Google is contractually prohibited from using your audio or transcripts to train their AI models. This isn't a preference or a policy of ours — it's written into Google's own published terms for their paid API services.

hearlog transcribes with Google's paid Gemini API — not the free tier. Google's terms draw a clear line between the two: on the paid API, Google states that it does not use your prompts or the responses (your audio and your transcripts) to train or improve its models. That guarantee is the foundation of this whole page.

You don't have to take our word for it — Google publishes these terms publicly:

Read Google's Gemini API terms

Verify it yourself: Google publishes their terms publicly. Search "Gemini API terms" and look for the data use section — it spells out that paid-service data isn't used for training.

When you record or upload audio in hearlog, that audio is saved in a private storage area, in a folder tied to your account. Only you can reach it: access is enforced by row-level security and served through short-lived, signed links that expire — never public URLs. Other users cannot see your files.

We keep your audio so the app can play your recordings back whenever you return to a note, and so a recording interrupted by a dropped network connection can finish transcribing once you're back online. In other words, your audio is stored to serve you — not to be mined.

The transcription itself sends your audio to Google's paid Gemini API to turn it into text, and Google can't train on it (Section 1). Our transcription services return the text without keeping a separate copy of your audio.

Honest note on retention: we currently keep your stored recordings so you can replay them — we do not automatically delete audio after it's transcribed. If you'd like a recording (or all of your data) removed, email us and we'll delete it.

• Your account — your email and sign-in details, so you can log in and find your notes.
• Your transcripts and notes — stored in our database, encrypted at rest, and protected by row-level security so you can only ever access your own data, never anyone else's.
• Your recordings — kept in the private, account-scoped storage described above, so you can play them back.

• Sell your data to anyone.
• Use your recordings or transcripts for advertising.
• Hand your recordings or transcripts to third parties for their own use. (Transcription sends your audio to Google's paid API to convert it to text — that's the processing you asked for, and Google can't train on it. We don't share your content with anyone else.)
• Train AI on your content.