MCP Server Scaffolding Script

A bash script to quickly scaffold new MCP (Model Context Protocol) server projects based on the memgraphdb-mcp-server template.

Features

• 🚀 Quick Setup: Create a complete MCP server project structure in seconds
• 📁 Complete Structure: Includes all necessary directories and files
• 🔧 Ready to Use: Pre-configured with TypeScript, Express, and MCP SDK
• 🐳 Docker Ready: Includes Dockerfile and .dockerignore
• 📝 Documentation: Auto-generates README and getting started guide
• ✅ Best Practices: Follows MCP server patterns and conventions

Prerequisites

• Bash shell (macOS, Linux, or WSL on Windows)
• Node.js 18+ (for running the generated project)

Installation

Option 1: Quick Install with curl (Recommended)

Download and run the script directly:

curl -fsSL https://raw.githubusercontent.com/playground/scaffold-mcp-server/main/scaffold-mcp-server.sh | bash -s -- my-mcp-server

Replace my-mcp-server with your desired server name.

undefinedOr download first, then run:undefined

# Download the script
curl -fsSL https://raw.githubusercontent.com/YOUR_USERNAME/YOUR_REPO/main/scaffold-mcp-server.sh -o scaffold-mcp-server.sh

# Make it executable
chmod +x scaffold-mcp-server.sh

# Run it
./scaffold-mcp-server.sh my-mcp-server

Option 2: Local Installation

If you already have the script locally:

chmod +x scaffold-mcp-server.sh
./scaffold-mcp-server.sh <server-name>

Usage

Basic Usage

./scaffold-mcp-server.sh <server-name>

Example

./scaffold-mcp-server.sh my-mcp-server

This will create a new directory my-mcp-server/ with the complete project structure.

One-Line Quick Start

Create and start a new MCP server in one command:

curl -fsSL https://raw.githubusercontent.com/YOUR_USERNAME/YOUR_REPO/main/scaffold-mcp-server.sh | bash -s -- my-mcp-server && cd my-mcp-server && npm install && cp .env.example .env && npm run dev

Server Name Requirements

• Must contain only lowercase letters, numbers, and hyphens
• Examples: my-mcp-server, api-connector-mcp, database-query-server

Generated Project Structure

your-server-name/
├── src/
│   ├── tools/              # MCP tools (functions callable by AI)
│   │   └── exampleTool.ts  # Example tool implementation
│   ├── services/           # Business logic and utilities
│   │   └── common.ts       # Common utility functions
│   ├── models/             # TypeScript interfaces and types
│   │   └── model.ts        # Data models and session types
│   ├── prompts/            # MCP prompts (pre-defined interactions)
│   │   └── example-prompts.ts
│   ├── scripts/            # Utility scripts
│   ├── mcp-server.ts       # MCP server configuration
│   └── server.ts           # HTTP server entry point
├── docs/
│   └── GETTING_STARTED.md  # Quick start guide
├── package.json            # Node.js dependencies and scripts
├── tsconfig.json           # TypeScript configuration
├── Dockerfile              # Docker container configuration
├── .dockerignore           # Docker ignore patterns
├── .env.example            # Environment variables template
├── .gitignore              # Git ignore patterns
└── README.md               # Project documentation

Quick Start After Scaffolding

1. undefinedNavigate to your project:

   cd your-server-name
   

2. undefinedInstall dependencies:

   npm install
   

3. undefinedConfigure environment:

   cp .env.example .env
   # Edit .env with your configuration
   

4. undefinedRun in development mode:

   npm run dev
   

5. undefinedTest the server:

   curl http://localhost:3000/mcp/health
   

Available NPM Scripts

The generated project includes these scripts:

• npm run dev - Run in development mode with hot reload
• npm run build - Build TypeScript to JavaScript
• npm start - Build and run in production mode
• npm run dev:port - Run on custom port: npm run dev:port --port=3001
• npm run inspector - Run MCP inspector for debugging
• npm run build:docker:arm64 - Build Docker image for ARM64 (Apple Silicon)
• npm run build:docker:amd64 - Build Docker image for AMD64 (Intel/AMD)

Customizing Your Server

1. Add Your Tools

Create new tool files in src/tools/:

// src/tools/myTool.ts
import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
import { z } from 'zod';

export function registerMyTool(server: McpServer) {
  server.tool(
    'my-tool',
    'Description of what this tool does',
    {
      param1: z.string().describe('Parameter description')
    },
    async (params: any) => {
      // Your tool logic here
      return {
        content: [{ type: 'text', text: 'Result' }]
      };
    }
  );
}

Then register it in src/mcp-server.ts:

import { registerMyTool } from './tools/myTool';
// ...
registerMyTool(server);

2. Add Your Prompts

Create prompt files in src/prompts/:

// src/prompts/my-prompts.ts
import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';

export function registerMyPrompts(server: McpServer) {
  server.prompt(
    'my-prompt',
    'Description of this prompt',
    async (args: any) => {
      return {
        messages: [
          {
            role: 'user',
            content: {
              type: 'text',
              text: 'Your prompt content'
            }
          }
        ]
      };
    }
  );
}

3. Add Your Services

Create service files in src/services/ for business logic, API clients, database connections, etc.

4. Update Configuration

• Edit .env for environment-specific settings
• Update package.json with your project details
• Customize README.md with your documentation

Docker Deployment

Build Docker Image

# For ARM64 (Apple Silicon)
npm run build:docker:arm64

# For AMD64 (Intel/AMD)
npm run build:docker:amd64

Run Docker Container

docker run -p 3000:8080 --env-file .env your-image-name

MCP Server Architecture

The generated server follows the MCP (Model Context Protocol) specification:

• undefinedStateful Sessions: Each client connection maintains its own session
• undefinedStreamable HTTP: Uses HTTP with Server-Sent Events (SSE) for bidirectional communication
• undefinedTools: Functions that AI assistants can call to perform actions
• undefinedPrompts: Pre-defined interaction patterns for common tasks
• undefinedResources: Data sources that can be queried

Environment Variables

The .env.example file includes:

PORT=3000                    # Server port
# Add your service-specific variables here

Health Check

The server includes a health check endpoint:

curl http://localhost:3000/mcp/health
# Response: {"status":"healthy"}

Troubleshooting

Script Errors

• undefinedPermission denied: Run chmod +x scaffold-mcp-server.sh
• undefinedDirectory exists: Choose a different server name or remove the existing directory
• undefinedInvalid name: Use only lowercase letters, numbers, and hyphens

Generated Project Issues

• undefinedModule not found: Run npm install
• undefinedTypeScript errors: Check tsconfig.json configuration
• undefinedPort in use: Change PORT in .env or use npm run dev:port --port=3001

Examples

Create a Database MCP Server

./scaffold-mcp-server.sh postgres-mcp-server
cd postgres-mcp-server
npm install
# Add pg dependency
npm install pg @types/pg
# Implement your database tools

Create an API Integration Server

./scaffold-mcp-server.sh github-api-mcp-server
cd github-api-mcp-server
npm install
# Add axios dependency
npm install axios
# Implement your API tools

Create a File System Server

./scaffold-mcp-server.sh filesystem-mcp-server
cd filesystem-mcp-server
npm install
# Implement file system tools

Template Source

This scaffolding script is based on the memgraphdb-mcp-server template, which includes:

• Express HTTP server with MCP protocol support
• Session management for stateful connections
• TypeScript configuration
• Docker containerization
• Logging with log4js
• Input validation with Zod

Contributing

To improve the scaffolding script:

1. Edit scaffold-mcp-server.sh
2. Test with: ./scaffold-mcp-server.sh test-server
3. Verify the generated project works
4. Clean up: rm -rf test-server

License

ISC

Made with Bob

This scaffolding script was created to streamline MCP server development and follows best practices from the Model Context Protocol specification.