Indices and environments
Keep documentation, blog and other search experiences independent inside one Project.
A named index is an independent search experience within a Project environment. Use separate indices when two parts of your website need different content, relevance, suggestions or reports.
For example, one Production environment can contain Documentation and Blog indices. Each search interface connects to its own index. Multiple sources inside Documentation can still feed one shared documentation search.
What stays separate
Each index owns its documents, publication history, crawlers, relevance settings, suggestions, connection credentials and Search analytics. Publishing to Blog does not replace Documentation. Creating a new index does not copy another index’s content or configuration.
Environments add another boundary. Production and preview can have the same index names without sharing data or credentials. Always check both selectors before changing an index.
Select an index in the application
Use the index switcher at the end of the header breadcrumb (Project › Search › index) to switch indices, open All indices or choose Create index to add one with a name and slug; Search in the breadcrumb also lists all indices. Follow its setup steps, publish content and connect the website using that index’s keys in Settings. Indices cannot currently be renamed or deleted, so choose a lasting name and slug.
For two search boxes on one website, keep their connections separate. A server proxy must route each interface to the intended index; do not silently reuse a default index or one private query key for both. When the CLI installs a key for the second search, add --env-prefix SITEOS_SEARCH_BLOG (or another readable prefix) so its variables do not collide with the first search:
The CLI reports the variable names it installed. Existing installations keep their names, and local installation does not set your hosting environment variables.
Select an index in the CLI
After signing in and selecting your Project, list the indices in the required environment:
Use the returned ID explicitly in index operations:
Replace INDEX_ID with the returned ID. Without --index, existing CLI integrations use the Default index; they do not infer your intended index from a display name. An invalid explicit ID fails rather than falling back to another index.
Keep the same explicit index in import, crawler, relevance, suggestions, credential and reporting operations. See Content and indexing and Website connection.