mcp-demo-refrigerator-inventory-mgmt

🧊 MCP Demo: Refrigerator Inventory Management

A demo project showcasing integration between an AI chatbot and a Spring Boot application using the Model Context Protocol (MCP) framework. It provides inventory management functionality for refrigerators by answering queries through an AI model, backed by a PostgreSQL database.

---

🛠️ Tech Stack

• Java / Spring Boot
• MCP Server / Client
• PostgreSQL / NeonDB
• Ollama with llama3.2 (local LLM)
• Swagger UI (for testing endpoints)

---

🚀 Project Setup

1. Set Up PostgreSQL

You can either:

• Run PostgreSQL locally
• Use Docker to spin up a PostgreSQL container
• Connect via a hosted DB service like NeonDB

Make sure your database is properly configured and running.

---

2. Set Up Local LLM with Ollama

Install and run Ollama for local LLM support:

> ollama list
> ollama pull llama3.2:latest
> ollama run llama3.2

💡 llama3.2 is around 2GB and runs locally for free—no API key required!

---

3. Run the Applications

Make sure you run all three components:

• refrigerator-inventory-core
• mcp-server
• mcp-client

Each should be correctly configured to communicate with the database and each other.

---

✅ Verifying MCP with Local Data

🔍 Step 1: Test AI Model is Running

Go to Swagger UI:

http://localhost:8082/swagger-ui/index.html#/ai-chat-controller

• Click Try it out on the chat endpoint

• Enter:

  {
    "message": "What is the capital of India?"
  }
  

• You should see:

  New Delhi
  

✔️ Confirms the local llama3.2 is up and integrated.

---

🔍 Step 2: Verify Response from Local Database

Navigate to:

http://localhost:8082/swagger-ui/index.html#/mcp-chat-controller/chat

• Click Try it out and enter:

  {
    "message": "Product below 100 rupees"
  }
  

• Sample response:

  Here are some products available below 100 rupees:
  
  1. Eggs - priced at 10.0 rupees
  2. Cucumber - priced at 40.0 rupees
  3. Spinach - priced at 19.99 rupees
  4. Gulaab Jamun - priced at 99.0 rupees
  

✔️ This confirms that the MCP server retrieves and responds using your own database, not internet data.

---

📸 Screenshot

---

🧾 Conclusion

The MCP Client connects to the MCP Server, which interacts with the refrigerator-inventory-core app and provides AI-generated answers based on your local database.

🎯 A practical demonstration of a local AI system working in real-time with your Spring Boot and PostgreSQL setup!