AWS MCP Servers Project Introduction

Project Overview

AWS MCP (Model Context Protocol) Servers is an open-source project developed by AWS Labs, designed to enhance the interaction between large language models (LLMs) and AWS services through dedicated MCP servers. The project address is https://github.com/awslabs/mcp. By providing standardized interfaces and tools, it helps AI assistants (such as Amazon Q Developer, Claude, Cline, etc.) execute AWS-related tasks more efficiently, while ensuring adherence to AWS best practices, security compliance, and efficient development processes.

Core Features

• Improve Model Output Quality: By injecting AWS service information into the model context, reduce hallucinations and improve response accuracy.
• Workflow Automation: Supports workflows like CDK, Terraform, and CloudFormation, simplifying complex tasks.
• Domain Expertise: Provides deep context of AWS services, compensating for the lack of model training data.
• Security First: Supports read-only mode, IAM permission control, and sensitive data restrictions.

Main Components

The project includes several independent MCP servers, covering various AWS services and features:

1. AWS CDK MCP Server

   • Functionality: Supports AWS CDK development, integrates CDK Nag to ensure security compliance.
   • Use Case: Rapid construction and management of cloud infrastructure.

2. AWS Terraform MCP Server

   • Functionality: Supports Terraform workflows, integrates Checkov for security scanning.
   • Use Case: Generating secure Terraform scripts.

3. AWS Serverless MCP Server

   • Functionality: Provides serverless development support for Lambda, API Gateway, etc., integrates SAM CLI.
   • Use Case: Rapidly building and deploying serverless applications.

4. AWS Documentation MCP Server

   • Functionality: Retrieves AWS documentation, converts it to Markdown, and provides content recommendations.
   • Use Case: Quickly finding AWS documentation or service recommendations.

5. Amazon ECS MCP Server

   • Functionality: Supports containerized application development, deployment, and troubleshooting.
   • Use Case: ECS deployment and containerized management.

More servers, including DynamoDB, Aurora, CloudWatch Logs, etc., can be found in the GitHub repository.

Technical Architecture

• Local Execution: Communicates with AI clients through stdio streams.
• AWS Lambda Adaptation: Supports deploying MCP servers as Lambda functions, suitable for cloud-based invocation.
• Security Control: Supports read-only mode, IAM integration, and sensitive data restrictions.

Example Configuration

{
  "mcpServers": {
    "awslabs.aws-serverless-mcp-server": {
      "command": "uvx",
      "args": ["awslabs.aws-serverless-mcp-server@latest"],
      "env": {
        "AWS_PROFILE": "your-aws-profile",
        "AWS_REGION": "us-east-1",
        "FASTMCP_LOG_LEVEL": "ERROR"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

Installation and Usage

Installation Methods

1. PyPI Installation:

   uv pip install awslabs.<server-name>-mcp-server
   

2. Source Code Execution:

   git clone https://github.com/awslabs/mcp.git
   cd mcp/src/<server-name>
   uv run main.py
   

3. Docker Execution:

   docker build -t awslabs/<server-name>-mcp-server .
   docker run --rm --interactive --env FASTMCP_LOG_LEVEL=ERROR awslabs/<server-name>-mcp-server:latest
   

Usage Steps

1. Configure AWS credentials.
2. Edit mcp.json to add server configurations.
3. Use an MCP-enabled AI client to interact with the server.

Advantages

• Standardized MCP protocol, compatible with various AI assistants.
• Integrates AWS best practices, ensuring high-quality output.
• Covers a wide range of AWS services.
• Open source, community-driven.

More Resources

• GitHub Repository: https://github.com/awslabs/mcp
• PyPI Package: https://pypi.org/project/awslabs.aws-documentation-mcp-server/
• AWS Blog: https://aws.amazon.com/blogs/
• Serverless Land: https://serverlessland.com/