BrandRisk

AI-powered video content analysis for brand safety compliance. Built with Next.js and powered by Memories.ai.

🎯 Features

• Instant Video Analysis: Upload videos and get brand safety analysis in seconds
• Interactive Timeline: Visual timeline with clickable risk markers
• GARM Categories: Detects profanity, violence, sexual content, drugs/alcohol, hate speech, sensitive issues, and sponsorship
• Multiple Export Formats: Export reports as JSON, CSV, or SRT subtitle files
• Real-time Processing: Live progress indicators and instant results

🚀 Quick Setup

1. Install Dependencies

npm install

2. Configure Memories.ai API

1. Get your API key from Memories.ai
2. Create account → API section → Create new API key
3. Copy .env.example to .env.local:

cp .env.example .env.local

4. Add your API key to .env.local:

MEMORIES_API_KEY=your_memories_ai_api_key_here

3. Run Development Server

npm run dev

Open http://localhost:3000 to see the application.

📝 Usage

1. Upload Video: Drag and drop a video file (MP4, AVI, MOV, etc.) up to 100MB
2. Wait for Analysis: The system will upload and analyze your video using AI
3. Review Timeline: Click on risk markers to jump to specific moments
4. Export Results: Download analysis reports in JSON, CSV, or SRT format

🛠️ Tech Stack

• Frontend: Next.js 15, React 19, TypeScript, Tailwind CSS
• Backend: Next.js API Routes
• AI Analysis: Memories.ai (video indexing, search, transcription)
• File Handling: react-dropzone
• Deployment: Vercel-ready

📊 Risk Categories (GARM Standards)

• 🤬 Profanity: Explicit language and swearing
• 💋 Sexual Content: Adult material and suggestive content
• 🍺 Drugs & Alcohol: Substance use and consumption
• 🔫 Violence: Weapons, fighting, and aggressive behavior
• ⚠️ Hate Speech: Discriminatory language
• 🗞️ Sensitive Issues: Political content and controversial topics
• 📣 Sponsorship: Undisclosed advertising content

🏗️ Project Structure

src/
├── app/
│   ├── api/memories/          # API routes for Memories.ai integration
│   │   ├── upload/           # Video upload endpoint
│   │   ├── status/           # Processing status check
│   │   ├── analyze/          # Brand safety analysis
│   │   └── export/           # Report export (JSON/CSV/SRT)
│   ├── globals.css           # Global styles
│   ├── layout.tsx            # Root layout
│   └── page.tsx              # Main application page
├── components/
│   ├── VideoPlayer.tsx       # Video player with timeline
│   └── UploadZone.tsx        # File upload component
└── lib/
    ├── memoriesClient.ts     # Memories.ai API wrapper
    └── riskCategories.ts     # GARM category definitions

🚀 Deploy on Vercel

1. Connect your repository to Vercel
2. Add environment variable: MEMORIES_API_KEY
3. Deploy automatically

🎪 Hackathon Demo Script

1. Problem: "Manual brand safety review takes 3+ hours per video"
2. Upload: Drop in sample ad video → processing spinner
3. Magic Moment: Timeline populates with risk markers instantly
4. Interaction: Click marker → jump to exact timestamp → show evidence
5. Value: "Export report → hand to editor → save 3 hours"

📄 License

Built for AI Engine Hackathon. Open source MIT license.