Aws Common MCP Servers

🚀 AWS Public MCP Server and CDK Deployment Guide

This project offers deployable Model Context Protocol (MCP) servers tailored for common AWS services. It comes with AWS Cloud Development Kit (CDK) code written in Python, which is used to provision the necessary infrastructure for running these servers on ECS Fargate.

The goal is to provide reusable components for AI developers (e.g., those using Amazon Bedrock, LangChain), enabling them to effortlessly deploy these servers to their AWS accounts and interact with AWS services through the standardized MCP interface.

✨ Features

Included MCP Servers

1. AWS Location Service (TypeScript): Offers functions such as searching for locations, retrieving location details, and calculating routes. Based on aws-location-server.
2. Amazon S3 (Python): Provides basic S3 operation tools, like listing buckets/objects, getting, uploading, and deleting objects.
3. Amazon Aurora PostgreSQL (via RDS Data API) (Python): Allows executing SQL statements to a specified Aurora PostgreSQL cluster using the secure RDS Data API.

📦 Installation

Prerequisites

• An AWS account.
• A locally configured AWS CLI with appropriate permissions (for CDK deployment). Credentials should be set via the ~/.aws/credentials file or environment variables.
• Node.js and npm (for the location service server's build process and CDK).
• Python 3.9+ (for CDK and Python-based MCP servers).
• AWS CDK CLI (npm install -g aws-cdk).
• Docker (for local running and CDK to build container images).
• Git.

Specific Prerequisites for Deploying the Aurora PostgreSQL Data API Server:

• An existing running Aurora PostgreSQL-compatible cluster in the target AWS region.
• The cluster must have the RDS Data API enabled.
• A secret containing database credentials (username, password) in AWS Secrets Manager. The secret ARN is required for deployment.

Project Structure

aws-mcp-infra/
├── mcp_servers/             # Source code for MCP servers
│   ├── location_service/    # TypeScript server for AWS Location Service
│   ├── s3/                  # Python server for S3
│   └── aurora_pg_data_api/  # Python server for Aurora PG data API
├── cdk_deployment/          # CDK deployment code and related configurations
│   ├── app.py              # Main application file
│   ├── stack.py            # Stack definition file
│   └── README.md           # Deployment instruction document
└── .env                     # Environment variable file for CDK configuration

Installation and Configuration

1. Install Dependencies

Run the following commands in the project directory to install the required dependencies:

pip install -r requirements.txt
npm install --location_service

2. Configure Environment Variables

Update the configuration in the .env file according to your AWS account and region:

AWS_ACCOUNT_ID=your_account_id
AWS_REGION=your_region
DB_CLUSTER_ARN=your_db_cluster_arn
DB_SECRET_ARN=your_db_secret_arn

Deployment

1. Initialize the CDK Environment

Run the following command in the project root directory to initialize the CDK environment:

cdk init --language python

2. Create and Deploy the Stack

Create a new CDK stack file stack.py, then run the following command to deploy the resources:

cdk synth && cdk deploy

💻 Usage Examples

1. Usage in AWS

• Lambda Functions or Bedrock Agents: Use the AWS SDK (e.g., boto3 in Python) to invoke ECS tasks, or interact via AWS Systems Manager Session Manager (if configured). The specific mechanism depends on the design of the client application.
• Local Testing: You can directly run these servers before deployment. For the location service, run node build/index.js; for the Aurora Data API server, run python aurora_pg_data_api_server.py. Ensure that the AWS credentials are configured in the local environment, and for the Aurora server, set the DB_CLUSTER_ARN and DB_SECRET_ARN environment variables.

Cleanup

To delete all deployed resources, run:

cdk destroy --all

Through the above steps, you can easily deploy and manage these MCP servers on AWS.