Skip to main content
The AI generator turns a topic, a sentence, or a full brief into a finished course through a short back-and-forth. You answer a few clarifying questions, pick a narrative angle, review a full proposed outline, and approve it. From there the lessons are written for you and appear one by one. You can ground the whole thing in your own PDFs and in your Knowledge sources so the result sounds like you, not like generic AI filler.
AI course generation runs on AI credits and is gated by a capability your plan must include. Paid plans (Pro and Scale) include it; see AI credits and access below.

Where to start

Open Courses in your academy and start a generation. You land on a full-page chat that asks “What course are we creating today?” Type your idea in your own words: a one-line topic or a longer brief both work. Three quick-start cards sit under the input:

From my catalog

Asks the agent to suggest a new course that complements what you already teach, grounded in your existing catalog.

From engagement

Turning what your students engage with into a course. Marked Soon.

From scratch

Skip AI and start with a blank course you build yourself.
You do not have to write a polished brief. A topic line is enough; the clarifying questions in the next step exist to fill in the rest.

The conversation

After you submit, the agent works through a guided sequence. At each step it either shows you selectable cards or waits for your reply. You can always answer in free text instead of clicking a card, and you can steer with a chat message at any gate.
1

Clarifying questions

The agent asks up to four short questions to sharpen the course: who it is for, scope, practice style, and anything specific to your topic. Each question renders as a card with selectable options, and every question is skippable. Your answers feed the rest of the build.
2

Narrative angles

The agent proposes three to four angles for the course, each a short bold label with a one-sentence framing. Pick the one that fits, or describe a different direction in chat. If you steer it, the agent proposes a fresh set.
3

Full outline

Once the angle is locked, the agent proposes the complete outline: a genre, a voice card, a structural card, a transformation promise, and the full lesson sequence with each lesson’s title and objective. Review it and either approve, or ask for changes (reorder, add or drop lessons, shift the focus). The agent re-proposes the whole revised outline until you are happy.
4

Approve and draft

When you approve the outline, the lessons are written and saved one at a time, appearing progressively as they complete. A full course can take roughly ten to thirty minutes depending on length. You can leave the page: drafting continues in the background, and reopening the generation reconnects you to the live progress.
The outline gate is where you do the steering. Nothing is written until you approve, and an approved outline can still be reworked by sending feedback before you approve a revision.

Ground it in your own material

The course is far more “yours” when the agent reads your real frameworks, terminology, and examples. You can attach source material from the + (Add source material) control in the chat input before you start, or seed it from what you already have saved.
Upload up to 2 PDFs per generation, with a combined limit of 80 pages and up to 25 MB per file. Password-protected PDFs are not supported: remove the password and re-upload.
The agent treats attached material as the authoritative voice and content backbone for the course: frameworks, terminology, examples, and stances in it carry through to the lessons rather than being replaced with generic equivalents. Your saved Knowledge identity (audience, voice, tone) is also folded in so the generation matches how you teach, the same context that powers AI customization.
Attached source material is read at generation time. Add or remove sources before you start a generation; sources cannot be attached after drafting has begun.

What you get

Generation creates a new draft course in your academy, populated with the drafted lessons. It is a normal course from there: edit any lesson, reorder them, set access, and publish when ready. Like every course you build, the result is your own editable copy, so refining it never affects anything in the Explore Library.
Cancel a running generation from the overflow menu at the top right. Cancelling keeps any lessons already saved, so a partially drafted course is preserved rather than thrown away.

AI credits and access

AI course generation is a capability gated by your plan. Paid plans (Pro and Scale) include it. On a plan without the capability, the chat input is walled with an unlock prompt and the generator will not run.
Generation draws from your monthly AI credit pool, roughly one credit per lesson drafted. Pro includes 100 AI credits and Scale includes 300. You can see your remaining balance from the usage gauge in the chat input before you start. If you are out of credits, the generation will not run until your pool refreshes or you upgrade.
Nothing about generation is visible to your students. They only ever see the finished, branded course once you publish it.