Press the hotkey
Press Ctrl+Shift+Space in any Windows app, and Voisee starts recording. You can change the shortcut in settings; Esc cancels the recording.
Dictation for WindowsTurn your voice
Turn your voice
into text
Voisee recognises Russian speech at the press of a hotkey. Press, speak, and the text is already on your clipboard, with periods, commas and capital letters.
RU
Russian speech
3.3 s
106 minutes of audio*
60 min
Every week at no cost
* 106 minutes of audio are recognised in 3.3 seconds. That is the model's processing time on a server with a single GPU, not counting upload and queueing.
Dictation · Ctrl+Shift+Space
How Voisee works
From keypress to text on your clipboard in three steps.
The model recognises your speech
The recording goes to our GPU server. The model adds punctuation and capital letters, and the replacement glossary fixes terms.
The text lands on your clipboard
Paste it with Ctrl+V into an AI chat, an editor or a messenger. Everything you dictate stays in the app history.
What Voisee does
Dictation for Windows and Russian speech recognition, powered by one model on our own GPU server.
Russian speech
For now, Voisee recognises Russian speech only. Numbers come out as digits, and common English words like YouTube and HD in Latin script.
Punctuation and capitals
The model adds commas, periods, question marks and capital letters itself, so the text arrives formatted rather than as one long line.
Replacement glossary
Does the model mix up a rare term? Add a replacement such as “GVT” → “JWT”. A shared glossary with common abbreviations is already on; your personal one lives in your account.
Hotkey dictation
The Windows app lives in the system tray. Ctrl+Shift+Space starts recording, and a second press puts the text on your clipboard. You can also hold the keys while you speak.
History and sync
Everything you dictate is saved in the app history.
How the replacement glossary fixes terms
Real model output on an excerpt from a webinar about 1C. The model heard “GVT” and “HDP”, and two glossary entries turn them into JWT and HTTP.
What the model recognised
В этом уроке мы рассмотрим, как настроить GVT-аутентификацию для HDP-сервиса в 1С. Фирма 1С очень своеобразно реализовала поддержку GVT-токенов.Replacement glossary
- GVT → JWTshared glossary
- HDP → HTTPpersonal glossary
Result · TXT
Click “Apply glossary” and the text after replacements will appear here.
Russian only, for now
Right now Voisee recognises Russian speech only. It writes numbers as digits and common English terms in Latin script, and the replacement glossary fixes whatever the model gets wrong. Other languages are not supported yet.
HDIPYouTubeJWTJSONURL1С902ещёвсё+ your replacement glossary
Frequently asked questions
Still have questions? Write to admin@nizamov.school.
How well does Voisee recognise speech?
It works best with calm speech in a quiet room: the text comes out with commas, periods and capital letters. Fast conversational speech and background noise lead to more mistakes. Terms and names the model gets wrong are fixed by the replacement glossary.
Which languages are supported?
Only Russian for now. Common English words inside Russian speech, such as IP, YouTube and HD, come out in Latin script, but the model handles whole English sentences poorly, so we do not promise English recognition.
What is the replacement glossary?
A list of “what the model writes → what it should be”, for example GVT → JWT. Replacements apply to every new transcript. A shared glossary with common abbreviations is already on; you keep your personal one in your account, on the glossary page.
How fast does the text arrive?
106 minutes of audio are recognised in 3.3 seconds. That is the model's processing time on a server with a single GPU, not counting upload and queueing. On top of that comes sending the recording to the server and waiting in the task queue, so the time until the text reaches your clipboard also depends on the recording length and your connection speed.
Sign up and get 60 min of dictation every week at no cost
No bank card required.