MCP & Agentic AI

Model Context Protocol servers, multi-agent systems, and AI-powered tools

Repositories exploring the Model Context Protocol (MCP) ecosystem — from running MCP servers on AWS infrastructure to building multi-agent systems with Strands and Bedrock. Includes document extraction, generative BI, and practical MCP patterns.

Repositories

strands-agents/sdk-python

PR #176
Mar 2026
Python

Added support for Amazon SageMaker AI endpoints as a Model Provider in the Strands Agents SDK, enabling developers to use SageMaker-hosted models with the Strands Agents framework.

Strands SDKSageMaker AIModel ProviderOpen Source

aws-document-extraction-hybrid

Feb 2026
Python

Hybrid document extraction architecture comparing Claude, BDA, and Textract for intelligent document processing.

Document ExtractionClaudeTextractBDA
View on GitHub

bedrock-strands-multi-agent-generative-bi-demo

Feb 2026
Python

AI-powered multi-agent data analytics platform using AWS Bedrock and Strands Agents SDK. Demonstrates generative BI with coordinated agent workflows.

Multi-AgentStrands SDKBedrockGenerative BI
View on GitHub

sagemaker-ai-mcp-server

Oct 2025
Python

MCP Server that uses SageMaker AI APIs to monitor and manage resources.

MCP ServerSageMaker AIResource Management
View on GitHub

mcp-server-with-fargate

Apr 2025
Python

Running an MCP Server on AWS Fargate on Amazon ECS. Production-ready pattern for hosting MCP servers on managed container infrastructure.

MCP ServerFargateECSDocker
View on GitHub

mcp-sse-client-server-docker

Apr 2025
Python

A simple implementation of MCP Client/Server architecture with HTTP SSE transport layer, dockerized.

MCPSSEDockerClient/Server
View on GitHub

llama-index-aws

Apr 2024
Jupyter Notebook

Learning Llama Index usage on AWS — both with Amazon SageMaker and Amazon Bedrock.

LlamaIndexRAGSageMakerBedrock
View on GitHub