n8n-nodes-postgres-vector-store-tool

import PostgresVectorStoreTool from 'n8n-nodes-postgres-vector-store-tool'; // Register the node with n8n // In your n8n workflow, the node appears as 'Postgres Vector Store Tool' // Configure connection: host, port, database, user, password (use credentials in n8n) // Choose mode: // Regular Retrieval: set tableName, vectorColumnName, contentColumnName, metadataColumnName, topK // Custom SQL: provide SQL query with $1 placeholder for embedding vector // Example custom SQL: // SELECT content, metadata, 1 - (embedding <=> $1) AS similarity // FROM documents // WHERE user_id = 'user123' // ORDER BY embedding <=> $1 // LIMIT 10 // The embedding is automatically passed from the connected embedding node // Output is JSON array in the 'text' field: [{"text": "...", "metadata": {...}}]