What is Mcp Server

The Model Context Protocol (MCP) is a framework designed for integrating AI capabilities into applications, facilitating seamless communication between software components and AI agents. It enables developers to create applications that can leverage AI functionalities efficiently.

Use cases

MCP is suitable for various applications, including chatbots, virtual assistants, automated decision-making systems, and other AI-driven applications requiring context management and communication. It can be used in customer service, healthcare, finance, and more to enhance user experiences and improve operational efficiency.

How to use

To use MCP with Spring Boot, developers need to build their application into a JAR file using Gradle, set up the Goose AI Agent according to its documentation, and create a shell script to run the application. This script can then be executed in conjunction with Goose sessions to allow the integration of AI functionalities.

Key features

Key features of MCP include contextual awareness, enabling applications to maintain and manage the context throughout interactions. It supports dynamic communication with AI agents and provides a structured way to define interactions, which significantly enhances the usability of AI in applications.

Where to use

MCP can be deployed in any environment where Spring Boot applications are utilized, including cloud platforms, enterprise systems, and local development setups. It is especially beneficial in scenarios that require adaptive AI features and real-time context management.

Model Context Protocol (MCP) + Spring Boot Integration

🔗 Useful Links

• MCP Introduction: modelcontextprotocol.io/introduction
• Spring AI MCP Documentation: docs.spring.io - MCP Client Boot Starter
• Goose (AI Agent) - Client Setup: block.github.io/goose

How to Run Your Spring Boot Application with Goose

Step 1: Create JAR File

Build your Spring Boot application using:

./gradlew clean build

The JAR will be located in target/.

Step 2: Install and Set Up Goose

Follow the official Goose documentation to install and configure the AI Agent.

Running Goose with WSL (Windows Subsystem for Linux)

If you’re using WSL and face issues running Goose sessions, follow these steps:

1. Create a Shell Script

Inside your WSL terminal, create a file named run-extension.sh:

nano run-extension.sh

Paste the following content:

#!/bin/bash
java -jar "FULL_PATH_OF_JAR_FILE/target/mcp_feature-0.0.1-SNAPSHOT.jar"

Replace FULL_PATH_OF_JAR_FILE with the actual path to your JAR file.

2. Ensure Java is Installed in WSL

If you encounter issues related to Java, ensure JDK is installed:

java -version

If not installed, you can install OpenJDK via:

sudo apt update
sudo apt install openjdk-21-jdk

3. Make the Script Executable

chmod +x run-extension.sh

4. Start Goose Session with Your Extension

goose session --with-extension="./run-extension.sh"

You’re now ready to run your Spring Boot application integrated with MCP and the Goose AI Agent! 🧠🚀

Simple flow diagram :