In the apps
you already use
Reply to messages, capture ideas, and switch back to your usual keyboard whenever you want.
A message, a note, or a longer story — just say what you want to write. Voice Keyboard turns your speech into text.
For messages, notes,
and ideas on the go.
the microphone
stop manually or let auto-stop finish the recording
the transcribed text
Animated illustration based on the app interface. Example shown in English, without sound.
Type less.
Say more.
Enable auto-stop with the timer button below Backspace or in the keyboard settings. When you pause, recording stops automatically. You can adjust the pause length or stop the recording manually at any time.
Auto-stop is off by default. Speech detection runs on your device and accounts for steady background noise; other voices and changing noise can affect it.
Use your voice where it suits you, refine your text when you need to, and return to your usual keyboard whenever you like.
Reply to messages, capture ideas, and switch back to your usual keyboard whenever you want.
Set up optional text processing to fix errors, shorten messages, translate them, or add emoji.
If your speech service is unreachable, the recording is saved on your device and retried when the service becomes available.
You will need an Android phone and a speech recognition service. You choose the service.
Download Voice Keyboard from F-Droid or GitHub and enable it in your phone’s settings.
Enter an API key from a service such as Groq or Mistral, or connect your own compatible server.
Open a text field, switch to Voice Keyboard, and tap the microphone.
An API key is your personal access key to a service. Terms and pricing depend on the provider; some offer free tiers.
Setup guide ↗You choose your speech recognition service.
Audio recordings are sent to the speech recognition service you choose. When you use text processing, the relevant text is sent to your chosen text provider.
You can use your own compatible server. The app has no ads or analytics, and its source code is open.
Privacy policy ↗What to know before you install.
Usually, yes: speech recognition is performed by your chosen service. You can also connect a compatible server on your local network. If the service is unreachable, the recording is saved and retried when it becomes available.
The app is free to download. Speech recognition requires a service, and its terms and potential costs depend on the provider. Some offer free tiers.
Yes. Use Voice Keyboard for dictation and switch back to your usual keyboard. How you switch depends on your keyboard and Android settings.
The app interface is available in 17 languages. Dictation languages depend on your chosen service and model; you can set up multiple languages for quick switching.
No, it is optional. To use the app’s text correction, shortening, or translation features, you will need to set up a text processing provider.
Install Voice Keyboard and set up a speech recognition service.