Privacy: STT for iPhone and Mac
Applies to STT (Speech to Text) by Herman Harazha: the iPhone app in the App Store and the Mac app from GitHub.
STT collects no personal data and sends nothing to its developer. The developer has no servers; STT has no accounts, no analytics, no ads and no tracking.
Speech is turned into text on the device itself, by a model stored on it. The microphone works only during a recording the person started. Audio and text are not sent anywhere.
Recordings and their texts are kept on the device, in the app's own folder, for as long as the History setting says, and are deleted in the app. On the iPhone they are part of the iPhone's backup, like any app's documents.
The network is used to download the chosen speech model from Hugging Face (huggingface.co); as with any download, that server sees the device's IP address. The Mac app also checks GitHub for updates once a day.
Language, the dictionary and settings such as how long history is kept move between the person's own devices through iCloud key-value storage, in that person's Apple Account. The developer cannot see them.
A text or a recording leaves the app only where the person sends it, through Share or the clipboard.
Questions: geragaraza@gmail.com or the Issues of the STT repository.