How the AI works here

How a lesson is made

From the photo we take the learning goal and the exercises: what your child needs to learn. Then AI makes a new lesson of its own to teach that goal, adapted to the condition and level. The photo itself is not kept: it's only processed in working memory and then destroyed straight away.

Where the AI runs

We use the AI model Claude, which runs on Amazon Web Services in the European Union (Frankfurt). Your data stays in the EU and is not used to train AI models.

Checking: a second AI, and you decide

AI-made content isn't automatically reliable. That's why a second, independent AI checks every lesson for accuracy, level and suitability. If that check is unsure about even one point, the lesson is flagged and you, as the parent, decide whether your child may see it. Every lesson notes that it was made with AI.

What the AI doesn't do here

AI doesn't copy the original exercise one-to-one as an explanation: we learn the concept and create our own explanation for it. AI does not diagnose and does not give medical advice. This is a learning aid, not care.

AI can make mistakes

Despite the check, a mistake can slip into a lesson. Stay involved and use your own judgement. If you see something that's wrong or doesn't fit, use the “Report this lesson” button. Your report helps us improve the check.