Tracks out of the box
One line — oelt.track.score(0.85) — works in SCORM 1.2, SCORM 2004, cmi5,
or no LMS at all. Completion rules are declared, not hand-wired.
An AI-first e-learning authoring toolkit
An open source, AI-first toolkit that lets an AI actually build real, ready-to-ship online courses — accessible, SCORM- and cmi5-compliant, and ready for the LMS you already have. Keep the bespoke. Standardize the plumbing.
You, to your AI assistant
Build me a 15-minute course on GDPR basics with a scored quiz…
oelt.track.score(0.85)
→ LMSSetValue("cmi.core.score.raw", "85")
→ LMSCommit("") ✓ accepted You've seen what Claude can do: interactive scenarios, beautiful one-off learning experiences, in minutes. Then reality hits — the LMS needs SCORM. Completion doesn't report. The quiz score goes nowhere. A screen reader can't navigate it. Legal asks about WCAG. Every generated course is a brilliant prototype that can't ship.
One line — oelt.track.score(0.85) — works in SCORM 1.2, SCORM 2004, cmi5,
or no LMS at all. Completion rules are declared, not hand-wired.
Every interaction component ships keyboard-ready and screen-reader tested to WCAG 2.2 AA. Validation catches what generation misses.
An MCP server, agent skills, and docs written for model consumption. Your AI assistant already knows how to use it — that's the point.
Describe the course to Claude (or any agent) with OELTKit connected.
Accessibility, tracking, and packaging checks run automatically; the AI fixes its own findings.
Download a SCORM or cmi5 package that imports anywhere.
Which standard does cmi5 build on?
Right — cmi5 is an xAPI profile.
Not quite — cmi5 is built on xAPI.
Apache 2.0. Plain web components — no framework, no build step, no subscription. Works with the LMS you already have and the AI assistant you already use.