly in the background and does not interfere with your work.
Yes. The recording goes directly from your device to the Groq API — we never see or store it. Your voice, your data.
Yes. An internet connection is required for voice processing, but the application itself runs locally.
After transcription, AI automatically corrects grammar, punctuation, and text structure, so the result sounds natural and clean without additional editing.
Register for free on Groq, create an API key, and add it to the application. The entire process takes approximately 2 minutes.