How to fork LearnKit AI and add your own lesson templates

LearnKit AI ships as a monorepo under Apache-2.0. Everything is editable. The lesson templates are plain TypeScript objects in packages/core/src/generate.ts. Forking the whole stack and adding your company's specific tools, terminology, and focus areas takes about an afternoon.

Start by cloning and installing: git clone https://github.com/learnkit-ai/learnkit && cd learnkit && pnpm install. The dev server starts with pnpm dev. You will see the full landing page at localhost:3000 and the interactive demo at /demo.

The lesson generator is in packages/core/src/generate.ts. The function buildLessonsForWeek() accepts a week number and returns three Lesson objects. The lesson titles, descriptions, and tool assignments are plain strings - search for "week === 1" to find week one, edit the title and description fields to match your stack. Run pnpm test after every change to catch regressions in the schema.

To add a new role, append a string to the SUPPORTED_ROLES array in packages/core/src/data.ts. To add a new tool, append to SUPPORTED_TOOLS. Both are just string arrays - no type changes required, because the schemas are Zod-validated at runtime.

For white-labeling the React components, the easiest path is to override CSS custom properties. Drop :root { --accent: #7C3AED; --paper: #FAFAF8; } into your global stylesheet and the warm theme will inherit your brand colors. For deeper changes, the component source is in packages/react/src/ - no Tailwind dependency, just inline styles and CSS custom properties.

The whole stack - schemas, core, react, and the Next.js demo app - is under Apache-2.0. Fork it, ship it commercially, embed it in a closed-source product. The only requirement is preserving the license notice. PRs back to main welcome.