Model Context Protocol moderate risk

Qdrant

Qdrant MCP server provides semantic memory via vector search, enabling storage and retrieval of information using natural language queries.

Connections & Capabilities

Connects To

GitHubDockerAWS

Capabilities

readwrite

Quickstart

Install

npx @smithery/cli install mcp-server-qdrant --client claude

Config

{
  "qdrant": {
    "command": "uvx",
    "args": ["mcp-server-qdrant"],
    "env": {
      "QDRANT_URL": "https://xyz-example.eu-central.aws.cloud.qdrant.io:6333",
      "QDRANT_API_KEY": "your_api_key",
      "COLLECTION_NAME": "your-collection-name",
      "EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2"
    }
  }
}

Exposed MCP Tools (2)

moderate

qdrant-store

Stores information with optional metadata in the Qdrant database.

Writes data to the Qdrant database, potentially overwriting existing information.

safe

qdrant-find

Retrieves relevant information from the Qdrant database based on a query.

Read-only operation; retrieves data but does not modify it.

Safety Assessment

This server is relatively safe for read and write operations to a vector database. The primary risks involve exposure of the Qdrant API key and potential network vulnerabilities. Using a local Qdrant instance mitigates network risks, but proper key management is still crucial.

Uses environment variables for configuration, avoiding hardcoded secrets.
Provides read and write operations within the Qdrant database, but doesn't execute code.
Can be configured with specific collection names, limiting scope.
Based on FastMCP, which offers some basic security features.

Requires network access to the Qdrant server if not using local mode.
API key for Qdrant server needs to be managed securely.
No explicit input validation or sanitization mentioned.
Potential for data exfiltration if Qdrant server is compromised.
Lack of RBAC within the MCP server itself.