Parents shouldn't have to take "safe for kids" on faith. This page describes, concretely, how Yoggi's safety mechanisms work, what data is collected, and who can see it. If something here ever changes, this page will change with it.
Content safety is built into every response, not filtered afterward
Yoggi's AI is instructed, as a core part of how it generates every single reply, to refuse violence, sexual content, drugs, and other age-inappropriate topics, and to redirect the conversation gently instead of returning a cold refusal. This applies to every message, from the first word of a conversation, because it is part of how the response is generated rather than a separate check layered on top.
Responses are also adapted to the child's age: children aged 3–5 get short, very simple answers; 6–10 get slightly longer, accessible answers; 11–15 get more developed, clearer explanations. The AI always identifies itself honestly as an AI if a child asks, and answers in the language the parent has configured, out of 12 supported languages.
Real-time safety alerts: what triggers an email, and what doesn't
Every message a child sends is classified, at the moment it's sent, into one of three levels: none (nothing to flag), warning (worth keeping an eye on), or urgent (needs a parent's attention now). This classification is a contextual judgment made by the same AI call that generates the reply to the child, not a keyword blocklist.
- Warning: recorded and shown as a badge on that day's conversation in the parent's history. No extra notification is sent, to avoid overwhelming parents with low-signal alerts.
- Urgent: in addition to the badge, an email is sent to the parent immediately, with a short explanation written by the AI for the parent. Only one urgent email is sent per child per day, even if several urgent messages occur, to prevent alert fatigue.
This mechanism runs independently of the conversation itself: it never blocks or changes the answer given to the child. It only informs the parent, in parallel, that a conversation may be worth a check-in.
What Yoggi's safety alerts are, and what they are not
Every alert email and every AI-generated summary states clearly that it was generated by AI. It is a signal to prompt a conversation, not a diagnosis, and it is not a substitute for a parent's own judgment or for professional medical or psychological advice. Yoggi is a conversational companion, not a clinical or crisis service.
Full visibility for parents, protected from the child
Parents get a read-only history of every conversation, organized by day, with a visible badge on any day that had a safety alert. Two additional layers are free on every plan, not just premium:
Daily AI insight
A nightly summary of the previous day's conversation: engagement level, topics, mood, and any genuine points of attention. Viewable in-app over a rolling 7-day window.
Weekly email recap
An optional email combining the week's daily insights into one summary sent to the parent, always labeled as AI-generated content, never presented as medical or psychological advice.
PIN-protected settings
An optional 4-digit PIN, stored securely on the device, prevents a child from changing the daily message limit, language, sensitive topics, or safety-related settings.
Sensitive topics awareness
Parents can flag family-specific sensitive topics (grief, separation, illness, and others) so the AI treats them with extra care if the child brings them up spontaneously. Yoggi never introduces these topics on its own, and this setting never lowers the baseline safety rules above.
Data privacy and account structure
A child does not have their own account. Only the parent authenticates, via Google, Apple, or a one-time email code, and that parent's authenticated account is the only place identity and billing data live. Every backend request is scoped to that authenticated parent, so one parent's data is never visible to another. Yoggi is designed to meet COPPA (Children's Online Privacy Protection Act) and GDPR, including GDPR-K requirements for children's data.
See it before you decide to trust it.
Yoggi is free to try, with real-time alerts and full parental visibility included from the first conversation.
Reliability: built on multiple AI providers
Each AI capability Yoggi relies on, conversation, voice transcription, and voice replies, is backed by more than one AI provider, with automatic failover to a backup provider if the primary one is unavailable or rate-limited. This is a resilience measure, not a safety measure: it exists so a provider outage doesn't interrupt a child's conversation, and every provider in the chain follows the same age-adaptation and safety instructions.
Who is behind Yoggi
Lucas G.
Founder of Yoggi. He built Yoggi out of a personal conviction: an AI for children should open up the world to them, with care, rather than expose them to it without any guardrails. He designs and reviews Yoggi's safety behavior directly.
This page is written and maintained by the Yoggi team directly. We do not publish invented expert quotes or unverified certifications: every claim on this page describes something the product actually does today.
Frequently asked questions
How does Yoggi keep children safe?
What happens if a child says something concerning?
Does Yoggi's safety alert replace a parent's judgment or professional advice?
Is Yoggi COPPA and GDPR compliant?
Can my child see or change the safety settings?
← Back to home · Safe AI for Children, explained → · Real-time parent alerts, explained →
