Qdrant MCP Server

Qdrant MCP Server is an MCP server that provides semantic code search capabilities across software repositories using the Qdrant vector database and OpenAI text embeddings. It connects MCP-compatible clients like Claude Desktop to an indexing engine that parses local codebases, generates vector representations using models such as text-embedding-3-small, and stores them in Qdrant collections. Software engineers, technical leads, and code analysts use this tool to locate functionality based on meaning rather than literal text matching. The server enables users to discover related classes, examine error handling logic across distributed files, and surface target API endpoints through natural language queries. By supporting background file monitoring and incremental updates, it maintains an up-to-date vector index without requiring full rescans. The server also respects ignore patterns and handles custom file extensions, allowing development teams to safely query internal implementations, inspect architectural patterns, and retrieve contextually relevant code snippets directly within conversational assistant interfaces.

Category: AI Memory & Context

Tags: embeddings, qdrant, semantic search, vector-database, vector-search

Visit Qdrant MCP Server

How to install and configure Qdrant MCP Server

1. Ensure prerequisites are met: Node.js 18+, Python 3.8+, an OpenAI API key, and a running Qdrant instance (local Docker or Qdrant Cloud). 2. Install the package globally via npm with npm install -g @kindash/qdrant-mcp-server or via pip using pip install qdrant-mcp-server. 3. Export environment variables: bash export OPENAI_API_KEY="your-api-key" export QDRANT_URL="http://localhost:6333" 4. Index your target repository using the CLI: bash qdrant-indexer /path/to/your/code 5. Open your Claude Desktop configuration file (~/.claude/config.json) and register the server: json { "mcpServers": { "qdrant-search": { "command": "qdrant-mcp", "args": ["--collection", "codebase"], "env": { "OPENAI_API_KEY": "your-api-key", "QDRANT_URL": "http://localhost:6333" } } } } 6. Restart Claude Desktop to start querying the indexed codebase.

What you can do with Qdrant MCP Server

• Locate cross-cutting security patterns by querying authentication and permission checks across distributed directories without relying on keyword matching. - Find architecturally similar implementations by querying the vector database for classes or modules comparable to specific services. - Discover relevant API endpoints and route definitions across multiple source files using plain natural language descriptions of the endpoint behavior. - Audit existing error handling mechanisms by searching for recurring exception management patterns and logging structures in the repository. - Maintain current index states through automatic background reindexing that tracks local file modifications during ongoing development work.

Key facts

• Open Source
• https://github.com/steiner385/qdrant-mcp-server
• AI Memory & Context, Databases & Data Stores
• embeddings, qdrant, semantic search, vector-database, vector-search

Part of MCP Servers

What is Qdrant MCP Server?

Qdrant MCP Server is an open-source tool that exposes semantic code search capabilities to AI agents through the Model Context Protocol. It indexes source code repositories into a Qdrant vector database using vector embeddings generated by OpenAI models.

How do I install Qdrant MCP Server?

You can install Qdrant MCP Server globally using npm via npm install -g @kindash/qdrant-mcp-server or through pip using pip install qdrant-mcp-server. You must also supply your OpenAI API key and set up a Qdrant vector database instance via Docker or Qdrant Cloud.

Which MCP clients work with Qdrant MCP Server?

It works with any client supporting the Model Context Protocol, including Claude Desktop. Once added to the client configuration file with the necessary environment variables, the model can execute semantic queries against your indexed repository.

Is Qdrant MCP Server open source?

Yes, Qdrant MCP Server is open source software released under the MIT License. The complete source code, installation utilities, and documentation are available on its official GitHub repository.

What embedding models does Qdrant MCP Server support?

By default, the server uses OpenAI text-embedding-3-small to process code snippets. It also provides programmatic options for Cohere embeddings and custom file processors for handling specialized file types.