YouTube Transcripts
AI-generatedSummary
Fetches subtitles for a YouTube video with yt-dlp and returns the transcript alongside the video ID, title, URL, and chosen format. Processes each incoming item using its configured URL and carries through its input JSON and binary data.
Inputs
- YouTube URL (required) — YouTube watch or youtu.be URL with an 11-character video ID.
- Language Priority — Comma-separated language codes passed to yt-dlp in preference order. Defaults to en,zh,zh-CN,ja,ko,ru,de,fr,es,auto.
- Include Auto-Generated — Whether yt-dlp should also fetch auto-generated subtitles. Defaults to true.
- Output Format — Select plain text (txt), SRT, or VTT. Defaults to plain text; plain text has subtitle numbering, timestamps, and tags stripped.
Output shape
One output item per successfully processed input item, containing JSON fields videoId, title, url, transcript, and format, merged with the input JSON; input binary data is preserved.
The title is fetched with yt-dlp and falls back to "Unknown Title" if that lookup fails. Subtitle retrieval requires yt-dlp to be installed or YT_DLP_PATH configured, and fails if no matching subtitle file is found. With continue-on-fail enabled, a failed item instead carries its input JSON and an error field.