🎓 CloudFran LMS — Theory of Operations

How the compliance pipeline turns "watched a video" into "regulated-industry audit-ready completion evidence."

The lifecycle

📥Enroll

▶️Play

📊Track

🔒Gate

🏆Certify

Every learner action moves through these five stages. The compliance pipeline runs continuously — every second of audio a learner actually hears adds to a unique-seconds counter, not a playhead position.

Anti-skip architecture

Unique-seconds tracker

Each audio element has a Set<seconds>. Only seconds that fire a timeupdate event while the audio is actually playing (not scrubbing) get added. Scrubbing forward to second 900 doesn't add seconds 0–899 — those have to be played.

Max-reached-time anti-skip

Every audio element has a maxReachedTime. On seeking, if the new time is more than 1 second past max-reached, the player snaps back to max-reached. Forward scrubbing is rate-limited; backward scrubbing is always allowed.

80% listen-to-continue gate

The "Next" button is disabled until listenedSegments.size / audioDuration ≥ 0.80 for the current audio. Lock icon + message updates in real time.

Per-audio independence

If a module has 3 audios, each is tracked separately. 80% on each is required, not 80% of the weighted total. Regulated industries require this.

Attribution to the Execra ledger

Every module completion emits an event to the Execra attribution ledger: event=ModuleComplete, actor=learner, course_id, module_id, duration_minutes_consumed, 80%_gate_passed=true. Managers see the rollup on the attribution dashboard; regulators can audit the raw event stream back 7 years.

Download / print gating

Per-module toggle

Every module row has CanDownload and CanPrint booleans. The Viewer UI hides the download + print buttons when the flags are false.

CE-credit compliance

Some certifications require proof that content was consumed in-app. Toggling off CanDownload enforces that without breaking the learner experience for non-restricted content.

AI Tutor

Every course has a right-rail tutor. When a learner asks a question, the tutor's system prompt is augmented with the current module's title + the first 400 words of the iframe content — scoping its answers. Tutor responses are logged (opt-out available) for course-improvement feedback.

Course authoring

AI Course Creator

Describe a topic, get a multi-module course draft — text, audio narration, quiz questions. Edit before publish. Accelerates author throughput 10×.

SCORM upload

Upload existing SCORM packages; the compliance pipeline wraps them with the unique-seconds tracker on import.

Audio-first content

Every course ships with a narrated version by default. Learners can read or listen.

Ready to see it running?
Start a 14-day pilot with up to 50 learners.

See pricing →