Blog › Inside Questzo
How the Questzo checks work
When you press Submit on a quest, a ✓ or a ✗ appears next to every instruction within a second or two. Here is what happens in between, and why we built it this way.
Your code runs for real, in your browser
Questzo loads Python into your browser using Pyodide, so nothing is installed and nothing is sent to a server to run. Each time you press Run or Submit, your code starts in a fresh, empty folder, so files from an earlier run can never make a check pass by accident.
One check for every instruction
Every quest has a small piece of Python written by us that looks at the result of your code. It can look at three things:
- What your bot printed, line by line.
- What your code made: the variables and functions it defined, so a check can call your function with new values and see what comes back.
- Whether Python stopped with an error, and on which line.
The check returns one result for each instruction, with a reason when it fails. The instructions on the left are written from those checks, so they always say the same thing.
When a quest asks for input
If a quest uses input(), the check types the answers for you from a fixed list, so the same code gets the same result every time. When you press Run instead, a box pops up and you type the answers yourself.
A guard against endless loops
Everyone writes a loop that never ends at some point. Questzo counts the lines your code runs, and after 200,000 it stops and tells you a loop may be missing its way out, instead of freezing your browser tab.
Why Pip always gives the same answer
Real AI models give a slightly different answer each time, which would make a fair check impossible. Pip is a tiny scripted practice model with fixed replies. It teaches the same patterns a real model needs (prompts, system prompts, JSON, chat history, tools), and your code can be checked exactly.
Ask Pip why
When a check fails, Ask Pip why reads the result and explains it: which function or variable is missing, the line where Python stopped and what that error usually means, or your output and the expected output side by side.
Every quest is tested before release
Our quest test page plays every quest and practice challenge twice: once with its answer, which must pass every check, and once with its starter code, which must not. It also confirms there is exactly one check for each instruction. You can open it and watch it run.
Learn it by building it
Questzo teaches Python by building an AI agent, one pixel quest at a time. The first two chapters are free, with no sign-up.
Start the first quest