YouTube Transcript

const fetch = require("node-fetch")

async function getYouTubeCaptions(videoUrl) {
  const response = await fetch('https://website-tools-dot-maestro-218920.uk.r.appspot.com/getYoutubeCaptions', {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'User-Agent': 'Mozilla/5.0 (Linux; Android 10; K) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/141.0.0.0 Mobile Safari/537.36',
      'Referer': 'https://maestra.ai/tools/video-to-text/youtube-transcript-generator?v=RY1zQHjUG_Q'
    },
    body: JSON.stringify({
      videoUrl: videoUrl
    })
  });

  const result = await response.json();
  const transcript = extractTranscript(result.selectedCaptions);

  return {
    videoID: result.videoID,
    defaultLanguage: result.defaultLanguage,
    transcript: transcript
  };
}

function extractTranscript(webvttText) {
  let lines = webvttText.split('\n').filter(line =>
    !line.startsWith('WEBVTT') &&
    !line.startsWith('Kind:') &&
    !line.startsWith('Language:') &&
    line.trim() !== ''
  );

  let transcript = [];
  let currentText = '';

  for (let line of lines) {
    if (line.includes('-->') || line.includes('align:') || line.includes('position:')) {
      continue;
    }

    if (line.trim() === '') {
      if (currentText.trim() !== '') {
        transcript.push(currentText.trim());
        currentText = '';
      }
      continue;
    }

    let cleanLine = line.replace(/<[^>]+>/g, '').trim();

    if (cleanLine !== '') {
      if (currentText !== '') {
        currentText += ' ';
      }
      currentText += cleanLine;
    }
  }

  if (currentText.trim() !== '') {
    transcript.push(currentText.trim());
  }

  return transcript.filter(text => text !== '[Musik]' && text !== '[Music]');
}

//use
const captions = await getYouTubeCaptions("https://youtu.be/FtHaMWRn8Mc?si=G503YW_gdLl4NWL_");
console.log(JSON.stringify(captions, null, 2));