TheftWatch - Advanced VMS with Augmented AI

Java 21 + Spring Boot 3 + Vaadin 24 video management system with AI-powered theft detection.

Prerequisites

• Java 21
• Maven 3.9+
• Docker & Docker Compose
• PostgreSQL 17
• FFmpeg
• Sige 7 (Armsom) for AI inference (optional for initial setup)

Quick Start

Default Credentials

A default Super Admin is created on first startup:

• Email: admin@theftwatch.local
• Password: admin123

Development (with Docker Compose)

# Start infrastructure
docker compose up -d postgres minio

# Run application
mvn spring-boot:run

Production Build

mvn clean package -DskipTests
docker compose up --profile production

Architecture

• Frontend: Vaadin 24 (Java-based UI)
• Backend: Spring Boot 3.3
• Database: PostgreSQL 17 with Flyway migrations
• Storage: MinIO (S3-compatible) for video/recordings
• AI: Sige 7 integration via REST API
• Streaming: FFmpeg RTSP → HLS transcoding

User Hierarchy

1. Super Admin: Full system control, creates Special Admins and End Users
2. Special Admin: Manages End User realms/dominions, partial camera setup
3. End User: Business-level access to assigned realms/dominions

Domain Model

• Realm: Top-level organizational unit (e.g., "Company A")
• Dominion: Location/branch within a realm (e.g., "Store #1", "Warehouse")
• Camera: RTSP-enabled camera assigned to realm/dominion
• Alert: AI-generated theft detection events

Key Features

• Multi-camera RTSP streaming
• HLS video delivery for browser playback
• AI-powered theft detection via Sige 7
• Role-based access control (RBAC)
• Realm/Dominion hierarchy for multi-tenant deployments
• Alert management with severity levels
• Docker containerization

Project Structure

src/main/java/com/theftwatch/theftwatch/
├── config/           - App configuration
├── domain/           - JPA entities
├── enums/            - Role, AlertSeverity, CameraStatus
├── repository/       - Spring Data JPA repos
├── service/          - Business logic
├── security/         - Auth & RBAC
├── ai/               - Sige 7 inference
├── streaming/        - FFmpeg RTSP/HLS
└── ui/               - Vaadin views

Development

# Run tests
mvn test

# Package
mvn clean package

# Run
java -jar target/theftwatch-0.1.0.jar

License

Proprietary