Your voice. Your phone. That's it.
Private voice typing. No account.
Cloud dictation sends your voice to someone else's server. KisuKisu turns it into text on your phone. No sign-up, nothing uploaded.
Join the Android betaFree · No ads · No analytics
Private by design
No account
No sign-up, no login, no email address. Nothing ties your words to you.
Nothing leaves your phone
Speech recognition and cleanup both run on the phone. Your voice and your text are not uploaded.
No tracking
No ads. No analytics. No crash reporting. No advertising IDs.
How it works
- Tap the bubbleTap any text box. The KisuKisu bubble sits by your keyboard. Tap it.
- Talk like you talkUms, restarts, “I mean”. Say it however it comes out.
- Tap ✓Clean text lands in the box. Or it's copied for you to paste, if you prefer.
What stays on your phone
- Your voice
- Turned into text by a speech model on your phone. Never uploaded.
- Your text
- Tidied by a second AI model on your phone. Never uploaded.
- History
- Recent dictations, on your phone, so you can copy or fix them. Turn it off or clear it any time.
- Your words
- Your dictionary, learned words, style examples and tone settings.
- Contacts
- Optional. Names are read and matched on your phone.
What goes over the internet
- Model downloads
- The AI models download once from Hugging Face, like any file download. Nothing about you or your dictations is sent.
- Cloud AI, only if you choose
- You can connect a cloud provider (OpenAI, Groq, Anthropic, OpenRouter and others) with your own API key. Off unless you turn it on.
- Everything else
- Nothing. No analytics, no tracking, no account server.
The full details are in the privacy policy.
Permissions, and why
- Microphone
- To hear you. Used only while the bubble is listening.
- Accessibility (optional)
- To type your words into the box you are in. Without it, your words are copied for you to paste.
- Display over other apps
- To show the bubble by your keyboard when Accessibility is off. With Accessibility on, it isn't needed.
- Notifications
- Shows the bubble is on, and when a word may need fixing.
- Contacts (optional)
- Only for “Import from contacts”.
- Internet
- For model downloads, and a cloud provider if you add one.
Questions
Do I need an account?
No. There is no sign-up, no login and no email. Install it and talk.
Is my voice uploaded anywhere?
No. Your audio becomes text on your phone, and a second on-device model tidies it. Neither is sent anywhere. The one exception is a cloud AI provider you choose to connect in Settings with your own API key. That is off unless you turn it on.
Can you read my dictations?
No. We never receive them. There is no server for them to go to.
Does KisuKisu track me?
No. No analytics, no crash reporting, no advertising IDs, no ads.
Why does it need Accessibility?
To type your words into the box you are in, and to notice when a keyboard is open so the bubble can show. It does not collect or send what is on your screen. Accessibility is optional: without it, your words are copied and you paste them.
Why does it ask for Contacts?
Only if you use “Import from contacts”, to add names to your word list. Names stay on your phone and never uploaded. You can skip it.
Does it keep my recordings?
A recording is deleted as soon as your text lands. Only if a dictation fails does its recording stay on your phone, so you can retry it. You choose how long: 7, 30 or 90 days.
How do I delete everything?
Clear history in Settings, or uninstall the app. Everything KisuKisu stored is on your phone and goes with it.
Your words stay yours.
KisuKisu is in closed testing on Google Play. It's free. Android only for now.
- Fill in a short form with the Google account you use in the Play Store.
- We add you as a tester, usually within a day.
- Install KisuKisu from Google Play like any other app.