The Search Configuration screen controls how pages are categorised in search results and how user queries are interpreted before being sent to the search engine. You can reach it from the Account menu in the admin console, or open it directly at admin.airgentic.com/edit_search.
Categories let you group indexed pages into named buckets (e.g. Products, Support, News) that appear as filters or labels in search results. Each category has matching rules that determine which pages belong to it.
When enabled, Airgentic analyses each page's URL and content during indexing and automatically assigns a category. AI-generated categories appear in the Auto Mappings section below the manual ones.
When disabled, the legacy URL Path Depth setting becomes available as a fallback.
Only shown when AI Auto-Categorization is disabled. Automatically categorises pages based on their URL path structure.
Set a number to pick the URL path segment used as the category name:
1 for https://example.com/services/complaints/page.html → category Services2 for the same URL → category ComplaintsLeave blank to disable URL-depth categorisation.
Manual mappings are human-defined rules evaluated in order. The first rule that matches a page's URL or metadata wins.
The fallback label applied to pages that don't match any manual or auto mapping. Set a Default Type Title (e.g. General) so uncategorised pages still have a meaningful label rather than appearing blank.
Click Add Type to create a new mapping. Each mapping has:
| Field | Description |
|---|---|
| Display Title | The category label shown to users (e.g. Products). |
| Search Boost | A ranking adjustment from −5 (demote) to +5 (promote) applied to all pages in this category. Use 0 for no change. |
| URL Contains | If any of these strings appear anywhere in the page URL, the mapping matches. Add multiple values — any match is sufficient. Entries can be reordered by drag-and-drop. |
| Metadata Contains | Matches if a specified metadata field contains a given value. Enter a field name (e.g. og:type) and a value (e.g. article). |
| Metadata Is | Matches if a specified metadata field is an exact match for a given value. |
| Metadata Matches (Regex) | Matches if a specified metadata field matches a regular expression (e.g. (?i)gazette for a case-insensitive match). |
Click Remove Type at the bottom of a mapping to delete it.
Visible only when AI Auto-Categorization is enabled. Shows categories automatically generated during the last crawl, including how many documents each covers.
Auto mappings are read-only — they are regenerated by the AI each crawl. Two actions are available for each:
Both actions require a reindex to take effect.
Expand Advanced Options to configure XPath expressions that extract category names directly from page HTML at crawl time.
//meta[@name='document-type']/@content reads the document-type meta tag.Synonym rules let you modify user queries before they reach the search engine, improving recall when users phrase things differently from how the content is written.
When a query contains term X, also search for term Y (in addition to X). Use this for abbreviations, alternative spellings, or related terms.
X = "FAQ" → Y = "frequently asked questions"When a query contains term X, replace it with term Y. X is removed from the query and Y is used instead.
X = "T&Cs" → Y = "terms and conditions"Remove a term from all queries. Useful for filtering out noise words specific to your domain.
"pdf" so searches for "application form pdf" become "application form".Click Save Configuration. If you have changed any category mappings (Manual Mappings or XPath-Based Categories), a confirmation dialog will appear with an option to Initiate index update now. Leave this checked to immediately start reindexing so the new categories take effect. Uncheck it if you want to batch changes and trigger reindexing manually later.
Changes to Search Synonyms take effect without reindexing.