Skip to main content

Add Web Page Knowledge

TL;DR
  • Add any publicly accessible web page by URL — your agent fetches and indexes the content automatically
  • Reference external documentation, status pages, runbook sites, and wiki articles during investigations
  • One-step setup: provide a URL, name, and optional description
  • Content is indexed immediately and available across all future conversations

The problem: external docs stay external

Your team's knowledge isn't all in one place. Runbooks live on wiki sites. Vendor documentation is on external portals. Architecture diagrams and status pages are scattered across different URLs. When your agent investigates an issue, it can't reference these external resources — unless someone manually copies the content and uploads it as a file.

That manual process is tedious and creates stale copies. The original page gets updated, but the uploaded file doesn't. Over time, your agent's knowledge diverges from the actual documentation your team relies on.

How web page knowledge works

When you add a web page as a knowledge source, the agent fetches the page content and stores it for reference:

  1. You provide a URL, a name, and an optional description
  2. The agent fetches the page content via an anonymous HTTP request
  3. The page content is stored and indexed
  4. Your agent can reference this content during conversations and investigations

The fetch happens at the time you add the URL. Content is stored as a point-in-time snapshot of the page.

Public pages only

The agent fetches pages anonymously — without authentication credentials. Pages that require login, VPN access, or corporate SSO cannot be indexed. If you need to add content from protected pages, copy the content and upload it as a file instead.

What gets indexed

The agent fetches the full page content from the URL you provide. Key details:

AspectBehavior
Content fetchedFull page content from the single URL provided
Link followingNo — only the specified URL is fetched, not linked pages
AuthenticationAnonymous — no credentials sent with the request
Supported protocolsHTTP and HTTPS
Fetch timeout30 seconds
RefreshManual — delete and re-add the URL to get updated content

When to use web page knowledge

Web page knowledge sources work best for:

  • Public documentation — vendor docs, API references, cloud service guides
  • Status pages — service health dashboards, incident history pages
  • Wiki articles — publicly accessible knowledge base articles
  • Architecture overviews — publicly hosted architecture diagrams and design docs
  • Runbook sites — external runbook repositories accessible without authentication

For internal or protected content that requires authentication, use file uploads instead.

Limitations

LimitationDetails
No authenticationCannot access pages behind login, VPN, or SSO
Single page onlyDoes not crawl or follow links to other pages
Point-in-time snapshotContent is not automatically refreshed when the source page changes
30-second timeoutPages that take longer than 30 seconds to load will fail
URL formatMust be an absolute HTTP or HTTPS URL
CapabilityWhat it adds
Connect Knowledge→ Overview of all knowledge source types — files, web pages, and repositories
Upload Knowledge Documents→ Upload files directly or let your agent create documents during conversations
File Attachments→ Share files in chat for immediate analysis context

Get started

Add a Web Page Knowledge Source — step-by-step guide to adding your first web page

Was this page helpful?