Video Download & Subtitle Generation

This skill downloads videos from 1800+ websites and generates subtitles using Faster Whisper AI.

Supported Websites

This skill supports downloading from virtually any video website thanks to yt-dlp. Some popular ones include:

Video Platforms:

• YouTube, YouTube Shorts
• Bilibili (哔哩哔哩), Niconico (ニコニコ動画)
• TikTok, Douyin (抖音)
• Twitter/X, Facebook, Instagram
• Vimeo, Dailymotion
• Twitch (clips, VODs), Kick
• Rutube, VK Video

Chinese Platforms:

• iQiyi (爱奇艺), Youku (优酷), MangoTV (芒果TV)
• Weibo Video, Douyu (斗鱼), Huya (虎牙)

International:

• Netflix, Disney+, HBO Max, Amazon Prime Video
• BBC iPlayer, ITV, Channel 4
• ARD, ZDF, Arte

For the complete list of 1800+ supported sites, see: yt-dlp supported extractors

Local document about supported sites: references/supportedsites.md

Important: If you see an error like Sign in to confirm you’re not a bot, you should use the cookiefile parameter for authenticated downloads. See the cookiefile usage section at the end of this document.

Prerequisites

Ensure the following Python packages are installed:

• yt-dlp - For downloading videos from any supported site
• yt-dlp-ejs - External JavaScript for yt-dlp supporting many runtimes
• ffmpeg-python - For audio extraction
• faster-whisper - For speech-to-text transcription (faster and more memory-efficient than openai-whisper). Note: The first run will download models from HuggingFace (default: small, ~3GB). A VPN is required for mainland China users.
• tqdm - For progress bar display during transcription

Install via pip:

pip install yt-dlp yt-dlp-ejs ffmpeg-python faster-whisper tqdm

ffmpeg must also be installed on your system

Usage

Command Line

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads"}'

Parameters

Output

The skill will:

1. Create a folder for each video (named after the video title)
2. Download the video file to that folder
3. Extract audio as WAV file
4. Generate transcript using Faster Whisper
5. Save subtitle as .txt file

Output structure:

downloads/
└── Video Title/
    ├── Video Title.mp4
    ├── Video Title.wav
    └── Video Title.txt

Examples

Download YouTube video:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],"output":"./my_videos"}'

Download Bilibili video:

python scripts/video_parser.py '{"urls":["https://www.bilibili.com/video/BV1xx411c7XD"],"output":"./downloads"}'

Download TikTok video:

python scripts/video_parser.py '{"urls":["https://www.tiktok.com/@username/video/1234567890"],"output":"./tiktok"}'

Download multiple videos from different sites:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO1","https://www.bilibili.com/video/BV1xx","https://twitter.com/user/status/123"],"output":"./videos"}'

Download Twitch clip:

python scripts/video_parser.py '{"urls":["https://www.twitch.tv//channel/clip/ClipName"],"output":"./clips"}'

Download only (without transcription):

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","transcribe":false}'

Generate SRT subtitle:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","subtitle_format":"srt"}'

Generate VTT subtitle:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","subtitle_format":"vtt"}'

Download video with built-in subtitles:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","download_subtitle":true}'

Download with custom Faster Whisper model:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","model":"large"}'

Skip transcription if subtitle already exists:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","overwrite_subtitle":false}'

Download video with Cookie:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","cookie":"sid=xxx; token=yyy"}'

Download video with cookies from browser:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","cookiesfrombrowser":"chrome"}'

Download video with cookie file:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","cookiefile":"/path/to/cookies.txt"}'

Only download subtitles:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","onlysubtitle":true,"cookiefile":"/path/to/cookies.txt"}'

cookiefile usage:

• Install the Get cookies.txt LOCALLY Chrome extension first.

URL:

• Log in to the target website, then use the extension to export cookies.txt to your local machine.
• Set cookiefile to that local file path, for example:

python scripts/video_parser.py '{"urls":["https://www.youtube.com/watch?v=VIDEO_ID"],"output":"./downloads","cookiefile":"/Users/yourname/Downloads/cookies.txt"}'

Troubleshooting

• ffmpeg not found: Install ffmpeg via brew install ffmpeg (macOS) or your system's package manager
• Faster Whisper model download fails: Models are downloaded from HuggingFace. Mainland China users need a VPN/proxy to download models. The default small model is ~75MB.
• Download fails: Some videos may be geo-restricted, age-gated, or unavailable. Check the video URL and try again.
• Cookie/auth required: Some sites need authentication. You can pass cookies via yt-dlp options if needed.