The AWS MCP Server Is Now Generally Available

For a while now, one question kept coming up when building with AI agents and MCP tools: how do you give an agent real, authenticated access to AWS without handing it the keys to the kingdom?

AWS just shipped an answer — the AWS MCP Server is now generally available. It’s a managed remote Model Context Protocol server that gives AI agents and coding assistants secure, authenticated access to all AWS services through a small, fixed set of tools.

The Problem It Solves

AI coding agents are useful for many tasks, but they run into trouble when working with AWS at any meaningful depth. Without access to current AWS documentation, agents rely on training data that may be months out of date. They don’t know about services like Amazon S3 Vectors, Aurora DSQL, or Bedrock AgentCore if those launched after their training cutoff.

When asked to build infrastructure, they reach for the AWS CLI rather than CDK or CloudFormation, and they produce IAM policies far broader than necessary. The result is infrastructure that works in a demo but isn’t production-ready.

What’s Included

The server exposes three core tools:

• call_aws — executes any of the 15,000+ AWS API operations using your existing IAM credentials. When new APIs launch, they’re supported within days.
• search_documentation / read_documentation — retrieves current AWS documentation and best practices at query time, so the agent always works from up-to-date information.
• run_script — lets the agent write a short Python script that runs server-side in a sandboxed environment. The sandbox inherits your IAM permissions but has no network access. This lets agents chain API calls, filter responses, and compute results in a single round-trip — faster and more context-efficient.

What’s New at GA

With the GA release, several new capabilities shipped:

• IAM context keys — you no longer need a separate IAM permission to use the server. Fine-grained access is expressed in a standard IAM policy.
• Documentation retrieval without authentication — anyone can look up docs without needing IAM permissions.
• Reduced token consumption — fewer tokens per interaction for complex, multi-step workflows.
• run_script tool — Python execution in a sandboxed, network-isolated environment.
• Skills (replacing Agent SOPs) — curated guidance maintained by AWS service teams for tasks where agents most commonly make mistakes.
• Enterprise audit — CloudWatch metrics under the AWS-MCP namespace and full CloudTrail capture provide the audit trail compliance teams require.

Setup

Using it with Claude Code looks like this:

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"]}'

The proxy bridges IAM SigV4 authentication to MCP’s OAuth 2.1. Once configured, you can ask your agent questions about AWS services and it will retrieve current documentation and make authenticated API calls.

Pricing and Availability

The AWS MCP Server is available today in US East (N. Virginia) and Europe (Frankfurt), and can make API calls to any region. There is no additional charge — you pay only for the AWS resources you create and any data transfer costs.

It works with Claude Code, Kiro, Cursor, Codex, and any MCP-compatible client. Check the AWS MCP Server User Guide to get started.