Your child’s messages, analysed by AI — so you don’t have to review them all manually.
ChatMinder’s AI scores every message for risk. You set the rules. Safe messages flow freely. Flagged ones come to you.
✨ Smart Approval is a Premium feature. Start a free trial →
Every message gets a risk score
ChatMinder’s AI assigns one of four risk levels to every message your child sends or receives:
- Safe — completely routine conversation
- Low — minor flags, likely fine in context
- Medium — worth reviewing
- High — requires your attention
The AI looks at language, context, patterns, and topics rather than just matching keywords — which means it catches nuanced content that a simple word-filter would miss, and avoids flagging normal conversation.
You decide what happens at each risk level
For each risk level, you choose one of four actions:
- Allow automatically — the message is delivered without any intervention
- Allow and notify guardian — the message is delivered immediately, but you’re notified and can review it afterwards. If you decide it shouldn’t have gone through, you can reject it after the fact and it will be recalled.
- Send for review — the message is held and sent to you for manual approval
- Reject automatically — the message is blocked before delivery
A typical setup:
- Allow for Safe
- Allow and notify guardian for Low
- Send for Review for Medium
- Reject for High
This means you only see the messages that genuinely warrant attention — everything else flows normally.
Different rules for different children
Each protected user has their own independent AI rule configuration. You can apply stricter rules for a younger child and more relaxed ones for a teenager — all from the same guardian account, without any crossover between children.
Different rules for different conversations ✨
AI rules are set per child by default — but with Premium, you can go one level further. Using per-conversation protection, individual channels can be assigned their own independent AI severity action rules.
This is useful when one conversation consistently needs a different level of scrutiny. A family group chat can be set to Trusted — messages pass through without any AI analysis. A chat with a contact you’re less sure about can use Advanced mode with its own stricter Allow / Review / Reject thresholds, completely separate from the child’s global rules.
Honest about what AI can and can’t detect
Smart Approval is good at detecting explicit content, grooming-associated topics, bullying language, self-harm themes, and coercive language.
It is not perfect — there will occasionally be false positives (normal messages flagged) and false negatives (concerning content missed). It’s a layer of protection, not a replacement for guardian judgement.
Even with these limitations, it can still be a safe ane reliable tool. If in doubt, use Allow and notify guardian for the Safe and Low risk levels — you’ll still see everything that happens, without holding up delivery — and reserve Send for Review for Medium and Reject for High.
Is AI reading my child’s messages?
Messages are sent to Azure OpenAI solely for scoring. They are not stored by the AI service, not used to train any models, and not shared with any third party for any purpose beyond analysis. The service is enterprise-grade and operates under a strict data processing agreement.
See risk scores when you read your child’s chats
With Premium, each message in the guardian’s chat history view shows its AI risk score alongside the message content. This gives you instant context when reviewing a conversation — you can see at a glance which messages the AI flagged and which it considered routine.
Double tap or long press a message to show Message Options and select Message Insights.
Ready to get started?
It takes minutes to set up. No phone number or email required for your child.