Courses as structured data, not documents
Every Coursecast draft is structured from the start. Export it as JSON today and map it into your LMS, wiki or internal tools. A documented schema, API and webhooks are on the roadmap.
// the shape of the JSON export (values shortened) { "course_title": "Incident response 101", "audience": "Support agents", "summary": "How to classify and escalate P1 incidents.", "learning_objectives": ["Escalate a P1 within the policy window"], "review_notes": ["The source doesn't define 'acknowledge'."], "lessons": [{ "title": "The 15-minute notification rule", "objective": "Notify Tier 2 within the window", "estimated_minutes": 3, "narration_script": "When a ticket is classified P1…", "key_points": ["The clock starts at classification"], "knowledge_check": [{ "question": "A P1 is classified at 10:02…", "options": ["By 10:32", "By 10:17"], "correct_option": 1, "explanation": "Policy §4.2: within 15 minutes." }], "review_notes": [] }] }
Field names match the app's JSON export; values are shortened examples. The format may change during early access.
JSON export
The whole course: title, audience, summary, objectives, review notes, and every lesson's script, key points and knowledge checks with the correct option and explanation.
Markdown export
Clean, structured Markdown for wikis, Notion, static sites and docs-as-code pipelines.
Learning IR schema
A published, versioned schema for courses with stable IDs and provenance.
REST API
Create drafts, read courses and releases, and trigger compiles from your systems.
Webhooks
Events for source changes, Training PRs, approvals and releases.
MCP server
Let AI assistants read approved, released course content through the Model Context Protocol, with citations.
Primitives, not just "generate course"
The lifecycle as resources you can program. Design notes, not a published API.
Sources
Register, version and own source documents.
Claims & obligations
Extracted statements with passage links.
Change events
Diffs, materiality and blast radius.
Courses (Learning IR)
Typed objects with stable IDs.
Evals
Run checks, read results, add your own evaluator packs.
Releases
Manifests, approvals, compiles and rollbacks.
Rules SDK
Write your own policy rules as code for the eval lab.
Renderers & deployment adapters
Add output formats and destinations.
Source connectors
Bring a new repository into the graph.
Map the export into anything
Teams already use the JSON export to seed LMS imports, generate wiki pages and feed internal knowledge tools. A tiny example in JavaScript:
No API key, no SDK: it's a file you own.
// turn an exported course into Markdown headings + quiz CSV rows import course from "./course.json" with { type: "json" }; for (const [i, lesson] of course.lessons.entries()) { console.log(`## Lesson ${i + 1}: ${lesson.title}`); for (const q of lesson.knowledge_check ?? []) { console.log([q.question, ...q.options, q.options[q.correct_option]].join(",")); } }
Questions
Is there an API today?
Will the JSON format change?
What is the MCP server for?
Which model does Coursecast use?
Building on top of training content?
Tell us what you'd do with an API. We read every message.