Cox's Bazar AI Itinerary MCP Server

A Model Context Protocol (MCP) server that provides travel planning tools and weather information for Cox's Bazar, Bangladesh.

Features

• Weather Tools: Get temperature forecasts and detailed weather information
• Itinerary Tools: Generate AI-powered travel itineraries
• Travel Prompts: Pre-configured prompts for travel planning

Installation

uv sync

Usage

Run Inspector Tool

• Need node version > 20.x.x

./scripts/run-inspector.sh

Run as installed command

uv run cox-mcp-server

# script 
./scripts/run-script.sh

Requirements

• Python 3.13+
• mcp[cli] >= 1.20.0
• python-dateutil >= 2.9.0
• requests >= 2.32.5

Project Structure

.
 src/
    mcp_server/
        __init__.py
        server.py           # Main server entry point
        mcp_instance.py     # MCP instance configuration
        models/             # Pydantic models and schemas
           __init__.py
           itinerary_models.py  # Itinerary data models
        components/         # MCP components
           __init__.py
           tools/          # MCP tools
              __init__.py
              auth_user.py    # User authentication tool
              itinerary.py    # Travel itinerary tools
           resources/      # MCP resources
              __init__.py
              weather.py      # Weather data resources
           prompts/        # MCP prompts
               __init__.py
               travel_prompts.py  # Travel planning prompts
        config/             # Configuration modules
           auth_provider.py    # Authentication provider
           custom_routes.py    # Custom routes configuration
        core/               # Core functionality
           __init__.py
           prompts/
               __init__.py
               get_travel_prompts.py  # Travel prompt handlers
        utils/              # Utilities
            __init__.py
            elicitation.py  # Elicitation utilities
            get_weather_forecast.py  # Weather forecast utilities
            helpers.py      # Helper functions
            types.py        # Type definitions
            register_mcp_components.py  # MCP component registration
 scripts/                    # Shell scripts
    run-inspector.sh        # Run MCP inspector
    run-script.sh           # Run server script
 tests/                      # Test directory
 _docs/                      # Documentation files
    remote-mcp-connect.md   # Remote MCP connection guide
 cox-mcp.jpeg                # Project image
 Dockerfile                  # Docker configuration
 glama.json                  # Glama configuration
 README.md                   # Project documentation
 license                     # MIT License
 pyproject.toml              # Project configuration and dependencies
 uv.lock                     # Dependency lock file

Connect to Claude Desktop

Option 1: Local Development (without Docker)

Add this configuration to your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "coxs-bazar-itinerary-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/Users/code4mk/Documents/GitHub/gumpper-group/mcp-explore/mcp-server-python-template",
        "run",
        "cox-mcp-server"
      ]
    }
  }
}

Option 2: Using Docker

First, build the Docker image:

docker build \
  --platform=linux/amd64 \
  --build-arg PYTHON_VERSION=3.12-slim \
  -t cox-mcp-server .

Then add this configuration to your Claude Desktop config file:

{
  "mcpServers": {
    "coxs-bazar-itinerary-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "cox-mcp-server"
      ]
    }
  }
}

Option 3: Using uvx (Run directly from GitHub)

Add this configuration to your Claude Desktop config file:

{
  "mcpServers": {
    "coxs-bazar-itinerary-mcp": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/code4mk/coxs-bazar-itinerary-mcp-server.git",
        "cox-mcp-server"
      ]
    }
  }
}

License

MIT