How sotonis handles your data
Exactly where event audio goes and what remains afterward — stated plainly. Feel free to print this page as-is for your church board.
Need a signed written confirmation? Just email us.
How audio is handled
- Event audio streams from the operator PC's browser to the interpretation engine (Google Gemini API) and is used solely to generate captions.
- sotonis servers never receive or store audio. The stream goes directly from the operator PC to the engine — our servers are not on that path.
- No recording is made. When the event ends, no audio remains.
Is it used to train AI?
- No. Under its terms, the paid Google Gemini API does not use inputs (audio) or outputs (captions) to improve products or train models.
- Google retains logs only for a limited period, solely to detect violations of its Prohibited Use Policy (per the Gemini API Additional Terms).
How captions are handled
- Captions pass through our server only to reach audience phones and screens (in memory). They are not stored, and vanish when the session ends.
- The only caption record after an event is the file you download on the operator PC. Keeping, sharing, or deleting it is entirely up to you.
What we do keep
- Account email, event info (name, date, join code), usage time, and payment records — the minimum needed to provide and bill the service.
- Glossaries you register for an event (names, terminology) are saved to your account and can be edited or deleted any time.
- Nothing spoken during the event is on this list — prayers, personal stories, and sermon content are simply not stored.
About your audience
- Audience members just scan a QR — no name, phone number, or email is ever entered. There is no such thing as an audience account.
Reference: Google Gemini API Additional Terms (Paid Services) — ai.google.dev/gemini-api/terms
Last updated 2026-08-25 · Operated by SORiGiO · ok@sotonis.com