WatchForge

WatchForge is a self-hosted, AI-powered video processing system built in Go. It detects people, animals, and insects in video feeds using YOLOv4-tiny and GoCV, and provides a cross-platform user interface via Fyne for mobile and desktop access.

🚀 Project Overview

WatchForge enables intelligent surveillance by processing video files or streams, detecting motion and objects of interest, and serving annotated results through a Go-based server. It is designed to run on resource-constrained devices like mini PCs and supports modular development with future extensibility in mind.

✨ Features

• 🎥 Video processing using GoCV (OpenCV bindings for Go)
• 🧠 AI detection with YOLOv4-tiny (ONNX format)
• 📦 Self-hosted Go server with REST API
• 🔒 User management with PostgreSQL + GORM
• 📱 Native UI for Android, Linux, Windows (via Fyne)
• 📤 Serve processed videos to mobile/desktop clients
• 🔄 Modular architecture for future model upgrades

🛠️ Tech Stack

• Language: Go
• UI: Fyne (mobile + desktop)
• AI: YOLOv4-tiny (ONNX) + OpenCV DNN via GoCV
• Database: PostgreSQL
• ORM: GORM
• Video I/O: GoCV (FFmpeg-enabled OpenCV)
• Server: Go HTTP server (net/http or Gin)

👨‍💻 Getting Started

Prerequisites

• Go 1.20+
• OpenCV with FFmpeg support
• GoCV (gocv.io/x/gocv)
• PostgreSQL (local or Docker)
• Git

Clone the Repository

git clone https://github.com/yourusername/watchforge.git
cd watchforge