vector-assist-generate-query

The “vector-assist-generate-query” tool produces optimized SQL queries for vector search, leveraging metadata and specifications to enable semantic and similarity searches.

3 minute read

About

The vector-assist-generate-query tool generates optimized SQL queries for vector search by leveraging the metadata and vector specifications defined in a specific spec_id. It serves as the primary actionable tool for generating the executable SQL required to retrieve relevant results based on vector similarity.

The tool contextually understands requirements such as distance functions, quantization, and filtering to ensure the resulting query is compatible with the corresponding vector index. Additionally, it can automatically handle iterative index scans for filtered queries and calculate the necessary search parameters (like ef_search) to meet a target recall.

Compatible Sources

This tool can be used with the following database sources:

Source Name
Cloud SQL for PostgreSQL Source

Requirements

Tip

Ensure that your target PostgreSQL database has the required vector_assist extension installed, in order for this tool to execute successfully.

Parameters

The tool takes the following input parameters:

Parameter	Type	Description	Required
`spec_id`	string	Unique ID of the vector spec for query generation.	No
`table_name`	string	Target table name for generating the vector query.	No
`schema_name`	string	Schema name for the query’s target table.	No
`column_name`	string	Text or vector column name identifying the specific spec.	No
`search_text`	string	Text string to search for; embeddings are auto-generated.	No
`search_vector`	string	Vector to search for; use instead of search_text.	No
`output_column_names`	array	List of columns to retrieve in the search results.	No
`top_k`	integer	Number of nearest neighbors to return (defaults to 10).	No
`filter_expressions`	array	List of filter expressions applied to the vector query.	No
`target_recall`	float	Target recall rate, overriding the spec-level default.	No
`iterative_index_search`	boolean	Enables iterative search for filtered queries to guarantee results.	No

Note Parameters are marked as required or optional based on the vector assist function definitions. The function may perform further validation on optional parameters to ensure all necessary data is available before returning a response.

Example

kind: tool
name: generate_query
type: vector-assist-generate-query
source: my-database-source
description: "This tool generates optimized SQL queries for vector search by leveraging the metadata and vector specifications defined in a specific spec_id. It may return a single query or a sequence of multiple SQL queries that can be executed sequentially. Use this tool when a user wants to perform semantic or similarity searches on their data. It serves as the primary actionable tool to invoke for generating the executable SQL required to retrieve relevant results based on vector similarity."

Reference

field	type	required	description
type	string	true	Must be “vector-assist-generate-query”.
source	string	true	Name of the source the SQL should execute on.
description	string	false	Description of the tool that is passed to the agent.

Feedback

Was this page helpful?

Glad to hear it! Please tell us how we can improve.

Sorry to hear that. Please tell us how we can improve.

Last modified April 13, 2026: chore(main): release 1.1.0 (#3024) (da6f5f8)