YouTube Transcript Extractor — AI-Ready Subtitles

Extract clean subtitle/transcript text from any YouTube video with subtitles. Designed for AI training data pipelines, content analysis, and LLM training.

Features

• 🎯 Input a YouTube URL or bare video ID
• 🌐 Supports manual and auto-generated captions
• 🌍 Multi-language — specify any ISO 639-1 language code (default: en)
• ⏱ Optional [MM:SS] timestamps in output
• 🧹 Clean, join-transcript format
• 📊 Rich metadata: video_id, duration, word count, language
• 🛡️ Robust error handling with descriptive error messages

Input

Field	Type	Required	Default	Description
video_url	string	✅	—	YouTube URL (any format) or bare video ID
language	string	❌	en	ISO 639-1 language code
include_timestamps	bool	❌	false	Add [MM:SS] before each subtitle line

Output (one item per run)

Field	Type	Description
video_id	string	11-char YouTube video ID
title	string	Video title (if retrievable)
duration	int	Approximate duration in seconds
language	string	Language code of the transcript
transcript_type	string	"manual" or "auto-generated"
transcript	string	Full clean text of the subtitles
word_count	int	Word count of the transcript
url	string	Full YouTube URL

Supported URL formats

• https://www.youtube.com/watch?v=VIDEO_ID
• https://youtu.be/VIDEO_ID
• https://www.youtube.com/embed/VIDEO_ID
• https://www.youtube.com/shorts/VIDEO_ID
• Bare VIDEO_ID (11 characters)

Use Cases

• AI/LLM Training Data — collect natural language text from millions of YouTube videos
• Content Analysis — analyze video content at scale for SEO, research, or moderation
• Accessibility — extract captions for further processing or translation
• Dataset Building — build large text corpora from video subtitles

---

Built with youtube_transcript_api ❤️