mcp-registry/aws-mcp-server

    ==================
      
       /// MCP ///
      /// AWS ///
        
    ==================
        
    [server:online]
    [protocol:ready]

aws-mcp-server

by alexei-led

Light-weight Model Context Protocol (MCP) server that lets AI assistants safely request AWS-CLI documentation and execute AWS commands (with Unix-pipe support) inside an isolated Docker container.

143
21
Open Source
GitHub

01

aws_cli_help

Retrieve detailed AWS CLI documentation for AWS services and commands.

02

aws_cli_pipeline

Execute AWS CLI commands (with optional Unix pipes) and return formatted results optimized for AI consumption.

Installation

1. Prerequisites:
• Docker installed and running
• An AWS IAM user or role with the minimal permissions you want the server to have
• Environment variables AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and (optionally) AWS_SESSION_TOKEN, AWS_REGION set
2. Clone the repository
git clone https://github.com/alexei-led/aws-mcp-server.git cd aws-mcp-server
3. Build the container image (replace tag as desired)
docker build -t aws-mcp-server:latest .
4. Launch the server (exposes MCP on port 3333 by default)
-e AWS_ACCESS_KEY_ID=$AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY=$AWS_SECRET_ACCESS_KEY \
-e AWS_SESSION_TOKEN=$AWS_SESSION_TOKEN \
-e AWS_REGION=${AWS_REGION:-us-east-1} \
-p 3333:3333 \
docker run -d --name aws-mcp-server \ aws-mcp-server:latest # OR with an AWS credentials file mounted # docker run -d -p 3333:3333 -v $HOME/.aws:/root/.aws aws-mcp-server:latest
5. Verify it is reachable
curl http://localhost:3333/healthz
6. Point your MCP-aware client (Claude, Cursor, etc.) at http://<host>:3333 to start issuing AWS CLI commands.

Documentation

License: MIT License
Updated 7/15/2025