About Qdrant Retrieve MCP Server
Qdrant Retrieve MCP Server lets MCP clients run semantic searches across Qdrant collections using one or more queries. It returns matching document text, scores, and collection sources, with a configurable result count.
Key features
- 01Search across multiple collections
- 02Submit multiple queries
- 03Configure result count
- 04Track source collection
Use cases
- 01Find relevant documents across Qdrant collections
- 02Compare results from several search queries
MCP support
Exposes the `qdrant_retrieve` tool to MCP clients for retrieving semantically similar documents from Qdrant collections. It connects to a Qdrant instance specified by URL.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y @gergelyszerovay/mcp-server-qdrant-retriveQDRANT_API_KEYSecretLocal servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEQdrant Retrieve MCP Server
MCP server for semantic search with Qdrant vector database.
Features
- Semantic search across multiple collections
- Multi-query support
- Configurable result count
- Collection source tracking
Note: The server connects to a Qdrant instance specified by URL.
Note 2: The first retrieve might be slower, as the MCP server downloads the required embedding model.
API
Tools
- qdrant_retrieve
- Retrieves semantically similar documents from multiple Qdrant vector store collections based on multiple queries
- Inputs:
collectionNames(string[]): Names of the Qdrant collections to search acrosstopK(number): Number of top similar documents to retrieve (default: 3)query(string[]): Array of query texts to search for
- Returns:
results: Array of retrieved documents with:query: The query that produced this resultcollectionName: Collection name that this result came fromtext: Document text contentscore: Similarity score between 0 and 1
Usage with Claude Desktop
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"qdrant": {
"command": "npx",
"args": ["-y", "@gergelyszerovay/mcp-server-qdrant-retrive"],
"env": {
"QDRANT_API_KEY": "your_api_key_here"
}
}
}
}
Command Line Options
MCP server for semantic search with Qdrant vector database.
Options
--enableHttpTransport Enable HTTP transport [default: false]
--enableStdioTransport Enable stdio transport [default: true]
--enableRestServer Enable REST API server [default: false]
--mcpHttpPort=<port> Port for MCP HTTP server [default: 3001]
--restHttpPort=<port> Port for REST HTTP server [default: 3002]
--qdrantUrl=<url> URL for Qdrant vector database [default: http://localhost:6333]
--embeddingModelType=<type> Type of embedding model to use [default: Xenova/all-MiniLM-L6-v2]
--help Show this help message
Environment Variables
QDRANT_API_KEY API key for authenticated Qdrant instances (optional)
Examples
$ mcp-qdrant --enableHttpTransport
$ mcp-qdrant --mcpHttpPort=3005 --restHttpPort=3006
$ mcp-qdrant --qdrantUrl=http://qdrant.example.com:6333
$ mcp-qdrant --embeddingModelType=Xenova/all-MiniLM-L6-v2
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.