Child Safety: live awareness around children

What it does

Svid already describes every frame of your footage, and one of the plainest things it can say is when a child has left a supervised area on their own. "Child left the play area alone" is exactly the kind of search Svid already answers; Child Safety raises it in the moment instead of after. Think of it as an extra pair of eyes for staff at pickup and playtime, not a system that decides anything on its own.

How it works without identifying anyone

No facial recognition, no biometric IDs and no people tracking, of anyone, children or adults, ever. Child Safety reads plain descriptions of what's in frame, the way a colleague at the gate would say "the girl in the yellow coat is heading for the car park". Nothing biometric is created or stored about any child, and no attempt is made to recognise a returning adult. It describes what's visible right now, nothing more.

What an alert looks like

A plain description and a link to the moment, on screen, by email, or by JSON webhook if the setting already has a system for staff messages. A typical alert reads: "Child in a yellow coat left the play area alone by the side gate, 15:41." The alert prompts a person to look and act; it never accuses anyone of anything.

Where it runs

Child Safety is switched on per camera, so the entrance and the outdoor play area can run it while the office is left alone. It works on RTSP feeds from existing cameras, on uploaded files, and on a live phone camera covering a gate during pickup. Recorded footage runs through the same pipeline. Cloud mode sends frames to the vision model for description; for settings that can't allow that, a fully on-prem appliance mode keeps footage inside the building, with 30-day retention by default.

What it costs

There's no rate card. The Store prices Child Safety from your own usage before you buy, shown as "+$X/mo at your usage" next to the check. It's added to a monthly subscription priced from your own usage with a $9 a month floor, and the whole derivation is on screen at signup. The pricing section walks through the arithmetic.

Does it use facial recognition on children?

No. Never, on anyone. Child Safety reads plain descriptions of what's visible, and no biometric data about any child is created or stored.

What happens when an alert fires?

A staff member gets an alert on screen, by email or by webhook with a description and a link to the moment. A person looks, and a person decides. The check flags; it doesn't judge.

What's the difference between this and Child Safety Review?

This is live awareness during the day. Child Safety Review is for afterwards, reconstructing what actually happened when a question or incident comes up.