All of it, in plain English. If you want the full policy text, email us and we’ll send it.
There is no ad business. There are no data brokers. Nobody pays us for your family’s information, because nobody can.
Your email and name. Your child’s nickname. Device info. We don’t ask for an address, a phone number, or a birth date.
Switches you turn on if you want them. You decide which ones to use, and when.
Anything we hold from a call log, an SMS, or a notification is gone after seven days. You don’t have to ask.
One tap in the app. Account and child data goes immediately. Database backups roll off within 35 days.
See how to delete your data →Children don’t have accounts. Children don’t see your details. Children don’t get marketed to — by us or anyone else.
Web content is filtered at the network level by AdGuard. AdGuard sees the domains your child’s device requests; it doesn’t know your name, your email, or who the child is — only an anonymous device ID.
Optional. Off unless you turn it on. When it’s on, your child’s messages go to OpenAI to generate the response — no names, no accounts, just the text and an age group.
Need the full policy? Email us at admin@childsafe.uk and we will send it to you.
No Firebase Analytics. No Google Analytics. No Mixpanel, Amplitude, AppsFlyer, Adjust or Branch. None.
No AdMob. No Facebook SDK. No advertising IDs collected. No ads on our site or in either app, ever.
We don’t build a profile of your child for any purpose — not for targeting, not for ranking, not for our own product analytics.
No face scans. No voice prints. No fingerprints. ChildSafe doesn’t need them and doesn’t collect them.
ChildSafe runs on UK servers (DigitalOcean, London). Connections between the apps and our servers are encrypted in transit.
Third parties we rely on to deliver the service: Google (push notifications, maps, Play Billing, Play Integrity), AdGuard DNS (web filtering), OpenAI (only if you turn on the AI Learning Buddy), Anthropic (only when keyword monitoring fires — see below), Zoho Mail (account verification and password reset emails). Each has its own privacy policy and may process data outside the UK.
If you’ve turned on keyword monitoring, ChildSafe sends each match to Anthropic’s Claude AI for a sanity check — to filter out song lyrics, news headlines, and homework topics that aren’t really about anything.
What gets sent: the matched keyword, its category, an age band, the app name, the URL if any, and about 300 characters of surrounding text. What is not sent: your name, your email, your child’s name, the device ID, or any account information.
We use Anthropic’s commercial API. Under their commercial terms, inputs are not used to train their models. If the AI fails or times out, the alert is sent anyway — we never silently drop a child-safety signal.