Privacy Policy

Last updated: 2026-08-29

Brain Dump has two tiers, and they handle your data differently. This page describes both. The free tier is on-device by default. The AI subscription sends text to a server. Nothing is sent unless you turn the subscription on.

Free tier: on-device, nothing leaves your phone

Audio never leaves your device. Apple's on-device speech recognition transcribes it locally, using the Neural Engine. The resulting text is saved as a Markdown file in your iCloud Drive. We do not operate a server that receives your audio or your transcripts, and we do not have a database of your notes.

AI subscription: text is sent to a server

The optional AI subscription (rewrite, summary, and action-item extraction) works differently. When you tap a polish button, the transcript text of that note, not the audio, is sent to a server for processing, and the result is returned to your device. This is off by default. It only happens when you actively use an AI feature on a note you have already transcribed on-device, and only that note's text is sent, never the audio and never your other notes.

Brain Dump's hosted AI provider is Cerebras. You can instead configure a custom, OpenAI-compatible endpoint in settings, in which case that provider's own terms apply, not the retention terms below. AI rewriting is configuration-dependent, and the app's public default configuration leaves backend URLs empty. In practice, this means most users, who never enable or subscribe to AI polish, never send any text to a server at all.

See Pricing for what each AI feature does, and How offline works for the full on-device data flow.

Payments

Subscription payments are processed by Apple. Apple collects the billing details needed to process payment and meet its own legal obligations. We do not receive or store your payment details.

Cookies

This website may use cookies to function correctly. You can decline them; some site features may not work as well if you do.

Data retention

Free-tier notes exist only as Markdown files in your iCloud Drive, under your control. If you email us a crash report or a support request, we keep that email for troubleshooting and nothing else.

When you use the AI subscription, the request goes to Cerebras, our hosted inference provider. Cerebras does not retain prompt content, API requests or responses, chat or transaction logs, user input, or model output. Cerebras does retain organization name, user emails, and usage metrics such as throughput and load, used for authentication, billing, and service performance. Source: Does Cerebras retain my data?

Cerebras also uses automatic prompt caching to speed up repeated requests. The cache is ephemeral: held in memory, never written to disk, and never shared between organizations. A cache entry is guaranteed to live for at least 5 minutes and may persist up to 1 hour under heavy load, after which it is cleared. Source: Prompt caching

This retention description covers only the hosted AI subscription. It does not apply to on-device transcription, which never contacts a server, and it does not apply if you configure a custom AI endpoint in settings, since that provider sets its own retention terms.

Security

We take reasonable steps to keep your data safe. No system is 100% secure.

Children's privacy

We do not knowingly collect information from children. If you are under 16, get a parent or guardian's permission before using this app.

Updates to this policy

We may update this policy. Material changes will be reflected in the date above.

Questions

Email [email protected] with questions about this policy.