Ferndesk
Search and AI optimization

Search your help center

Readers search your public help center from the search dialog. Results can include both articles and API reference operations in one ranked list, and off-topic queries can return fewer matches or none.

On your public help center, open search with ⌘K (Mac) or Ctrl+K (Windows/Linux). The dialog shows a search field at the top.

The placeholder is Search for an article or ask AI... when Ask AI is available, or Search for an article... otherwise.

This is different from thread search in the Fern workspace, which only searches Fern threads, not published documentation.

What appears in results

Matching results can mix two types on the same page:

  • Articles: the article title and a highlighted snippet

  • API reference operations: the HTTP method, path, and an optional highlighted snippet

Only content that is public and visible to the reader is eligible. Locked or private content does not appear in public search.

Article and API results are ranked together, so the strongest match for the query can be either type. Relevant API operations can appear above articles when they fit the query better, and the reverse is also true.

When search returns fewer results

Search keeps results that are relevant to the query. It does not always fill the page.

Queries about your product should surface the closest articles and API operations. Queries with no matching help content can return a short list or none. When nothing matches, the dialog shows No articles found.

If results look different from an earlier visit, ranking may have changed which matches appear first, or weak matches may no longer appear.

Language

Search follows the language the reader is viewing. When a reader uses a non-default enabled language, article candidates are limited to that language before results are ranked.

Improve what readers find

Search quality depends on published content. To strengthen results:

  • Publish clear articles for the jobs readers search for

  • Keep API references public when readers should discover those operations in search

  • Review missed searches in analytics to spot content gaps

For how search engines and AI tools discover your help center outside the in-product dialog, see SEO & GEO (AEO) optimization for your help center.

Was this helpful?