OpenMAIC is a free, open-source platform that turns any topic or document into an interactive AI-taught course with slides, quizzes, and a live AI teacher, and it just added 1,625 stars in a day on GitHub trending. This guide self-hosts it in about 10 minutes with Node.js, pnpm, and a free Gemini API key.
Read the full story: OpenMAIC Setup: Self-Host a Free AI Course Builder →
Transcript
OpenMAIC turns a single topic, or a pile of your own documents, into a full interactive classroom: slides, quizzes, a whiteboard, and an AI teacher who actually talks you through it. It just landed on GitHub's trending page with over sixteen hundred stars in a single day, on top of almost twenty four thousand total, and the project just shipped a new agent workbench that lets you chat with it while it builds. Here's how to get it running yourself in about ten minutes. First, check your prerequisites. OpenMAIC needs Node point J S version twenty or newer, and P N P M version ten or newer, so run node dash dash version, pnpm dash dash version, and git dash dash version to confirm you have them. Next, clone the repository and install dependencies. Git clone the T H U dash M A I C slash OpenMAIC repo, move into the folder, and run pnpm install. Then configure your environment. Copy dot env dot example to dot env dot local. OpenMAIC needs at least one AI provider key to generate anything, and the fastest free option is Google's Gemini. Sign up at A I studio dot google dot com slash A P I key, grab a key, and add it as GOOGLE underscore A P I underscore KEY in your env file, along with the default Gemini model. Finally, run it. Type pnpm dev, open localhost port three thousand in your browser, and type a topic into the one click generator, or upload a document. Watch OpenMAIC's agents plan, build, and hand you back a full interactive lesson. A few things worth knowing before you lean on it. Every lesson still burns tokens against whichever provider you configured, so it isn't free at scale, just free to install. The default local database token is baked into the public code, which is fine on your own machine but not for a public deployment. And exporting an actual video file needs a separate Docker container, not just the base install. That's the setup.