suggest

Real-time autocomplete suggestions for search-as-you-type UIs are provided by returning query suggestions with response times under 100ms as you type. This Skill handles the fast query refinement needed for responsive interfaces.

Typo-resilient autocomplete handles common typing mistakes directly during the suggestion process without requiring a separate spellcheck step. This ensures users still receive accurate query suggestions despite minor input errors.

Rich entity suggestions enhance search queries by returning metadata like titles, descriptions, and images alongside the autocomplete text. This allows you to build preview cards and detect entities directly from user input.

An API key is required to access rich suggestions and retrieve the enhanced entity information for your search queries. You must configure this key to enable the full autocomplete and entity detection features.

Refining queries for RAG systems using autocomplete involves passing user input to generate typo-resilient suggestions before execution. This ensures the retrieval system receives accurate, entity-rich queries for better generation results.

The suggest Skill works with curl as its core dependency for fetching autocomplete results. You can use curl scripts to send queries and receive the real-time suggestion data for your applications.