Qdrant MCP Server

By thumbsupordown

The official Qdrant MCP server turns a Qdrant vector database into a semantic memory layer for AI agents. Agents can store information with optional metadata into collections and retrieve the most relevant entries through semantic search — a minimal, official building block for embedding-based memory and retrieval workflows.

Install Qdrant

Published by: Official (vendor)
Transport: Local process (stdio)
Authentication: API key
Package: mcp-server-qdrant

Claude Desktop, Claude Code and Cursor

Add this to the mcpServers object in your client's config file, then restart the client.

{
  "mcpServers": {
    "qdrant": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "${QDRANT_URL}",
        "QDRANT_API_KEY": "${QDRANT_API_KEY}",
        "COLLECTION_NAME": "${COLLECTION_NAME}"
      }
    }
  }
}

VS Code uses a different key — show that config

Identical entry, filed under servers rather than mcpServers. Put it in .vscode/mcp.json for one workspace.

{
  "servers": {
    "qdrant": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "${QDRANT_URL}",
        "QDRANT_API_KEY": "${QDRANT_API_KEY}",
        "COLLECTION_NAME": "${COLLECTION_NAME}"
      }
    }
  }
}

Before it will answer

QDRANT_API_KEY is only needed for a remote cluster. Set QDRANT_URL or QDRANT_LOCAL_PATH — one or the other, never both, which is the misconfiguration the server rejects on startup. EMBEDDING_MODEL defaults to sentence-transformers/all-MiniLM-L6-v2 and is downloaded on first run.

Documentation ↗ Source ↗ Config checked against vendor docs August 25, 2026 · how we verify

Tools

qdrant-find

Retrieve relevant information from Qdrant via semantic search.

qdrant-store

Store text with optional metadata as vectors in Qdrant.

Want to create your own MCP server?

Create an account to add and manage your own MCP servers.