What is the best coding language for a 7-year-old?

Scratch, or plain-language direction of an AI inside a guided tool. At seven, typing speed and spelling are real barriers, so anything requiring exact syntax spends the child's attention on punctuation rather than logic. Both of these routes let a seven-year-old build genuine cause-and-effect logic — loops, conditions, events — without that tax.

Should my child learn Python or JavaScript first?

Python, in most cases: its syntax is closer to English, its error messages are more readable, and a beginner can write something useful in ten lines. Choose JavaScript instead if what your child specifically wants is a web page or browser game they can share with friends — the motivation of a shareable result outweighs the rougher on-ramp.

Is Scratch real coding or just a toy?

It is real coding. Scratch has variables, loops, conditionals, event handling and concurrency — the same concepts as any text language, with the punctuation removed. What it doesn't teach is the discipline of exact syntax and the habit of reading error messages, which is why most kids should move on around ten to twelve.

Is it still worth learning a language if AI writes the code?

Yes, but for a different reason than before. You no longer need to memorize a language to produce code — but you absolutely need to read it to check that what the AI produced is correct and does what you asked. Reading fluency in one language is now worth more than writing fluency in three.

My child tried Scratch and got bored. What now?

Boredom at that stage is almost always a project problem, not a language problem. Ask what they actually want to build and go straight at it, even if the project looks too hard. A too-ambitious project the child cares about beats an achievable one they don't.

How long before I can tell if it is working?

About three months — and the signal isn't skill level, it's whether they start a project on their own without being asked. Kids who have found the thing tinker unprompted. If after a term coding only happens during the scheduled class, the format is wrong even if the reports are glowing.

Is Python too hard for a fourth grader?

Usually, yes — but not for the reason parents assume. The concepts are fine; a nine-year-old who has used Scratch already understands loops and variables. What is hard is that Python punishes typing errors immediately and cryptically, and a child who cannot yet touch-type spends their session hunting a missing colon instead of thinking. If your fourth-grader types fluently and wants Python specifically, let them. If they hunt and peck, stay in Scratch another year; nothing is lost.

Should my child learn Java for AP Computer Science A?

Not in elementary or middle school. AP Computer Science A does use "a subset of the Java programming language," but it is a high-school course with high-school prerequisites, and the Java subset is a few weeks of syntax for a student who already knows how to structure and debug a program. Spend the elementary years on the structuring and debugging. Note also that AP Computer Science Principles, the broader course most students take first, is language-agnostic — so for many students Java never comes up at all.

Is Scratch actually free, or is that a trial?

Free. Scratch is a project of MIT and costs nothing to use, with no paid tier to be upsold into. That makes it the correct default for grades 3–5 in almost every American household, and it means any paid product aimed at that age band owes you a specific answer about what it does that Scratch does not.

Does the language choice affect what my child’s school will place them into?

Rarely, at this age. US elementary and middle school computer science standards derived from the CSTA framework are written in terms of what a student can do rather than what they can type, and placement decisions at that level are almost never language-specific. High school is different, but by then the choice is the school's, not yours.

What does a US family actually pay for this?

The floor is zero and it is a genuinely good floor: Scratch and Python are both free, and the ten AI activities on our activities guide need no account. Paid kids' coding and AI products in the US mostly sit between roughly $8 and $30 a month per child. Luchi Family is one plan for the whole family: $99 a month ($49 for founding families), every kid included, with your first 7 days for $5 — cancel before it renews.

Coding for kids

The best coding language for kids, by age

There's no single best coding language for kids — the right first step depends on age and, honestly, on what your child wants to make. Here's a straight answer by age, plus the shift that changes the whole question in the AI era. If you are looking for the best programming language for kids in a US classroom context, the answer below is mapped to CSTA grade bands, AP Computer Science, and what the Bureau of Labor Statistics actually projects.

Flat pastel illustration of three stacks of rounded building bricks of different heights, the tallest topped with a small flag, with two empty speech bubbles beside them on a cream background.

A straight answer by age

Ages 5–7: blocks — ScratchJr and drag-and-drop tools. The point is cause-and-effect and sequencing, not typing.
Ages 7–10: Scratch — Blocks with real logic — loops, variables, events. Enough to build genuine games without syntax pain.
Ages 10–12: Python (or AI-directed) — A clean first text language — or, increasingly, directing an AI in plain language and reading the code it writes.

The same answer, mapped to US grade bands

Kindergarten through grade 2 — CSTA Level 1A, in the K–12 Computer Science Standards that most US state curricula are written against — is blocks, and typing is the enemy. ScratchJr and similar drag-and-drop tools let a five-year-old express a sequence without fighting a keyboard. The learning objective is that instructions run in order and produce a predictable result. A child at this level does not have a programming language; they have cause and effect, which is the correct thing to have.

Grades 3 through 5 — Level 1B — is Scratch, and this is the least controversial recommendation in children’s computing. Scratch, from MIT, is free, and it has real logic: loops, variables, conditionals, events, message passing. A fourth-grader can build a game with a score and a losing condition without ever mistyping a semicolon. Level 1B asks students to compare algorithms and to test and refine programs, and Scratch removes every obstacle between a child and doing exactly that.

Grades 6 through 8 — Level 2 — is where a text language becomes reasonable, and Python is the usual and correct answer. Its syntax stays out of the way, the error messages are comprehensible, and the gap between "I know Scratch" and "I wrote a Python script" is small enough to cross in a summer. JavaScript is a defensible alternative if your child’s motivation is specifically making things that live in a browser. High school picks up at Level 3A and 3B, where the language is usually the school’s choice rather than yours.

And there is now a real fourth answer that did not exist five years ago: plain, precise English, used to direct an AI, with the child reading and correcting the code it produces. We make the case for it below, because it is the one that changes the shape of the question.

What the US labor data actually says — and what it does not

Parents ask which language because underneath it is a question about the future, so let us answer that one with numbers rather than vibes. The Bureau of Labor Statistics publishes projections for both of the occupations people mean when they say "coding job," and they point in opposite directions.

For computer programmers, BLS projects that employment "is projected to decline 7 percent from 2025 to 2035," with a median annual wage of $100,390 in May 2025. For software developers, quality assurance analysts and testers, BLS projects that employment "is projected to grow 10 percent from 2025 to 2035, much faster than the average for all occupations," with a median annual wage for software developers of $135,980 in May 2025.

A seventeen-point spread between two jobs that both involve writing code is the most useful fact in this entire article, and it is worth being precise about what separates them. The occupation defined mainly by translating a specification into working code is the one shrinking. The occupation defined by deciding what should be built, structuring it, and verifying that it works is the one growing. That is the distinction, and it is exactly the distinction that AI tooling has sharpened.

Now the honest part. None of this tells you which language to start a nine-year-old on. A projection is a projection, made before anyone knows what the next decade of tooling looks like, and the child you are choosing for will enter the labor market around 2036. What the data supports is a much narrower claim: the durable half of the skill is specification and verification, not syntax recall. Choose the option that gets your child to specifying and verifying fastest, and the language question mostly answers itself.

Where each language shows up in American schools

If you want your choice at home to connect to school, it helps to know what school is actually going to ask for.

In elementary and middle school, most US computer science instruction follows state standards adapted from the CSTA K–12 framework, and those standards are deliberately language-agnostic. They describe what a student should be able to do — decompose a problem, compare algorithms, test and refine a program — not what they should type it in. In practice that means block languages in the elementary grades and a text language somewhere in middle school. Your state’s standards are published; search your state department of education for "computer science standards" and read your child’s grade band.

High school is where a specific language finally matters, and only for one course. AP Computer Science A has students "design and implement computer programs using a subset of the Java programming language." AP Computer Science Principles — the broader, more popular course — is language-agnostic and can be taught in almost anything.

The practical implication surprises most parents: Java is a high-school problem, not an elementary-school one. Starting a fourth-grader on Java to prepare for an exam they might take at sixteen is a way of making them quit at ten. If your child reaches high school able to decompose a problem and debug their own work, the Java subset in AP CS A is a few weeks of syntax on top of skills they already have. If they arrive with Java syntax and no debugging instinct, the course is miserable.

The first "language" that matters now

Here's the shift: the most valuable language for a child today may be plain, precise English (or their home language) — because that's how you direct an AI to build software. A child who can describe exactly what they want, notice when the result is wrong, and refine the instruction has the core skill. Scratch and Python are still worth learning; they're just no longer the only door in.

This isn't a prediction about the future of software. It's already how a lot of code gets written: a person describes the intent, a machine drafts it, and the person checks and corrects. The bottleneck moved from "can you type the syntax" to "can you say what you mean and tell whether you got it." That second skill is teachable at seven. The first one mostly wasn't.

The realistic options, side by side

Every language below is a defensible first choice. The differences that actually matter to a child are how fast they see something working, and whether the thing they make is something they wanted to make.

Starting point	Best age	First working thing	The catch
ScratchJr	5–7	A character that moves, same afternoon	Runs out of room fast — most kids outgrow it inside a year.
Scratch	7–10	A playable game in a week or two	Block dragging gets tedious on bigger projects, and it doesn't transfer to typed code as directly as people assume.
Python	10–12	A text adventure or small tool in a few sessions	The first month is mostly punctuation errors — which is where a lot of kids quietly decide they're "not a computer person."
JavaScript + HTML	10–12	A web page or browser game they can send to a friend	Three languages entangled at once (HTML, CSS, JS). Powerful, but a rough solo start.
Directing an AI in plain language	5–17	A real, custom project in the first session	Only teaches coding if the child has to test, correct, and explain — otherwise it's just watching.

How to actually choose, in one question

Ask your child what they want to make. Not "do you want to learn to code" — that question gets you a shrug. "If a computer could build you anything, what would it be?" gets you a specific answer: a game where my cat dodges cars, a quiz about dinosaurs, a website for my club.

Then pick the shortest route to that thing. A child who wants a platformer should start in Scratch or by directing an AI, because both give them a playable game quickly. A child who wants to automate something boring — rename a hundred photos, track their reading — should start in Python, because that's what Python is genuinely best at. A child who wants something other people can visit should start with a web page.

The reason this beats the standard advice is motivation compounding. The first project a child finishes decides whether there's a second one. Choosing the "best" language and losing the child in week three is a worse outcome than choosing a merely fine language and having them still building in month six.

What good looks like in the first three months

1. Week one: something runs The child ends the first session with a thing that works and that they chose. Not a tutorial output every kid in the class also produced — their idea, however small and lopsided.
2. Weeks two to four: something breaks This is the part parents mistake for failure. A project that doesn't work and has to be diagnosed is the single most valuable moment in the subject. Resist fixing it for them — and resist a tool that fixes it silently.
3. Month two: they explain it Ask them to walk you through how it works. If they can, the learning is real. If they can only show you the output, the tool did the thinking and it's time to change how they're using it.
4. Month three: they want more control When a child starts saying "but I want it to do X and it won't," they've earned the underlying language. That's the moment Python or JavaScript stops being homework and starts being a tool. Introducing it earlier is a guess; introducing it here answers a need they can feel.

Three traps worth avoiding

The first is optimizing for the job market. Nobody knows which languages will matter when a current nine-year-old turns twenty-two, and the parents who confidently picked Java for their kids in 2010 weren't wrong so much as answering a question that stopped mattering. Optimize for whether the child keeps going — that you can actually observe.

The second is syntax-first instruction. A course that spends six weeks on variables and loops before the child makes anything is teaching in the order convenient for the curriculum, not the order that keeps a ten-year-old in the room. Real programmers learn a new language by building something badly in it and looking things up. Kids can learn the same way.

The third is treating the first choice as permanent. Almost every professional developer uses languages they didn't start with, and switching is mostly painless because the hard part — decomposition, precision, debugging — is language-independent. There is no wrong first choice, only a wrong-for-your-child amount of time before something works.

Quick answers

What is the best coding language for a 7-year-old?: Scratch, or plain-language direction of an AI inside a guided tool. At seven, typing speed and spelling are real barriers, so anything requiring exact syntax spends the child's attention on punctuation rather than logic. Both of these routes let a seven-year-old build genuine cause-and-effect logic — loops, conditions, events — without that tax.
Should my child learn Python or JavaScript first?: Python, in most cases: its syntax is closer to English, its error messages are more readable, and a beginner can write something useful in ten lines. Choose JavaScript instead if what your child specifically wants is a web page or browser game they can share with friends — the motivation of a shareable result outweighs the rougher on-ramp.
Is Scratch real coding or just a toy?: It is real coding. Scratch has variables, loops, conditionals, event handling and concurrency — the same concepts as any text language, with the punctuation removed. What it doesn't teach is the discipline of exact syntax and the habit of reading error messages, which is why most kids should move on around ten to twelve.
Is it still worth learning a language if AI writes the code?: Yes, but for a different reason than before. You no longer need to memorize a language to produce code — but you absolutely need to read it to check that what the AI produced is correct and does what you asked. Reading fluency in one language is now worth more than writing fluency in three.
My child tried Scratch and got bored. What now?: Boredom at that stage is almost always a project problem, not a language problem. Ask what they actually want to build and go straight at it, even if the project looks too hard. A too-ambitious project the child cares about beats an achievable one they don't.
How long before I can tell if it is working?: About three months — and the signal isn't skill level, it's whether they start a project on their own without being asked. Kids who have found the thing tinker unprompted. If after a term coding only happens during the scheduled class, the format is wrong even if the reports are glowing.
Is Python too hard for a fourth grader?: Usually, yes — but not for the reason parents assume. The concepts are fine; a nine-year-old who has used Scratch already understands loops and variables. What is hard is that Python punishes typing errors immediately and cryptically, and a child who cannot yet touch-type spends their session hunting a missing colon instead of thinking. If your fourth-grader types fluently and wants Python specifically, let them. If they hunt and peck, stay in Scratch another year; nothing is lost.
Should my child learn Java for AP Computer Science A?: Not in elementary or middle school. AP Computer Science A does use "a subset of the Java programming language," but it is a high-school course with high-school prerequisites, and the Java subset is a few weeks of syntax for a student who already knows how to structure and debug a program. Spend the elementary years on the structuring and debugging. Note also that AP Computer Science Principles, the broader course most students take first, is language-agnostic — so for many students Java never comes up at all.
Is Scratch actually free, or is that a trial?: Free. Scratch is a project of MIT and costs nothing to use, with no paid tier to be upsold into. That makes it the correct default for grades 3–5 in almost every American household, and it means any paid product aimed at that age band owes you a specific answer about what it does that Scratch does not.
Does the language choice affect what my child’s school will place them into?: Rarely, at this age. US elementary and middle school computer science standards derived from the CSTA framework are written in terms of what a student can do rather than what they can type, and placement decisions at that level are almost never language-specific. High school is different, but by then the choice is the school's, not yours.
What does a US family actually pay for this?: The floor is zero and it is a genuinely good floor: Scratch and Python are both free, and the ten AI activities on our activities guide need no account. Paid kids' coding and AI products in the US mostly sit between roughly $8 and $30 a month per child. Luchi Family is one plan for the whole family: $99 a month ($49 for founding families), every kid included, with your first 7 days for $5 — cancel before it renews.

Skip the "which language" paralysis

In Luchi, kids start by directing an AI in plain language and meet real code along the way — no wrong first choice.

Get on the list

Back to coding for kids

Part of the series: Coding for kids

Related guides in this series

Coding with AI for kids — The new way kids build real games and apps — by telling an AI what to make and checking what comes back. What “coding with AI” actually means, and why it’s the skill that lasts.
Coding games for kids — Kids learn to code fastest when they’re building a game they want to play. The best game-first ways to start — free, on-screen and off.