Privacy Policy

Hearsay (“the app”, “we”, “us”) is a personal life-logging tool that records audio, transcribes it, and helps you identify recurring speakers in your day. This policy explains what data we collect, where it goes, and what control you have over it.

Who runs Hearsay

Hearsay is operated by Gavin Vickery. Contact: gavin@geekforbrains.com.

What we collect

Audio recordings.When you start a recording session, the app captures audio from your iPhone's microphone in short chunks (~30 seconds each) and uploads them to Hearsay's servers over HTTPS.

Transcripts and metadata.Each chunk is transcribed by OpenAI's transcription API (gpt-4o-transcribe-diarize). The resulting text, timestamps, speaker turns, and confidence scores are stored in our database.

Voice fingerprints. A short numeric voice embedding (192-dim vector) is extracted per speaker turn so that recurring voices can be clustered and named. Embeddings are stored alongside transcripts; they cannot be played back as audio.

Optional data, only if you opt in:

• Calendar events for the calendars you select, so conversations can be tagged with the matching meeting.
• Location at the time of recording, so conversations can be tagged with where they happened.
• Photos you attach to a person.

Account data. Your email, a hashed password, and login tokens.

What we do NOT keep

Raw audio is deleted. After transcription and embedding extraction (typically within minutes of upload), the original audio file is permanently deleted from object storage. We retain only the transcript, the voice embedding, and a short representative voice clip per identified speaker (used by you to confirm cluster identity).

No analytics on conversation content. We do not mine your transcripts to build advertising profiles, train models, or share content with third parties beyond the transcription provider.

Who we share data with

• OpenAIreceives audio chunks and returns transcripts. OpenAI's API terms apply to that processing; we do not opt your data into OpenAI's model training.
• DigitalOcean Spaces hosts the audio object storage during the brief window between upload and deletion.
• Sentry receives crash reports and error traces from the app and our servers. These reports do not include transcript content.
• Apple receives standard App Store / TestFlight diagnostics if you opt into sharing them.

We do not sell or rent your data, and we do not use it for advertising.

Where data is stored

Servers are operated in the United States by DigitalOcean. Database and object storage are encrypted at rest and in transit (TLS).

Your rights and controls

• Delete a recording or conversation: in the app, swipe a card → Delete. The transcript, embeddings, and any retained voice clip are removed immediately.
• Delete your account: Settings → Account → Delete account, or email gavin@geekforbrains.com. All data tied to your account is purged within 30 days.
• Revoke permissions: in iOS Settings → Hearsay you can revoke microphone, calendar, location, photos, and notifications at any time. Hearsay degrades gracefully when permissions are missing.

Recording responsibility

Hearsay is a tool. You are responsible for complying with the laws and norms in your jurisdiction regarding recording conversations — particularly two-party consent requirements. If you are in doubt, ask the people you're with before recording.

Children

Hearsay is not directed at children under 13 and we do not knowingly collect data from them. Do not create an account if you are under 13.

Changes to this policy

We will post updates here with a new “Last updated” date. Material changes will be communicated by email when reasonable.

Contact

Questions, deletion requests, or anything else: gavin@geekforbrains.com.