Pebblemind
For ages 6–11 · free while we're new

You can't predict their future. You can prepare their mind.

Pebblemind teaches four skills every subject rests on — logical reasoning, critical thinking, problem solving and first principles — and adapts every question to hold your child right at the edge of what they can do.

Sessions are short and end with something you do together, off-screen. We deliberately don't maximise engagement.

80%
success rate we hold every child at, question by question
6–11
one system that grows with them for five years
15 min
a session, then it ends on purpose
0
streaks, ads, trackers or child-to-child contact
The simple version

School teaches answers.
We teach working it out.

Maths Reading Science Coding A job in 2045
Can you work it out?

All of it sits on one skill. If a child has it, every subject gets easier. If they don't, no amount of practice sticks.

What we teach

Four skills. That’s the whole curriculum.

Each one starts as a question a 7-year-old can answer out loud — and turns into something a 27-year-old gets paid for.

01

Logical reasoning

Working out what must be true.

“Every red one sank. So what will this red one do?”

7 yrs Child in a puzzle room

Works out which block is making the tower fall — and moves that one.

27 yrs As an adult at work

Finds the single false assumption inside a plan, a spreadsheet, a diagnosis.

02

Critical thinking

Asking whether something is actually right.

“You said it’s the biggest. How do you know?”

7 yrs Child in a puzzle room

Asks how you know the taller jar holds more, instead of taking your word.

27 yrs As an adult at work

Isn’t fooled by a headline, a salesperson, or a very fluent chatbot.

03

Problem solving

Getting from stuck to a first step.

“It won’t reach. What could I change?”

7 yrs Child in a puzzle room

Tries a smaller ramp when the marble won’t reach, then a smaller one again.

27 yrs As an adult at work

Is the person who names the first step when everyone else is still stuck.

04

First principles

Breaking a thing down to its parts.

“What is a wheel actually for?”

7 yrs Child in a puzzle room

Takes the idea apart to find out what it’s really doing.

27 yrs As an adult at work

Builds the thing that didn’t exist, instead of copying the thing that does.

Nobody outgrows these. A 7-year-old uses them on a tower of blocks; a doctor, a lawyer and a founder use the same four on a Tuesday afternoon.

What the research found

Teach a child to think, and they think better.

Not our opinion. Three findings in plain words — and the one honest catch that shaped how we built this.

As per the research of Science 2011 Psychological Bulletin 2019 Frontiers in Psychology 2020 Harvard 1960
Finding 1
They solved problems nobody had taught them.
The puzzles they practisedbetter
Puzzles they had never seenbetter too

Most brain games only make a child better at that one game. Here the gain carried over to brand-new problems — which is the whole reason we teach thinking instead of answers.

Finding 2
Head to head, the taught child usually comes out ahead.
6 — the taught child does better 4 — the other child does better

Same new puzzle, one child taught this way and one not, 10 times over: the taught child wins about 6 of them, not the 5 you would get from a coin toss. A small, steady, repeatable edge — measured across 8,925 children.

Finding 3
What improved is exactly what we teach.
Logical reasoning
Critical thinking
Problem solving
First principles

All four moved when they were taught on purpose, in children aged 4 to 12 — which is why our whole curriculum is these four, and why we start at 6 and stay until 11.

The honest catch
Six weeks of it wears off.
startend of a short coursemonths later

Retested months after a short programme, most of the gain had gone. Thinking behaves like fitness: stop and it slips. That is why we run five years, not six weeks.

Diamond & Lee, Science, 2011 · Takacs & Kassai, Psychological Bulletin, 2019 — 90 studies, 8,925 children, a medium pooled gain (g = 0.34) · Scionti et al., Frontiers in Psychology, 2020 — transfer to untrained tasks · Bruner, The Process of Education, Harvard University Press, 1960. Bars above show direction, not exact per-skill figures.

Five years, one thread

It grows with them

Ages 6–7

Sparks

Sorts, notices, names

Everything is a thing you can pick up. Nothing needs reading.

Ages 7–8

Explorers

Predicts before acting

The rule is withheld. They test for it instead of being told.

Ages 9–10

Investigators

Weighs evidence

Two sources disagree. Which do you trust, and why?

Ages 10–11

Strategists

Carries a method across

Scruffy, open-ended problems. The win is a method that travels.

What you’ll actually notice

You won’t see a score. You’ll hear the difference.

“She argued with me for twenty minutes. With reasons. That has never once happened over homework.”

Priya
two children, 7 and 10 — nine weeks in
Week two
“Wait — that can’t be right.”

She stopped herself. Nobody told her to check.

Week five
“But how do you know?”

Asked of a teacher. Then of an advert. Then of me.

Week nine
“I’ll try the smaller one first.”

Eleven attempts at a marble run, no adult in the room.

These are the three sentences parents report back most. None of them are about a score, a level, or a streak — and all three are the sound of a child thinking on their own.

Made by children, last week

Not answers. Things.

Stories that branch, games with rules they wrote, machines that only work once the logic does.

ending one her ending ending three
The lighthouse that argued

Three endings, and she had to make every one of them earn the choice before it.

Story builder · Maya, 8 years
attempt 10 · one degree too shallow
11 attempts
The marble run that finally worked

Ten failures first. The tenth lost it on one slope that was a single degree too shallow.

Marble run · Idris, 10 years
all three, or nothing
Three switches, one bulb

Nothing lights until the whole loop makes sense — which means it can't be guessed.

Circuit builder · Noor, 7 years
Growth, not grades

Nothing in here is scored

What we never collect
Grades
Percentile ranks
Class leaderboards
Daily streaks
Comparison with other children
What we watch instead
How they explain it↗
What they do when wrong↗
Whether they check↗
What they pick when nothing is scored↗

Movement in these is the only progress we report — in one short note a week, written in sentences.

Make-your-own

Where they stop answering
and start building

Five build-your-own studios. A child makes something that actually runs, then works out why it doesn't.

What the world is asking for

Employers want exactly these four

Not our framing — theirs. We’ve named the survey term under each bar so you can check us.

Logical reasoning 70%
#1 most-cited core skill reported as “analytical thinking” — named by ≈ 7 in 10 employers
Critical thinking Top 5
top-five core skill reported within “analytical & creative thinking”
Problem solving Top 5
top-five core skill reported as “complex problem-solving”
First principles Rising
fastest-rising cluster reported within “creative thinking & curiosity”

Source: World Economic Forum, Future of Jobs Report 2025 — a survey of employers representing over 14 million workers. Bar length shows each skill’s prominence in that ranking; the report’s own wording differs from ours, so we print both.

Whoever they turn out to be

Four skills, every career

You can't know at 7 whether you're raising a surgeon, a set designer or someone with a job that doesn't exist yet. You don't have to. Underneath all of them sits the same four moves.

Doctor

Which symptom doesn't fit the story? That's odd-one-out, with a life on the other end of it.

Critical thinking Logical reasoning
Engineer

What is actually true about this load, this material, this constraint? First principles, eight hours a day.

First principles Problem solving
Lawyer

Two accounts. Both credible. One holds up. Deciding which is the entire profession.

Critical thinking
Entrepreneur

Nobody hands you the problem, the instructions or the permission. You work out the first step, then take it.

Problem solving First principles
Creators

Why would a person do that? A film, a song, a game, a story — the craft is reasoning about people all the way down.

Logical reasoning First principles
Whatever comes next

Most of the jobs your child will hold don't exist yet. These four moves outlive the invention.

All four

We are not teaching your child a curriculum, and we are not teaching them to be better at grades.

We're teaching them to be better at growing — at meeting something unfamiliar, thinking it through, being wrong out loud, and going again. That habit outlives every syllabus, every exam, and every job title they'll ever hold.

The one number that runs everything

What 80% means

Out of every 10 questions, your child gets about 8 right and 2 wrong. That mix isn’t an accident — it’s chosen, question by question, and it’s the whole trick.

10 out of 10 Too easy

They already knew all of it. Nice feeling, nothing learned.

8 out of 10 Just right

Mostly steady, with two that make them stop and think. This is the one we aim for.

5 out of 10 Too hard

Half wrong. A 7-year-old decides they’re bad at it and stops trying.

One real session, question by question
8 in 10
question 1difficulty, held near 8 in 1014

Tired day? It gets easier within three questions. Flying? It moves up. Your child never sees any of this happen — they just find it interesting.

How that same session went
11
worked out
3
got wrong, then retried
79%
right — a good day
Put simply

Easy enough that they keep going. Hard enough that they have to think. Those 2 wrong answers are the reason it works — and they’re what your child will still be doing tomorrow.

Why not 10 out of 10?
All right means they already knew it. Two sessions like that and the questions step up.
Why not half?
Half wrong doesn’t feel hard, it feels like “I’m bad at this”. Children stop trying long before they run out of ability.
Nobody picks a level
No test to sit, no difficulty setting for you to guess at. Each question is chosen from how the last few went.

The idea isn’t ours: research on learning rates puts the fastest progress just short of always getting it right (Wilson et al., Nature Communications, 2019 — the “85% rule”). The caveat: that work was on machines and simple tasks, so we treat 80% as a target to steer by, not a law — the engine follows the child.

What it costs

Nothing, for now — and there is nothing to pay with

There is no checkout on this site. You make a parent account, add your children, and they start. Here is what it will cost later, printed now so it can never be a surprise.

$7 per child, per month
Free
for everyone, for a limited first stretch — every child on your account, everything included
No card, no expiry date, no countdown clock.
When charging starts, you get a month’s notice — and your child keeps everything they made.
If money is the reason a child can’t use this, write to us and it stays free.
What a child gets — all of it, today
All four skills — logical reasoning, critical thinking, problem solving, first principles
The whole five-year arc, ages 6 to 11, adapting question by question
The Create hub: stories, games, laser puzzles, marble runs, circuits
Your weekly note and the growth dashboard, for every child you add
No ads, no tracking, no streaks, no child-to-child contact

If you want to support the work while it’s free, there’s a Ko-fi tip jar in the footer. It’s entirely optional, it buys nothing extra, and your child’s access never depends on it.

One question, as a child sees it

Which one doesn't belong?

Odd-one-out · level 4 of 12

Nothing is timed and nothing is scored. Pick one and it will tell you what it was reading.

What parents ask first

Is there anything in here designed to keep my child hooked?

No, and that's a product promise rather than a preference. No streaks, no daily rewards, no notifications, no next-episode autoplay. Sessions end and suggest something to do together off-screen.

My 6-year-old can't read much yet.

Sparks is narrated end to end and works on pictures and quantity alone. Reading load only rises when a child's own profile shows they're ready for it.

What happens to my child's data?

We store the minimum needed to run the account, never sell it, show no ads, and carry no third-party trackers. Children have a display name and nothing else.

Does this replace maths practice?

No — and we'd be suspicious of anything claiming to. This trains the reasoning that makes practice land, which is a different job.

Will it work on an old phone?

Yes. The atmosphere and depth degrade to a simpler flat version on weak hardware or when reduced motion is on, and the page is never blank without them.

Fifteen minutes a day, and it ends on purpose

Everything is free while we're in early access — no card, no countdown, nothing to pay with. You'll get one short note a week about what actually clicked.