Plan requirement
| Subscription | Suite Team or higher, Knowledge Professional or higher |
| Access | Admin |
There are exactly two ways to get another system's content into help centre search: the web crawler, which needs no developers, and the Federated Search API, which needs a middleware layer somebody builds and maintains. Which one you can use is usually decided by the content rather than by preference.
Choosing between them
| Web crawler | Point it at public pages. No developer resources. Bounded by 500 URLs per crawler, four levels deep, 50 crawlers per account. |
| Federated Search API | Your developers push records of external content into the search indexes. No public-page requirement, but the integration is yours to build and keep working. |
If the content is on a public site, start with the crawler. If it is not, the API is the only route, and that is a project rather than a setting.
The connector you may be looking for is a different feature
Signing in to a system such as Confluence and picking spaces or folders is a separate capability that feeds an AI agent's generated answers. It does not put that content into help centre search for end users. Do not go looking for a sign-in flow in your search settings, because there is not one.
That older way of connecting sources went into legacy status on 31 July 2025 and stops working on 10 December 2026. If an AI agent of yours draws on content connected that way, the move to the current knowledge sources is a dated job rather than an open one.
What the current connectors ask of you
An account can hold up to 50 external content sources. Zendesk supplies dedicated connectors for the systems it supports, and a universal connector for anything else with a usable REST API; that universal route needs a Knowledge admin to set it up and does a full refresh every twenty-four hours, so a correction made in the source system is not visible the same afternoon.
For Confluence, sign in with an account that has the right permissions on the content you want, and name the site. Personal spaces cannot be connected, which catches out anyone who drafted the documentation in their own space and never moved it.
Every record carries a source and a type
External content is associated with a content source, meaning where it came from, and a content type, meaning what kind of thing it is. Those two are not bookkeeping: search results group external content under its source, and readers can filter their search by type. You can also configure which external sources are included or excluded.
So decide the source and type names as though a customer will read them, because a customer will.
Scope it narrowly
The step that determines whether this helps or hurts. A crawler aimed at a whole site brings in drafts, old campaign pages and the careers section, and all of that then competes with your articles.
Point it at the part of the site that holds current, reviewed content, and nothing else.
The question to settle first
Would you be comfortable sending a customer to that content as it stands? If the honest answer is no, connecting it puts the content in front of them anyway. A source worth connecting is one somebody maintains.
Watch what it displaces
Every external result competes with your own for the same attention. If your articles start appearing below documentation from elsewhere, either narrow the source or improve the article.
Comments
0 comments
Article is closed for comments.