yichen-douyin-fetcher

Douyin Fetcher

Douyin video download skill, uses Playwright to control the browser and intercept API responses, bypasses Douyin's msToken/X-Bogus signature verification, and obtains the watermark-free video direct link and downloads it.

Core method

1. Launch headless Chromium (Playwright)
2. Visit the Douyin video page
3. Intercept aweme/detail API response
4. Extract from the JSON response play_addr.url_list[0](watermark-free direct link)
5. Download with requests, with Referer to simulate the browser source

How to use

Download a single video

The user provides a Douyin link (the following formats are supported):

• https://www.douyin.com/video/7611845735025364265
• https://www.douyin.com/jingxuan?modal_id=7611845735025364265
• Any video URL under douyin.com

Run the script

Run the download script:

python3 yichen-douyin-fetcher/scripts/download.py "<抖音链接>" [输出路径]

Parameter description:

• Douyin link (required): video page URL
• Output path (optional): local save path, defaults to ~/Downloads/douyin_<video_id>.mp4
• --metadata-only(optional): only scrape metadata, do not download the video, used to quickly check whether the link can be parsed
• --timeout 60(optional): wait aweme/detail seconds for the response

Example:

python3 yichen-douyin-fetcher/scripts/download.py "https://www.douyin.com/video/7611845735025364265"
python3 yichen-douyin-fetcher/scripts/download.py "https://www.douyin.com/video/7611845735025364265" "/tmp/my_video.mp4"
python3 yichen-douyin-fetcher/scripts/download.py "https://www.douyin.com/video/7611845735025364265" --metadata-only

Output information

After a successful download, returns:

• Video title (desc)
• Video size (MB)
• Save path
• Author information
• Generates in the same directory <视频文件名>.metadata.json, containing concise metadata such as title, author, publish time, duration, and interaction data

Dependencies

• Python 3.9+
• playwright (pip install playwright; if Chromium is missing on first run, the script will automatically execute python3 -m playwright install chromium)
• requests

Notes

• The video link must contain a valid video_id
• Recommended page links need to be converted to the video detail page format (/video/ prefix)
• The downloaded video is in watermark-free MP4 format
• Some videos may fail to download due to copyright or regional restrictions
• This skill is only responsible for scraping the video and metadata; when a spoken script is needed, continue calling yichen-volc-asr for transcription