1
Profile snapshot
Age, guardian constraints, and optional demographic context are resolved before any model call.
Learning architecture
Every turn follows a governed path.
The platform concept keeps the learning agent inside a safety envelope. The child sees a helpful tutor; the system keeps the profile, rules, and audit trail attached to every interaction.
2
Input governance
Prompts are screened for personal data, unsafe content, adult themes, and attempts to bypass rules.
3
Learning agent
The agent teaches toward a lesson goal using age-appropriate vocabulary, examples, and session limits.
4
Output governance
Responses are checked before display, then logged with the profile snapshot and governance decision.
Non-negotiable guardrails
The safety rules are product requirements.
iKidsLearn starts from governance, not from an open chat box. These requirements must be preserved by future prompts, model calls, storage, and UI flows.
Kid profile in context
Every model call includes the child's basic profile and parent-configured constraints.
Input governance
All child input is screened before the learning agent receives it.
Output governance
All tutor output is screened before the child sees it.
Profile-driven rules
Reading level, vocabulary, topic range, and session caps adapt to the child profile.
Guardian visibility
Every interaction is inspectable by the account guardian. There is no private mode.
Auditability
Input/output pairs, profile snapshots, and allow/redact/block decisions are stored for review.
AI governance model
Built for safer AI practice, not unrestricted chatting.
The product goal is to teach kids how to use AI effectively while making the safety layer visible to guardians and enforceable by the system.
Identity
Guardian-managed child profile and constraints. Data minimization and consent rules stay central to the future implementation.
Policy
Age-specific restrictions for reading level, vocabulary, topic boundaries, session length, and cultural sensitivity.
Agent
Learning objectives, guided hints, Socratic questions, and refusal or redirect behavior when a request falls outside the safe envelope.
Audit
Guardian dashboard, review history, governance decisions, and deletion/export workflows to be defined in the PRD.
Failure
If the guardrails service is unavailable, the platform refuses new prompts, surfaces a clear notice to the child and a flag to the guardian, and queues a service-status entry in the audit log. Open chat is never the fallback.
Who it serves
A learning surface for kids, a control surface for guardians.
For kids
- Ask questions inside guided lessons instead of open-ended chat.
- Learn how to compare, challenge, and improve AI answers.
- Get explanations matched to age and reading level.
- Practice curiosity without sharing personal information.
For guardians
- Configure child profiles, constraints, and topic boundaries.
- Review full session history and governance decisions.
- Set session caps and lesson goals.
- Request export, deletion, and retention controls when implemented.
For schools and educators
Classroom-ready governance, school-controlled rosters.
iKidsLearn is being designed to drop into classrooms without forcing teachers to write prompts or audit conversations by hand. The same governance layer that protects a single child scales to a class, a grade, or a district.
For teachers
- Assign lessons with bounded topics and reading-level targets.
- Inspect every AI session a student had during class time.
- Set classroom-wide constraints that override student preferences.
- Export an audit trail per student for parent conferences and IEP review.
For administrators
- SSO with school identity, roster sync, and least-privilege access.
- District-level retention, deletion, and data-residency controls.
- Compliance posture aligned with COPPA, FERPA, and applicable state laws.
- Reporting on engagement, governance decisions, and learning outcomes.
Pilot interest from schools or districts: hello@websphone.com.
Planning roadmap
Next steps before scaffolding the app.
This first website establishes the public product direction. The application should not be scaffolded until the policy and compliance decisions are written down.
Website concept
Static public site with the core positioning, safety model, and local governance demo.
One-page PRD
Define MVP scope, personas, success metrics, domain model, and safety acceptance criteria.
Policy docs
Resolve consent, retention, deletion, and profile-driven cultural sensitivity requirements.
Prototype app
Scaffold the learning UI, guardrails service, parent dashboard, and audit log data model.
Children's privacy
Built to meet COPPA and GDPR-K from day one.
iKidsLearn is being designed to satisfy the U.S. Children's Online Privacy Protection Act (COPPA), the EU/UK GDPR provisions for children's data, and applicable U.S. state laws (including FERPA for the educator channel). Before any child uses the platform, an account guardian completes verifiable parental consent.
- Data minimization. The kid profile carries only what governance needs (age band, reading level, optional demographic context, guardian-set constraints).
- No behavioral advertising to children. Ever.
- No model training on identifiable children's content. Ever.
- Guardian visibility, no private mode. Every interaction is inspectable.
- Right to deletion and export. Guardians can request full deletion or a portable session export.
Audience note. This marketing site is for parents, educators, and school administrators evaluating the planned platform — not for children. It uses Google Analytics 4 to count visits from these adult audiences. The eventual kids product will operate under a separate, stricter regime with verifiable parental consent before any data is collected.
Read the full intent in our privacy commitments.
Early access
Join the iKidsLearn planning list.
Share an email to receive product updates. The form opens your email client; we don't store data on this site yet.