Your agent acts as your personal tutor, generating progressively harder problems. It tracks your skill level on each concept and picks the next problem based on where you are.
Current courses: Algorithms & Data Structures and Concurrency.
Get started
Clone the project and cd into it:
git clone git@github.com:zuzuleinen/algotutor.git && cd algotutor
Run initial set-up to choose your agent and optionally enroll in a course:
Train > Check > Review
Start a training session with make train.
Submit your solution by telling your tutor check. It's OK to say, "I don't know". They will pick an easier problem for you.
Start a daily review session with make review.
Read more here about the motivations behind the project.
Commands
Local commands — terminal
| Command | What it does |
|---|---|
make init |
First-time setup — enroll in courses, pick a default agent |
make enroll |
Add another course to your enrollment |
make train |
Launch your agent in training mode for the active course |
make train <slug> |
Switch active course to <slug> and launch the agent there |
make review |
Open the review TUI across every enrolled course |
make review <slug> |
Open the review TUI scoped to one course |
make list |
List all available courses and your enrollment status |
make run |
Sanity-check your solution before check (active-course aware) |
Agent commands — chat
Type these in your agent session:
| Command | What it does |
|---|---|
train |
Get the next problem in the active course |
train <course> |
Switch to <course> and start training there |
check |
Submit your solution for evaluation (grading, mistake logging, level updates) |
I don't know |
Break the problem into simpler sub-problems |
I want to solve [problem desc] |
Request a specific problem with description or screenshot |
reset |
Wipe progress in the active course (with confirm reset gate) |
reset all |
Wipe progress in every enrolled course (with confirm reset all gate) |
Requirements
- An AI coding agent — see docs/agents.md for the supported list and setup
- Go ≥ 1.26