MCP CLI Tool

April 4, 2025 ยท View on GitHub

A command-line interface for interacting with MCP (Model Control Protocol) servers. This tool allows you to easily interact with various AI models through a simple command-line interface.

Features

  • Interactive CLI interface
  • Support for multiple MCP servers
  • Easy-to-use argument prompts
  • Progress tracking for long-running operations
  • Configurable server settings

Installation

You can install the package directly from GitHub using pip:

pip install git+https://github.com/JckHoe/python-mcp-client.git

Or from PyPI:

pip install mcp-cli

Configuration

Create a servers_config.json file with your MCP server configuration:

{
  "mcpServers": {
    "mcp-hfspace": {
      "command": "npx",
      "args": [
        "-y",
        "@llmindset/mcp-hfspace",
        "shuttleai/shuttle-jaguar",
        "styletts2/styletts2",
        "Qwen/QVQ-72B-preview"
      ]
    }
  }
}

Usage

Run the CLI tool:

mcp-cli

Or specify a custom configuration file:

mcp-cli -c path/to/config.json

Enable verbose logging:

mcp-cli -v

Available Tools

The tool will display all available tools from your configured MCP servers. Each tool will show:

  • Name
  • Description
  • Required and optional arguments

Interactive Usage

  1. The tool will display a list of available tools
  2. Enter the name of the tool you want to use
  3. Follow the prompts to enter required and optional arguments
  4. View the results of the tool execution
  5. Type '/bye' to exit the program

Development

To set up the development environment:

git clone https://github.com/yourusername/mcp-cli.git
cd mcp-cli
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -e .

License

MIT License