90+ languages, one hotkey

Offline Speech to Text in Multiple Languages

Kokukeyboard is offline speech to text for people who do not only write in English. Press a hotkey, speak in your own language, press it again, and the text is typed at your cursor in whatever app you are using. Transcription runs on OpenAI's Whisper, locally on your computer, and it recognizes 90+ languages. Every language is included in the one price: there is no language pack to buy, no per-language fee, and nothing is uploaded.

or download the installer directly (.exe)

No card for the trial. 30-day refund if you buy and change your mind.

Dictation in your language, offline

Many dictation tools are built around one language first and treat the rest as add-ons, or only work while connected to a cloud service. Kokukeyboard uses Whisper, a speech recognition model trained on many languages at once. The same model that writes your English email can write your Spanish, German, French, Portuguese, Polish, Japanese or Hindi one, without switching products.

By default, the language of each dictation is detected automatically. Speak Dutch to a colleague in one window and English to a client in the next, and each comes out in the language you spoke. If you always dictate in one language, you can pin it in Settings so the app does not need to guess.

Because Whisper runs on your own CPU or GPU, dictation in your language works offline. Once the model is downloaded during setup, no internet connection is needed to transcribe. That is useful on a flight, in a building with poor reception, or simply when you would rather your words did not travel to a server in another country.

Works in every app you write in

Kokukeyboard sits in the system tray and listens for a global hotkey - Ctrl+Shift+D by default, changeable in Settings. When you press it again, the transcription is typed into whichever window has focus. Email, documents, chat, browser forms, a translation tool you are checking your work in: if the window accepts typed text, your dictated words land there.

It also helps when your keyboard layout does not match the language you are writing. Accented letters, special characters and other scripts are typed for you as part of the transcription, so there is no need to switch keyboard layouts or hunt for dead keys mid-sentence.

The app transcribes once, after you finish speaking, instead of showing guesses word by word. You say the whole thought in your own rhythm and get the full sentence typed in one pass.

No per-language fee, no subscription

Kokukeyboard is $14.99 once for two computers, and that price covers every language Whisper recognizes. There is no add-on for a second language and no monthly plan that grows with how much you dictate. Because the recognition happens on your own hardware, there is no per-minute server cost to pass on. The dictation app runs on Windows and on macOS (see offline dictation for Mac).

The 14-day free trial includes every language and every feature, with no card required. Try it in the languages you actually write in, on the work you actually do. If you buy and it is not right for you, there is a 30-day refund.

Recordings in other languages too

If you have interviews, lectures or voice notes already recorded in several languages, the Pro tier transcribes whole folders of audio files offline and detects each file's language on its own. See offline AI transcription for whole folders of audio.

For a walk-through of everyday dictation, read offline speech to text for Windows. If you are comparing pricing models, read why Kokukeyboard is a one-time payment. Everything else is on the homepage.

or download the installer directly (.exe)