MCP Lambda SAM logo

MCP Lambda SAM

by markvp

This project provides a serverless implementation of the Model Context Protocol (MCP) using AWS Lambda and SAM. It offers two distinct interfaces: System Configuration (Administrative) and System Usage (Client).

View on GitHub

Last updated: N/A

What is MCP Lambda SAM?

MCP Lambda SAM is a serverless implementation of the Model Context Protocol (MCP) built using AWS Lambda and SAM. It allows for the registration and management of MCP tools, resources, and prompts, as well as establishing SSE connections, sending commands, and receiving streaming responses.

How to use MCP Lambda SAM?

Administrators can deploy the application using various methods, including AWS Serverless Application Repository (SAR), npx, or local development. They can then register MCP tools, resources, and prompts using the Registration API. Clients can connect to the server via SSE and send commands using the message endpoint, ensuring they have the necessary IAM permissions.

Key features of MCP Lambda SAM

  • Serverless architecture

  • Administrative and client interfaces

  • Dynamic Lambda tool invocation

  • SSE connection support

  • IAM permission management

  • Registration API

Use cases of MCP Lambda SAM

  • Orchestrating serverless AI tools

  • Building conversational AI applications

  • Creating event-driven systems

  • Managing and invoking Lambda functions dynamically

  • Implementing a streaming API

FAQ from MCP Lambda SAM

What is MCP?

Model Context Protocol (MCP) is a protocol for managing and invoking models and tools in a context-aware manner.

What AWS services are used?

AWS Lambda, DynamoDB, IAM, SQS, and API Gateway (via Lambda Function URLs).

How do I deploy the application?

You can deploy using AWS Serverless Application Repository (SAR), npx, programmatic usage with install, or local development and deployment.

How do I grant permissions to clients?

You can grant access using either an IAM policy or aws lambda add-permission to invoke the mcp function URL.

What are common client errors?

Common client errors include 401 (Invalid/missing AWS credentials), 403 (Insufficient permissions), 404 (Invalid session ID), and 429 (Rate limit exceeded).