AWS MCP Servers

A collection of open-source MCP servers connects AI assistants to AWS documentation and cloud services, with locally run servers and AWS-hosted options.

AWS MCP Servers connects coding assistants and AI agents to AWS documentation, infrastructure tools, and cloud services. It's for developers who want their assistant to work with AWS resources and service-specific guidance.

Documentation and guidance: Give assistants access to AWS documentation and API references, plus CloudFormation guidance, CDK examples, security validation, and deployment troubleshooting.
Infrastructure and deployment: Let agents manage EKS clusters, deploy applications to ECS, work on serverless applications with SAM CLI, and build containers locally with Finch.
Data and knowledge retrieval: Connect assistants to DynamoDB and Aurora database operations, or query Amazon Bedrock Knowledge Bases with citations and search Amazon Kendra indexes.
Operations and costs: Analyze CloudWatch metrics and logs, troubleshoot operational issues, retrieve AWS pricing estimates, and work with billing and cost data.
Workflow execution: Expose Lambda functions and Step Functions workflows as agent tools, alongside SNS and SQS messaging operations.

The repository's servers run locally on macOS, Windows, and Linux, or in containers, and use stdio connections with MCP clients such as Cursor, Cline, Kiro, and Claude Code. They're open source under Apache 2.0. Local execution doesn't make AWS operations offline: service requests go to AWS. The separate AWS Knowledge MCP Server runs on AWS infrastructure and provides documentation and architectural guidance.

Agent Toolkit for AWS is the project's successor and the maintainers' recommended choice for production agent development. This collection remains maintained and accepts contributions.