Search
AI-generatedSummary
Execute a vector search query against a specified knowledge base to retrieve the most relevant documents or chunks.
Inputs
- Knowledge Base ID (required) — ID of the knowledge base to search
- Query (required) — Text string to perform the search query with
- Top K — Number of top results to return (minimum 1, maximum 20, default 5)
- Search Skip — Number of results to skip for pagination (minimum 0, default 0)
- Metadata Filter — Optional JSON object to filter results by metadata fields
- Coalesce Neighbors — Override the knowledge base's default setting for number of neighboring chunks to coalesce (integer 0-5)
Output shape
a list of search result items from the knowledge base, each containing metadata and relevance information
Returns up to Top K results matching the query text. Supports pagination via Search Skip. Metadata Filter narrows results by matching metadata fields. Coalesce Neighbors influences context chunk grouping as an override.