Create a course

The four ways to build a course in Coassemble: transform a document, convert a presentation, generate from a prompt, or start from scratch.

There are four ways to create a course. Pick whichever matches what you're starting with, then refine the result in the course editor. Most people mix them depending on the content they have.

Transform a document

Turn a document you already have into a course. Built-in AI reads your content and breaks it into screens. There's a 50MB file size limit for document transformation.

Convert a presentation

Turn your existing slides into a course, so you can reuse a deck you've already built rather than starting over.

Generate from a prompt

Describe your topic and let AI generate a structured course. You set the audience, tone, and length, then review the outline before it builds. If the first outline isn't quite right, use Regenerate to create a different version.

Start from scratch

Build screen by screen yourself, adding screens and content as you go. This gives you the most control and works well when you know exactly what you want.

After you've created a course

However you start, you refine everything in the editor: edit text, swap images, add or reorder screens, and use undo and redo. Your work autosaves as you go.

Frequently asked questions

Which method is fastest?

Starting from a document or a prompt is usually quickest, since both give you a full draft to refine rather than a blank course.

Can I change the content after AI builds it?

Yes. Everything AI generates is fully editable, you can rewrite text, replace images, and add, remove, or reorder screens.

What file types can I transform?

You can transform PDFs, Word documents (.doc, .docx), PowerPoint files (.ppt, .pptx), text files (.txt), and Rich Text Format (.rtf). Document transformation has a 50MB file size limit. For the best result, use well-structured, text-based files: headings and bullet points help the AI break your content into logical screens, and for PDFs use real text rather than scanned images.