# awslabs/mcp [Health: Active]

**Category:** ☁️ Cloud Platforms  
**Repository:** https://github.com/awslabs/mcp  
**GitHub Stars:** 9836  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/awslabs-mcp

## Description
AWS MCP servers for seamless integration with AWS services and resources.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What awslabs/mcp does

awslabs/mcp groups open-source MCP servers for AWS into a single repository. The servers give MCP-compatible applications access to AWS documentation, contextual guidance, and best practices, along with specialized capabilities for working with AWS services and resources.

The collection covers a broad set of cloud-development needs. Its catalog includes infrastructure and deployment, container platforms, serverless and functions, AI and machine learning, databases, search and analytics, backend APIs, caching, messaging, cost and operations, healthcare, and developer support. Individual servers are organized around particular AWS workflows rather than presenting one universal tool surface.

This makes the repository relevant for developers using AI-assisted cloud development, infrastructure management, application development, data operations, and conversational assistants that need AWS-specific context.

## How it works

Each MCP server is a lightweight process that maintains a client connection through the Model Context Protocol. An MCP-compatible host can then discover and use the capabilities exposed by the selected server alongside the model’s normal conversation or coding workflow.

The servers in this repository use stdio transport. Communication takes place through standard input and standard output, rather than through a hosted HTTP endpoint. The repository also describes an AWS Lambda handler module for running MCP server functionality in Lambda-based environments.

The collection is intended for use with local AI applications such as coding assistants, IDEs, and chatbot applications. It can provide AWS reference material and service-specific context while an agent helps with cloud-native development or operations.

## Setup and configuration

Installation depends on the specific server selected from the repository. The project provides setup guidance for macOS, Linux, and Windows, as well as container-based execution. Client configuration examples are available for Claude Desktop, Cline, Cursor, Windsurf, VS Code, Claude Code, Kiro, and fx.

Because the servers use stdio, the client configuration needs to launch the chosen server as a local process. Select the server that matches the AWS workflow you want to support, then add its command and settings to the configuration format used by your MCP client. The repository also includes examples for project-level configuration in clients such as Cursor, VS Code, and Claude Code.

AWS access and permissions should be configured according to the requirements of the selected server and the AWS resources it accesses. Keep those permissions scoped to the actions required by the agent workflow.

## Tools and capabilities

The repository’s server collection supports capabilities such as:

- Retrieve AWS documentation and related technical context.
- Provide AWS-focused guidance and development best practices.
- Assist with infrastructure and deployment workflows.
- Support serverless, container, database, analytics, and machine-learning use cases.
- Extend coding assistants and chat applications with AWS-specific MCP tools.
- Run MCP functionality locally through stdio or through the documented Lambda handler module.

## Limitations and notes

The repository is a collection of specialized servers, so available tools depend on which server you install. Its standard transport is stdio; SSE support was removed from the latest major versions, and streamable HTTP is described as future work. Applications that require SSE need to use a previous major version of the relevant server.

AWS permissions, client configuration, and applicable policies remain the operator’s responsibility. The repository also identifies the Agent Toolkit for AWS as the successor for production software built with coding agents or customer-facing agents, while this project continues to work and accept contributions.

_Full upstream README: https://allmcps.com/mcp/awslabs-mcp/readme_