Last updated: 4 August 2026

This policy describes exactly what the app does with your data. It is written from the source code, not from a template.

The short version

How you are identified

On first launch the app generates a random identifier (a UUID) and stores it in the iOS Keychain. Everything the app remembers is filed under that identifier.

It is not linked to you, your Apple ID, your device hardware, or anything else. We cannot tell who you are. If you delete and reinstall the app you get a new identifier and start over — your previous data is no longer reachable by anyone, including us.

What is stored on our server

To provide cross-session memory and a study plan — the entire point of the app — the following is stored, filed under the random identifier above:

What Why
Your stated goal, self-reported level, and interests To generate and adapt your study plan
Transcripts of your practice conversations (both your side and the coach's) To analyse mistakes and compute progress metrics
Mistakes you made, including the exact words you said To build the error log and track whether each pattern is improving
Vocabulary and expressions, with spaced-repetition scheduling state To decide what comes back for review and when
Moments where you couldn't express something, and the phrasing you were given To turn them into review items
Per-session metrics (speaking rate, sentence length, talk share, retention rate, etc.) To show progress over time
Session timestamps, duration, token counts and cost To show usage and enforce the budget cap

Transcripts contain whatever you chose to talk about. If you discuss personal matters during practice, that text is stored. Delete it from Settings › Data if you would rather it not be.

What is never stored on our server

Third parties