AWS MCP Server

What is this?

The AWS MCP Server is a fully managed remote Multi-Channel Protocol (MCP) server designed by Amazon Web Services and packaged as @aws-sdk/mcp-server. It provides AI agents with secure, scalable access to over 15,000 AWS APIs and pre-built Agent SOPs, simplifying multi-step workflows across AWS services.

By unifying the AWS API MCP and AWS Knowledge servers into a single interface, it reduces complexity and accelerates development for AI-driven automation.

In addition to comprehensive API coverage, the AWS MCP Server integrates with IAM for fine-grained authorization, CloudTrail for audit logging, and handles hosting, scaling, and high availability—making it an enterprise-ready middleware solution for orchestrating infrastructure, monitoring, and compliance tasks.

Quick Start

Install the server using npm:

npm install @modelcontextprotocol/server-aws-mcp-server

Then add it to your MCP client configuration:

{
  "mcpServers": {
    "aws-mcp-server": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-aws-mcp-server"],
      "env": {
        "API_KEY": "your-api-key-here"
      }
    }
  }
}

Key Features

Feature 1: Comprehensive API coverage with unified access to over 15,000 AWS service APIs.

Feature 2: Pre-built SOPs for common workflows, reducing setup time and ensuring best practices.

Feature 3: Security-first design with IAM-based permissions and CloudTrail audit logging for full traceability.

Example Usage

For example, you can list all S3 buckets in a single execute call:

// Example code
const result = await client.callTool({
  name: "aws-mcp-server.execute",
  arguments: {
    steps: [{ name: "ListBuckets", api: "S3.ListBuckets", params: {} }]
  }
});

This code sends an MCP execute request to list S3 buckets and returns an array of bucket names in the response.

Configuration

The server accepts the following environment variables:

API_KEY – Your AWS API key for authenticating with the AWS MCP Server.

SETTING_1 (optional) – Configuration setting for specifying a custom VPC endpoint or region override.

Available Tools/Resources

mcp.execute: Main tool to perform MCP execute operations against AWS services.

Agent SOP Library: Collection of pre-built workflows like launch-ec2-instance and configure-cloudwatch-alarm.

Who Should Use This?

This server is perfect for:

Use case 1: Teams automating infrastructure provisioning and management with AI agents.

Use case 2: Enterprises needing secure, audited multi-step workflows across AWS services.

Use case 3: Developers integrating AWS operations into CI/CD pipelines and ChatOps platforms.

Conclusion

The AWS MCP Server offers a secure, scalable, and managed MCP interface for orchestrating AWS operations with minimal overhead. Give it a try to streamline your AWS automation workflows.

Check out the GitHub repository for more information and to contribute.