How to use the website

Start here

Welcome. You do not need to know how to code. Level 1 teaches you to use and question AI; Level 2 guides you through training an image model; Level 3 turns those ideas into programs made of blocks; Level 4 teaches the Python needed for a handwritten-digit recognizer.

This introduction is strongly recommended, but optional. You can start the first AI lesson now and return here later. Skipping this page does not prevent you from finishing a level.

Find your place

1. Open Courses or the level overview. Choose Level 1 if you are starting from zero. Open its outline to see the readings, exercises, and quizzes in order.
2. Open a reading. Its introduction says what you will learn; the worked example shows how the idea is used. Read with the example open, rather than trying to memorise all the terminology first.
3. Use the previous and next links to continue. On a small screen, open the level outline when you need to revisit a page. A short detour to a prerequisite is part of learning.
4. Sign in when you want to submit exercises and keep account progress. Readings and quizzes can be explored without a saved account session. If you work signed out, do not assume results will appear on another device.

A reading, an exercise, and a quiz do different jobs

A reading teaches with worked examples. Practical readings include a mini lab or editor beside the explanation. On a small screen, Practice opens it over the page. A build walks through successive states of the same project: inspect each illustrated stage, try it in your own working copy, and explain the result before adding the next part.

An exercise has a task, a place to answer, and a check or submit action. It is normal to refer back to the preceding reading. A quiz asks you to choose answers and then learn from the explanations. Passing a quiz does not mean you must remember every technical word forever.

For an article, mark it complete after you can explain its example in your own words. Completion is a record of your progress, not a judgment about your ability.

Run, inspect, then check

In a coding exercise, Run executes your program. Output is what it prints, and an error is the runtime explaining where it stopped. Running successfully is not the same as producing the required result. Read the task again, inspect the output, and then use the exercise's check or submit button.

For example, a task might ask you to count images. A program that prints hello runs successfully, but does not solve that task. Change the code to do the counting; do not replace the work with a guessed answer.

The practical articles keep their tools here: the image lab in Level 2, blocks in Level 3, and a Python text editor in Level 4. Try this stage loads a complete runnable example into your mini editor; changing a fictional name, age, or other parameter makes it your own. When a program asks for input, Terminal pauses; type an answer and press Enter or Send to resume it. Files supplies the lesson's data, and Terminal shows what actually ran. Exercises have their own controls and will be developed further separately.

What is saved?

There are several separate kinds of work:

• Course progress: completion records associated with your signed-in account when a save succeeds.
• Block workspaces: stored by the browser for that workspace. A different browser or device is not the same storage location. Clearing site data can remove them.
• Article mini editors: your code, block work, and text files save per lesson in this browser. Terminal answers are used by the active run only. Read any saving warning and keep a downloaded copy of important work. This is separate from exercise submission state.
• Standalone editor files: use its download action to keep a copy. Downloaded Python is source code; it is not automatically a saved block arrangement or a trained image model.
• Image-lab projects: photos, labels, split roles, and notes save together across Level 2 readings. Download project preserves these editable inputs; Download model preserves learned image weights for prediction after a refresh. The training itself runs here in the browser.

Read any save error before closing the page. Saving code does not necessarily save training photos, a model's learned weights, or the current runtime's variables.

When you get stuck

First describe what you expected and what appeared instead. Check the last completed checkpoint. Look for a missed step, a different name, or a block in a different position. Read the first useful error message before repeatedly pressing Run.

Use a hint when you need one. If you ask Guide for help, describe the task, what you tried, and the result. Ask for an explanation of the next step; do not treat an AI answer as proof that your program works. Run and inspect the result yourself.

Keep a small learning note: “I expected __; I saw __; changing __ helped because __.” Explaining a fix is a better sign of understanding than never seeing an error.

Read a practical article without losing your place

Follow one stage at a time. The article shows the current blocks or Python lines, explains what each does, states what to look for, and suggests a small change. Try this stage provides the whole program needed at that point, including earlier steps. It is a starting copy you can edit, not a check that you have learned the idea.

In the blocks course, drag and connect real blocks in the mini editor. When Python begins, you will express the same familiar instructions as typed text. A line such as assigning a name still names a value; only the way you write the instruction changes. Run, inspect, explain, and then personalise.

The full reference solution stays collapsed at the end of the article. Build first, then open it to compare the complete program and recover a missed detail. It is normal to revisit similar states many times: the small differences show how a program grows.

Your first small success

You are ready when you can find a lesson, identify where its example runs, and explain the difference between Run and Check. Continue to your first AI conversation. You will build confidence one visible result at a time.