ClipJev 🎬

ClipJev is an AI-powered web application that automatically analyzes YouTube videos, finds the most engaging moments, and prepares them as 9:16 vertical clips perfect for YouTube Shorts, TikTok, or Instagram Reels.

✨ Features

• Automated YouTube Transcripts: Uses Supadata API to reliably fetch YouTube transcripts, bypassing common datacenter blocking issues.
• Intelligent Clip Selection: Leverages the JEV model (~typesafe/jev-latest) to semantically read transcripts in chunks, score moments for virality/engagement, and extract perfect start/end timestamps.
• Live Preview Simulator: Simulates a 9:16 vertical crop directly in the browser using the YouTube iframe API, letting you preview the exact crop without waiting for server processing.
• Background Export (Local): When running locally or on a VPS (like Railway), ClipJev uses yt-dlp and ffmpeg to download the high-res video in the background and precisely crop it to 9:16.

🚀 Getting Started

Prerequisites

• Node.js 18+
• Supadata API Key (Free tier gives 100 requests/month)
• (Optional) ffmpeg and yt-dlp installed on your system if you want to enable the local download/export feature.

Installation

1. Clone the repository and install dependencies:

   git clone <your-repo-url>
   cd clipjev
   npm install

2. Set up your environment variables. Create a .env.local file in the root directory:

   # Required: Supadata API Key for reliable YouTube transcript fetching
   SUPADATA_API_KEY=your_supadata_api_key
   
   # Required: The URL to your JEV model instance
   JEV_URL=your_jev_api_url
   
   # Optional: Set to "true" to disable server-side ffmpeg/yt-dlp downloads 
   # (Required when deploying to serverless environments like Vercel)
   NEXT_PUBLIC_DISABLE_DOWNLOADS=true

3. Start the development server:

   npm run dev

4. Open http://localhost:3000 in your browser.

🏗️ Architecture

• Framework: Next.js 15 (App Router)
• Styling: Tailwind CSS + Radix UI Primitives
• Icons: Lucide React
• Video Player: Native YouTube Iframe API
• AI Processing: JEV Decision Model (via standard HTTP requests)

☁️ Deployment (Vercel)

ClipJev's frontend and AI processing can be easily deployed to Vercel. However, because Vercel's serverless functions have strict size limits and execution timeouts, the background video downloading (yt-dlp + ffmpeg) is not supported on Vercel.

To deploy to Vercel:

1. Push your code to GitHub.
2. Import the project in the Vercel dashboard.
3. Add your SUPADATA_API_KEY and JEV_URL environment variables.
4. Crucial: Add NEXT_PUBLIC_DISABLE_DOWNLOADS=true to your Vercel environment variables. This hides the export UI and prevents the serverless functions from trying to spawn ffmpeg.

📜 License

MIT License