AlgoStudio
Teach programming with real, auto-graded code. Students write solutions in the browser (or in VS Code), hit run, and get per-test feedback in seconds — every run sandboxed in its own container. Teachers build courses by hand or by asking the built-in AI assistant.
**Author problems**
Write a statement, starter code, and tests — static pairs or a reference solution with a random input generator. Python and C# out of the box.
[Make a problem →](problems/index.md)
**Build with the assistant**
"Create a recursion course with five problems" is a sentence, not an afternoon. The assistant creates courses, problems, files, and pages for you.
[Meet the assistant →](assistant.md)
**Manage access**
Groups are your classes. Assign a course to a group (with an optional due date) and every member gets access — and an email.
[Courses & access →](courses/index.md)
**Review student work**
Per-student rosters, every submission with per-test results, an activity calendar, and the attempt where it finally clicked.
[Review submissions →](students/submissions.md)
Pick your path
| You are a… | Start here |
|---|---|
| Teacher building a course | Getting started, then Problems |
| Student solving problems | Solving problems |
| Admin running the server | Installation, then the Admin reference |
How it fits together
- A course holds problems (auto-graded coding challenges) and content pages (reading material).
- Groups of students get courses assigned to them; maintainers edit a course's content.
- On submit, the student's code runs in an isolated Docker container against the problem's tests — see Limits & sandboxing.
- Words in italics unclear? There's a glossary.
What's not done yet
AlgoStudio is young. The honest list of gaps lives in one place: Roadmap & limitations.