How we use AI

Last updated October 26, 2026

Draft. Kinovi is in a small, invite-only test. A privacy lawyer reviews these pages before Kinovi opens to everyone. Words marked in yellow are not final yet.

Kinovi is AI

It says so when a session starts. It never claims to be a person or a friend. It is a building tool, like a senior developer who explains everything.

Which AI

Kinovi uses Anthropic's Claude models through their business service.

No training

Our AI provider does not train on business data by default, and we never opt in.

Rules we follow

We follow Anthropic's guidelines for organizations that serve minors:

• follow children's privacy law
• tell users they are talking to AI
• use safety filters and monitoring
• follow their child-safety guidance

Kids stay in charge

Kinovi suggests and explains. Kids decide, and approve every build step. Kinovi does not do schoolwork, and it does not quiz kids.

What can go wrong

AI can make mistakes, and safety filters are not perfect. That is why we have several layers of safety, a Report button, and a grown-up review before anything goes public. See Safety.

Your own AI keys

If you add your own AI key, those requests go under your own agreement with that company.