How to introduce your kid to coding at home: the first three sessions

You introduce your kid to coding at home in three short sessions: one away from the screen, one snapping blocks together, and one that ends as something another person plays. Each sitting runs 20 to 30 minutes, and you do not need to know how to code. Advice on how to teach kids to code usually stops at a tool list. Below is the first click, what to say when it breaks, and what to aim at.

Why do most "coding for kids" pages end with "find a teacher"?

Because most of them sell classes. Seven of the ten pages dominating this search belong to camps, courses, or tutoring companies, and the category's most common last step is "find a great teacher". None scripts the first twenty minutes or the first stuck moment.

They are also out of date. On June 2, 2026, Code.org rebranded as CodeAI, and the Hour of Code became the Hour of AI. The free tools, with the ages their makers publish:

• ScratchJr, ages 5 to 7. A tablet app, no reading needed, no web version.
• Scratch, "designed especially for ages 8 to 16". Any browser, and the only one here whose project becomes a link.
• CodeAI (formerly Code.org), free in the browser. A structured puzzle sequence.
• Blockly Games, for children with no programming experience. Pond Tutor pairs blocks with JavaScript.
• Hour of AI (formerly Hour of Code), free, year-round. A good first hour, not a course.

Blockly Games and the CodeAI courses end in a score. ScratchJr ends in a project that stays on the tablet. Only Scratch ends in something a cousin can click. Kubrio's ten free Code with Blocks guides sit beside those: each is one question your kid answers by snapping blocks and running them.

What age should your kid start, and with what?

Blocks before typed text, at every age. Around 6 or 7 with icon blocks on a tablet, 8 and up with Scratch. Ignore "the earlier the better".

Bers and colleagues (2014) studied 53 kindergartners building robot programs: 76% reached the target level on sequencing commands, but only 53% mastered loops and 41% mastered the "If Not" half of a conditional. A six-year-old who hits a wall at "repeat" is on track. Those children also scored highest on the project they chose themselves.

Nobody has missed a window at ten, either. Resnick and colleagues (2009) report that the Scratch community runs from ages eight to sixteen, "peaking at 12", and gives the test for a first tool: a "low floor", a "high ceiling", and "wide walls". Marina Bers, who co-created ScratchJr, sharpens it in Coding as a Playground (2021): "Playgrounds are open-ended. Playpens are limited."

Kubrio's World 1 opens at "What is a command?", then "What is a program?" One browser tab, no sign-up.

What are the first three sessions?

Unplugged, then first blocks, then a thing someone else plays. Stop while your kid still wants more. Scherer and colleagues (2020) analyzed 139 programming interventions and found physicality-based ones especially effective (g = 0.72), alongside Scratch.

Ages 6 to 8

1. Unplugged, 15 minutes. Your kid tells you how to make a sandwich and you follow it literally, so "put jam on bread" puts the jar on the loaf. Run it again with the steps fixed: "you debugged it."
2. First blocks, 20 minutes. In ScratchJr, drag three blocks: green flag, move right, say something. Tap the flag. Then: "make it do something different."
3. The thing that gets played, 25 minutes. A two-scene story with one interaction: tap the dragon, it hides. Then a grandparent taps it on a video call.

Ages 9 to 11

1. Unplugged, 15 minutes. Write the rules of a game your kid knows, precisely enough for a stranger to play. Then follow them exactly and break them on purpose. The edge cases they find are conditionals.
2. First blocks, 25 minutes. At scratch.mit.edu, select Create, add a sprite, and make it move with the arrow keys. Six blocks, and the moment a kid feels they are driving.
3. The thing that gets played, 30 minutes. A catch game: a sprite you steer, a sprite that falls, a score that rises. Then Share, and send the link to one person by name.

Set the Scratch account up first: sharing needs verification, and Scratch emails the link to a parent or guardian.

Ages 12 to 13

1. Unplugged, 15 minutes. Take something that works and sabotage one line. Predict what breaks, run it, then explain the gap. Your kid is diagnosing, not receiving.
2. First blocks with a foot in text, 25 minutes. Blockly Games Pond Tutor flips between blocks and real JavaScript on one problem, which a twelve-year-old who thinks blocks are for little kids needs to see.
3. The thing that gets played, 45 minutes. A game with a losing condition: lives, a timer, a game-over screen. Share it, ask the player for one change, then make it.

In Kubrio's guides, World 2 includes "Can I make a tapping game?", the first guide that ends as a game someone else can tap.

What do you do when your kid is stuck?

Keep the problem with your kid and take away the panic. Do not take the mouse.

Grolnick and colleagues (2002) had 60 mothers and their third-grade children work on homework-like tasks, with mothers randomly assigned to pressure or no pressure. Working together, there was no difference. The children of pressured mothers performed significantly worse afterwards, working alone (F(1, 53) = 4.27, p < .05). The cost of taking over shows up when you are not in the room.

Struggle also does work you cannot see. Kapur (2008) gave 309 eleventh-graders physics problems, either well-structured or hard with no support: the unsupported groups produced worse-looking solutions, then beat the others on individual post-tests. It is a learning principle, not a coding study, but it holds.

Five sentences to say:

1. "Show me what you expected it to do." Half the time they find the bug mid-sentence.
2. "Which part do you know is working?" Werner and Denner (2009), watching 126 girls aged 10 to 14 program in pairs, saw comparing a broken instance to a working one keep them unstuck.
3. "Want to try one thing and see what happens?" A choice, not an instruction.
4. "This is the boring part. It's boring for everyone who does this."
5. "Do you want help, or do you want me to sit here?" A stuck kid often picks company.

Three not to say: "Here, let me do it." "It's easy, you only need to..." "You're so smart, you'll get it." The first takes ownership, the second says the problem is your kid, the third makes the next failure read as a missing trait.

That is Tek's job, one of Kubrio's thinking partners. Tek finds the next step and asks where it went sideways, and never hands over the answer. Two free guides, "What is a bug?" and "Find the event bug", make the broken program the point.

Does coding make kids better at other things?

Some things, moderately, and not reading. Scherer and colleagues (2019) pooled 105 studies and 539 effect sizes on what transfers from programming. Overall transfer was g = 0.49: creative thinking 0.73, math 0.57, reasoning 0.37, other school subjects 0.28, and literacy −0.02, a flat zero. Against untreated controls it was 0.65; against active controls doing another structured activity, 0.16.

So run this for the made thing, not the IQ points. That is the only promise Kubrio makes too: a game other kids play, not a smarter kid. On what makes making work, read the science of hands-on learning.

Why does the first game need a player?

Because an audience turns an exercise into a project. Papert wrote in Mindstorms (1980) that in most software "the computer is being used to program the child", whereas in his vision "the child programs the computer" and gains a sense of mastery. A score screen is not a witness.

The second reason matters most if you have a daughter. Master, Meltzoff and Cheryan (2021) found the stereotype that girls are less interested in computing is measurable by third grade (d = 0.35). Offered the same activity described two ways, 20% of 50 eight-year-old girls took home the version with the stereotype attached, and 80% took the version without (d = 1.10). Introduce it as making the thing she already wants to make.

There are two kinds of projects for kids. The first kind ends on the fridge door. The second kind ends as a real thing other kids can read, watch, or play. Kubrio makes the second kind. Kids make them by directing the same AI the pros use. In Create a Rescue Mission Game Other Kids Can Play, your kid invents a crew of characters, gives each one power, and builds grid levels solvable only by combining them. A solver proves each level is beatable before it ships, and other Kubrio kids play it and post scores. Every season closes with Demo Week, and the next runs December 1 to 7, 2026.

What to do at home this week

1. Book three sittings of 25 minutes, and decide who plays the finished thing, by name.
2. Run the unplugged session for your kid's age.
3. Create the Scratch account yourself, with your own email for verification.
4. Try one free Code with Blocks guide, starting at "What is a command?"
5. Put the five stuck sentences on a note beside the screen.
6. Send the finished link the same day the third session ends.

Ingenious projects that inspire kids to make things — until they find the thing only they could make.

For a bigger build than a catch game, Create a Rescue Mission Game Other Kids Can Play ends as a puzzle game other kids beat and score. Or explore all the projects, and start here when you're ready.

FAQ

What age should my kid start coding? Around 6 or 7 with icon blocks on a tablet, and 8 or older with Scratch in a browser. There is no evidence-backed optimal age.

Scratch or ScratchJr? ScratchJr if your kid is 5 to 7 and you have a tablet, because it needs no reading. Scratch from 8, because a finished project becomes a link you can send.

Does my kid need to learn Python? No. Blocks before typed text at every age in this band. Typing syntax is the part your kid fights, and not the part that matters. If a twelve-year-old wants real code, Pond Tutor puts blocks and JavaScript side by side.

Is Roblox Studio coding? Yes, it uses a real scripting language, and it hooks a kid who lives in Roblox. It is a bigger first step than Scratch, so it suits session five, not session one.

My kid gets frustrated and quits. What do I do? Shorten the sittings and stop earlier than feels natural, then ask "which part do you know is working?"