Sample Model Context Protocol Demos

This project provides a series of example modules using the AWS Model Context Protocol (MCP), covering multiple languages and technology stacks, including TypeScript, Python, Spring AI, etc., demonstrating the application of MCP in scenarios such as client-server communication, ECS deployment, and RAG integration.

Artificial intelligence chatbots Developer tools #MCP protocol #AWS integration #Multi-language examples #ECS deployment .TypeScript

rating : 2.5 points

downloads : 5.2K

update time : 2025-04-28

Open Site

What is the MCP server?

The MCP server is middleware that follows the Model Context Protocol standard. It serves as a bridge connecting AI clients (such as Amazon Bedrock) and business logic servers, supporting two-way communication and context management.

How to use the MCP server?

Through the standardized protocol interface, developers can quickly integrate AI capabilities into existing systems, supporting containerized deployment and local development modes.

Applicable scenarios

It is suitable for scenarios where large language models need to be integrated into business systems, such as AI-enhanced applications like intelligent customer service, document processing, and appointment management.

Main features

Multi-protocol support

Supports both Server-Sent Events (SSE) and standard input/output (stdio) communication protocols simultaneously.

Cloud platform integration

Provides an AWS ECS deployment template for seamless docking with Amazon Bedrock.

Multi-language implementation

Provides examples in multiple languages such as TypeScript, Python, Java/Kotlin, etc.

RAG enhancement

Some examples include the Retrieval Augmented Generation (RAG) function implemented by pgVector.

Advantages

The standardized protocol ensures compatibility between different components.

Flexible deployment options (local/container/cloud)

Abundant example code accelerates the development process.

Supports integration with mainstream AI service platforms

Limitations

Infrastructure knowledge is required for production deployment.

Some advanced features depend on specific cloud services.

Beginners may need time to understand the protocol details.

How to use

Select the implementation language

Choose the TypeScript, Python, or Java/Kotlin version according to your technology stack.

Configure the environment

Install the corresponding language environment and configure AWS credentials (if you need to use Bedrock).

Run the example

Start the client and server components according to the module instructions.

Usage cases

Intelligent customer service system

A customer service agent implemented using Spring AI to handle user inquiries and manage conversation states.

Pet adoption assistant

A dog adoption advisor based on RAG to retrieve matching pet information from the knowledge base.

Frequently Asked Questions

How is the MCP protocol different from ordinary API calls?

Is it necessary to use AWS services?

How to extend custom functions?

Related resources

MCP protocol specification document

The complete technical specification of the protocol

GitHub repository

All example code and deployment templates

AWS Bedrock documentation

The official guide of the Amazon Bedrock service

Module Name	Language	Description
Server-Client MCP/SSE Demo	TypeScript	This full-fledged demo creates an Amazon Bedrock MCP client using the converse API and an MCP server. The sample is deployed via containers and communicates with each other using MCP/SSE.
Server-Client MCP/stdio Demo (Local)	Python	This is a simple demo that shows an Amazon Bedrock MCP client using the converse API and MCP standard input/output (stdin/stdout). The sample can be run locally and connected to Amazon Bedrock.
Server-Client MCP/SSE on ECS	Spring AI & Kotlin	Provides an example of a Spring AI MCP server running on ECS; the server is used by a Spring AI Agent also running on ECS, which is publicly accessible via a load balancer.
Server-Client MCP/SSE with pgVector RAG in Bedrock Converse Client	Spring AI & Java	A Bedrock-based Spring AI dog adoption agent that uses PostgreSQL and pgvector for Retrieval Augmented Generation (RAG), along with an MCP server to manage adoption appointments.
MCP/SSE Server on ECS	Spring AI & Kotlin	A very basic Spring AI MCP server running on ECS using Server-Sent Events (SSE).