CookiesForAI

AWS MCP server

The AWS MCP servers give AI agents access to AWS: a managed AWS MCP Server for API calls and docs, plus more than 50 open-source servers for services like CDK, EKS, Lambda and DynamoDB.

By Amazon Web Services Official · Updated

Connect AWS

Claude Code, managed AWS MCP Server through the AWS proxy

Terminal
claude mcp add-json aws-mcp --scope user '{"command":"uvx","args":["mcp-proxy-for-aws==1.6.0","https://aws-mcp.us-east-1.api.aws/mcp","--metadata","AWS_REGION=us-west-2"]}'

Claude Code, open-source AWS Documentation server

Terminal
claude mcp add aws-docs uvx awslabs.aws-documentation-mcp-server@latest

Claude Code, open-source infrastructure as code server

Terminal
claude mcp add aws-iac uvx awslabs.aws-iac-mcp-server@latest

Any client with a JSON config, one open-source server

Config
{"mcpServers": {"awslabs.aws-documentation-mcp-server": {"command": "uvx", "args": ["awslabs.aws-documentation-mcp-server@latest"], "env": {"FASTMCP_LOG_LEVEL": "ERROR", "AWS_DOCUMENTATION_PARTITION": "aws"}}}}

Configure AWS credentials before adding a server: the open-source servers read your local AWS profile, and the proxy signs managed-server requests with the same credentials. Each awslabs server README has one-click installs for Kiro, Cursor and VS Code. Set AWS_REGION in the metadata to the Region your resources are in.

What the AWS MCP server does

AWS publishes MCP servers as a family, not a single server. The awslabs/mcp repo holds more than 50 open-source servers, each for one area: AWS documentation, infrastructure as code with CloudFormation and CDK, EKS and ECS containers, serverless with SAM, DynamoDB, Aurora and other databases, IAM, pricing, support cases and more. Each runs locally over stdio with uvx and your AWS credentials, and you add only the ones you need.

The AWS MCP Server is the managed option, generally available since 2026-05-06 in US East (N. Virginia) and Europe (Frankfurt). AWS hosts it, and its call_aws tool can run any of more than 15,000 AWS API operations under your IAM permissions, with CloudTrail logging every call. It also searches and reads current AWS docs and runs Python scripts in a sandbox. Clients reach it through mcp-proxy-for-aws, an open-source proxy that signs requests with your IAM credentials.

AWS now calls the Agent Toolkit for AWS, launched 2026-05-06, the successor to the awslabs servers, plugins and skills; the repo still works and accepts contributions, and its most used projects will move into the toolkit over time. The awslabs/mcp repo had 9,746 GitHub stars on 2026-10-01. The managed server is listed in the MCP Registry as io.github.aws/aws-mcp.

Tools it gives your agent

  • call_aws
  • search_documentation
  • read_documentation
  • run_script

When to use AWS

  • You want the agent to check current AWS docs before writing CDK or CloudFormation.
  • You want to ask about resources in your account, such as Lambda functions or S3 buckets, in plain words.
  • You want the agent to debug an EKS or ECS deployment with real cluster data.
  • You want a cost estimate for an architecture before you build it.

When to pick something else

  • Broad credentials: call_aws can change and delete resources with whatever your IAM role allows, so give the agent a scoped role and keep CloudTrail on.
  • Picking one install: the family has dozens of servers and some are deprecated, such as the Cloud Control API server, so check each server's README first.

What AWS needs

  • An AWS account and credentials configured locally (AWS CLI profile or IAM role)
  • uv and Python 3.10 or newer for the open-source servers and the proxy

What the AWS MCP server costs

No charge for the AWS MCP servers, and the open-source ones are Apache-2.0; you pay for the AWS resources your agent creates or uses, and AWS has a free tier for some services.

Current prices: aws.amazon.com . We do not list prices: vendors change them often, so check the publisher's own page.

Questions people ask

Is the AWS MCP Server free?

Yes. AWS charges nothing extra for the managed AWS MCP Server or the open-source servers; you pay only for AWS resources the agent creates or uses and any data transfer (AWS announcement read 2026-10-01).

Which AWS MCP server should I start with?

AWS recommends the managed AWS MCP Server for broad, audited access to AWS APIs and docs. Add an awslabs server when you need a specific service, such as the EKS server for Kubernetes clusters.

  • Cloudflare: The Cloudflare MCP servers give AI agents access to a Cloudflare account through hosted endpoints: read Workers logs and builds, manage bindings, check DNS analytics, search the docs and render web pages.
  • Kubernetes: The Kubernetes MCP server gives AI agents direct access to Kubernetes and OpenShift clusters: list and edit any resource, read pod logs, exec into pods, view events and install Helm charts.
  • Sentry: The Sentry MCP server gives AI coding agents access to your Sentry account: search errors and events, read issue details and stack traces, look at performance data and triage issues while fixing code.
  • Terraform: The Terraform MCP server gives AI agents current Terraform Registry data on providers, modules and policies, and with a token manages HCP Terraform and Terraform Enterprise workspaces, runs and plans.