Qdrant

by Qdrant

v0.8.1

Repository Docs

Databases & Data

vector-databasememorysemantic-searchrag

Summary

Qdrant's official MCP server — semantic memory for AI assistants backed by a real vector database.

Features

Store and retrieve memories in a Qdrant collection
Semantic search over stored information
Doubles as a persistent memory layer for agents

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

Open your agent's MCP configuration file.
Merge the snippet below into it, filling in the environment variables with your own values.
Restart the agent — the "Qdrant" tools will be available.

{
  "mcpServers": {
    "qdrant-mcp": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "COLLECTION_NAME": "my-collection"
      }
    }
  }
}

Claude Code

Add the server with a single CLI command.

claude mcp add qdrant-mcp -e QDRANT_URL=http://localhost:6333 -e COLLECTION_NAME=my-collection -- uvx mcp-server-qdrant

Claude Desktop

Add the server to claude_desktop_config.json (Settings → Developer → Edit Config).

{
  "mcpServers": {
    "qdrant-mcp": {
      "command": "uvx",
      "args": [
        "mcp-server-qdrant"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333",
        "COLLECTION_NAME": "my-collection"
      }
    }
  }
}

Cursor

{
  "command": "uvx",
  "args": [
    "mcp-server-qdrant"
  ],
  "env": {
    "QDRANT_URL": "http://localhost:6333",
    "COLLECTION_NAME": "my-collection"
  }
}

Windsurf

{
  "command": "uvx",
  "args": [
    "mcp-server-qdrant"
  ],
  "env": {
    "QDRANT_URL": "http://localhost:6333",
    "COLLECTION_NAME": "my-collection"
  }
}

VS Code

{
  "command": "uvx",
  "args": [
    "mcp-server-qdrant"
  ],
  "env": {
    "QDRANT_URL": "http://localhost:6333",
    "COLLECTION_NAME": "my-collection"
  }
}

Description

Qdrant's official MCP server turns a vector collection into a memory layer an assistant can write to and search, which is a neat inversion of the usual retrieval setup.

The two tools

Store. Save a piece of information with optional metadata. The server embeds it and writes it to the collection.
Find. Retrieve semantically similar entries for a query.

That is deliberately minimal, and it is enough: an assistant that can store what it learns and retrieve it later has persistent memory across sessions without any application code.

Uses

Agent memory — preferences, decisions, and context that should survive a new conversation
Code snippet libraries searchable by description rather than by name
Team knowledge captured incrementally as questions get answered
RAG retrieval over a collection populated by a separate ingestion pipeline

Setup

uvx mcp-server-qdrant pointed at a Qdrant URL, or run fully local with an in-memory instance for testing. Tool descriptions are configurable so the assistant knows what this particular collection is for.