Douyin Video Downloader

Downloads Douyin videos by opening the page in a headless browser, extracting the source URL, and downloading via curl. This bypasses yt-dlp's cookie issues with Douyin's anti-scraping.

Prerequisites

• agent-browser (npm i -g agent-browser)
• curl

Usage

python3 scripts/douyin_download.py <URL> [--output-dir DIR] [--filename NAME]

Supported URL Formats

• https://www.douyin.com/video/ — direct video page
• https://www.douyin.com/search/...?modal_id= — search results with video modal
• https://v.douyin.com/ — share short links
• https://www.douyin.com/note/ — note/image posts with video

Examples

# Basic download to ~/Downloads
python3 scripts/douyin_download.py 'https://www.douyin.com/video/7577715519366576522'

# Custom output directory and filename
python3 scripts/douyin_download.py 'https://www.douyin.com/video/7577715519366576522' \
  -o ~/Videos -f my_video

# From search page URL
python3 scripts/douyin_download.py 'https://www.douyin.com/search/关键词?modal_id=7577715519366576522'

How It Works

1. Normalize URL → extract video ID, construct direct video page URL
2. Open page in agent-browser (headless Chromium)
3. Extract element's currentSrc (CDN direct link)
4. Close browser
5. Download MP4 via curl with proper Referer header

Notes

• No login required — fresh browser session is sufficient
• Video title auto-detected from page title for filename
• Large videos may take 30-60s to download depending on network
• CDN links are temporary (~2h validity); download promptly after extraction