About Whizbee · Parents first, builders second
The team behind safe AI for kids
Whizbee is made by a small team of parents and builders. We build AI for children ages 7-11+ that asks a child to think first, then coaches — the opposite of tools that hand kids the answer. We back our research honestly, sell no data, run no ads, and prove learning without surveillance.
Who we are
We are a small team of parents and builders. We are parents first and builders second, and Whizbee is the product we wanted to exist for our own kids before we wanted it to exist for anyone else. That order matters: every decision runs through the question a parent asks, not the question a growth chart asks.
You will not find a founder-hero story or a headshot on this page, and that is deliberate. Whizbee is not one person’s brand — it is a set of commitments a team keeps. We would rather you judge us by those commitments and by how the product actually behaves than by a face.
Why Whizbee exists
Most AI tools do one thing when a child asks a question: they hand over the answer. It feels helpful, and it quietly removes the exact moment where learning happens — the trying. We built the opposite. Whizbee asks the child to attempt first, coaches them when they are stuck, and only counts something as learned once the child can explain it back in their own words.
General-purpose chatbots were never designed for a seven-year-old. They are open-ended, they can wander anywhere, and they reward passive consumption. A tool for children needs different bones: finite missions, a closed and moderated space, and an honest line between what a child mastered and what they merely looked up.
What we believe (and the evidence for it)
Our design choices are grounded in named cognitive science, quoted honestly rather than cherry-picked.
Effort is where learning lives
A 2025 MIT Media Lab study on "cognitive debt" found that people who leaned on AI to write remembered less and felt less ownership of their own ideas. That study was run with adults, not children, and we will not pretend otherwise — but the direction it points matches decades of learning science, so we designed against it: Whizbee asks the child to attempt before it ever helps.
A little struggle makes learning stick
Cognitive scientist Robert Bjork calls them "desirable difficulties" — the small effortful moments, like retrieving an answer from memory or wrestling with a problem, that feel harder but build durable understanding. Whizbee is built around that friction on purpose, not around the frictionless shortcut.
Deep reading is a skill worth protecting
Maryanne Wolf, who studies the reading brain, warns that the deep-reading circuit is shaped by practice and can weaken when everything is skimmed or summarised for us. We take that seriously: Whizbee coaches a child toward understanding in their own words rather than delivering a tidy paragraph to copy.
Our commitments to your family
Privacy by design, not bolted on
Whizbee is built to be COPPA- and GDPR-conscious from the first line of code. A child signs in with a temporary Magic PIN, not an email we collect, and the space is closed — no open web, no external links.
We never sell your child’s data, and we run no ads
There is no advertising model and no data broker on the other side of this product. Children are not the thing being sold, and their data is never sold either.
No dark patterns, no engagement traps
No points, streaks, badges, or leaderboards. No confetti loops engineered to keep a child hooked. Sessions end with a clean Hard Stop and return the child to the real world.
Proof, not surveillance
Parents receive a structured Proof Pack of what a child learned, explored, and created — not a live keystroke feed of every word typed. Honest proof for you; a calm, private space for your child.
A 14-day refund, no argument
The Founder Year is a one-time $49 for 12 months for the first 200 families. If Whizbee is not right for your family within 14 days, we refund you — no interrogation, no dark-pattern retention flow.
How we handle research content
When we publish about AI and children, we hold ourselves to editorial standards a parent can check:
- We cite primary sources — the original study, paper, or regulation — not a summary of a summary.
- We caveat honestly. When a finding came from adults, we say so; when evidence is early or contested, we say that too, instead of dressing a guess as settled science.
- We date what we publish and review it. Every research page carries a "last updated" date so you can see how current it is.
- We do not overclaim our own product. A claim on a Whizbee page maps to a mechanism that actually ships, checked on the server — never a brochure promise.
Last reviewed: July 12, 2026.
Join the Founder Year
A safe AI learning space for your child and honest proof for you, from a team that builds for kids first. The Founder Year is a one-time $49 for 12 months, open to the first 200 families — with a 14-day refund if it is not right for you. Not a subscription.
Claim your Founder spot