A demonstration application consisting of an Angular chat client and a Spring Boot microservice that interfaces with Ollama's cloud-based LLM.
| Component | Description |
|---|---|
| llm_chat_client | Angular/Node.js chat client |
| llm_microservice | Spring Boot microservice that communicates with the Ollama cloud model "gpt-oss:120b" and streams responses to the client via SSE |
- Client → Microservice: User submits a question via the Angular frontend
- Microservice → Cloud Model: The microservice forwards the question to the Ollama API
- Cloud Model → Microservice → Client: The response is streamed back using Server-Sent Events (SSE)
- Node.js (v18+)
- Java 17+ (for Spring Boot)
- An Ollama account (free) to obtain an API key
git clone https://github.com/yourusername/your-repo.git
cd your-repo
export OLLAMA_API_KEY=your_api_key_here
cd llm_microservice
./mvnw spring-boot:run
cd llm_chat_client
npm install
ng serve
Navigate to http://localhost:4200 in your browser.
This demo is not production-ready! Please address the following before deploying to production:
| Communication Path | Status |
|---|---|
| Microservice ↔ Ollama Cloud | ✅ Secured (JWT + TLS) |
| Frontend ↔ Microservice |
Note on TLS: This application is designed to be deployed on cloud platforms (e.g., Render.com, Heroku, Railway) that provide TLS termination automatically. No additional TLS configuration is required when using such platforms.
- Write tests
- Document the code
- Allow users to select from available cloud models
- Add conversation context (MCP functionality) — currently, the model has no memory of previous questions
- Improve UI/UX
- Implement conversation persistence (with appropriate security considerations)
- Add caching, preprocessing, and postprocessing
- Complete this documentation
MIT License
Whoever wants to contribute can contribute