What is a full-text payload index used for, and how does it differ from an exact-match keyword index?

A full-text payload index is designed for searching textual content by its words or tokens. It is useful when a query should match text content rather than require the entire payload value to equal a specific string.

A keyword index is intended for exact categorical matching. A field such as status=published or category=electronics should normally be treated as a keyword because the application cares about equality, not token-level text search.

The trade-off is search semantics and index behavior. Full-text indexing provides text-oriented matching but has to tokenize and process text, while keyword indexing is simpler for exact values. If a field needs both exact and full-text semantics, separate representations can be clearer than forcing one index to do both jobs.

A common mistake is assuming that a full-text payload index turns Qdrant into a general-purpose search engine. Its text capabilities are designed for payload filtering and should be evaluated against the application's language-analysis requirements.