Quick Start

This guide will help you create your first MCP server using egile-mcp-starter.

Generate Your First MCP Server

Interactive Mode

The simplest way to get started is to run the command interactively:

egile-mcp-starter

You’ll be prompted for various configuration options:

project_name [My MCP Server]: Weather MCP Server
project_description [A Model Context Protocol server built with FASTMCP]: Weather data MCP server with forecast tools
author_name [Your Name]: John Doe
author_email [your.email@example.com]: john@example.com
github_username [yourusername]: johndoe
server_type [full]: tools
python_version [3.11]: 3.11
use_docker [y]: y
use_github_actions [y]: y
include_examples [y]: y

Non-Interactive Mode

For automation or CI/CD, you can use the --no-input flag with a configuration file:

# Create a config file
cat > my-config.yaml << 'EOF'
default_context:
  project_name: "My Weather Server"
  project_slug: "my_weather_server"
  project_description: "A MCP server for weather data"
  author_name: "Your Name"
  author_email: "your@email.com"
  github_username: "yourusername"
  server_type: "tools"
  python_version: "3.11"
  use_docker: "y"
  use_github_actions: "y"
  include_examples: "y"
EOF

# Generate project
egile-mcp-starter --config-file my-config.yaml --no-input

Quick Generation with CLI Options

For quick project generation without configuration files, use CLI options:

# Generate with custom project name
egile-mcp-starter --project-name "weather_api_server" --no-input

# Choose template and customize name
egile-mcp-starter --template rag --project-name "document_search_server" --no-input

# Generate in specific directory with custom name
egile-mcp-starter --output-dir ./servers --project-name "my_custom_server" --verbose

CLI Options Reference

Option	Description	Example
--project-name	Override project name (affects directory and package names)	--project-name "my_server"
--template	Choose template (mcp, rag)	--template rag
--output-dir	Output directory	--output-dir ./projects
--no-input	Skip interactive prompts	--no-input
--verbose	Show detailed output	--verbose
--list-templates	List available templates	--list-templates

Note: The --project-name option is particularly useful for CI/CD pipelines where you need predictable directory names.

Project Structure

After generation, your project will have this structure:

my_weather_server/
├── src/
│   ├── my_weather_server/
│   │   ├── __init__.py
│   │   ├── server.py          # Main MCP server
│   │   ├── config.py          # Configuration management
│   │   ├── tools/             # Tool implementations
│   │   ├── resources/         # Resource handlers
│   │   ├── prompts/           # Prompt templates
│   │   └── utils.py           # Utility functions
│   └── main.py                # Entry point
├── tests/                     # Test suite
├── Dockerfile                 # Docker configuration
├── docker-compose.yml         # Docker Compose setup
├── .github/workflows/ci.yml   # GitHub Actions CI/CD
├── pyproject.toml             # Project configuration
└── README.md                  # Project documentation

Set Up Your Development Environment

Navigate to your new project:

cd my_weather_server

Install dependencies:

# With Poetry (recommended)
poetry install

# Or with pip
pip install -e ".[dev]"

If you enabled pre-commit hooks:

poetry run pre-commit install

Run Your Server

Start your MCP server:

# With Poetry
poetry run python src/main.py

# Or activate the virtual environment
poetry shell
python src/main.py

Test Your Server

Run the test suite:

# Run all tests
poetry run pytest

# Run with coverage
poetry run pytest --cov=src --cov-report=html

Code Quality Checks

Format and check your code:

# Format code
poetry run black src tests

# Check linting
poetry run flake8 src tests

# Type checking
poetry run mypy src

# Run all pre-commit checks
poetry run pre-commit run --all-files

Docker Support

If you enabled Docker support:

# Build the image
docker build -t my-weather-server .

# Run with docker-compose
docker-compose up

Next Steps

1. Customize your server: Edit the files in src/my_weather_server/

2. Add your tools: Implement custom tools in src/my_weather_server/tools/

3. Configure settings: Edit config.example.yaml

4. Write tests: Add tests in the tests/ directory

5. Deploy: Use the included Docker and CI/CD configurations

Integration with Claude Desktop

To use your server with Claude Desktop, add it to your configuration:

{
  "mcpServers": {
    "my-weather-server": {
      "command": "poetry",
      "args": ["run", "python", "/path/to/my_weather_server/src/main.py"],
      "env": {
        "LOG_LEVEL": "INFO"
      }
    }
  }
}

CLI Options

Common Options

• --output-dir DIR: Specify output directory (default: current directory)

• --config-file FILE: Use a configuration file

• --no-input: Don’t prompt for input (use defaults or config file)

• --verbose: Enable verbose output

• --version: Show version information

• --help: Show help message

Examples

# Generate in a specific directory
egile-mcp-starter --output-dir ./my-projects

# Use custom configuration
egile-mcp-starter --config-file custom-config.yaml

# Automated generation
egile-mcp-starter --no-input --output-dir ./servers

# Verbose output for debugging
egile-mcp-starter --verbose