Corpospeak Privacy Policy

Last updated: 11 September 2026

Corpospeak does not collect, store, transmit, or share any personal data. There is no account, no sign-in, no sync and no server of mine. The app makes no network connections at all, so nothing you say reaches me or anyone else.

What the app does with your microphone

Corpospeak listens through the microphone so it can turn what you say into corporate English and read it back to you. Your device does all of it: Apple’s Speech framework writes down what you said, Apple Intelligence rewrites it, and one of the voices below speaks the result. Corpospeak holds the audio and the transcript in memory while it is working on them and never writes either to disk. The one thing it saves is the identifier of the voice you picked, so it uses the same one next time.

Corpospeak only ever recognizes speech on your device. If your device cannot recognize your language by itself, the app tells you so and does not listen. It never falls back to Apple’s servers.

Voices

Corpospeak speaks with the Personal Voice you record in Accessibility → Personal Voice, in the Settings app on an iPhone or iPad and in System Settings on a Mac, and only after you allow it. The voice stays on your device and under the system’s control. Take Corpospeak’s access back whenever you want, in the same place.

Until you have a Personal Voice, Corpospeak speaks with Kokoro, an open-weights speech model that ships inside the app and runs on your device. It is never downloaded and it sends nothing anywhere. Apple’s built-in voices are in the same menu, and you can pick any of them instead.

Network

Corpospeak makes no network connections. It has no analytics, no advertising, no accounts, and no third-party services. The Kokoro speech model and the open-source library that loads it are built into the app, so nothing is fetched to make it work.

Contact

Questions about this policy: email me at alex@alexecollins.com.

Help with the app itself is on the support page.