AWS Serverless MCP Server

0.1.19 verified Fri May 01 auth: no python

An AWS Labs Model Context Protocol (MCP) server for AWS Serverless, enabling Claude and other AI agents to interact with AWS serverless services. Current version 0.1.19, pre-1.0, actively developed by AWS Labs. Release cadence is irregular.

pip install awslabs-aws-serverless-mcp-server

Common errors

error ModuleNotFoundError: No module named 'awslabs' ↓

cause Trying to import from old package name 'aws_serverless_mcp_server' directly.

fix

Install the new package: pip install awslabs-aws-serverless-mcp-server and use from awslabs.aws_serverless_mcp_server import ...

error AttributeError: module 'awslabs.aws_serverless_mcp_server' has no attribute 'create_mcp_server' ↓

cause Outdated version of the library where the function did not exist.

fix

Upgrade to latest: pip install --upgrade awslabs-aws-serverless-mcp-server

error botocore.exceptions.NoCredentialsError: Unable to locate credentials ↓

cause AWS credentials not configured in environment or config file.

fix

Set environment variables AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY, or run 'aws configure'.

Warnings

breaking Package name changed from 'aws-serverless-mcp-server' to 'awslabs-aws-serverless-mcp-server' in version 0.1.0. Update imports and install commands accordingly. ↓

fix Use 'awslabs-aws-serverless-mcp-server' and 'from awslabs.aws_serverless_mcp_server import ...'

deprecated MCPServerConfig.aws_profile parameter deprecated in 0.1.15, use explicit credentials in config instead. ↓

fix Pass aws_access_key_id and aws_secret_access_key directly.

gotcha Requires Python >=3.10. Installing on older Python versions will fail silently or with cryptic errors. ↓

fix Ensure you have Python 3.10+ (python --version) before installing.

gotcha AWS credentials are required even for local testing. The server will fail to run without them. ↓

fix Set AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables, or configure a credentials file.

Imports

create_mcp_server

from awslabs.aws_serverless_mcp_server import create_mcp_server

MCPServerConfig

from awslabs.aws_serverless_mcp_server.config import MCPServerConfig

Quickstart

Creates and runs an MCP server with AWS credentials from environment variables.

import os
from awslabs.aws_serverless_mcp_server import create_mcp_server

config = MCPServerConfig(
    aws_region=os.environ.get('AWS_REGION', 'us-east-1'),
    aws_access_key_id=os.environ.get('AWS_ACCESS_KEY_ID', ''),
    aws_secret_access_key=os.environ.get('AWS_SECRET_ACCESS_KEY', '')
)
server = create_mcp_server(config)
server.run()