Qdrant Vector Search
Official Qdrant Model Context Protocol server for semantic vector search, knowledge retrieval, and persistent agent memory.
About this Server
The official Qdrant Model Context Protocol (MCP) server connects AI assistants and coding agents directly to Qdrant vector database instances—whether hosted on Qdrant Cloud, self-hosted via Docker, or running locally on disk. It provides automatic embedding generation via FastEmbed models (e.g. all-MiniLM-L6-v2), allowing agents to store conversational memories, contextual facts, and documents, and retrieve them via semantic similarity search.
Direct low-latency RPC protocol communication with structured JSON schemas.
Restricted capability bounds to prevent accidental modifications or leakage.
Security & Best Practices
When to Use (and When Not To)
- •Semantic search, document retrieval, and persistent agent long-term memory
- •Connecting AI assistants to Qdrant Cloud vector clusters or local fast embedded collections
- •RAG workflows requiring automatic sentence-transformer embedding generation without extra API calls
- •Relational SQL transactions with foreign keys and multi-table joins (use DBHub or Neon MCP instead)
- •Document databases with complex MongoDB aggregation pipelines (use MongoDB MCP instead)
- •OLAP columnar analytical aggregations across huge Parquet files (use DuckDB / MotherDuck MCP instead)
Install & Setup
claude mcp add qdrant -- uvx mcp-server-qdrantExposed Tools & Endpoints
2 toolsFunctions your AI agent can invoke via this Model Context Protocol server
Stores information in the Qdrant vector database with automatic embeddings and optional JSON metadata.
Retrieves the most semantically relevant information from the Qdrant database using natural language queries.