How Luminara uses AI

1. When AI is used

Two optional grown-up tools use AI:

• Plan wizard — drafts a suggested daily schedule from the shape of day you describe.
• Mission wizard — suggests hands-on, real-world activities matching the categories you choose.

Both live in the grown-up dashboard and run only when started by a signed-in grown-up. Our server verifies that the requester is a parent member of the family (or a grown-up who has just cleared the family PIN on a shared device) before any request is made. Requests from a child's device are refused. Every other part of Luminara — schedules, missions, messages, rewards, the Guardian extension — works without AI, and sends nothing to an AI provider.

2. What is sent

Only the short brief the wizard builds from your choices is sent. For a day plan, that is:

• The children the plan is for, by first name only.
• An age range you type, such as “6 to 12”.
• Whether it is a school day, school break, or weekend, and the pace you want.
• Wake-up and wind-down times.
• Focus areas you select, and fixed events you list, such as “Swim class 4:00pm”.
• Any free-text notes you choose to type.

For missions, the brief is the activity categories, indoor/outdoor, solo or group, how much grown-up help is right, an age range, how many ideas you want, and your notes.

Because the notes field is free text, please avoid typing anything you would not want processed by a third-party provider — for example a full name, address, phone number, school name, or health information. Nothing in Luminara requires it.

3. What is never sent

The AI features do not receive, and cannot see:

• Family or child messages, announcements, or reactions.
• Photos, including task-proof photos.
• PINs, password hashes, sign-in credentials, or pairing codes.
• Progress data — XP, coins, badges, streaks, submissions, or approval history.
• Guardian activity — time on apps, opened apps, or blocked domains.
• Browsing history, precise location, contacts, email addresses, or last names.

4. Who processes it

AI requests are processed by Anthropic (the Claude API), acting as a service provider to Luminara. Requests are sent from Luminara's server, never directly from your browser or from the Guardian extension, and our provider credentials are held server-side only.

Under Anthropic's commercial API terms, the inputs and outputs of these requests are not used to train its models. The provider may retain request content briefly for safety and abuse monitoring under its own terms. Anthropic is disclosed as a recipient in our Privacy Policy and in the Children's Privacy Notice.

5. You stay in control

AI output is a draft, never an action. A generated plan or mission list is shown to the grown-up who asked for it, and nothing is added to a child's day, assigned, or shared until that grown-up reviews it and chooses to save it. You can edit or discard any draft. AI cannot approve work, award rewards, message a child, change a PIN, alter the Guardian allow-list, or adjust any safety setting.

To limit cost and misuse, each family can run a capped number of generations per day.

6. Children and AI

There is no child-facing AI in Luminara. Children cannot chat with an AI, cannot start a generation, and are not shown AI output directly — they see the plan or missions only after a grown-up has reviewed and saved them. No child profile is used to personalize an AI model, and no AI is used to profile, score, evaluate, or make automated decisions about a child.

7. Accuracy and limits

AI-generated suggestions can be wrong, unsuitable for a particular child, or simply unappealing — that is why a grown-up reviews them before anything is saved. Suggestions are ideas for family activities and daily structure. They are not educational assessment, medical, psychological, developmental, or safety advice, and they are not a substitute for your judgment about what is right and safe for your child. Please assess any suggested activity for your child's age, ability, and setting, and supervise accordingly.

8. Your choices

Using AI is entirely optional. If you prefer not to use it, simply do not open the plan or mission wizards — you can build schedules and assign missions by hand, and nothing is sent to an AI provider. Because these requests are not stored as a child's personal record, there is no separate AI history to delete; your rights over family data are described in the Privacy Policy and Data Retention Policy.

9. Changes to this page

If Luminara adds an AI feature that handles a new category of information, uses a different provider, or changes who can start it, we will update this page and the effective date, and provide notice and obtain consent where required.

10. Contact

Luminara Privacy
support@luminarakids.com

Questions about how these features work, or a request to review what a grown-up sent, are welcome at this address.