> For the complete documentation index, see [llms.txt](https://docs.varbase.vardot.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.varbase.vardot.com/content-designers/search-forms-and-email/site-search.md).

# Site Search

Not every visitor uses the menu; many go straight to search. Varbase ships a search page with filters, in every language, built on **Search API**. This page shows how it works for visitors and editors, and what to plan.

## Searching as a Visitor

The search control in the header opens a box, and results appear on the search page (`/search`). Each result shows the title, a short extract and the type of content. Filters on the side narrow results by **Tags**, **Content Type** and **Date Published**, and show only values that exist in the current results.

<figure><img src="https://410909369-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FzBcuWLZ87gvtX5vA2YGD%2Fuploads%2Fgit-blob-dbfb0db240310d58107e688173873561c0053232%2Flearn-17-search-page.png?alt=media" alt="The Search page with a Search by keyword box and button, and filters for Tags, Content Type and Date Published with options such as Any time, Past 24 hours, Past week, Past month and Past year"><figcaption><p>The Search Page and Its Filters</p></figcaption></figure>

Search respects the language of the page: on the Arabic site it finds Arabic content.

## How It Works

* Content is copied into an **index** designed for fast searching. On a Varbase Starter site the index lives in the site's own database, the right choice for most sites.
* When content is published or changed, it is queued and the site's scheduled tasks (cron) update the index, usually within minutes. Administrators see indexing status under **Configuration** \ **Search and metadata** \ ***Search API***: the **Content** index (`/admin/config/search/search-api/index/content`) shows **Index status**, with **Index now** and **Queue all items for reindexing** for a re-index after a large change, such as a migration. The second index there, **Content vector (Ask AI)**, holds content for Ask AI.
* The search page is a Drupal Canvas page made of components: the search box, the results and each filter. Editors can add an introduction or reorder filters without a developer. See [Live Components](https://docs.varbase.vardot.com/content-designers/building-pages-with-drupal-canvas/component-catalog/live-components.md).

<figure><img src="https://410909369-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FzBcuWLZ87gvtX5vA2YGD%2Fuploads%2Fgit-blob-1a6edc77b98ba779a655b7672d114b62dc8cc78a%2Flearn-17-search-api-indexes.png?alt=media" alt="The Search API page listing the Database server and Content index used by site search, and a Content vector (Ask AI) server and index used by AI answers"><figcaption><p>Search Servers and Indexes</p></figcaption></figure>

## Scenario: A New Post Does Not Appear in Search

1. Check the post is published. Site search shows only published content, as visitors see it; drafts are found in the content list filters instead.
2. Check that **Prevent this node from being indexed.**, under **Search API Exclude** on the post's form, is not ticked. It keeps an item out of site search, such as a thank-you page (it does not affect outside search engines).
3. Wait for the scheduled tasks to run, or ask an administrator to check the index status.
4. If the problem continues, ask the development team to check cron on that server.

## Learning From What Visitors Search For

**Reports** \ ***Top search phrases*** (`/admin/reports/search`) shows what visitors type, in their own words, and what they cannot find in the menus. Review it monthly: improve titles and descriptions, add missing content, or rename menu items. If the report is empty on a new site, ask the development team to check that search phrases are being logged.

## Good Results Start With Good Content

Clear titles, a short description, the main words near the start of the text, and accurate tags make results easy to scan and filters useful. When search seems poor, the fix is often in the content, not the software.

## Planning Notes

* Decide which content is searchable and which is not, and which filters visitors need. Filters come from the content model.
* Does search need to cover documents or content from other systems? Larger or multilingual sites may need Solr or AI search; AI search adds a vector index (see [AI Agents, Answers and Automation](https://docs.varbase.vardot.com/content-designers/ai-in-varbase/ai-agents-answers-and-automation.md)).
* Test search in each language with real phrases.
