Chronulus Model Context Protocol Integration

Engage Chronulus AI Predictive & Forecasting Agents within Claude

Initial Setup: Utilizing Claude for Desktop

Acquisition

The Claude for Desktop utility is presently accessible for both macOS and Windows operating systems.

Install the application package here

Configuration Protocol

Refer to the standardized guidance documentation here for configuring the Claude desktop client environment.

Your specific configuration file for Claude can typically be located at one of these system paths:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Subsequently, select one of the deployment methodologies below that aligns best with your operational requirements and incorporate it into your claude_desktop_config.json manifest.

Via Python Package Installer (pip)

(Method A) Secure the officially released version from the Python Package Index (PyPI)

pip install chronulus-mcp

(Method B) Obtain and install directly from the Github repository source

git clone https://github.com/ChronulusAI/chronulus-mcp.git
cd chronulus-mcp
pip install .

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

**Advisory Note:** Should an execution failure arise, such as "MCP chronulus-agents: spawn python ENOENT", it is highly probable that the fully qualified path to the Python executable must be specified. For illustration, use `/Library/Frameworks/Python.framework/Versions/3.11/bin/python3` rather than the unqualified term `python`.

Via Containerization (docker)

This process establishes a Docker image, named 'chronulus-mcp', intended for recurrent use within the Claude configuration setup.

git clone https://github.com/ChronulusAI/chronulus-mcp.git
cd chronulus-mcp
 docker build . -t 'chronulus-mcp'

Ensure that the terminal argument used for the image name in the build command precisely matches the identifier referenced in your Claude configuration section.

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

Via Unified Virtual Execution (uvx)

The `uvx` utility automates the retrieval from the PyPI registry, subsequent installation of the `chronulus-mcp` package, and immediate execution thereof.

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

**Troubleshooting Guidance:** If an error like "MCP chronulus-agents: spawn uvx ENOENT" is encountered, corrective action requires one of two steps: 1. [Install the uv package](https://docs.astral.sh/uv/getting-started/installation/) if it is absent, or 2. Specify the absolute system path for `uvx`. For example, use `/Users/username/.local/bin/uvx` instead of the simple executable name `uvx`.

Supplementary Service Endpoints (Filesystem, Network Fetch, etc.)

Our demonstration environment utilizes supplementary external servers such as fetch and filesystem.

For comprehensive details concerning the installation and configuration parameters of these auxiliary server components, consult the documentation provided by the respective component maintainers.

The subsequent JSON structure illustrates the integration of the filesystem and fetch capabilities alongside the Chronulus service within the claude_desktop_config.json file:

{
  "mcpServers": {
    "chronulus-agents": {
      "command": "uvx",
      "args": ["chronulus-mcp"],
      "env": {
        "CHRONULUS_API_KEY": "<YOUR_CHRONULUS_API_KEY>"
      }
    },
    "filesystem": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-filesystem",
        "/path/to/AIWorkspace"
      ]
    },
    "fetch": {
      "command": "uvx",
      "args": ["mcp-server-fetch"]
    }
  }
}

Claude Operational Preferences

To optimize the user experience when interacting with Claude across diverse toolsets, it is strongly advised to incorporate your specific preferences under the Claude Settings pane.

You may update your preferences through these avenues:

Within Claude Desktop: Navigate to Settings -> General -> Claude Settings -> Profile (tab)
Via the web interface at claude.ai/settings: Select the Profile (tab)

These preference settings are synchronized across both the Claude Desktop application and the Claude.ai web environment, necessitating instructions that are universally applicable.

The preference configuration utilized in our demonstration outputs is provided below:

## Tool-Dependent Protocols
The directives below are applicable exclusively when MCP Servers are actively available.

### Filesystem Tool Directives
- Refrain from invoking 'read_file' or 'read_multiple_files' operations against binary file formats (e.g., JPEG, PDF, DOCX).
- When dealing with proprietary binary structures (e.g., images, PDF documents, Word files), employ the 'get_info' function instead of any 'read_*' utilities for preliminary inspection.

### Chronulus Agents Tool Directives
- When invoking Chronulus services, prioritize input mechanisms such as TextFromFile, PdfFromFile, and ImageFromFile over direct, unsorted file content scanning.
- For any visualization of forecasts generated by Chronulus, mandate the inclusion of the accompanying analytical explanation provided by Chronulus immediately following the graphic representation, clearly demarcated under the heading "Chronulus Explanation."