AWS MCP Servers

Official AWS Labs collection of Model Context Protocol servers for connecting AI coding assistants and agents to AWS services and documentation.

Free· Free and open source (Apache 2.0). AWS service usage billed at standard AWS rates. Optional AWS-hosted 'remote managed' servers included at no additional charge beyond consumed AWS services.MCP Servers

Visit website →

Best for

AWS-centric engineering teams using MCP-aware assistants (Cursor, Cline, Claude Desktop, Amazon Q Developer) who want their AI agents to read real AWS docs, query live resources and generate idiomatic infrastructure code.

Skip if

Teams on GCP or Azure, users of chat UIs that don't support MCP, or anyone unwilling to configure scoped IAM roles before letting an agent hold AWS write credentials.

AWS MCP Servers is an official, actively maintained collection of Model Context Protocol servers published by AWS Labs on GitHub. Each server is a small, focused program that exposes a specific slice of AWS — a service, a documentation set, or a workflow — through the standard MCP interface, so any MCP-aware client (Claude Desktop, Cursor, Windsurf, Cline, Kiro, Amazon Q Developer, custom agents built on the Anthropic, OpenAI or Bedrock SDKs) can call it as a tool.

The catalogue is broad. Infrastructure-as-code servers cover CloudFormation, CDK, Terraform and Serverless workflows; runtime servers touch EKS, ECS, Lambda, and Step Functions; data servers wrap DynamoDB, the RDS family (Postgres, MySQL, Oracle), DocumentDB, Neptune, Keyspaces, Redshift, S3 Tables and Timestream; AI/ML servers reach Bedrock knowledge bases, Bedrock Agents, and SageMaker; and cross-cutting servers surface the AWS documentation, pricing catalogue, cost explorer, CloudWatch logs and IAM. A general 'AWS API' server handles anything without a dedicated wrapper.

Typical workflows include asking a coding assistant to scaffold a CDK stack against real, current AWS docs; letting an agent query a DynamoDB table and iterate on a Lambda handler in one loop; grounding a chatbot in a Bedrock knowledge base without hand-writing retrieval code; or giving an ops copilot read-only access to CloudWatch and cost data for incident review. Servers ship with pre-built 'Agent SOPs' encoding AWS best practices, syntactic validation of API calls before execution, and IAM-scoped permissions so an agent inherits the operator's blast radius rather than a wildcard role. Both stdio and streamable-HTTP transports are supported (SSE was removed in May 2025), and AWS also offers remote, managed versions of many servers with audit logging built in.

Editor's take

This is the closest thing to a canonical MCP stack for AWS, and the fact that it comes from AWS Labs itself matters — coverage tracks the real APIs and docs rather than a third party's snapshot. Treat it as a menu, not a monolith: pick the two or three servers you actually need, scope the IAM tightly, and it becomes one of the more genuinely useful additions you can bolt onto a coding agent.

— The AI Tool Bible editorial team

Pros

✅ First-party, actively maintained by AWS Labs — coverage of new services lands quickly and stays in sync with real AWS APIs and docs
✅ Very broad surface area: compute, storage, data, AI/ML, IaC, observability, cost and documentation servers in one repo
✅ Apache 2.0 licensed and open source; runs locally over stdio or as a hosted remote server
✅ IAM-scoped permissions and syntactic validation reduce the risk of an agent issuing destructive or malformed API calls
✅ One-click install buttons for Cursor, Cline, Windsurf, Kiro and Amazon Q Developer lower setup friction significantly
✅ Pre-built Agent SOPs encode AWS Well-Architected patterns so agents produce closer-to-idiomatic infrastructure
✅ Grounding servers (AWS docs, pricing, knowledge bases) meaningfully reduce hallucinated service names and outdated API shapes

Cons

⚠️ AWS-only — no value if your stack is on GCP, Azure, or a non-hyperscaler
⚠️ Sprawling repo with dozens of servers; picking, configuring and updating the right subset takes real effort
⚠️ Powerful write-capable servers are dangerous without carefully scoped IAM roles — an over-permissive setup can let an agent create billable or destructive resources
⚠️ Requires MCP-aware client tooling; not usable from vanilla chat UIs that don't speak MCP
⚠️ Some servers are early / experimental and quality varies between the mature and newer entries
⚠️ SSE transport removal in May 2025 broke older client integrations that hadn't moved to streamable HTTP

Use cases

AWS infrastructure-as-code scaffolding with CDK or CloudFormationGrounded answers from live AWS documentationDynamoDB and RDS query and schema exploration from an IDE agentBedrock knowledge base retrieval for RAG chatbotsEKS and ECS cluster inspection and troubleshootingCloudWatch log search and incident triageAWS cost and pricing lookups for FinOps agentsLambda function development and deployment loopsTerraform plan review against AWS best practicesS3 Tables and Redshift analytical query workflows