Support

We are glad to help.

Contact us

To help us reproduce the issue, please include: device model, OS version, Comma version (Settings → About), a description and the time it happened. We usually reply within 2 business days.

FAQ

I released mid-thought — do I lose it?

No. Releasing is a pause, not the end: the button turns amber meaning the segment is kept. Hold again to continue, then tap Finish for the whole round.

Where is the transcript? Can I edit it?

Finished records are in History, where you can read the full text and copy all of it or a selection. There is no in-app text editor — transcripts are drafts, so copy them out to edit.

How is online recognition billed?

By seconds of audio actually recognized. Retrying the same audio over the network does not double-charge. You can check quota in the app.

Does it work without a network?

Yes. Choose on-device recognition in Settings and download a model tier that fits your device; recognition then works offline. System speech recognition is also available (iOS 26 or later).

Is my audio uploaded and stored?

In online mode audio is processed in server memory and released after use; it is not stored long-term, and transcripts are not written to the database. Recordings and history stay on your device. See the privacy policy.

How do I delete my account and data?

In the app: Settings → About → Delete Account, which removes the server-side account and credentials. On-device history and recordings are yours to delete in the app. You can also follow the account deletion page to request deletion by email.

Recognition keeps getting a word wrong

Add the word (a name, a term, a company name) to Hotwords so recognition prefers it. You can also switch recognition mode or model tier in Settings.

App information

App	Comma
Developer	Wang Shaoxing
Platform	iOS (App Store)
Support email	support@wangshaoxing.com
Privacy Policy	comma.wangshaoxing.com/en/privacy.html
Terms of Service	comma.wangshaoxing.com/en/terms.html