YouTube Transcript Ninja

Extracts transcripts (subtitles/captions) from YouTube videos. This Apify Actor is built with Python and processes a list of video URLs, outputting structured transcript data.

Key Features

• Fetches available transcripts and captions from standard YouTube videos and Shorts.
• Processes single or multiple video URLs in one run.
• Implements retry logic with exponential backoff for handling failed requests.
• Supports Apify Proxy (including residential proxies) and custom proxy configurations.
• Outputs both the raw API response and a clean, normalized array of transcript segments.

How to Use

Configure the actor's input via JSON. The primary requirement is the startUrls array containing the YouTube video URLs you want to process.

Input

Example Input:

{
  "startUrls": [
    { "url": "https://www.youtube.com/watch?v=BthfXVCRWEQ" },
    { "url": "https://www.youtube.com/watch?v=FWAdfuPpLOc" }
  ],
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Output

The actor outputs dataset items for each processed video. Each item contains metadata and the transcript segments.

Output Fields:
* video_id: The YouTube video ID.
* title: The video title.
* url: The full video URL.
* lang: The transcript language code (e.g., 'en').
* format: The transcript format (e.g., 'vtt').
* segments: An array of transcript objects with start time, end time, and text.
* inputUrl: The original URL processed.
* fetchedAt: ISO timestamp of when the transcript was fetched.
* success: Boolean indicating if the fetch was successful.
* raw_response: The raw transcript API response object.

Example Output Item:

{
  "video_id": "BthfXVCRWEQ",
  "title": "I Hatched the World's Biggest vs Smallest Eggs",
  "url": "https://www.youtube.com/watch?v=BthfXVCRWEQ",
  "lang": "en",
  "format": "vtt",
  "segments": [
    {
      "start": 0.12,
      "end": 5.2,
      "text": "Here is the world's biggest egg...\nand here is one of the smallest."
    }
    // ... more segments
  ],
  "inputUrl": "https://www.youtube.com/watch?v=BthfXVCRWEQ",
  "fetchedAt": "2023-10-26T10:30:00.000Z",
  "success": true,
  "raw_response": { /* Raw API data */ }
}