01Move one
Noticing
The chicken feed is low. Nobody said anything.
A letter to parents
How do you raise a kid in the age of AI?
Your kid is growing up with AI in the room. Sunny Street is a small town where they practice the things AI can’t do for them: noticing, caring, deciding, and owning the choices they made.
They grow up with AI, not from it.
Free during beta. Parent consent verified. No ads, ever.A day in Sunny Cove
Sixty seconds of actual gameplay.
No cutscenes, no marketing polish. A kid asks Eli, an AI neighbor who runs his own routines, about a supply shortage and chooses what to do next.
The frame
This screen is a place.
Left column, AS-IS: what today’s AI chat products do when a kid reaches for them. Right column, TO-BE: what Sunny Street does instead.
When your kid is stuck on a hard problem
hands over the answer in one secondgives them a reason to sit with it one more minute
When your kid is bored in a quiet moment
fills the silence with an infinite feedfills the silence with a person to walk over to
When your kid is upset after a fight with a friend
tells them whatever they want to heargives them neighbors who sometimes disagree
Small shifts compound over time.
Four things your kid practices
The things AI cannot do for them.
Noticing, caring, deciding, and owning the choices they made. One loop, rehearsed every day your kid plays.
02Move two
Caring
Dale’s truck broke down again. I brought him coffee.
03Move three
Deciding
Eli said plant corn. I saw the wind. I planted squash.
04Move four
Owning
I was right about the wind. He told me so the next morning.
One loop. Four things AI cannot replace. Practiced every day your kid plays.
Why now
We are not alone in this argument.
The idea that kids need productive friction, not more answers, is converging fast across research labs, policy groups, and parent advocates. Sunny Street is the first product designed around it.
The convergence, in four lines
Capability. Safety. And now, developmental impact.
FrictionIt’s the feature.
AgencyThe kid must drive.
CompanionshipNot from a bot.
BoredomGive permission.
Friction in human exchange is a good thing. It comes from bringing our whole selves to the encounter.
A forklift can lift the weights for you, but the whole point is the lifting.
The best thing I am doing is giving my children permission to be bored.
AI products for children should be evaluated on a third dimension beyond capability and safety: developmental impact.
Doubt is the driver of curiosity. Uncertainty, not certainty, is what makes learning feel alive.
Adults lose skills to AI. Children never build them.
Institutions in conversation
MITHarvard GSEUC BerkeleyBrookingsCommon Sense MediaWorld Economic Forum
We did not invent this thesis. We are the first team shipping a product around it.
For parents, our commitments
What we build in. What we keep out.
Trust is not a badge row. It is what we built in on purpose, and the things we gave up to keep your kid’s environment clean.
- We build
Evidence in your kid’s voice, not our dashboard.
Every week, a short prose chronicle of who they are becoming. Not a stat page. Not a compliance report. A keepsake.
- We build
For the year, not the session.
A 28 day arc with an ending. Your kid finishes, not quits. No streaks. No daily urgency.
- We refuse
No AI pretending to be a friend.
Characters in a town, not companions. They follow their own routines whether your kid shows up or not.
- We refuse
No ads. No IAP. No data sold.
Permanent. Not a phase we will grow out of.
- We refuse
Not used to train AI.
Conversations are never used to train any model without a separate parent opt in.
- We refuse
COPPA Tier 2 parent consent, by design.
Locked until a parent verifies consent. Reviewable, revokable. Built before beta, not after.
Ages 8 to 11, on purpose.
Not for all kids. Designed for the window when patterns of attention and self-direction are setting.
No streaks. No daily urgency.
Stories end on purpose. The town does not chase your kid past their attention.
More commitments
- Ages 8 to 11, on purpose.
Not for all kids. Designed for the window when patterns of attention and self-direction are setting.
- No streaks. No daily urgency.
Stories end on purpose. The town does not chase your kid past their attention.
Join the waitlist
We are still early enough to design this well.
Closed beta opens early May 2026. 50 to 100 families. The arc on this page is the design we are building toward, not a finished game yet.
We are still early enough to get this right. That is the reason we are building now.