Skip to main content

Knowledge

Website Knowledge Sources

Website sources let you crawl one page or an entire domain and turn that content into searchable knowledge for your agents.

Crawl Modes

KriyaOS supports two crawl modes:

  • single: index only the provided page URL.
  • domain: follow links within the same domain (bounded by limits).

Crawl Limits

To keep ingestion predictable, website sources have limits such as:

  • max depth (how many link-hops)
  • max pages (how many pages to index)

Rendered Crawling

Some pages require JavaScript rendering to show real content. If you have a modern marketing site or a SPA, enable render to improve extraction.

Recommendation

Start with single + a small page limit to verify content quality, then expand to a domain crawl if results look good.

Include and Exclude Patterns

Use include/exclude patterns to control what gets indexed. This helps you avoid low-value pages like tag archives, privacy pages, or parameterized URLs.