How to build a SCORM course with AI that actually tracks

Quick answer

Have your AI scaffold a course, install a coursetemplates block for each interaction, and wire completion with @coursetemplates/scorm. Then run coursekit check to catch silent failures and coursekit pack to build the LMS zip. The blocks handle the SCORM/xAPI tracking for you.

Key facts

You keep the learning-design decisions; the AI does the execution.
Blocks report completion + score through one API over SCORM 1.2/2004/cmi5/xAPI.
coursekit check catches the tracking failures AI usually ships.

1. Install a block

Ask your AI (Claude Code / Cursor) for the interaction you need:

npx shadcn add knowledge-check

2. Wire completion

The block calls the tracker; you decide when the learner passes:

import { createTracker } from '@coursetemplates/scorm'
const course = createTracker()
// ...on finish:
course.passed({ scaled: 1 })

3. Check + pack

Verify it tracks, then build the SCORM zip:

npx coursekit check .
npx coursekit pack .

FAQ

Which AI?

Any. The blocks are model-agnostic — Claude, GPT, Gemini — via the coursetemplates MCP server or a plain copy-paste.

Start with a course that already works. Download it, tell your AI what to change, and it tracks in your LMS without you configuring anything.

Browse the templates

Updated 2026-07-01