Skip to main content
A website import copies the text of your site’s pages into Aide’s knowledge, so Aide can answer from them. Start one from Knowledge > New content > Import your website. This article covers the detail: which pages an import reads, what it skips, and what a re-import keeps. For the other kinds of knowledge, see Add knowledge from your help center or website.

​
Start an import

  1. Open Knowledge and click New content, then Import your website.
  2. Enter the address under Your website, starting with https://.
  3. Optional: under Skip pages under, add the paths you don’t want, such as /glossary or /blog/tag.
  4. Click Import pages.
The dialog shows progress and, when the import ends, how many pages it brought in. Aide can answer from them once they are indexed, usually within a minute. Nothing changes on your website. An address can be added once. Adding the same address again is refused with “This address has already been added.”

​
Which pages an import reads

  • Up to 300 pages per import.
  • The address sets the scope. Importing https://acme.com reads the whole site. Importing https://acme.com/help reads only the pages under /help.
  • One site only. Links to other sites aren’t followed. acme.com and www.acme.com count as the same site.
  • Your sitemap is read too. Aide reads the sitemap your robots.txt names, or /sitemap.xml, so it finds pages that nothing links to. A missing or unreadable sitemap is skipped.
  • Support pages come first. The first eight pages Aide finds whose address points to FAQs, help, support, shipping, delivery, returns, refunds, contact, about or pricing are read before the rest.
  • Sections share the limit. Aide takes pages from each section of your site in turn, such as /blog, /customers and /glossary, so one large section can’t use up the limit.
  • Pages with almost no text are left out, and don’t count toward the limit.
A site that shows its text only once a browser runs its JavaScript gives Aide little to read. Aide learns from what the site publishes for search engines, such as its description, contact details and FAQ answers. If that is all the site has, the import says so.

​
What is skipped when Shopify is connected

When your Shopify store is connected, a website import skips product, collection, cart, checkout, account and search pages, including under a market prefix such as /en-ca/. Your products already come from Shopify with current prices, stock and variants. Skipped pages don’t count toward the 300.

​
Excluded paths

An excluded path covers the page at that path and every page under it. /glossary skips /glossary/seo but not /glossary-of-care.
  • A website can have up to 50 excluded paths. Each starts with / and is at most 200 characters.
  • To change them later, open the website source, open the menu above its list of pages and choose Excluded paths. Click Save when you’re done.
  • Adding a path removes the pages already imported under it straight away. Aide stops answering from them. You don’t need to re-import.
  • Removing a path lets later imports read that section again. The pages that were removed when you excluded it stay removed.

​
Re-import

A website doesn’t update in Aide on its own. Open the source and click Re-import. The button reads Importing… while it runs, and its tooltip says when the site was last imported.
  • Each page is matched by its address and updated in place with the site’s current text. Nothing is duplicated.
  • New pages are added.
  • A page you deleted in Aide stays deleted.
  • A star you gave a page stays.
  • A page that has gone from your site isn’t removed by a re-import. Delete it in Aide.

​
Add pages from one address

To bring in a new section without a full re-import, open the menu above the list of pages and choose Add pages from a URL. Enter an address on the same site, such as https://acme.com/customers, and click Import pages. Aide reads that page and the pages under it into the same source. Pages it already has are updated. An address on another site, or under an excluded path, is refused. This partial import doesn’t change the “last imported” time on Re-import.

​
Text inside images

On imported pages, Aide reads the text inside images, such as size charts, price lists and nutrition panels, and answers from it like any other text on the page.
  • Images are read a few moments after the pages are imported.
  • JPEG, PNG, GIF and WebP images are read, up to the first eight on a page.
  • Not read: images smaller than 100 pixels on a side (icons, buttons), SVG graphics, and images embedded in the page’s code.
  • An image is read once. A re-import reuses what was read, unless the image’s address changed.
  • Pages imported before this was available get their images read at the next re-import.
  • If your workspace reaches its daily AI usage limit, or AI is paused, the remaining images wait until the next day (see What counts as a billable AI reply).
Images in articles you write in Aide are not read. Put that information in the article’s text.

​
When an import finds nothing

The source shows why: the address doesn’t exist, the site’s secure connection failed, the site didn’t answer, the site refused Aide (it blocks automated readers or asks for a sign-in), it answered with an error, the address isn’t a web page, or the site has no text Aide can read. Try another address, such as your help center, or write the articles in Aide (see Write articles and snippets, and organise collections).