Deploys a Model Context Protocol (MCP) server designed to run on AWS Lambda, integrating custom tools with AI clients.
This project provides a robust implementation of a Model Context Protocol (MCP) server engineered to operate efficiently within a serverless AWS environment. It enables seamless integration with MCP-compatible clients, such as Claude Desktop, by exposing custom AI-driven functionalities. Built with AWS CDK for infrastructure-as-code, it leverages AWS Lambda for serverless execution, AWS Cognito for secure OAuth 2.0 authentication, AWS DynamoDB for data storage, and AWS API Gateway for HTTP endpoints. The framework supports the addition of custom MCP tools, exemplified by an included weather forecasting utility, and features a Hono.js-based RESTful API.
Key Features
01Serverless MCP server on AWS Lambda
02Extensible architecture for custom MCP tools (e.g., WeatherTool)
03Hono.js-based RESTful API endpoints
04AWS Cognito OAuth 2.0 authentication
05Infrastructure as Code with AWS CDK
060 GitHub stars
Use Cases
01Deploying custom AI agent tools accessible via MCP clients like Claude Desktop
02Building scalable and secure serverless backend services for AI applications
03Integrating custom business logic or data sources with Model Context Protocol
