Sources

Website

Your site is where your positioning is already written down. Adding it makes your own words searchable — useful when someone asks how you describe a service or what you claim publicly.

Add a site

Sources → Website, paste the address, and start. Vythos finds the site's pages — from its sitemap where there is one, otherwise by following links within the same domain — reads the main content of each, and puts the text into your memory.

By default it takes up to 30 pages, and never more than 100. It stays on the one domain, crawls politely with a pause between requests, and respects robots.txt.

What it's for, and what it isn't

Good for: your service descriptions, your about page, published case studies, your public FAQ.

Not a rule source. Pages from a website can't become rules. A website is a description of what you sell, written for buyers; your rules are the documents your team is held to, and they should come from Drive, Notion or an upload.

Not for other people's sites. Nothing stops you crawling a public site, but memory is for material you're entitled to use, and your own words are what make answers about your agency correct.

Keeping it fresh

Crawls happen when you ask for them. Re-run the crawl after a site redesign or a repositioning; pages that changed replace their old text.

Troubleshooting

Problem	Fix
Very few pages found	The site may have no sitemap and little internal linking. Add key pages individually
A page's text looks thin	Heavy JavaScript rendering can leave little for a crawler to read. Paste the copy into a document and upload it instead
The crawl stopped early	Page limits apply — raise the limit, or crawl a section at a time