400 hands-on exercises. Zero setup. Runs in your browser. Meet Pytor, your built-in AI Python tutor.
Start Learning → pytor.mwmai.no
Mirror: matswm86.github.io/pylearn
PyLearn is a free, interactive app that teaches Python programming from absolute zero. No installation, no accounts — open the link and start coding. Along the way, Pytor — a friendly Python snake AI tutor — lives in the corner of the screen, ready to answer questions, give Socratic hints, and generate custom lessons on demand.
- Works on phones, tablets, and desktops
- Python runs directly in your browser (via Pyodide)
- Instant feedback — write code, click Run, see if you got it right
- Progress tracking — streaks and completion stats saved locally
| # | Topic | Exercises | What you'll learn |
|---|---|---|---|
| 1 | Variables | 50 | Storing data, naming, assignment, f-strings, scope |
| 2 | Data Types | 50 | int, float, str, bool, type conversion, string methods |
| 3 | Conditionals | 50 | if/elif/else, boolean logic, ternary, match/case |
| 4 | Functions | 50 | def, arguments, return, *args/**kwargs, closures, decorators |
| 5 | Lists & Sets | 50 | Indexing, slicing, comprehensions, sorting, set operations |
| 6 | Dictionaries | 50 | Key-value pairs, nesting, comprehensions, Counter, grouping |
| 7 | FastAPI | 50 | Pydantic models, validation, CRUD APIs, auth, middleware |
| 8 | API Calling | 50 | JSON, HTTP headers, auth, pagination, rate limiting, webhooks |
Each topic has a lesson with plain-English explanations, analogies, runnable examples, and common mistakes — plus 50 progressive exercises from Beginner to Challenge.
Pytor is the friendly Python snake who lives in the bottom-right corner of PyLearn. Click him (or the 🐍 Pytor button in the top bar) to open a chat sidebar where you can:
- Ask questions about any Python concept — Pytor uses Socratic questioning to guide you to the answer instead of just handing it over
- Get a hint on any exercise you're stuck on — Pytor reads your current code and points at what to try next
- Generate a custom lesson on any topic — type "decorators" or "asyncio" and Pytor builds an interactive HTML lesson on the fly
Under the hood:
- Primary: Groq with
llama-3.3-70b-versatile(~1-2s responses) - Fallback: local Ollama with
llama3.1:8b - Backend: tiny stdlib-only Python HTTP bridge at
pytor.mwmai.no/api/tutor - Frontend: vanilla JS, no framework — graceful no-op if the backend is unreachable
The Explain Like I'm 5 mode is on by default. It adds simplified explanations using everyday analogies:
- Variables = labeled jars
- Functions = recipes
- Lists = shopping lists
- Dictionaries = phone books
- APIs = restaurant waiters
Toggle it with the ELI5 button in the top-right corner.
- Pick a topic — start with Variables if you're brand new
- Read the lesson — concepts explained with analogies and live examples
- Do the exercises — write Python in the built-in editor
- Click Run & Check — your code runs in-browser and gets auto-graded
- Use hints if stuck — each exercise has progressive hints
- Check the solution — learn from the answer when needed
pylearn/
├── docs/ # Static site (GitHub Pages)
│ ├── index.html # Main page
│ ├── style.css # Responsive styles
│ ├── app.js # SPA logic + Pyodide integration
│ ├── lessons.js # Lesson content for all 8 topics
│ └── exercises.json # 400 exercises (generated)
├── pylearn/ # Python source
│ └── exercises/ # Exercise definitions with auto-grading
├── build.py # Generates exercises.json from Python source
└── pyproject.toml
The Python source in pylearn/exercises/ is the source of truth. Each exercise has test mechanisms (check lambdas, function test cases, or expected output). The build script converts these to web-compatible assertion strings:
python build.py # Regenerates docs/exercises.jsoncd docs && python -m http.server 8000
# Open http://localhost:8000- Pyodide — CPython compiled to WebAssembly, runs Python in the browser
- CodeMirror 5 — Code editor with Python syntax highlighting
- Vanilla JS — No framework, hash-based SPA routing
- LocalStorage — Progress persistence, no backend needed
Exercise data lives in pylearn/exercises/*.py. Each exercise is a make_exercise() call with description, hints, solution, and test mechanism. PRs welcome for new exercises, better descriptions, or additional topics.
MIT