Universal Media Extraction Guide

1. What is yt-dlp?

yt-dlp is a feature-rich command-line media downloader. It is a fork of the original youtube-dl project, optimized for higher speeds and better compatibility with modern streaming protocols like DASH and HLS.

2. Setting up on Hugging Face

To run a yt-dlp website on the free tier, you need three files in your repository:

• README.md: Set sdk: gradio and python_version: 3.11.
• packages.txt: Include ffmpeg to handle video/audio merging.
• requirements.txt: Include yt-dlp and gradio.

3. The FFmpeg Backbone

High-quality web video is often served as separate Video-only and Audio-only streams. Without FFmpeg, your app is limited to low-quality legacy formats (usually 720p or lower).

FFmpeg acts as the "Muxer," joining these streams into a single MP4 or MKV file after the download is complete.

4. Working Website Implementation

Below is a production-ready app.py for a Hugging Face Space. It uses Gradio for the UI and yt-dlp for the backend.

import gradio as gr
import yt_dlp
import os

def download_video(url):
    # Setup options with FFmpeg integration
    ydl_opts = {
        'format': 'bestvideo+bestaudio/best',
        'outtmpl': 'downloads/%(title)s.%(ext)s',
        # Ensure FFmpeg is used to merge streams
        'postprocessors': [{
            'key': 'FFmpegVideoConvertor',
            'preferedformat': 'mp4',
        }],
    }
    
    with yt_dlp.YoutubeDL(ydl_opts) as ydl:
        info = ydl.extract_info(url, download=True)
        filename = ydl.prepare_filename(info).replace('.webm', '.mp4').replace('.mkv', '.mp4')
        
    return filename

# Gradio Interface
with gr.Blocks(theme=gr.themes.Soft()) as demo:
    gr.Markdown("# 🚀 Private Video Downloader")
    url_input = gr.Textbox(label="Paste Link (YouTube, TikTok, Instagram, etc.)")
    btn = gr.Button("Fetch Video", variant="primary")
    output_file = gr.File(label="Download Ready")

    btn.click(fn=download_video, inputs=url_input, outputs=output_file)

demo.launch()