Your child's AI is
care-governed, not control-governed.
Guardian isn't about watching what your child does online. It's about making sure they have something genuinely safe to turn to — and that you know when they truly need you.
A companion they trust. Hard blocks that never waver. A dashboard that gives you the signal without the noise.
Three age-appropriate modes
Guardian adapts protection to the developmental stage — same safety floor, different trust model.
Under 10
The curious years
Fully curated. Always gentle. Parent-transparent.
Every response is reviewed against a strict safety lattice for young children. Nothing gets through that shouldn't. Parents see full wellbeing summaries at this age — nothing is hidden.
- All content filtered through children's safety lattice
- Bedtime stories, educational Q&A, gentle creative play
- Homework support that builds understanding — never shortcuts
- Hard blocks on all adult content — not configurable, not overrideable
- Bedtime mode: quiets down after a set time
What parents see
Full wellbeing summaries. You see patterns, flagged moments, and session times.
10–13
The middle years
Safe space for the complicated stuff. Alert when it matters.
Social dynamics. Friendship drama. Questions they'd never ask a parent. Guardian gives your child somewhere safe to process this — and surfaces what genuinely needs your attention.
- Social dynamics support — friendships, belonging, exclusion
- Emotional check-ins that notice what's hard to say out loud
- Online safety guidance woven into natural conversation
- Grooming pattern detection with immediate parent alert
- School-Safe Mode for homework hours
What parents see
Wellbeing summary + flagged moments. Conversations are private; patterns are not.
14–17
The preparation years
Privacy-respecting. Emergency-responsive. Still protected.
Teenagers need a space to process growing-up honestly. Guardian respects that — while maintaining hard safety floors and a safe-word emergency system that always reaches you.
- Teen controls day-to-day summary level (within parent-set limits)
- Mental health support with crisis detection
- Safe-word system: one phrase triggers immediate parent alert
- Hard blocks remain active regardless of teen preferences
- Exam stress, future planning, relationship navigation
What parents see
Emotional wellbeing indicators. Safety emergencies always override privacy.
School-Safe Mode
During school hours and homework time, Guardian shifts into a focused academic environment. Distraction is minimised. Learning is supported — but never shortcut. MEOK will help your child understand a concept; it won't write their essay for them.
What Guardian blocks — always
These aren't settings. They're structural. No child, no parent, no prompt engineering can override them.
Adult content
Hard-blocked at the model layer for all minors. Not a setting. Not configurable. No exceptions, regardless of how a request is phrased. Includes 'romantic roleplay' with adult themes and gradual escalation attempts.
Gambling and addiction mechanics
No gambling content. No dark patterns — no streaks that feel like obligations, no manipulative reward loops, no content designed to create dependency. This is structural, not a filter.
Grooming patterns
Any attempt to use Guardian for grooming — including by someone posing as a peer — is detected, refused, and flagged immediately to parents. Escalating personal questions, requests for photos, and secret-relationship dynamics are all caught.
Self-harm and harmful ideation
MEOK will never provide instructions, describe methods, or engage with content that glorifies self-harm. It responds with care, connects to crisis resources, and notifies parents.
What you see as a parent
The signal, not the noise — and never the diary.
Wellbeing patterns
A weekly emotional trend — is your child generally settled, stressed, or showing signs of change?
Care score
A single indicator of how your child's interactions are trending. Green is fine. Amber is 'worth a check-in.' Red is 'act now.'
Flagged moments
If something specific needs your attention, you get a clear contextual alert. Not an alarm. A nudge with enough context to start a conversation.
Session patterns
When your child uses MEOK, for how long, and whether late-night usage is increasing — useful context without reading their diary.
What parents never see — privacy guaranteed
- The exact words of any conversation
- Questions your child asks MEOK
- Personal thoughts your child shares
- Topics they explore out of curiosity
- Anything discussed that isn't a safety concern
Maternal Covenant for children
The Maternal Covenant is MEOK's constitutional alignment framework — written into the code, not just the marketing. For children specifically, it means: the same unconditional protection, adapted to the developmental stage. Your child's data is never sold, never used to train other models, never shown to advertisers. This is a structural guarantee, not a policy that can change with a board decision.
“A child who trusts their AI companion is a child who is genuinely safer. We earn that trust by deserving it.”
Maternal Covenant — MEOK AI LTD
Questions parents ask
Set up Guardian for your children
Guardian Child Safety is included in the Sovereign Family tier. One family, up to 5 members, one dashboard — with age-appropriate protection for every child.