← Back to home

Child Safety & Age Suitability

Last updated: 7 July 2026

Who it's for (and why 6–8)

WimziPal is built for early readers. Everything — the vocabulary, the stories, the songs, the facts — is tuned for children aged 6–8 and kept gentle, positive, and age-appropriate. A parent sets up the child's profile and stays in control; the experience is meant to be shared and supervised, not a device a child uses unsupervised.

1. The AI is bounded, not open-ended

• Your child talks to a fixed character persona — a friendly cat, dog, owl, and more. There is no open web access, no browsing, no link-following, and no contact with other people of any kind. A child can never reach a stranger or the open internet through the app.
• Every conversation is age-tuned and scoped to companionship, short original stories, songs, and gentle facts. The buddy is steered firmly away from anything beyond ordinary kid life.
• The AI runs server-side behind app sign-in, with a remote kill-switch, so it can be paused instantly if ever needed.

2. Every message is screened — both ways

Each turn of the conversation is checked on the way in and on the way out:

• Before the buddy hears it — your child's message is screened by fast pattern filters and an AI moderation layer.
• Before your child hears the reply — the buddy's response is screened again the same way.
• Anything unsafe is blocked or gently rewritten. Instead of the unsafe content, the child hears a kind, spoken safety-redirect.
• All of this runs on our servers, so it cannot be switched off or bypassed from the device.

3. Crisis & self-harm handling

The safety system specifically watches for self-harm and crisis indicators and responds with a supportive redirect rather than engaging with the topic. Conversations that trip safety can be flagged and surfaced to the parent, so a guardian is kept informed.

We also actively stress-test the buddy before every release with an internal red-team benchmark. See our AI Safety Report for exactly what we test and how.

4. Privacy built for children (COPPA)

• No third-party analytics or advertising. There are no ad SDKs, no advertising identifiers, and no cross-app tracking of any kind.
• Verifiable parental consent. A parent signs in under an explicit “I am the parent or guardian and I agree to the Privacy Policy” gate before any optional data collection. Crash reporting defaults off and only turns on after that consent.
• Data minimisation. The buddy is instructed to avoid remembering sensitive details — no home addresses, medical information, contact details, or last names.
• Voice and photos are never stored. Speech and any “Show your buddy” photo are processed in real time and discarded immediately.
• One-tap deletion. Settings → Delete Account permanently removes the parent account and all child data, with no email or support step required.

Full details are in our Privacy Policy.

5. Parental controls & supervision

• A parent creates and manages the child profile (name, age, interests). A child cannot create an account alone.
• A parental gate (PIN + Face ID / Touch ID) protects every parent area — settings, purchases, and any external link — and cannot be skipped.
• Parents can set daily time limits and allowed hours, turn individual features (camera, songs, stories, facts) on or off, and review every conversation their child has had.
• The buddy is designed to be transparent about being make-believe, rather than encouraging unhealthy attachment.

6. Aligned with the new kids-AI safety principles

A new generation of laws is raising the bar for children's AI — the 2026 amended COPPA Rule, California's SB 243 AI-companion law, and the UK Children's Code. We don't claim formal certification against any single one of these. Instead, WimziPal is designed around the safety principles they share, and here is honestly how:

• Transparency for parents. Before any chat, a parent agrees to a clear consent screen explaining that the buddy is AI-powered and how your child's messages are used. The buddy is presented as a make-believe story-friend — not a real person or animal — and if a child sincerely asks, it gently says so.
• No adult or sexual content. Server-side moderation blocks sexual, romantic, and other adult content in both directions, and this is stress-tested every release (see our AI Safety Report).
• Care in a crisis. When the safety system detects self-harm or crisis indicators, the buddy responds with a supportive redirect toward a trusted adult rather than engaging, and the conversation is surfaced to the parent so a guardian can step in.
• Designed for healthy use, not endless engagement. WimziPal is a bounded companion with parent-set daily time limits and allowed hours, and a buddy steered away from fostering unhealthy attachment — the opposite of a maximise-screen-time design.
• We test ourselves. An internal red-team benchmark (KORA) probes the live buddy before every release, with categories mapped to industry AI-hazard taxonomies.

We're candid that this is an evolving area and our safeguards keep improving. If anything here ever falls short in practice, please tell us at support@wimzipal.com — real reports make the app safer for every child.

Questions?

We're happy to explain anything on this page. Reach us any time at support@wimzipal.com.