Notion and online course

Sharing settings decide what a course assistant is able to quote

A course provider running on this platform keeps the sales material and the product in the same tree, often a few levels apart. Publishing is inherited downwards, so the decision is made at the top of a subtree and the consequences arrive at the bottom. Nothing about the crawl configuration protects you here, because a crawl reads exactly what an anonymous visitor reads, and the sharing setting decided who that is.

Why this pairing is its own job

The Notion install guide covers the tag, and the online course guide covers what the assistant has to know. What follows is the part that belongs to neither.

  • The syllabus a buyer needs and the lessons a buyer paid for are usually branches of one page tree, which is convenient to write and dangerous to publish.
  • Publishing to the web is inherited by child pages, so one setting at the top of the tree can make an entire course readable without anybody choosing that page by page.
  • A database published to the web publishes its rows, which for a course provider means the cohort you have not announced and the module still being drafted.
  • The install and the material sit on opposite sides of the wall on this platform, which is what makes the sharing question the whole of the work: you are never configuring a widget here, you are deciding what a public reader can see.

What changes about the install here

<script src="https://cdn.askably.xyz/w.js" data-key="pk_live_YOUR_KEY" defer></script>

The tag is the same one on the Notion guide. Everything below is about where it goes on a online course site specifically.

The widget goes on the sales site, and the workspace is only ever a source

Because there is nowhere in a published workspace page to put a script, the two halves of this job separate completely. The install belongs to whatever serves your sales pages and your checkout. The workspace contributes material and nothing else.

That is a comfortable arrangement for a course provider, because the pre sale questions that decide a purchase, what is included, how long access lasts, whether there is a refund window, are all typed on the sales page, which is exactly where the widget can live.

You cannot have the assistant inside the lessons and keep the lessons private

The two ways to put a tag anywhere near this content are to serve the workspace through a hosting layer on your own domain, or to move the lessons somewhere else. The first one makes the lesson pages public, which is the thing you were trying not to do.

So the honest shape is an assistant on the public side answering pre sale and access questions, and a support route for learners who are already inside. Trying to have it both ways is how paid material ends up published.

If your lessons are on a separate learning platform already, none of this applies to them and the workspace is just where your policies happen to live. Index those pages and leave the rest alone.

Where the answerable material lives

Source material on a Notion online course site
The published curriculum pageModule by module, with the format and the realistic length of each part. This is the page that answers whether the course fits somebody's week, and it is the one page in the workspace that unambiguously should be public.
Access terms and the cancellation policy, if they exist as pages at allOn many providers these live in a checkout footer or in an email template rather than in the workspace. Wherever they are, the compliant wording is what has to be indexed, because a paraphrase of a cancellation right is a misstatement of it.
Lesson pages, which have to be out rather than merely unindexedOut means not published to the web. Leaving them published and hoping a crawl misses them is not a control: the pages are readable by anybody with the URL regardless of what any crawler does.
A cohort database, where publishing the database publishes every rowDates, times and time zones for cohorts that have been announced are useful material. The same database usually holds the ones that have not, and publishing is a property of the database rather than of the rows in it.

The first thing to get right

Do this first
Move the lesson tree out from under anything that is published, so that no published page is an ancestor of a paid one.

Sharing is inherited downwards, which means per page settings are a promise about the current shape of the tree rather than a rule. Somebody moving a page during a redesign changes what is public without touching a single sharing setting. Separating the trees makes the guarantee structural instead.

The failure that belongs to this combination

Module three, quoted to somebody who has not bought the course

A prospect asks what the third module covers. The answer describes it properly, because the lesson page was published as a child of the syllabus page months ago and a crawl read it like any other public page.

The citation is the worse half. It is a working link to the lesson, so the visitor does not need the answer, they can simply open the material. From the assistant's side nothing went wrong: it quoted a public page and cited it, which is what it is for.

Check this before launch by opening your own published pages in a private window while signed out of the workspace. Whatever you can read there is what the assistant can read, and what it can read it can quote with a link attached.

Before you go live

  • The origin the page is served from has to be on the allowlist for that assistant, or nothing renders and the browser console says which origin was refused. An apex domain and its www are two different origins to a browser, so list both, along with any staging or preview host you want it to work on.
  • Open the site as a visitor would, on the pages a online course visitor actually lands on, and ask it something only your own material could answer. A widget that renders is not the same as a widget that has read anything.

Questions

Can it read the lessons without quoting them?
No. Indexed material becomes answers, and every answer carries numbered citations to the pages it used, so there is no state where a page informs an answer without being quotable and linkable. If it must not be quoted, it must not be indexed, and on this platform that means it must not be published.
We turned off search engine indexing on the site. Is that enough?
No. That setting is about search engines, not about access. The page is still served to anybody who has the URL, and URLs travel: in a refund thread, in a forwarded email, in a screenshot. Decide this with sharing settings and page structure, which control who can read the page at all.
Where does a learner get help once they are inside the course?
From a person, in most cases. The assistant cannot see enrolments, progress, marking or certificates, and on this platform it is not on the lesson pages either. Point it at the published access terms and technical requirements so it clears the common pre sale and setup questions, and route anything about a specific learner to support with their email already captured.

Keep reading

Try it on your own material

Upload a document or point it at your site, paste one line of HTML, then ask it something only your business could answer.