edumints SCORM MCP

An MCP server that assembles interactive, SCORM-compliant e-learning courses. You (or an AI client like Claude) are the author; this server is the assembler. Describe a course as a structured spec — the server validates, renders, and packages a self-contained SCORM zip that runs on any LMS (Moodle, SCORM Cloud, …).

🌐 Languages: English · Türkçe · Español · Русский · 简体中文 · Azərbaycanca · Қазақша · Кыргызча

Open-source, developed by the edumints.com platform. Built to be self-hosted — run it on your own computer or your own server — and open to contribution.

---

The idea (a different approach)

Most e-learning is built by hand in heavyweight desktop tools. Here, an AI client describes the course (objectives, screens, quizzes, branching, media) through the Model Context Protocol, and the server does the hard part: validation, premium theming, accessible HTML rendering, the SCORM runtime bridge, and packaging. The result is a standards-compliant SCORM package — no vendor lock-in.

Author = the MCP client · Assembler = this server.

Features

• 28 screen types — title, content, MCQ, true/false, fill-in-blank, drag & drop, hotspot, branching scenario, accordion, tabs, flashcards, matching, sorting, timeline, lottie, guided software simulation, video, summary, decision scenario, term match race, escape room, labeled diagram, data chart, image compare, results breakdown, poll, composable game, adaptive practice.
• Composable game engine — the game screen composes mechanic primitives (score/lives/timer/ hints) + declarative when event if cond then action rules + branching nodes; the adaptive_practice screen estimates competency (Elo or Bayesian Knowledge Tracing) to calibrate difficulty to the learner. Optional xAPI/cmi5 telemetry, an anti-slop quality gate (lint_course), and game accessibility (WCAG 2.2.1). All deterministic — no server-side LLM. See docs/GAME-ECD.md, docs/GAME-ADAPTIVE.md.
• Slide-stage player — fixed 16:9 stage that scales to any screen, a player bar (play/seek/ captions/menu/replay), and timed timeline reveal synced to narration. Section-grouped outline menu. Adjustable stage size; fully responsive/mobile; inline SVG icons (no emoji).
• Logic & gamification — variables/state, conditional visibility, branching, points & timer HUD.
• Assessment — aligned questions with feedback on correct/incorrect, scoring written to SCORM.
• Media — cross-MCP ingestion (bring audio/image/video from your own MCPs → add_asset), ffmpeg processing, programmatic motion-graphic/data-viz video (HyperFrames), and a built-in Turkish TTS (Piper, offline) for quick narration.
• Theming & accessibility — light/neutral/high-contrast presets, brand tokens, WCAG-minded, prefers-reduced-motion respected.
• SCORM 1.2 & 2004, deterministic packaging, cost guardrails, opt-in/lazy heavy features (nothing loads unless a course uses it).

Quickstart (self-hosting)

Docker (recommended)

git clone https://github.com/kemalyy/edumints-scorm-mcp.git
cd edumints-scorm-mcp
docker build -t edumints-scorm-mcp .
docker run -p 8000:8000 -v "$PWD/data:/data" edumints-scorm-mcp
# MCP endpoint: http://localhost:8000/mcp   ·   health: http://localhost:8000/health

The image includes everything for the optional features (ffmpeg, Node + HyperFrames for video, Piper + a Turkish voice for TTS).

Apple Silicon + Docker Desktop: if docker build/run crashes with Illegal instruction (SIGILL), that's an upstream native-ARM64 issue in the cryptography package's Rust bindings (pyca/cryptography#14733), not this repo. Workaround: docker build --platform linux/amd64 -t edumints-scorm-mcp . (runs under emulation).

Local (Python)

python -m venv .venv && source .venv/bin/activate
pip install ".[tts]"          # ".[tts]" adds the offline Turkish TTS (Piper); omit to skip
python server.py              # serves the MCP over HTTP

For video generation, also install Node 22+ and HyperFrames (npm i -g hyperframes) + ffmpeg.

Configuration

Copy .env.example and adjust (data dir, quotas, base URL, TTL). See the file for all options. No secrets are required to run locally.

Connecting an AI client

Point any MCP client at http://<your-host>:8000/mcp:

• Claude (desktop/web/Code) — add it as a connector / MCP server.
• Antigravity and other MCP clients — same endpoint (HTTP/Streamable).

Then ask: "Create a 6-minute interactive course about X with a quiz and a summary." The client calls the tools below; you get a downloadable SCORM zip.

Pairs with the authoring skill (a Claude Agent Skill that teaches an AI client how to author high-quality courses with this server): https://github.com/kemalyy/edumints-scorm-skill

Example

A full course is one build_from_spec call with a JSON spec (this is examples/small.en.json, trimmed for brevity):

{
  "title": "Introduction to SCORM",
  "scorm_version": "1.2",
  "language": "en",
  "tracking": { "completion_rule": "viewed_all_and_passed", "passing_score": 50 },
  "screens": [
    { "type": "title_slide", "id": "t1", "title": "Introduction to SCORM", "subtitle": "Basic concepts in 5 minutes" },
    { "type": "content_slide", "id": "c1", "title": "What is SCORM?", "body_html": "<p><strong>SCORM</strong> is a standard that lets e-learning content talk to an LMS.</p>" },
    { "type": "mcq", "id": "q1", "title": "Mini Quiz", "prompt_html": "<p>What is the purpose of SCORM?</p>",
      "options": [
        { "id": "a", "text_html": "Content–LMS communication", "correct": true },
        { "id": "b", "text_html": "Video editing" }
      ], "points": 10 },
    { "type": "summary", "id": "s1", "title": "Congratulations", "body_html": "<p>You have learned the basics.</p>" }
  ]
}

build_from_spec(spec) → { project_id, screens: 4, warnings: [] }
build_package(project_id) → downloadable SCORM 1.2 zip
                              ├─ imsmanifest.xml
                              ├─ index.html          (self-contained player + runtime)
                              └─ assets/

The zip imports into any SCORM 1.2/2004 LMS (Moodle, SCORM Cloud, Rustici Engine, …) — no server dependency at runtime. The full runnable spec (with quiz feedback, more screens) is at examples/small.en.json; more real-world specs (games, branching, themed courses) live under examples/.

Key tools (MCP)

Tool	Purpose
build_from_spec	One JSON spec → validated project + packaged SCORM zip (the main path)
create_project / add_screen / update_screen / …	Granular, incremental editing
set_theme / set_tracking	Theming + completion/scoring rules
add_asset	Ingest audio/image/video (data-URI or https, SSRF-guarded)
synthesize_speech	Built-in Turkish narration (Piper, offline) → audio asset
make_video_from_image_audio / render_motion_video / render_screen_video	Video (ffmpeg / HyperFrames)
preview / validate_package / build_package	Preview, validate, download the SCORM zip

Architecture

MCP client (author)  ──►  scorm-mcp (assembler)
                            ├─ core/        models (Pydantic), packaging, storage
                            ├─ components/  HTML renderer + runtime engine + video compiler
                            ├─ auth/        API-key + OAuth, SSRF guards
                            ├─ themes/      design tokens / presets
                            ├─ runtime/     vendored SCORM runtime (scorm-again, MIT)
                            └─ server.py    FastMCP tools (HTTP)

Output: a self-contained index.html + imsmanifest.xml + assets + SCORM runtime, zipped.

Contributing

Issues and PRs welcome. The codebase favors small, focused modules, additive changes, and backward compatibility. See CONTRIBUTING.md.

Testing

Run tests with pytest.

Licenses

• This project: MIT — see LICENSE.
• Bundled third-party components (scorm-again, lottie-web): see THIRD_PARTY_NOTICES.md.

Developed by edumints.com. SCORM is a trademark of ADL; other product names referenced are trademarks of their respective owners (nominative use only).