ntua-unit-of-control-and-informatics/jaqpot-mcp-server

Jaqpot MCP Server

Model Context Protocol (MCP) server for Jaqpot platform integration. This server allows LLMs to interact with the Jaqpot API for machine learning model predictions and discovery.

Features

• Model Predictions: Execute predictions using Jaqpot models with the jaqpot-python-sdk
• Model Search: Search and discover models by description and features
• Authentication: Secure API access using client key and secret
• MCP Integration: Full Model Context Protocol compliance for seamless LLM integration

Installation

Prerequisites

1. Install uv:

   curl -LsSf https://astral.sh/uv/install.sh | sh
   

2. Install Python 3.10+ using uv:

   uv python install 3.10
   

3. Configure your Jaqpot API credentials:

   • Get your API key and secret from Jaqpot platform
   • Set them as environment variables (see Configuration section)

Install the Server

# Clone the repository
git clone https://github.com/ntua-unit-of-control-and-informatics/jaqpot-mcp-server.git
cd jaqpot-mcp-server

# Install with uv
uv sync

Configuration

MCP Client Configuration

Add the server to your MCP client configuration. For Claude Desktop, add to your claude_desktop_config.json:

{
  "mcpServers": {
    "jaqpot": {
      "type": "stdio",
      "command": "uv",
      "args": [
        "--directory", "/path/to/jaqpot-mcp-server",
        "run", "server.py"
      ],
      "env": {
        "JAQPOT_API_KEY": "your-api-key",
        "JAQPOT_API_SECRET": "your-api-secret"
      }
    }
  }
}

Replace /path/to/jaqpot-mcp-server with the actual path to your cloned repository.

Alternative: Direct Python Execution

{
  "mcpServers": {
    "jaqpot": {
      "type": "stdio",
      "command": "python",
      "args": ["/path/to/jaqpot-mcp-server/server.py"],
      "env": {
        "JAQPOT_API_KEY": "your-api-key",
        "JAQPOT_API_SECRET": "your-api-secret"
      }
    }
  }
}

Usage

Running the Server

# With uv (recommended)
uv run server.py

# Or directly with Python
python server.py

Testing the Installation

Run the example usage script:

uv run python examples/usage_example.py

MCP Tools Available

1. jaqpot_predict

Make predictions using a Jaqpot model.

Parameters:

• model_id (int): The ID of the Jaqpot model
• dataset (list): Input data for prediction

2. jaqpot_search_models

Search for models by criteria.

Parameters:

• query (str): Search query
• page (int, optional): Page number for pagination
• size (int, optional): Number of results per page

3. jaqpot_get_model_info

Get detailed information about a specific model.

Parameters:

• model_id (int): The ID of the Jaqpot model

Development

Local Development

1. Clone the repository and install dependencies:

git clone https://github.com/ntua-unit-of-control-and-informatics/jaqpot-mcp-server.git
cd jaqpot-mcp-server
uv sync --dev

2. Set environment variables:

export JAQPOT_API_KEY="your-api-key"
export JAQPOT_API_SECRET="your-api-secret"

3. Run the server:

uv run server.py

Testing

Run the example usage:

uv run python examples/usage_example.py

License

See file.