Postgres Vector Store Tool icon

Postgres Vector Store Tool

Vector similarity search tool for AI agents with pgvector support

Actions2

Custom SQL

AI-generated

Summary

Execute a custom SQL query on a PostgreSQL vector database using a single embedding vector as a parameter to perform vector operations such as similarity search.

Inputs

  • agentDescription — Optional description to guide how the AI agent should use this tool.
  • sqlQuery (required) — Custom SQL query string that must include a $1 placeholder for the embedding vector parameter.
  • customOptions.debug — Enable detailed logging for troubleshooting by setting debug mode to true.

Output shape

a single JSON string representing the query result set

The SQL query must include exactly one parameter ($1) representing the embedding vector serialized as a JSON array string. The output is a JSON string formatted with indentation, containing the rows returned by the query. Errors will throw exceptions with descriptions. Debug logs are available when debug mode is enabled.

Examples

Example 1: Perform a vector similarity search with custom SQL

Specify the SQL query ordering rows by similarity to the provided embedding vector using the <=> operator, for example: SELECT * FROM n8n_vectors ORDER BY embedding <=> $1 LIMIT 10. Set debug mode to true to receive detailed execution logs.

Discussion