Smart Troubleshooter Docs

An intelligent tool for automatically generating structured troubleshooting documentation from free-text input — powered by Angular, Spring Boot, and Ollama with the Mistral LLM.

Project Overview

Smart Troubleshooter Docs is a modern AI-assisted system that transforms unstructured problem descriptions into structured troubleshooting documentation. It is designed to assist support teams by turning free-form input (e.g., ticket texts, error logs) into useful, standardized output.

Tech Stack:

• Frontend: Angular
• Backend: Spring Boot (Java)
• LLM Integration: Ollama with the Mistral model
• Containerization: Docker & Docker Compose

Project Structure

smart-troubleshooter-docs/
├── backend/       # Spring Boot Backend
├── frontend/      # Angular Frontend

Getting Started

1. Prerequisites

• Node.js (recommended: ≥ 18.x)
• npm
• Java JDK (recommended: 17+)
• Maven
• Ollama (installed and running)
• Mistral LLM (must be pulled once):

ollama pull mistral

2. Starting the Project

Option 1: Using Docker (Recommended)

# Start all services with Docker Compose
docker-compose up -d

# Pull Mistral model (if not already done)
docker exec ollama ollama pull mistral

The app should now be running at:
http://localhost:4200 (Frontend)
http://localhost:8080 (Backend API)

Option 2: Manual Setup

Backend (Spring Boot)

cd backend
./mvnw spring-boot:run

Frontend (Angular)

cd frontend
npm install
npm start

The app should now be running at:
http://localhost:4200

How It Works

1. User submits free-text description of a problem.
2. Backend sends the input to Mistral via local Ollama instance.
3. Structured output is returned and rendered in the frontend.

Development

Backend

• Java 17+
• Spring Boot REST API
• Integration with local Ollama

Frontend

• Angular 16+
• HTTP communication with backend
• UI for input and output visualization

Testing

• Backend: Run mvn test
• Frontend: Run ng test

Roadmap

• Multilingual support (EN/DE)
• Save/export troubleshooting documents

License

MIT License — free to use with attribution.

Contributing

Pull requests, suggestions, and bug reports are welcome!