AWS Bedrock MCP Client and Server

What is AWS Bedrock MCP Client and Server?

A Spring Boot application that serves as a client for the Model Context Protocol (MCP), enabling interaction with Amazon Bedrock AI models with tool-using capabilities. It provides a REST API for calculations, weather information, and web page retrieval, and runs as a serverless AWS Lambda function behind API Gateway.

How to use AWS Bedrock MCP Client and Server?

Clone the repository, configure AWS credentials with privileges for Lambda, API Gateway, Bedrock, and the Anthropic Sonnet 4.0 model. Build with mvn clean package install -DWEATHER_API_KEY=<your_key> and deploy using AWS CDK (cdk bootstrap, cdk deploy --context WEATHER_API_KEY=<your_key>). After deployment, use the provided API Gateway URL to call endpoints like /hello, /tools, /calculate, /weather, /webpage, and /query via curl.

Key features of AWS Bedrock MCP Client and Server

• REST API for interacting with Amazon Bedrock models
• MCP tools for calculations, weather, and web page retrieval
• Handles complex AI queries that may use multiple tools
• Serverless and scalable via AWS Lambda and API Gateway
• Deployed and managed with AWS CDK
• Uses Java 21, Spring Boot, and Maven multi-module structure

Use cases of AWS Bedrock MCP Client and Server

• Ask a question like “What is the weather in London and calculate 45 divided by 9?”
• Summarise a web page by providing its URL
• Perform individual tool operations: add numbers, get weather, convert a web page to markdown
• Test the API with a simple hello-world endpoint