home / mcp / chronulus mcp server

Chronulus MCP Server

MCP Server for Chronulus AI Forecasting and Prediction Agents

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "chronulusai-chronulus-mcp": {
      "command": "python",
      "args": [
        "-m",
        "chronulus_mcp"
      ],
      "env": {
        "CHRONULUS_API_KEY": "<YOUR_CHRONULUS_API_KEY>"
      }
    }
  }
}

You set up the Chronulus MCP Server to enable Claude to communicate with Chronulus AI Forecasting and Prediction Agents. This server acts as a bridge, allowing you to run Chronulus agents locally or remotely and connect them to Claude via MCP, with optional filesystem and fetch integrations for data access.

How to use

To use the Chronulus MCP Server, run one or more MCP endpoints that point to Chronulus agents. You can run them as local stdio processes or via containerized or managed runtimes, then point your MCP client to the appropriate command configuration. The Chronulus agents provide forecasting capabilities, and when you plot forecasts, you should include Chronulus explanations with the plots for clarity. If you want to extend capabilities, you can add additional servers like a filesystem workspace or a fetch service to your MCP configuration.

How to install

Prerequisites: you need Python installed to run the Chronulus MCP server as a stdio process, or you can use Docker or uvx to manage the runtime.

Option 1: Run Chronulus MCP with Python (stdio) directly. Use the following command configuration in your MCP client setup to start the agent.

Option 2: Build and run with Docker. Use a Docker image named chronulus-mcp and run it with an MCP config to pass the API key.

Option 3: Use uvx to pull the latest Chronulus MCP and run it automatically.

Configuration examples

{
  "mcpServers": {
    "chronulus_agents": {
      "command": "python",
      "args": ["-m", "chronulus_mcp"],
      "env": {
        "CHRONULUS_API_KEY": "<YOUR_CHRONULUS_API_KEY>"
      }
    }
  }
}

Additional runtime options (examples)

{
  "mcpServers": {
    "chronulus_agents": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "-e", "CHRONULUS_API_KEY", "chronulus-mcp"],
      "env": {
        "CHRONULUS_API_KEY": "<YOUR_CHRONULUS_API_KEY>"
      }
    }
  }
}

Using uvx to run the MCP

{
  "mcpServers": {
    "chronulus_agents": {
      "command": "uvx",
      "args": ["chronulus-mcp"],
      "env": {
        "CHRONULUS_API_KEY": "<YOUR_CHRONULUS_API_KEY>"
      }
    }
  }
}

Notes on environment and troubleshooting

If you encounter errors related to the Python executable, provide the absolute path to Python in the command configuration, for example "/Library/Frameworks/Python.framework/Versions/3.11/bin/python3" instead of just python.

Filesystems and data fetch (optional)

You can extend the MCP setup with additional servers for data access. For example, you can add a filesystem server to access your AIWorkspace path and a fetch server to retrieve data. Configure your Claude/Claude Desktop setup to include these MCP servers alongside Chronulus agents.

Available tools

chronulus_agents

Interface to Chronulus AI Forecasting & Prediction Agents. Use input types like TextFromFile, PdfFromFile, and ImageFromFile for file inputs. Include Chronulus explanations when plotting forecasts.

filesystem

Filesystem tool to interact with your AIWorkspace path. Do not use read_file on binary files; use get_info for binary files.

fetch

Fetch tool to retrieve data from remote sources as part of the MCP setup.