Skip to main content

Index documents (content payload)

POST 
/content-assistance/index/documentJson

Indexes text content supplied directly in the request body, rather than crawling documents from URLs. The texts are chunked and stored for use in retrieval-augmented generation.

Use this endpoint instead of Index documents when the content is not available at a public URL — for example internal documents or text generated on the fly.

  • For GENERAL type: only documents is required.
  • For CATEGORY / CATEGORY_SHARED type: provide documents plus either scopeIds (category concept IDs) or breadcrumbs.
  • For PRODUCT type: provide documents plus either scopeIds (product IDs) or identifiers.
  • For CONDITIONAL / CONDITIONAL_SHARED type: provide documents plus scopeIds (the IDs of the Conditional Documents the content should be attached to, as listed under AI Knowledge Base in Data Platform).

The _SHARED variants index content across all locales (the project group); the non-shared CATEGORY and CONDITIONAL variants apply to the current locale only. GENERAL and PRODUCT are always locale-specific.

This operation is asynchronous and returns a task that can be monitored via the Tasks endpoint.

Request​

Responses​

OK — indexing task started