YouTube Transcript Extractor

Extracts transcripts and captions from YouTube videos in bulk. Outputs clean, structured text optimized for AI/RAG pipelines, content analysis, translation, and research.

What it does

Takes YouTube video URLs (or channel/playlist URLs), extracts available transcripts and captions, and outputs structured text with timestamps. Supports multiple languages and automatic/manual caption types.

Key Features

• Bulk extraction - Process hundreds of videos in one run
• Multi-language support - Extract transcripts in any available language
• Channel/Playlist support - Automatically discover videos from channels and playlists
• AI-ready output - Clean text format optimized for RAG, embeddings, and LLMs
• Timestamp preservation - Keep or remove timestamps based on your needs
• Chunking options - Split transcripts into configurable chunks for embedding pipelines

Input

Field	Type	Default	Description
urls	string[]	required	YouTube video, channel, or playlist URLs
languages	string[]	["en"]	Preferred transcript languages (ISO 639-1 codes)
includeAutoGenerated	boolean	true	Include auto-generated captions
includeTimestamps	boolean	true	Include start/duration timestamps
chunkSize	integer	0	Split transcript into chunks of N characters (0 = no chunking)
chunkOverlap	integer	200	Overlap between chunks in characters
outputFormat	string	"structured"	Output format: "structured", "plain_text", "srt", "vtt"
maxVideosPerChannel	integer	50	Max videos to process per channel/playlist
proxyConfiguration	object	{}	Apify proxy settings

Output

Each video produces a dataset item:

{
  "videoId": "dQw4w9WgXcQ",
  "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "title": "Video Title",
  "channelName": "Channel Name",
  "language": "en",
  "isAutoGenerated": false,
  "availableLanguages": ["en", "es", "fr"],
  "transcript": [
    {
      "text": "Hello and welcome",
      "start": 0.0,
      "duration": 2.5
    }
  ],
  "fullText": "Hello and welcome to this video...",
  "wordCount": 1523,
  "duration": 612.5,
  "chunks": [
    {
      "index": 0,
      "text": "Hello and welcome to this video...",
      "startTime": 0.0,
      "endTime": 45.2
    }
  ],
  "extractedAt": "2026-02-23T07:00:00Z"
}

Use Cases

• RAG Knowledge Bases - Build searchable knowledge bases from educational YouTube channels
• Content Research - Analyze competitor content, extract key topics
• Translation - Extract source text for translation workflows
• SEO Analysis - Analyze video content for keyword research
• Podcast Transcription - Many podcasts are uploaded to YouTube with captions
• Training Data - Collect text data for fine-tuning language models
• Accessibility - Generate text versions of video content

Pricing

Pay per result:

• $0.15 per 1,000 videos processed
• Free tier: 100 videos/month

Example Usage

Extract transcripts from specific videos

{
  "urls": [
    "https://www.youtube.com/watch?v=VIDEO_ID_1",
    "https://www.youtube.com/watch?v=VIDEO_ID_2"
  ],
  "languages": ["en"],
  "outputFormat": "structured"
}

Extract from a channel with chunking for RAG

{
  "urls": ["https://www.youtube.com/@ChannelName"],
  "languages": ["en"],
  "chunkSize": 1000,
  "chunkOverlap": 200,
  "maxVideosPerChannel": 100,
  "outputFormat": "structured"
}

Get plain text transcripts in multiple languages

{
  "urls": ["https://www.youtube.com/playlist?list=PLAYLIST_ID"],
  "languages": ["en", "es", "fr"],
  "includeTimestamps": false,
  "outputFormat": "plain_text"
}