Media Downloader

Single-page web application to download video and audio from YouTube, Facebook, Instagram, TikTok, X, Threads, and Reddit.

Architecture

• Frontend: Vanilla HTML5, CSS3, and JavaScript with PWA install and share-target support.
• Backend: Express.js server (server.js) executing yt-dlp in a local Python virtual environment.
• Dependencies: Node.js (version 20+), Python (version 3.10+), and ffmpeg.

Prerequisites

Ensure ffmpeg is installed on the host system.

Ubuntu/Debian

sudo apt update && sudo apt install -y ffmpeg

macOS

brew install ffmpeg

Local Installation

1. Install Node.js dependencies:

    npm ci

2. Create a local Python virtual environment and install/upgrade yt-dlp:

   python3 -m venv venv
   ./venv/bin/pip install --upgrade yt-dlp yt-dlp-ejs

3. Start the Express server:

   npm start

   The application runs at http://localhost:3300.

4. Run the regression tests:

   npm test

Docker Deployment

Build and run the application using the provided Dockerfile.

1. Build the Docker image:

   docker build -t media-downloader .

2. Run the container:

   docker run -d -p 3300:3300 --name media-downloader media-downloader

Operations

• Health check: GET /health
• API requests are limited to 30 requests per client every 10 minutes.
• At most three downloads run concurrently by default. Override with MAX_CONCURRENT_DOWNLOADS.
• Set TRUST_PROXY=1 only when the app is behind one trusted reverse proxy.

Design Details

• Install Preference: Stores PWA install-banner dismissal in localStorage.
• Signature Decryption: Uses node.js runtime with --remote-components ejs:github challenge solver to handle YouTube signature decryption.
• Format Fallback: Uses format_id+bestaudio/format_id/best format selector to download requested resolution with best audio, falling back to standalone format if needed.