MCP FastMCP Server

An MCP (Model Context Protocol) server built with FastMCP, optimized for AWS Lambda deployment using the Lambda Web Adapter.

Features

• Serverless Ready: Optimized for AWS Lambda with stateless_http=True
• Lambda Web Adapter: Uses AWS Lambda Web Adapter for seamless Lambda integration
• Streamable HTTP: Uses streamable-http transport for MCP over HTTP
• Multiple Tools: Includes greeting, math operations, and server details tools
• Resources & Prompts: Includes sample resources and prompts
• One-Click Deployment: Automated deployment script for AWS Lambda

Quick Start

1. Clone and Deploy to AWS Lambda

# Clone the repository
git clone https://github.com/mishramohit437/fastmcp-lambda.git
cd fastmcp-lambda

# Configure AWS CLI (first time only)
aws configure

# Deploy to AWS Lambda
./deploy_lambda.sh

The script will:

• Build the Docker image
• Push to Amazon ECR
• Create Lambda function
• Create Function URL
• Output your MCP endpoint

2. Configure Claude Desktop

Add to your MCP configuration (mcp.json or claude_desktop_config.json):

{
  "mcpServers": {
    "fastmcp-lambda": {
      "url": "YOUR_FUNCTION_URL/mcp",
      "transport": "streamable-http"
    }
  }
}

Replace YOUR_FUNCTION_URL with the URL output by the deployment script.

Local Development

# Install dependencies
pip install fastmcp

# Run in stdio mode (for local Claude Desktop testing)
python -m mcp_fast_mcp.server --stdio

# Run in HTTP mode (for web testing)
python -m mcp_fast_mcp.server --port 8080

Using Docker

# Build and run with Docker Compose
docker-compose up --build

# Or build manually
docker build -t mcp-fast-mcp .
docker run -p 8080:8080 mcp-fast-mcp

Available Tools

Tool	Description
hello_world	A simple hello world tool for testing
get_server_details	Returns server environment details
greet	Greet someone by name
add	Add two numbers together
multiply	Multiply two numbers together

Resources

Resource	Description
greeting://hello	A simple greeting resource

Prompts

Prompt	Description
greet_prompt	Generate a greeting prompt for a given name

AWS Deployment

For detailed deployment instructions, see DEPLOYMENT.md.

Prerequisites

• AWS Account with appropriate permissions
• AWS CLI installed and configured
• Docker installed

Quick Deploy

# Make script executable (if needed)
chmod +x deploy_lambda.sh

# Deploy with default region (eu-north-1)
./deploy_lambda.sh

# Or specify your region
AWS_REGION=us-east-1 ./deploy_lambda.sh

Architecture

Client → Lambda Function URL → Lambda Web Adapter → FastMCP Server → Response
         (HTTPS)              (Container)          (streamable-http)

Components:

• ECR: Container registry for Docker images
• Lambda: Serverless compute running FastMCP
• Function URL: Direct HTTPS endpoint
• IAM Role: Basic execution role with CloudWatch logs

Configuration

Key Settings

Setting	Value	Description
PORT	8080	Server port (Lambda Web Adapter default)
Transport	streamable-http	MCP transport for Lambda compatibility
Stateless	True	Serverless optimization
Platform	linux/amd64	Lambda architecture

Environment Variables

Variable	Default	Description
PORT	8080	Server port
HOST	0.0.0.0	Server host
AWS_REGION	eu-north-1	AWS region for deployment

Development

Project Structure

fastmcp-lambda/
├── mcp_fast_mcp/
│   ├── __init__.py
│   └── server.py          # Main server with tools
├── Dockerfile             # Lambda Web Adapter setup
├── deploy_lambda.sh       # Deployment script
├── docker-compose.yml     # Local Docker testing
├── DEPLOYMENT.md          # Detailed AWS deployment guide
├── pyproject.toml         # Project config
└── README.md              # This file

Adding New Tools

@mcp.tool()
def your_tool(param: str) -> str:
    """Tool description."""
    return f"Result: {param}"

Adding New Resources

@mcp.resource("your-resource://path")
def your_resource() -> str:
    """Resource description."""
    return "Resource content"

Maintenance

Update the Deployment

# Make changes to your code
# Rebuild and redeploy
./deploy_lambda.sh

View Logs

# CloudWatch Logs (replace with your region)
aws logs tail /aws/lambda/mcp-fast-mcp --region us-east-1 --follow

Cost Estimation

Service	Free Tier	Paid
Lambda Requests	1M requests/month	$0.20 per 1M requests
Lambda Duration	400K GB-seconds/month	$0.00001667/GB-second
Function URL	1M requests/month	$0.20 per 1M requests

Estimated cost for 10K requests/month: ~$0 (within free tier)

License

MIT

Links

• FastMCP Documentation
• MCP Protocol
• AWS Lambda
• Lambda Web Adapter