Privacy, in plain words

The formal version lives in our privacy policy. This is the same design explained the way we'd explain it across a kitchen table.

Messages pass through; they don't stay

When your kid sends a message, it travels through YouGoodAI on its way to the AI, the safety rules run on it in memory, and the answer comes back the same way. We never store your family's conversations, and we never ask for your family's sensitive personal data, so neither can leak. Chat history lives on your family's own devices, where you can keep it or delete it; it's yours, not ours.

One honest nuance: very long answers can take minutes to write, so they wait in an encrypted holding space on our service until your kid's app picks them up, and are erased the moment that happens.

What we do keep

• Account basics: names, emails, your team's settings and rules.
• Safety-event labels: records like “Block Violence matched, 3:42pm”. A label and a timestamp, never the message.

What parents see, and don't

Parents see the activity feed: safety events, broad topics, trends, digests. Parents do not see conversations, ever; there is no transcript view to unlock. The single exception is the short preview on an ask-first approval, shown because nobody can make a judgment call blind, and always visible to the kid as a held request.

Promises, not just settings

• No secret monitoring. When a rule acts, your kid sees it act. There is no covert mode, by design.
• No health surveillance.We don't detect or report health-related concerns at any granularity; the only carve-out is imminent safety.
• Beliefs stay off-limits. Rules can cover topics and behavior, but we never report which direction a kid leans on politics, religion, or identity, in any direction.