Learning data should not become attention inventory.

What Skrolz uses

We use an account or anonymous session, topic and learning preferences, Arc progress, saved content, predictions, teach-backs, and memory-check history to operate the learning experience.

What analytics never includes

Raw predictions, teach-back text, transcripts, source excerpts, email, handle, exact date of birth, and notification tokens are excluded from product analytics.

AI processing

OpenAI requests are made on the server, limited to the immediate learning task, and sent with response storage disabled. We use a privacy-preserving identifier for abuse prevention.

Retention and control

Raw prediction and teach-back text expires after 30 days by default. Users can request an export or delete their account. Permanent accounts reauthenticate before sensitive operations.

Ages 16–17

Profiles are private and undiscoverable by default, analytics consent is off by default, and creator publishing is unavailable.

Contact

Privacy contact and controller details will be added before public production launch. The hackathon demo should use guest or prepared non-personal data.