Claude connectors
Claude and AWS: what the AWS MCP connector does and what a cloud team should settle first
A plain guide for cloud, platform and security leads. Every fact about the connector comes from Anthropic's directory page and Amazon Web Services' own pages, read on 10 November 2026.
The Claude AWS integration is a connector called AWS MCP, made by Amazon Web Services and listed in Anthropic's connector directory with an Anthropic verified badge. AWS describes it as a managed server that gives agents access to AWS through the Model Context Protocol, an open standard for connecting AI to tools. Searching AWS documentation needs no sign-in, while API calls use your existing IAM credentials, and AWS says CloudTrail logs all API calls. This page covers what it is and what a cloud team should decide before turning it on.
plain answer
What is the AWS MCP connector for Claude?
The Claude AWS integration is a connector named AWS MCP, made by Amazon Web Services. Anthropic's directory describes it as "a managed server that gives agents access to AWS through MCP." MCP is the Model Context Protocol, an open standard that connects AI models to outside tools. The listing carries an Anthropic verified badge, says it was added in August 2026 and shows no beta label.
The directory lists nine tools, among them aws___call_aws, aws___search_documentation, aws___read_documentation and aws___run_script. AWS's own page says agents "can execute AWS CLI commands, search AWS documentation, and follow curated skills."
Source: Anthropic connector directory, AWS MCP (https://claude.com/marketplace/connectors/aws-mcp), and AWS, Agent Toolkit for AWS (https://aws.amazon.com/products/developer-tools/agent-toolkit-for-aws), read 10 November 2026.
plain answer
Who uses AWS with Claude?
The listing's categories are Communication, Data & analytics, Developer tools and Productivity, and its tools cover documentation, regional availability and API calls. Our reading, not a claim from AWS or Anthropic: cloud engineers and platform teams are the main users, and developers who want AWS documentation answered in a chat are the lightest users, since AWS says documentation search needs no authentication.
Source: Anthropic connector directory, AWS MCP (https://claude.com/marketplace/connectors/aws-mcp), and AWS documentation (https://docs.aws.amazon.com/agent-toolkit/latest/userguide/mcp-server.html), read 10 November 2026.
how to
How do you connect Claude to AWS?
Anthropic's directory page shows an Add to Claude button and two connector addresses, one in us-east-1 and one in eu-central-1. It marks sign-in as not required on the listing. AWS's documentation explains why: agents can search documentation and retrieve service information without authentication, but to run AWS API calls, run Python scripts in a sandbox or follow curated skills, "agents authenticate through your existing IAM credentials."
Your AWS administrators own the IAM side. AWS's documentation is the one to follow for it.
Source: Anthropic connector directory, AWS MCP (https://claude.com/marketplace/connectors/aws-mcp), and AWS documentation (https://docs.aws.amazon.com/agent-toolkit/latest/userguide/mcp-server.html), read 10 November 2026.
security
What should a cloud team decide before turning it on?
AWS says all capabilities sit behind "a single endpoint with CloudWatch metrics and IAM-based access controls" and that "CloudTrail logs all API calls for audit visibility." AWS's product page FAQ says IAM policies can limit which actions agents take. The same page says Agent Toolkit for AWS is available at no additional charge, with you paying only for the AWS resources your agent provisions or interacts with.
Four questions are ours, not AWS's. Which IAM role will Claude act under, and what may it do? Which accounts and regions are in scope? Who approves any change Claude proposes before it runs? Who reads the CloudTrail record afterwards?
Source for the statements: AWS documentation and AWS, Agent Toolkit for AWS, read 10 November 2026. The four questions are Kriv's own suggestions.
kriv role
Where Kriv helps
Kriv AI runs Claude training for teams, including a Claude Code track for engineering teams that covers MCP integrations and a governance module on who may run what and what gets logged. A cloud team can start there, then set the IAM rules above. Kriv also publishes an AI governance consulting page for AWS.
Kriv did not build the AWS MCP connector and does not speak for Amazon Web Services. Your AWS administrators hold the access and the settings.
limits
What this page could not confirm
The pages we read do not say which Claude plans include the connector. We did not test it, and we did not read anything about using it from Claude Code, so this page makes no claim about that. For anything not cited above, check AWS's and Anthropic's current pages. Read 10 November 2026.
Related resources
Continue exploring
Straight answers
Frequently asked questions about Claude and AWS: what the AWS MCP connector does and what a cloud team should settle first
What is AWS MCP?
AWS describes it as a managed server that gives agents access to AWS through the Model Context Protocol (AWS documentation, read 10 November 2026).
Do I need to sign in to use it?
AWS says documentation search needs no authentication. API calls, sandboxed scripts and curated skills use your existing IAM credentials (read 10 November 2026).
Does AWS log what Claude does?
AWS says CloudTrail logs all API calls and that CloudWatch metrics are available (read 10 November 2026).
Does it cost extra?
AWS's page says Agent Toolkit for AWS is available at no additional charge and you pay only for the AWS resources your agent provisions or interacts with (read 10 November 2026).
Who made it?
Amazon Web Services, shown with an Anthropic verified badge in Anthropic's directory (read 10 November 2026).
Talk to the team that would do the work
Bring your requirements to a working session with the person who'll actually deliver.
Book a Discovery Call