Build Custom Qdrant Vector Store MCP Server

This n8n workflow demonstrates building your own Qdrant MCP server, going beyond the official implementation by exposing additional APIs like facet search, grouped search, and recommendations. It uses an MCP server trigger connected to five custom workflow tools, combining n8n's Qdrant nodes for inserts and similarity searches with HTTP requests for advanced features, plus Edit Field and Aggregate nodes for proper MCP responses.

Benefits include easy customization and maintenance of a Qdrant collection manager supporting select and create operations. It's ideal for AI-driven applications needing semantic search, customer feedback analysis, or recommendation systems, saving significant development time compared to building from scratch.

Use cases span business intelligence, such as querying customer sentiments on products by company, listing available entities, or comparing feedback across groups. Connect any compatible MCP client (e.g., Claude Desktop) after setting up a Qdrant collection via prerequisites. Example queries: 'List available companies', 'Customer feedback on deliveries from company X', or 'Ease of use feedback from companies X and Y'.