Open-source course engine

Interactive courses built around your goals

Build a course that meets the learner where they are: practical lessons, exercises, flashcards, spaced repetition, and progress tracking. Describe the audience and outcome, let an AI coding tool generate the content, then ship a static course site.

Read the guide Open live course GitHub

terminal

$ gh repo clone jmilne22/vibe-learn
$ npm run new-course -- python
$ claude "Create a Python course"
$ npm run build
# Course ready at dist/python/

Available courses

Infrastructure Development with Go

From DevOps engineer to infrastructure developer. Build CLI tools, API clients, concurrent services, networking utilities, container internals, Kubernetes controllers, and algorithm practice.

13 modules · ~10-20 hours

Sample Course

A starter template demonstrating every platform feature. Fork this as a base for your own course.

Template

3 modules · ~2-4 hours

How it works

Scaffold

Create a course directory with the right lesson, exercise, and asset structure.

npm run new-course -- python

Generate

Give your AI tool the starter prompt, learner profile, and target outcome.

claude "Create a Python course"

Build and ship

Compile markdown and YAML into a static site that can be deployed anywhere.

npm run build

What you get

Lessons & code comparisons

Markdown with syntax highlighting and side-by-side code comparisons. Show the same concept in two languages.

[ ]

Exercises & challenges

Warmups, challenges, multiple variants, difficulty modes, progressive hints, and annotated solutions.

SRS

Spaced repetition & flashcards

SM-2 scheduling surfaces exercises and flashcards when you're about to forget. Review, discover, and weak-spot modes.

|||

Analytics & progress

Mastery dashboard with health cards, heatmaps, streaks, action items, and full data export. All local, no accounts.

Zero config

Content files drive features — no registries, no config files, no wiring. Drop files in the right place and the build figures out the rest.

→

Drop a course folder — it appears on the site

The build scans courses/ for any directory with a course.yaml. Add a folder, rebuild, done.

courses/python/ + course.yaml → auto-discovered

→

Add a content file — the feature activates

Each plugin declares a content pattern. If a course has matching files, it gets the feature. No files? No dead links.

content/flashcards/flashcards.yaml → flashcards page active

Start building your course

Works with Claude Code, Cursor, Copilot, Augment, or any AI coding tool. See STARTER_PROMPT.md to get started.

Fork on GitHub