AWS MCP Servers
AWS Labs' official collection of 80+ MCP servers, one per AWS service area (documentation, IaC, databases, serverless, cost, observability), each installable independently via uvx or Docker.
Overview
AWS MCP Servers is the open source collection maintained by AWS Labs to connect coding agents to real AWS documentation, pricing and service APIs instead of relying on a model's training data, which is often out of date for a fast-moving cloud platform.
How it works
- The repository is organized as one server per AWS service or workflow
area: documentation, CloudFormation/CDK, DynamoDB, Aurora, ElastiCache,
Lambda, ECS/EKS, Bedrock, CloudWatch, pricing and dozens more, each in
its own
src/folder with its own README. - Every server is installed and run independently through
uvx, so an agent only pulls in the servers relevant to what it's doing rather than one large, all-purpose tool. - Servers run locally by default, using your own AWS credentials, but several (like AWS Knowledge and the in-preview AWS MCP Server) are also offered as fully managed, remote servers hosted by AWS.
- AWS Labs also publishes Docker images for each server in a public ECR
registry, for teams that prefer running these in containers over
uvx. - As of mid-2026, AWS is migrating the most-used servers here into a successor product called Agent Toolkit for AWS, aimed at production agent workloads; this repository keeps working and accepting contributions in the meantime.
Examples
"Use the Terraform MCP server and the AWS Pricing MCP server to estimate
the monthly cost of this CDK stack before I deploy it."
Installation
You need uv (for uvx) or Docker, valid AWS credentials configured
locally, and an MCP client such as VS Code, Claude Code/Desktop or Cursor.
- Install
uvif you don't have it, or use AWS Labs' Docker images from its public ECR registry instead. - Configure AWS credentials locally (
aws configure, or environment variables) so the server can call real AWS APIs on your behalf. - Pick the server you need, for example the AWS Documentation server, and add its config block (see the sidebar) to your MCP client.
- Restart your client and confirm the new tools appear.
- Repeat step 3 for any other AWS service server you want; each one installs and runs independently.
Related assets
Awesome MCP Servers
The largest community catalog of MCP servers: thousands of entries across dozens of categories (databases, browser automation, cloud, communication and more), tagged by language and scope.
Context7
Pulls updated, version-specific documentation and code examples straight from the source into your prompt, so agents stop citing outdated APIs or hallucinating ones that don't exist.
Agentic Awesome Skills (AAS Core)
A 2,100+ skill catalog plus AAS Core, a local CLI/MCP control plane that lets Codex or Claude search, compose and validate an exact, reviewable skill stack before anything is installed.