Amazon ECS MCP Server

Official

AWS's official server for managing and troubleshooting Amazon ECS container deployments

How to install ↓ Compare GitHub source repository ↗

Category: Other #10 of 230
Stars: ★ 9.8k Very popular
Transport: stdio (local process)
Runtime: Python
Credentials: API key / credential required
License: Apache-2.0
Last commit: 5 days ago
Tools: 7

70FMRS · B

Maintained by the official AWS Labs team; the ALLOW_WRITE=false safe-by-default design is a real positive. It involves creating and deleting real cloud resources, so risk is higher than a typical local tool — all write-related tools are clearly listed here. This review is a static check of source and docs, not an actual deployment test.

Strongest · Maintenance 16/20 Weakest · Reliability 12/20

Reliability

12/20

Security and permissions

15/20

Maintenance

16/20

Documentation

15/20

Setup experience

12/20

Test out of dateListed 2026-07-24

Read the FMRS scoring method →

Fit and risk

What it can accessRuns commands or codeUses the networkChanges third-party account data

Best for

Developers and operators who want AI assistance across the whole AWS containerization deployment flow
Scenarios that need to quickly root-cause an ECS deployment issue

Not for

Teams without an AWS account or not using ECS
Teams highly cautious about automated production deployments that don't want an AI directly creating cloud resources (keep ALLOW_WRITE=false for read-only troubleshooting)

Required permissions

Requires valid AWS credentials (an AWS CLI profile); actual scope depends on what that IAM identity is granted
With ALLOW_WRITE=true it can create/modify/delete ECS services, ECR repos, and other cloud resources; ALLOW_SENSITIVE_DATA gates access to logs and other sensitive data

Risks and side effects

With write access enabled, the assistant's actions can directly incur cloud costs or accidentally delete production resources — test in a non-production account or with an approval workflow first
delete_app removes all resources created by an Express Mode deployment and is irreversible — confirm the target application carefully

Setup

Before you start

Runtime：Python

AWS_PROFILE required AWS CLI credential profile name defined in ~/.aws/config or ~/.aws/credentials.

AWS_REGION required Target AWS region, e.g. us-east-1.

Other optional settings (3)

FASTMCP_LOG_LEVEL optional FastMCP logging level; README examples set it to ERROR.

ALLOW_WRITE optional Whether write operations are allowed; examples set it to false for a read-only safety mode.

ALLOW_SENSITIVE_DATA optional Whether sensitive data access is allowed; examples set it to false.

Recommended: uvx --from awslabs-ecs-mcp-server ecs-mcp-server (requires an AWS CLI profile with appropriate permissions configured beforehand). ALLOW_WRITE=false and ALLOW_SENSITIVE_DATA=false by default — a deliberate safe default requiring explicit opt-in for write access and sensitive data. AWS also offers a fully managed remote version; see docs.aws.amazon.com.

claude_desktop_config.json

{
  "mcpServers": {
    "awslabs.ecs-mcp-server": {
      "command": "uvx",
      "args": ["--from", "awslabs-ecs-mcp-server", "ecs-mcp-server"],
      "env": {
        "AWS_PROFILE": "<your-aws-profile>",
        "AWS_REGION": "<your-aws-region>",
        "FASTMCP_LOG_LEVEL": "ERROR",
        "ALLOW_WRITE": "false",
        "ALLOW_SENSITIVE_DATA": "false"
      }
    }
  }
}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "awslabs.ecs-mcp-server": {
      "command": "uvx",
      "args": [
        "--from",
        "awslabs-ecs-mcp-server",
        "ecs-mcp-server"
      ],
      "env": {
        "AWS_PROFILE": "<your-aws-profile>",
        "AWS_REGION": "<your-aws-region>",
        "FASTMCP_LOG_LEVEL": "ERROR",
        "ALLOW_WRITE": "false",
        "ALLOW_SENSITIVE_DATA": "false"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add awslabs.ecs-mcp-server -e 'AWS_PROFILE=<your-aws-profile>' -e 'AWS_REGION=<your-aws-region>' -e FASTMCP_LOG_LEVEL=ERROR -e ALLOW_WRITE=false -e ALLOW_SENSITIVE_DATA=false -- uvx --from awslabs-ecs-mcp-server ecs-mcp-server

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After install, the client tool list should show seven tools including ecs_troubleshooting_tool; running the read-only validate_ecs_express_mode_prerequisites confirms the connection and AWS credentials work.

Troubleshooting

Permission errors: check that the IAM identity behind the AWS CLI profile has the required ECS/ECR/IAM permissions
Express Mode deployment fails: run validate_ecs_express_mode_prerequisites first to check prerequisites are met
Service stays not-ready: use wait_for_service_ready or ecs_troubleshooting_tool to find the specific blocker

Things to try

Once connected, you can ask your AI assistant things like:

Containerize this application and deploy it to ECS Express Mode
Check whether my environment meets the ECS Express Mode prerequisites
Help me troubleshoot why my ECS service keeps becoming unstable
Delete this ECS application and its related resources

Tools 7

containerize_app writes

Provide best-practice guidance for containerizing an application

build_and_push_image_to_ecr writes

Create ECR infrastructure and build/push a Docker image

validate_ecs_express_mode_prerequisites read-only

Validate required IAM roles and images exist before deployment

ecs_resource_management writes

List/view/create ECS resources (clusters, services, tasks, task definitions) and ECR images

wait_for_service_ready read-only

Track ECS deployment progress and wait for the service to become ready

delete_app destructive

Clean up all resources created by an Express Mode deployment

ecs_troubleshooting_tool read-only

Diagnose and troubleshoot common ECS deployment problems

Use cases

Containerize a web app and deploy it to ECS without hand-writing a Dockerfile or CloudFormation template

Have an AI assistant diagnose why an ECS service deployment failed or is unhealthy

List the ECS clusters, services, and task definitions in the current account

Clean up an ECS Express Mode deployment and its associated resources that are no longer needed

Supported clients

Claude Desktop

Claude Code

VS Code

Cursor

Windsurf

Kiro

Cline

Listed from the project's documentation, not tested by this site.

Overview

An MCP server officially maintained by AWS Labs (the ECS sub-project within the awslabs/mcp monorepo), covering the full lifecycle from containerizing an app to ECS deployment and troubleshooting: containerization best-practice guidance, ECR image build/push, one-click ECS Express Mode deployment (auto-configured load balancer, auto-scaling, CloudFormation infrastructure), resource listing, and deployment diagnostics. AWS also offers a fully managed remote ECS MCP service; this entry covers the locally-runnable open-source implementation.

Similar servers

Hevy MCP 77 · B

Manage your Hevy workout data from AI assistants via MCP.

★ 517 · Tools 26 Compare with this →

Firecrawl MCP Server 75 · B

Firecrawl's official MCP server for web search, scraping, and structured extraction for AI agents

★ 7.6k · Tools 11 Compare with this →

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 659 · Tools 11 Compare with this →

Compartment 75 · B

Encrypted, fully offline long-term memory for AI agents

★ 583 · Tools 14 Compare with this →

Compare with other MCP servers

Source revision d534d0fa3c06 Data synced 2026-10-19 Read the FMRS scoring method