A custom iOS keyboard that transcribes your voice entirely on-device — no cloud, no account, no waiting. Speak. Stop. Done.
Unlike other dictation tools, SkipKey never sends your audio to a server. Speech recognition runs locally — privately, reliably, even offline. Optional AI post-processing uses your own cloud provider, if you choose to enable it.
The average person types ~40 words per minute but speaks at ~200 wpm.
Voice dictation lets you get your thoughts out up to 5× faster — with less effort.
Your audio is transcribed on your iPhone using a bundled AI model. Nothing is sent to a server. No account required.
SkipKey is a real iOS keyboard. Switch to it in Messages, Notes, Mail — anywhere you can type on your iPhone.
Tap Stop Recording. The transcript is processed on-device and placed into your text field — no copy-paste required.
Speak naturally in English, German, French, Spanish, Italian, and 21 more. Language is detected from your voice automatically.
* Audio transcription is always local. If you enable AI post-processing, transcript text — not your audio — is sent to your chosen third-party LLM provider. This feature is off by default.
Settings → General → Keyboard → Keyboards → Add New Keyboard → SkipKey. Enable Allow Full Access.
Tap the globe icon on your keyboard to switch. SkipKey appears alongside your regular keyboard.
SkipKey opens briefly to start recording. Return to your app and speak naturally in any language.
Tap Stop Recording. Your words are transcribed on-device and placed into your text field.
Optionally connect your preferred LLM provider and define custom modes that transform every transcript the moment you stop recording. Name the mode, write a one-line instruction — SkipKey applies it hands-free.
"Hey auntie, just wanted to say I'll be visiting in July — super excited to see you and the whole family!"
↓"Hallo Tante, ich wollte nur kurz sagen, dass ich im Juli vorbeikomme — ich freue mich so sehr, dich und die ganze Familie zu sehen!"
"yeah so we need to push the deadline back like a week or two, the API stuff just isn't done yet"
↓"We need to extend the deadline by one to two weeks, as the API integration is not yet complete."
Set up any mode in the app — give it a name, write a prompt, and the button appears in the keyboard. Switch modes with a single tap between recordings.
SkipKey detects your language from your voice — no selector, no setting, no switching.