Qdrant MCP Server

What is Qdrant MCP Server?

A server implementation that supports both Qdrant and Chroma vector databases for storing and retrieving domain knowledge. It uses Qdrant's built-in FastEmbed for efficient embedding generation without external API keys, and is designed for integration with AI IDEs like Cursor and Claude Desktop.

How to use Qdrant MCP Server?

Clone the repository, install dependencies, create a .env file with your database type, URL, API key, and collection name, then build and start the server with npm start. Use the API endpoints /api/store and /api/query for domain knowledge, or run npm run store-doc <file-path> to store documentation files.

Key features of Qdrant MCP Server

• Supports both Qdrant and Chroma vector databases
• Uses Qdrant's built-in FastEmbed for embedding generation
• Domain knowledge storage and retrieval via API
• Documentation file storage with metadata (PDF and TXT)
• Configurable database selection via environment variables
• Cursor and Claude Desktop integration support

Use cases of Qdrant MCP Server

• Store domain-specific knowledge in a vector database for AI IDE contexts
• Retrieve relevant knowledge context from past conversations or documentation
• Ingest PDF and TXT documentation files with file metadata into the database
• Integrate knowledge retrieval directly into Cursor or Claude Desktop workflows

FAQ from Qdrant MCP Server

What vector databases does Qdrant MCP Server support?

It supports both Qdrant and Chroma vector databases, selectable via the DATABASE_TYPE environment variable.

What are the prerequisites for using Qdrant MCP Server?

Node.js 20.x or later, npm 10.x or later, and a running Qdrant or Chroma vector database instance.

How does embedding generation work in Qdrant MCP Server?

It uses Qdrant's built-in FastEmbed, which employs quantized model weights and ONNX Runtime for inference, eliminating the need for external embedding API keys.

What file formats are supported for documentation storage?

PDF and TXT files are supported, with automatic content extraction and metadata storage (source, file name, extension, file size, last modified date, creation date, and content type).

How do I configure Qdrant MCP Server for a remote Qdrant instance?

Set the QDRANT_URL with your instance's full URL including port (e.g., https://your-instance-id.region.gcp.cloud.qdrant.io:6333) and provide your QDRANT_API_KEY in the .env file.