Website scraping | ElevenLabs Documentation

Website scraping lets your receptionist learn from your existing website: FAQs, service descriptions, policies, and anything else published on your site.

Source types

When adding a website source, choose:

Type	What it does	Best for
Single page	Reads one specific URL	FAQ page, pricing page, specific policy
Entire website	Finds and reads the pages on your site that contain business information	Your main website

URLs must use HTTPS.

How an entire website is read

Reception.ai reads your sitemap and homepage links, then selects the pages most likely to contain business information, such as services, pricing, team, locations, and help articles. Pages such as galleries, blog posts, and legal notices are skipped.

Scan	When it runs	Pages read
Quick scan	During onboarding	Up to 15
Deep scan	When you add an entire website, or read the rest of your site	Up to 80

Booking platforms and social profiles, such as Facebook, Yelp, Vagaro, Booksy, or Calendly pages, are always read as a single page.

Reading the rest of your site

Onboarding runs a quick scan only, so it’s done in under a minute. To import the remaining pages, select Read the rest of your site on the Home page, or Read more pages on the website source. A deep scan can take several minutes.

Status

While processing, a source shows its current step, such as crawling or extracting. When done, it shows Ready or Error.

Select Show scraped pages on a source to see which pages were read.

If a website can’t be read, check the address. Some sites block automated access, for example with anti-bot protection. In that case, add individual pages or upload the content as a file.

Keeping content current

Websites change over time. For a single page, select Reprocess to read it again. For an entire website, select Read more pages to refresh everything. Do this after changing pricing, services, or policies on your site.