MCP server
AWS MCP server
The AWS API MCP server from AWS Labs lets AI run AWS CLI commands, suggest commands and plan changes, with read-only and consent modes.
- Official server
- Cloud and infrastructure
At a glance
- Maintained by
- AWS Labs (Amazon Web Services)
- Tools exposed
- 3 listed below
- Reviewed
- September 24, 2026
Why it matters
What it does and why it matters
AWS runs a large share of business infrastructure. The AWS API MCP server, from AWS Labs, gives an assistant three tools: run AWS CLI commands, suggest the right commands for a task and produce an execution plan. That covers almost any AWS service, from listing resources to reading logs and, if allowed, making changes.
AWS now also offers a managed AWS MCP Server, generally available since May 2026, and its repository recommends it first for many uses. The API server remains useful for teams that want to run it themselves.
Two settings matter most. Read-only mode blocks any change, and consent mode asks for confirmation before each write. We use read-only mode by default, with a dedicated IAM role that has minimal permissions, and never connect admin credentials.
Use cases
Best business use cases
Resource inventory
Incident help
Cost questions
Command help
Tools
Available tools
Tool names as published by the maintainer at review time. Versions change, so check the repository for the current list.
The AWS API server only exposes three tools; call_aws reaches almost every AWS API through the CLI.
Read and write: Read and write across AWS APIs. Set READ_OPERATIONS_ONLY=true for read-only, or REQUIRE_MUTATION_CONSENT=true to confirm each write.
-
call_awsRun a validated AWS CLI command and return the result.
-
suggest_aws_commandsSuggest AWS CLI commands from a plain-language request.
-
get_execution_planExperimental. Give step-by-step guidance for complex AWS tasks.
Workflows
Example workflows
Find idle resources
-
1
Connect in read-only mode.
-
2
Ask for stopped instances and unattached volumes.
-
3
The assistant runs list commands.
-
4
It totals likely monthly cost.
-
5
An engineer decides what to remove.
Plan a change safely
-
1
Describe the change.
-
2
The assistant produces an execution plan.
-
3
An engineer reviews it.
-
4
With consent mode on, each write is confirmed.
-
5
The change is logged in CloudTrail.
Prompts
Example prompts
“List EC2 instances in us-east-1 that have been stopped for over 30 days.”
“Which S3 buckets allow public access?”
“Show me the plan to add a tag to all production instances, but do not run it.”
Setup
Installation and configuration
Examples for Claude Desktop and Claude Code. Replace the placeholder values with your own keys, and store secrets outside version control.
Claude Desktop: claude_desktop_config.json
{
"mcpServers": {
"awslabs.aws-api-mcp-server": {
"command": "uvx",
"args": [
"awslabs.aws-api-mcp-server@latest"
],
"env": {
"AWS_REGION": "us-east-1"
},
"disabled": false,
"autoApprove": []
}
}
}Claude Code: terminal
claude mcp add awslabs-aws-api-mcp-server -e AWS_REGION=us-east-1 -- uvx awslabs.aws-api-mcp-server@latest- Hosted server: Self-deploy to Amazon Bedrock AgentCore (see DEPLOYMENT.md). AWS also runs a managed AWS MCP Server (preview) at https://aws-mcp.us-east-1.api.aws/mcp..
- Local server: PyPI: awslabs.aws-api-mcp-server (run with uvx, stdio).
- Needs: Standard AWS credentials (profile via AWS_API_MCP_PROFILE_NAME, or default credential chain). IAM permissions limit what it can do..
Security
Security considerations
-
Read-only first
Start with read access and add actions only for reviewed tasks.
-
Scoped credentials
Use a dedicated role or token with the least access, never an owner or admin key.
-
Separate environments
Keep production changes out of assistant sessions, or require approval for them.
-
Audit
Use AWS CloudTrail to review every call.
Related servers
Servers that pair well with it
AWS S3
Amazon S3 is handled through the managed AWS MCP Server, which lets AI run scripted AWS actions, create presigned links and search AWS docs.
Azure
Microsoft's Azure MCP server lets AI work with many Azure services, from storage and Key Vault to SQL, monitoring and AKS, with a read-only mode.
Google Cloud
Google's gcloud MCP servers let AI run gcloud commands, read logs, metrics, alerts and traces, and work with Cloud Storage.
Keep exploring
Solutions and services that use it
Related services
View all related services- MCP integration Connect AI assistants to your CRM, files, databases and tools through Model Context Protocol servers, with safe permissions.
- Product scaling Keep a growing product fast and stable as users, data and features increase.
- Multi-tenant architecture One product serving many customer accounts, with each account's data kept separate and secure.
- User management and authentication Sign-up, login, roles and permissions done properly, including single sign-on for business customers.
- API integration Make two systems share data reliably, with retries, logging and alerts when something goes wrong.
- AI-Powered Web Development Business websites, customer portals, dashboards and web applications that load fast, rank well and are easy to change.
Solutions
View all solutionsIndustries
View all industries- Property management Tenant portals, maintenance tracking, owner statements and inbox automation for property managers.
- Insurance Claims intake, document and photo processing, policy knowledge assistants and customer portals for brokers and insurers.
- Logistics and transportation Driver apps, order and shipment tracking, document processing and integrations for logistics and delivery companies.
Case studies
View all case studiesGuides and articles
View all guides and articles- How we review AI-generated code The checks we apply to every AI-assisted change, from tests and security to readability and business rules.
- How to connect AI to your database safely Let staff ask questions of your data in plain English without risking production systems or sensitive records.
- MCP security best practices How to choose, configure and monitor MCP servers so AI assistants get the access they need and nothing more.
- How to automate reporting with AI and MCP Build reports that assemble themselves from your systems through MCP, with a written summary people can trust.
- Multi-tenant SaaS architecture explained How SaaS products serve many customers from one system, the main design options and how to choose.
- How to set up automated reporting Replace manual weekly and monthly reports with dashboards and summaries that update themselves.
Glossary terms
View all glossary terms- API integration An API integration connects two or more systems through their APIs so data and actions flow between them automatically.
- Multi-tenant Multi-tenant describes software where one application and database serve many customers, called tenants, while keeping each customer's data separate.
- Cloud hosting Cloud hosting runs websites and applications on servers rented from a provider, which can grow or shrink with demand.
- Reverse proxy A reverse proxy is a server that sits in front of your applications, receiving visitors' requests and passing them to the right app, often adding security and caching.
- VPS A VPS, or virtual private server, is a rented virtual machine with its own resources, on which you can install and run your own software.
- API An API, or application programming interface, is a defined way for one piece of software to request data or actions from another.
FAQ
Questions people ask us
Have a question that is not here? Ask us directly.
Yes. It comes from AWS Labs. AWS also offers a managed AWS MCP Server.
Yes. Read-only mode blocks writes, and consent mode asks before each write.
One tool runs AWS CLI commands, which covers almost every AWS service.