A repository of specialized MCP servers that give AI applications access to AWS documentation, contextual guidance and best practices. Each server is a separate package, usually launched with uvx. The README notes that the Agent Toolkit for AWS is the successor and that this repo continues to work and accept contributions.
What it does
Where AWS MCP Servers (awslabs/mcp) runs and what it adds, from its documentation
Ships many separate MCP servers (documentation, API, IaC, pricing, support, Bedrock, data and analytics, healthcare and life sciences)
AWS API MCP Server runs AWS CLI commands with validation and security controls
AWS Documentation MCP Server fetches and searches current AWS docs
AWS IaC MCP Server covers CloudFormation docs, CDK guidance and security validation
Provides an MCP AWS Lambda handler module for building your own servers
When to use it
Good fit
You want an AI assistant to read current AWS docs and API references
You want natural-language AWS operations or IaC help from your coding agent
You want to pick only the AWS servers relevant to one workload
Not for
Remote or HTTP transport: the servers are designed to support stdio only
New production work where AWS recommends the Agent Toolkit for AWS instead (per README tip)
How to install AWS MCP Servers (awslabs/mcp)
You need: uv and Python 3.10 or newer; AWS account and configured credentials for most servers
Install uv from Astral and Python 3.10+
Configure AWS credentials for the services you need
Add the server entry to your MCP client config (see config_example)
Claude Code: claude mcp add aws-docs uvx awslabs.aws-documentation-mcp-server@latest
Restart the client and see each server's own README for specific requirements