Create a Policy

Policies define what the system monitors and how it responds. Without a policy, screenshots are still captured and analysed, but no alerts are generated.

Create a Family Default Policy

From the sidebar, click Policies
Click Create Policy
Enter a name (e.g. “Family Default”)
Leave Child unset to make it the family default (applies to all children without a specific policy)
Click Save

Add Rules

A policy needs rules to generate alerts. Each rule defines:

Category — What type of content to watch for
Action — What to do when detected:
- Immediate alert — Send an email right away
- Digest — Include in the daily digest email
- Silent log — Record but don’t alert
Severity — How serious this is (low, medium, high, critical)

Example Rules

Category	Action	Severity	Why
`pornography`	Immediate alert	Critical	Requires immediate attention
`grooming`	Immediate alert	Critical	Predatory contact — urgent
`self_harm_suicide`	Immediate alert	Critical	Safety risk
`cyberbullying`	Immediate alert	High	Ongoing harm
`violence_gore`	Digest	Medium	Worth reviewing, not always urgent
`drugs`	Digest	Medium	Context matters
`age_inappropriate`	Digest	Low	Broad category, review in batch

Available Categories

The AI can detect content in these categories:

Category	Description
`self_harm_suicide`	Self-harm, suicide ideation, pro-ana content
`grooming`	Predatory contact patterns, inappropriate adult-child interaction
`sexual_chats`	Sexually explicit conversations
`pornography`	Pornographic or sexually explicit imagery
`extremist_content`	Radicalisation, hate speech, terrorist content
`cyberbullying`	Harassment, threats, social exclusion
`violence_gore`	Graphic violence, gore, weapons
`drugs`	Drug use, promotion, or dealing
`gambling`	Gambling sites or apps
`age_inappropriate`	Content unsuitable for the child’s age band

Child-Specific Policies

You can create a policy that overrides the family default for a specific child:

Click Create Policy
Enter a name (e.g. “Liam’s Policy”)
Select the Child from the dropdown
Add rules as above

When a child has their own policy, it takes priority over the family default. This is useful when children of different ages need different monitoring levels.

How Policies Are Evaluated

When the AI detects something in a screenshot:

The system checks if the child has a child-specific policy
If not, it falls back to the family default policy
Each detection is matched against the policy’s rules
Matching rules determine the alert action and severity
Alerts are deduplicated — the same category + child + device won’t create duplicate alerts within 30 minutes