AWS MCP Server

Interact with Amazon Web Services through your AI assistant - manage EC2, S3, Lambda, and hundreds of AWS services.

April 14, 2026
MCP ServerDevelopment Tools & DevOpsAWS MCP Server

AWS MCP Server

Manage AWS infrastructure and services directly from your AI assistant. The AWS MCP server provides natural language interfaces for managing EC2, S3, Lambda, IAM, CloudWatch, and hundreds of other AWS services across multiple regions and accounts.

This server is essential for cloud architects managing large AWS environments, DevOps teams automating deployment workflows, and security engineers performing compliance audits across their AWS organization.

What it Does

  • Resource Management: List, create, and manage AWS resources
  • Cost Analysis: Query billing and cost data
  • Security Audit: Review IAM policies and security configurations
  • Monitoring: Access CloudWatch logs and metrics
  • Multi-Region: Work across multiple AWS regions

Installation

# Using npx
npx -y @anthropic-ai/mcp-server-aws

# Using Docker
docker run -i --rm -v ~/.aws:/root/.aws mcp/aws

Configuration

{
  "mcpServers": {
    "aws": {
      "command": "npx",
      "args": ["-y", "@anthropic-ai/mcp-server-aws"],
      "env": {
        "AWS_REGION": "us-east-1",
        "AWS_PROFILE": "production"
      }
    }
  }
}

Available Tools

ToolPurposeCategory
list_ec2List EC2 instances with filteringCompute
list_s3_bucketsList S3 bucketsStorage
list_lambdasList Lambda functionsServerless
list_iam_rolesList IAM rolesSecurity
get_costsQuery AWS billing and cost dataCost
get_cloudwatch_logsRetrieve CloudWatch log groupsMonitoring
check_public_s3Identify publicly accessible S3 bucketsSecurity
list_rdsList RDS instancesDatabases

Common Workflows

Security Audit

  1. List all IAM roles and check for overly permissive policies
  2. Identify S3 buckets with public access enabled
  3. Review CloudTrail logs for unusual API activity
  4. Check security group rules for open ports to 0.0.0.0/0
  5. Generate a security findings summary

Cost Investigation

  1. Query cost data for the current month
  2. Identify the top-spending services
  3. List EC2 instances and check for idle or underutilized resources
  4. Review RDS instances for appropriate instance sizes
  5. Generate cost optimization recommendations

Resource Inventory

  1. List EC2 instances across all regions
  2. Check Lambda functions and their recent invocation counts
  3. Inventory S3 bucket sizes and object counts
  4. Review RDS instance configurations
  5. Compile a complete resource inventory report

Example Prompts

  • "List all EC2 instances in us-east-1"
  • "Show me S3 buckets with public access"
  • "What Lambda functions did I deploy this week?"
  • "Check IAM policies for security issues"
  • "Show CloudWatch logs for the api-service"
  • "What are my top 5 cost centers this month?"

Troubleshooting

IssueLikely CauseSolution
"Access denied"IAM credentials lack permissionsVerify the IAM policy allows the requested action
"Region not found"Unsupported or disabled AWS regionCheck AWS_REGION and verify the region is enabled in your account
"Resource not found"Resource doesn't exist in the specified regionVerify the resource exists and you have the correct region
"Throttling"API rate limit exceededImplement exponential backoff or request a rate limit increase

Security Notes

  • Use least-privilege IAM credentials
  • Prefer read-only operations when possible
  • Enable CloudTrail for audit logging
  • Use AWS SSO for credential management