Privacy policy

Last updated: June 12, 2026

TrainScribe is built to know as little about you as possible. This page explains, in plain language, what data the app handles and where it goes.

No account, no email

You never create an account. The app signs in anonymously with a random identifier (via Supabase). We do not ask for, or store, your name, email address, or any other identity. That random identifier is used to enforce free-tier limits and to link your subscription.

Your training data stays on your device

Your sessions, custom activities, and profile are stored locally on your phone. They are not uploaded to a server and we cannot read them.

Voice recordings

When you dictate a recap, the audio is sent to OpenAI to be transcribed, then deleted. The app asks for your consent before the first recording. Typing instead of speaking always works.

Recap text

The text of your recap is processed by an AI model (OpenAI, via our backend) to reconstruct it into a structured session. It is used for that reconstruction and not for advertising or profiling.

Analytics, only if you opt in

The app can share anonymous usage statistics (via PostHog) to help improve it. This is off unless you turn it on, and it never includes your training content or audio. You can change your choice anytime in Settings.

Crash reports

If the app crashes, a technical crash report is sent (via Sentry) so the bug can be fixed. Crash reports do not include your training content.

Purchases

Subscriptions are handled by Google Play and RevenueCat. We never see your payment details; we only learn whether your anonymous identifier has an active subscription.

Your controls

• Export: download all your data as JSON from Settings.
• Erase: delete all your sessions, custom activities, and profile from Settings. This is immediate and irreversible.
• Analytics: opt in or out anytime in Settings.

Contact

Questions about privacy? Write to contact.