Unlock New Possibilities: Java, Spring AI, and Cloud Intelligence

What is MCP (Model Context Protocol)?

•       A standard protocol for communication between AI models and external tools.

•       Provides a common language for AI apps and services.

•       Think of MCP as a "universal translator" that helps AI models (like ChatGPT) intelligently interact with various external tools and data sources.

•       Or Think of MCP like a USB-C port for AI applications. Just as USB-C provides a standardized way to connect electronic devices, MCP provides a standardized way to connect AI applications to external systems.

•       MCP servers provides methods which we call as Tools.

MCP Client and Server Architecture:

As we know API works on Client Server Architecture; Similarly, MCP also need client and server instances depending on use cases.

MCP Host: The AI application that coordinates and manages one or multiple MCP clients

MCP Client: A component that maintains a connection to an MCP server and obtains      context from an MCP server for the MCP host to use

MCP Server:

·       MCP servers are programs that expose specific capabilities to AI applications through standardized protocol interfaces.

·       There are some ready-to-use MCP servers available which we can directly plugin to our AI applications / MCP Hosts with some minimal configuration

e.g FileSystem, Git, GitHub,etc…

·       But for customized use cases we can build MCP server in some languages such as Python, Java, Node, C#

MCP Transports:

·       Stdio --- communication over standard in and standard out. The server reads JSON-RPC messages from its standard input (stdin) and sends messages to its standard output (stdout). Mainly used in AI Agents such Claude Desktop

·       Streamable HTTP --- The server operates as an independent process that can handle multiple client connections.

Note: This replaces Old HTTP + SSE tranport

Java MCP Client – Server Applications:

·       Build MCP Client server applications with the help of below resources:

o   https://docs.spring.io/spring-ai/reference/api/mcp/mcp-client-boot-starter-docs.html

o   https://www.baeldung.com/spring-ai-model-context-protocol-mcp

·       To Test MCP Client-Server Architecture build applications from above resources and test in Postman or any test tool in the similar way as API Testing.

·       Build Custom MCP server Tool as below in the service layer and registered in the main class:

Service class implementation:

Main Class Tool Registration:

·       Testing with Postman:

o   Click on New Project and select MCP

o   Select HTTP Mode from the drop down and provide the URL as shown. By Default MCP server connects to SSE transport

o   Click Connect and then it will show the listed tools from the server

o   Select any tool and provide the required parameter and click Run; it will provide the result in JSON format.