YouTube Transcript API: Extract Video Captions & Speech-to-Text

Need to get the text content from YouTube videos? The Transcript API extracts captions and subtitles from any YouTube video—perfect for content analysis, SEO, and accessibility.

Why Use YouTube Transcripts?

• Content Repurposing - Turn videos into blog posts
• SEO Optimization - Extract keywords from video content
• Research - Analyze video content at scale
• Accessibility - Provide captions for all users
• Translation - Get multi-language transcripts
• AI Training - Build datasets from video content

Using the Transcript API

const response = await fetch('https://api.sociavault.com/youtube/video/transcript', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer YOUR_API_KEY',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    url: 'https://www.youtube.com/watch?v=dQw4w9WgXcQ'
  })
});

const transcript = await response.json();

Sample Response

{
  "videoId": "dQw4w9WgXcQ",
  "title": "Video Title",
  "duration": 212,
  "transcript": {
    "text": "Welcome to today's video. Today we're going to discuss...",
    "segments": [
      {
        "start": 0.0,
        "end": 3.5,
        "text": "Welcome to today's video."
      },
      {
        "start": 3.5,
        "end": 7.2,
        "text": "Today we're going to discuss..."
      }
    ],
    "language": "en",
    "isAutoGenerated": false
  },
  "availableLanguages": ["en", "es", "fr", "de", "ja"]
}

Use Cases

Convert Videos to Blog Posts

Extract content for written articles:

const { transcript } = await getTranscript(videoUrl);

// Clean up for blog format
const blogContent = transcript.text
  .replace(/\s+/g, ' ') // Remove extra whitespace
  .split(/[.!?]+/) // Split into sentences
  .filter(s => s.trim().length > 0)
  .map(s => s.trim() + '.')
  .join('\n\n');

console.log(blogContent);

Extract Keywords for SEO

Analyze video content for keyword research:

const { transcript } = await getTranscript(videoUrl);

const words = transcript.text.toLowerCase()
  .replace(/[^\w\s]/g, '')
  .split(/\s+/)
  .filter(w => w.length > 3);

const wordFreq = {};
words.forEach(w => {
  wordFreq[w] = (wordFreq[w] || 0) + 1;
});

const topKeywords = Object.entries(wordFreq)
  .sort((a, b) => b[1] - a[1])
  .slice(0, 20);

console.log('Top keywords:', topKeywords);

Create Timestamps

Generate chapter markers:

const { transcript } = await getTranscript(videoUrl);

// Find key phrases for chapters
const chapters = [];
let currentChapter = { start: 0, text: '' };

transcript.segments.forEach(segment => {
  // Detect topic changes (simplified)
  if (segment.text.includes('moving on') || 
      segment.text.includes('next') ||
      segment.text.includes('now let\'s talk about')) {
    if (currentChapter.text) {
      chapters.push(currentChapter);
    }
    currentChapter = {
      start: segment.start,
      text: segment.text
    };
  }
});

chapters.forEach(ch => {
  const time = formatTimestamp(ch.start);
  console.log(`${time} - ${ch.text}`);
});

Multi-Language Support

Get transcripts in different languages:

const data = await getTranscript(videoUrl);

console.log('Available languages:', data.availableLanguages);

// Get Spanish transcript if available
if (data.availableLanguages.includes('es')) {
  const spanishTranscript = await getTranscript(videoUrl, { language: 'es' });
  console.log(spanishTranscript.transcript.text);
}

Content Analysis at Scale

Analyze multiple videos:

const videoUrls = ['url1', 'url2', 'url3'];

const transcripts = await Promise.all(
  videoUrls.map(url => getTranscript(url))
);

// Find common topics
const allText = transcripts.map(t => t.transcript.text).join(' ');

// Run through your analysis pipeline
const insights = analyzeContent(allText);

Related Endpoints

• YouTube Video Details - Video metadata
• YouTube Videos Scraper - Channel videos
• YouTube Comments - Video comments
• TikTok Transcript - TikTok transcripts
• Instagram Transcript - Instagram transcripts

Frequently Asked Questions

What if a video has no captions?

If no captions exist, the API uses speech-to-text to generate a transcript. The isAutoGenerated field indicates this.

How many languages are supported?

The API supports 100+ languages for both manual and auto-generated captions.

Are timestamps included?

Yes, the segments array includes start/end timestamps for each phrase, useful for caption generation.

Can I get transcripts for live streams?

Live streams must end and process before transcripts are available. Recent live streams may not have transcripts yet.

How accurate are auto-generated transcripts?

Auto-generated transcripts typically achieve 90-95% accuracy for clear English speech. Accuracy varies with audio quality and language.

Can I download SRT/VTT files?

The API returns JSON data. You can convert the segments to SRT/VTT format:

function toSRT(segments) {
  return segments.map((seg, i) => 
    `${i + 1}\n${formatTime(seg.start)} --> ${formatTime(seg.end)}\n${seg.text}\n`
  ).join('\n');
}

Get Started

Create your account and start extracting YouTube transcripts.

Documentation: /docs/api-reference/youtube/video-transcript