Spring AI MCP

This repository is a playground for learning and experimenting new ideas in MCP using Spring AI

Github Actions Build status

Required Software and Tools

• Java Version: 21 (Execute java -version in command line after installation)
• Ollama: latest (Execute ollama -v in command line after installation)
• Integrated Development Environment: Any version IntelliJ Idea or Eclipse

Pull Ollama Models

ollama pull llama3.2:3b

Build command

./mvnw clean install

Skip all tests and generate artefacts

./mvnw clean install -DskipTests

Run MCP Servers and Client applications

./mvnw clean spring-boot:run -pl spring-ai-mcp-server1

./mvnw clean spring-boot:run -pl spring-ai-mcp-server2

./mvnw clean spring-boot:run -pl spring-ai-mcp-client

or to run the client against OpenAI

export OPEN_AI_KEY={open_ai_key}

./mvnw clean -Dspring-boot.run.profiles=openai spring-boot:run -pl spring-ai-mcp-client

or to run the client against AWS LLM Model

export AWS_AI_ACCESS_KEY_ID={aws_key_id}

export AWS_AI_SECRET_ACCESS_KEY={aws_secret_key}

./mvnw clean -Dspring-boot.run.profiles=aws spring-boot:run -pl spring-ai-mcp-client

Run MCP Servers and Client applications using docker compose

export SPRING_PROFILES_ACTIVE="aws"

or

export SPRING_PROFILES_ACTIVE="openai"

./docker-compose-run.sh

Run E2E test with MCP Host and MCP Servers

./mvnw clean install -DskipE2E=false -pl spring-ai-mcp-e2e