youtube-transcribe-skill

Subtitle extraction from YouTube videos uses yt-dlp to automatically download transcripts in your preferred language, saving them as a timestamped text file. The Skill applies to standard YouTube URLs (watch?v= or youtu.be/ formats) and falls back to browser automation if CLI extraction fails.

Yes. This Skill automates transcript extraction using yt-dlp CLI for speed, with browser automation as a fallback. It requires yt-dlp availability and browser cookies, then outputs a local text file named after the video, eliminating manual copying or third-party transcript sites.

The Skill prioritizes common subtitle languages—Simplified Chinese, Traditional Chinese, and English—selecting the first available match on the video. If your preferred language isn't available, it retrieves whatever captions the video offers.

The Skill automatically switches to browser automation via chrome-devtools-mcp to retrieve transcripts when CLI methods fail, ensuring extraction works even when yt-dlp encounters restrictions or unsupported video configurations.

You need yt-dlp installed and browser cookies available for fallback extraction. The Skill handles the rest—it detects YouTube URLs, retrieves subtitles, and writes a formatted text file with timestamps and content to your local directory.

Extracted transcripts save as a text file named <Video Title>.txt with lines formatted as Timestamp Text, making the output ready for reference, analysis, translation, or import into other workflows without additional parsing.