Privacy

Last updated 16 September 2026

Your voice stays on the Mac

Speech is recognized locally by the model you downloaded. The app does not send audio or transcripts to us, and there is no account to sign in to.

Downloading a model

When you install a model, the app connects to the host of those files, including Hugging Face. That host receives ordinary network data such as your IP address. Once the model is installed, dictation no longer needs the network.

Data on your device

Settings, your personal dictionary and downloaded models are stored on your Mac. A recording exists only while recognition runs and is deleted afterwards; after a crash a temporary file may remain.

The recognized text is placed in the clipboard, which other apps can read according to your macOS settings. The app keeps no transcript history.

Permissions

The microphone is switched on by your action. Accessibility is used to check the field you started in and to paste the text there. Password and other secure fields are excluded, and Enter is never pressed for you.

This website

No email, password or payment details are needed to download the app. The site sets no cookies of its own and includes no advertising or analytics scripts. Our hosting provider processes ordinary technical connection data, such as IP address and user agent, to deliver the page and protect it from abuse.