MCP Server Qdrant

🚀 MCP Server: Qdrant

MCP Server Qdrant is a reference implementation designed for the Model Context Protocol (MCP), enabling Qdrant as a model storage service.

✨ Features

MCP Server Qdrant is a reference implementation designed for the Model Context Protocol (MCP), which uses Qdrant as a model storage service. It supports the following features:

1. Model Upload: Supports uploading model files in ONNX and TensorFlow (PB format).
2. Model Download: Allows downloading uploaded model files by specifying the model ID.
3. Model Deletion: Provides the function to delete a specified model.

🚀 Quick Start

Install Dependencies

Install the necessary Python packages using pip:

pip install "qdrant-client>=2.0.1" "onnxruntime>=1.9.0" "tensorflow>=2.4.0"

Run the Server

Run the MCP server on Qdrant:

# Run Qdrant using Docker and bind the port
docker run -p 6333:6333 qdrant/qdrant

# Start the MCP Server Qdrant
uvicorn mcp_server_qdrant.server:app --reload

Access the API

The MCP server provides the following RESTful API interfaces:

• Upload a Model: POST /api/models
• Download a Model: GET /api/models/{model_id}
• Delete a Model: DELETE /api/models/{model_id}

📚 Documentation

Global Configuration

Add the following content to settings.json in the global settings of VS Code:

{
  "mcp.servers": {
    "qdrant": {
      "command": "uvicorn",
      "args": ["mcp_server_qdrant.server:app"],
      "env": {
        "QDRANT_URL": "http://localhost:6333"
      }
    }
  }
}

Workspace Configuration

Create a .vscode/mcp.json file in the root directory of the workspace with the following content:

{
  "servers": {
    "qdrant": {
      "command": "uvicorn",
      "args": ["mcp_server_qdrant.server:app"],
      "env": {
        "QDRANT_URL": "http://localhost:6333"
      }
    }
  }
}

💻 Usage Examples

Upload a Model to Qdrant

curl -X POST http://localhost:8000/api/models \
     -F "model=@/path/to/model.onnx" \
     -F "model_type=onnx"

Download a Model with a Specified ID

curl -X GET http://localhost:8000/api/models/{model_id}

Delete a Model with a Specified ID

curl -X DELETE http://localhost:8000/api/models/{model_id}

📚 Documentation - VS Code Configuration

Global Configuration

Add the following content to settings.json in the global settings of VS Code:

{
  "mcp.servers": {
    "qdrant": {
      "command": "uvicorn",
      "args": ["mcp_server_qdrant.server:app"],
      "env": {
        "QDRANT_URL": "http://localhost:6333"
      }
    }
  }
}

Workspace Configuration

Create a .vscode/mcp.json file in the root directory of the workspace with the following content:

{
  "servers": {
    "qdrant": {
      "command": "uvicorn",
      "args": ["mcp_server_qdrant.server:app"],
      "env": {
        "QDRANT_URL": "http://localhost:6333"
      }
    }
  }
}

Docker Configuration

Create a .vscode/mcp.json file in the root directory of the workspace with the following content:

{
  "servers": {
    "qdrant": {
      "command": "docker",
      "args": [
        "run",
        "-p", "8000:8000",
        "-i",
        "--rm",
        "-e", "QDRANT_URL=http://localhost:6333"
      ],
      "env": {
        "QDRANT_URL": "http://localhost:6333"
      }
    }
  }
}

🤝 Contribution

If you have any suggestions for improving mcp-server-qdrant or find any issues, please feel free to submit a PR or an Issue on GitHub.

📄 License

MCP Server Qdrant is licensed under the MIT License. Please refer to LICENSE for detailed information.

Thank you for your support of the MCP project!