Skip to content

SafeStack

SafeStack analyzes warehouse video feeds with Gemini to detect violations of OSHA's safety policies, allowing operators to find when there are issues with their warehouse operations.

Core trick

Sends warehouse video chunks to Gemini with OSHA policy text, returning cited violations that trigger alert emails.

Tags

Description

SafeStack analyzes warehouse video feeds with Gemini to detect violations of OSHA's safety policies, allowing operators to find when there are issues with their warehouse operations. We provide detailed analysis, send emails to alert operators, and use nano banana to generate corrected images of what should have happened. Policies used: https://www.osha.gov/sites/default/files/publications/OSHA3885.pdf
Pkama07/safe-stack★ 01 forksTypeScriptCreated Dec 6, 2025 · last push Dec 7, 2025

README excerpt

# SafeStack 🛡️

**AI-Powered Workplace Safety Monitoring System**

SafeStack is a real-time video analysis platform that uses Google Gemini AI to detect OSHA safety violations in workplace environments. It monitors camera feeds, identifies policy violations, and alerts safety managers with evidence and AI-generated remediation suggestions.

![SafeStack Dashboard](https://img.shields.io/badge/Status-Active-brightgreen) ![Next.js](https://img.shields.io/badge/Next.js-14-black) ![Python](https://img.shields.io/badge/Python-3.10+-blue) ![Gemini](https://img.shields.io/badge/AI-Gemini_3_Pro-4285F4)

---

## ✨ Features

### 🎥 Real-Time Video Monitoring
- Multi-camera grid view with live video feeds
- Automatic video chunk capture and analysis (configurable intervals)
- Camera status indicators (idle, analyzing, alert)

### 🤖 AI-Powered Violation Detection
- **Gemini 3 Pro** analyzes video footage for safety violations
- Automatic frame extraction at violation timestamps
- Detects 30+ OSHA safety policies across 3 severity levels
- Deduplication to report only the most significant instance of each violation type

### 🚨 Smart Alert System
- Real-time alert stream with severity-based categorization
- Evidence images captured at the moment of violation
- **AI-generated "fixed" images** showing how the scene should look after remediation
- Detailed violation explanations with reasoning

### 📧 Notifications
- Email alerts via Resend for critical violations
- Configurable alert recipients

Team