🤖 Spring Boot 3.4.2 + Ollama + DeepSeek-R1 Integration (Java 21)

This project demonstrates how to run a local Large Language Model (LLM) using Ollama and integrate it with a Spring Boot 3.4.2 application using DeepSeek-R1. No need for cloud APIs — run everything locally for maximum performance and data privacy.

🎯 Challenge Focus: DeepSeek-R1 Integration

The primary objective was to demonstrate a seamless integration of the DeepSeek-R1 model using Ollama, highlighting:

• Local AI inference
• Minimal resource consumption
• Easy-to-implement AI chat functionality
• Cross-platform compatibility

🧩 Features

• 🔗 Connect Spring Boot with Ollama's local LLM API
• 💬 AI-powered chat endpoint using DeepSeek-R1 model
• 🛡️ Keeps your prompts and data local
• ⚡ Low latency inference via REST
• ☕ Powered by Java 21 and Spring Boot 3.4.2
• 🅰️ Integrated Angular Frontend

✨ Preview

Here's what the chat UI looks like when the Angular frontend is running:

📦 Prerequisites

• Java 21
• Maven 3.8+
• Ollama installed locally (WSL/Ubuntu/macOS/Linux)
• DeepSeek-R1 model pulled via Ollama

🚀 Getting Started

1. Install Ollama

curl -fsSL https://ollama.com/install.sh | sh
ollama serve

2. Pull DeepSeek-R1 Model

ollama pull deepseek-r1:1.5b

3. Build and Run the Application

# Clone the repository
git clone https://github.com/HenryXiloj/demo-ollama-deepseek-r1
cd demo-ollama-deepseek-r1

# Clean, build, and run the application
mvn clean install 
spring-boot:run

Note: The Maven build will:

• Install Node.js and npm
• Install frontend dependencies
• Build the Angular frontend
• Compile and run the Spring Boot backend

📁 Project Structure

📁 demo-ollama-deepseek-r1/
├── angular-ui/
│   └── src/app/
│       ├── app.component.ts
│       ├── app.component.html
│       ├── app.config.ts
│       └── chat/
│           ├── chat.component.ts
│           ├── chat.component.html
│           └── chat.component.scss
│
├── src/main/java/com/henry/ollama/
│   ├── Application.java
│   ├── config/OllamaProperties.java
│   ├── controller/ChatController.java
│   ├── record/OllamaRequest.java
│   ├── record/OllamaResponse.java
│   └── service/OllamaService.java

🧪 Testing the API

After starting the application, you can test the chat endpoint:

curl -X POST -H "Content-Type: text/plain" \
  -d "Explain AI in simple terms" \
  http://localhost:8080/api/chat

🪟 Bonus: WSL on Windows Setup

WSL Side

sudo ufw enable
sudo ufw allow 11434
sudo systemctl stop ollama
sudo lsof -i :11434
export OLLAMA_HOST=0.0.0.0
ollama serve

Windows (PowerShell)

$wsl_ip = (wsl hostname -I).Split()[0]
netsh interface portproxy add v4tov4 listenport=11434 listenaddress=0.0.0.0 connectport=11434 connectaddress=$wsl_ip
New-NetFirewallRule -DisplayName "Ollama-WSL" -Direction Inbound -Protocol TCP -LocalPort 11434 -Action Allow

📚 Resources

• Ollama Documentation
• DeepSeek on Hugging Face
• Spring Boot 3.4.2 Docs
• Java 21 Features
• LinkedIn Article: Integrating Ollama DeepSeek-R1 with Spring Boot

🙌 Author
Created with ❤️ by Henry Xiloj
🔗 GitHub Repo: github.com/HenryXiloj/demo-ollama-deepseek-r1
📝 Blog: jarmx.blogspot.com