Skip to main content

Documentation Index

Fetch the complete documentation index at: https://watermelon.ai/docs/llms.txt

Use this file to discover all available pages before exploring further.

Using the Website source, you can easily “crawl” the pages on your website (using a sitemap, your root domain, or manually added URLs) and integrate that information into your AI Agent. Once crawled, your AI Agent can use this knowledge to answer questions quickly and accurately, and keeps your AI Agent up to date with the latest information on your website without manual updates or documents.
Website is available in all plans. Limits differ per plan, for more information, check our Pricing page.

​
Adding a Website Source

1

Open your AI Agent

Go to Agents in the main menu and select the relevant AI Agent.
Website Menu
2

Navigate to Website

In the Agent menu, click Website to view your website settings, URL list, and crawling options.
Website Website
3

Fetch URLs

Website Sitemap
You can add URLs in three different ways:
  • Option A: Add your sitemap (recommended)
  • Option B: Fetch URLs from your root domain
    • Add your homepage (e.g., https://website.com) and the system will attempt to discover pages across the site.
  • Option C: Add URLs manually
    • Use this for specific pages you want to include without crawling the whole site.
After selecting the type of URL, click **Fetch URL’s. **Fetched URLs appear in a table. This includes:
  • The URL
  • The date added
  • Status indicators
  • Toggles for including/excluding and for link-sharing
Depending on your site size, fetching may take a few moments
If you reach the size limit of your Agent, you’ll need to remove sources to syncronize.
4

Customize URL usage

Website Customize Urls
After a URL is fetched, you can choose at any time to:
  • Include → content will be added to the AI Agent
  • Exclude → page content will be ignored
  • Enable/Disable link usage → controls whether the AI Agent may share this URL in answers
  • Delete URL → Removing a URL also removes its stored content.
These settings let you fine-tune which parts of your website your AI Agent uses.
5

Syncronize Agent

Once your URLs are prepared, select Synchronize Agent to update your Agent.
Website Synchronize

​
Crawl statuses

  • Crawled – Content added successfully
  • Not Crawled – Not processed yet
  • Queued – Waiting to be processed
  • Excluded – Skipped by your choice

​
JavaScript rendering

Some websites only load their content after JavaScript runs, for example when information appears after a click, tab change, or dynamic component. If your website works this way, enable Render JavaScript so the system can correctly load and read that content.
Website Render Javascript Pn
With JavaScript rendering enabled, the system can process content that appears behind elements such as:
  • Buttons
  • Tabs
  • Dropdowns
  • Shadow DOM components
  • Login walls (as long as the content is publicly accessible and not blocked by authentication or security measures)
Keep in mind that rendering JavaScript takes more time and resources, so crawling may be slower when this option is turned on.

​
Important behavior notes

  • The crawling process may take up to 24 hours to complete
  • You do not need to stay on the page or stay logged in; the crawling process will continue automatically
  • Failed pages will be retried up to 50 times before marking a URL as failed
  • If crawling a sitemap or domain finishes within seconds, it may indicate access issues or a technically difficult site.
  • If progress appears stuck at 90–100%, the system is still retrying a small number of slower or temporarily unavailable URLs.
Be sure to re-crawl after website updates to refresh your AI Agent’s knowledge!

​
Troubleshooting

Common reasons a URL cannot be crawled include:
  • robots.txt restrictions
  • The page blocks crawlers.
  • Incorrect or inaccessible URLs
  • Typos, wrong protocol, or pages that do not load.
  • Anti-bot protection
  • CAPTCHAs or bot detection systems block access.
  • Server errors
  • 404, 500, 504, or temporary downtime.
  • IP or geographic restrictions
  • Some sites block scraping or certain regions.
  • Server overload
  • Too many requests can cause timeouts.
Try correcting the URL, batching crawls, or retrying later.