MagicCuts 🎬✂️

MagicCuts is a powerful application that transforms long videos into viral short-form content. The platform uses AI to analyze video transcriptions, identify the most engaging moments, and automatically cut them into vertical short format videos optimized for social media platforms.

🚀 Project Overview

This project was created during a 12-hour YouTube challenge to build a tool that could help content creators repurpose their long-form content (like YouTube videos and livestreams) into viral shorts without manual editing.

MagicCuts:

1. Analyzes video transcriptions to find the most engaging segments
2. Automatically cuts these segments into vertical short-format videos
3. Provides an intuitive dashboard to review and manage your viral clips

🏗️ Project Structure

The project consists of two main components:

Frontend (Remix.js + React)

• Modern UI built with Remix, React, and TailwindCSS
• User authentication and project management
• Video preview and management dashboard

Backend (Node.js + Hono)

• RESTful API built with Hono
• Video processing and transcription with Deepgram
• AI analysis with OpenAI
• Video transformation with FFMPEG
• Storage with AWS S3
• Database management with Supabase

🛠️ Technologies Used

Frontend

• Remix.js: React framework for building modern web applications
• React: UI library
• TailwindCSS: Utility-first CSS framework
• HeroUI: UI component library
• TypeScript: Type-safe JavaScript
• Zustand: State management
• Remotion: JavaScript library for creating videos programmatically
• Framer Motion: Animation library

Backend

• Node.js: JavaScript runtime
• Hono: Lightweight, fast web framework
• TypeScript: Type-safe JavaScript
• Deepgram: Audio transcription service
• OpenAI: AI for content analysis
• FFMPEG: Video processing tool
• AWS S3: Cloud storage
• Supabase: Backend-as-a-Service with PostgreSQL database
• Docker: Containerization

🚀 Deployment

• Frontend: Deployed on Vercel
• Backend: Deployed on Railway

🏁 Getting Started

Prerequisites

• Node.js (v20+)
• PNPM package manager
• FFMPEG installed on your system
• AWS, Supabase, Deepgram, and OpenAI accounts

Environment Setup

Frontend

1. Navigate to the frontend directory and copy the example environment file:

cd frontend
cp .env.example .env

2. Update the .env file with your API keys and endpoints.

Backend

1. Navigate to the server directory and copy the example environment file:

cd server
cp .env.exemple .env

2. Update the .env file with your API keys and service credentials.

Installation

Frontend

cd frontend
pnpm install
pnpm run dev

Backend

cd server
pnpm install
pnpm run dev

Alternatively, you can use Docker for the backend:

cd server
docker compose up -d --build
docker compose logs -f

🔍 How It Works

1. Upload: Users upload their long-form video content
2. Transcription: The system transcribes the video using Deepgram
3. Analysis: AI analyzes the transcription to identify viral-worthy moments
4. Processing: The system cuts the video into short segments
5. Delivery: Users can preview and download the generated short videos

💡 Features

• AI-powered identification of viral-worthy moments
• Automatic video cutting and formatting
• User-friendly dashboard
• Video preview
• Transcription review
• Cloud storage integration
• Customizable output format

🤝 Contributing

Contributions are welcome! This project was built during a 12-hour challenge but is open to improvements and new features.

📄 License

This project is licensed under the MIT License.