Video File Summarizer API

{
  "success": true,
  "data": {
    "filename": "sample.mp4",
    "summary": "This walkthrough shows how teams extract structured data from social video at scale without standing up scrapers or OAuth apps. It covers a single API key across platforms, shared caching, and why clean JSON beats brittle HTML parses.\n\nThe speaker contrasts fresh fetches with cache hits: the same profile call can drop from multi-second billed work to a sub-second free response. They also stress verifying per-minute Whisper billing via durationSeconds and creditsCharged in the response.\n\nClosing takeaways focus on bringing your own media files for transcript-plus-summary in one POST, and using request logs to prove cache savings.",
    "keyPoints": [
      "One Captapi key works across supported platforms — no per-network OAuth.",
      "Pass cache=true for a free 24h shared cache hit on social endpoints.",
      "Whisper file endpoints bill per minute; read durationSeconds and creditsCharged.",
      "Summarize returns the full transcript alongside summary/keyPoints/topics.",
      "Short clips yield shorter summaries; longer audio aims for 2–3 paragraphs and 4–8 bullets.",
      "Account request history exposes cacheHit and responseTimeMs side by side.",
      "Use multipart POST (-F file=@path) — never put the file in the query string.",
      "Empty/no-speech audio returns HTTP 422 on summarize (cannot summarize silence)."
    ],
    "topics": [
      "APIs",
      "data extraction",
      "automation",
      "billing",
      "caching"
    ],
    "sentiment": "positive",
    "transcript": "Hey everyone, welcome back to the channel. Today we are breaking down structured data APIs for social video. First, one API key across platforms beats stitching OAuth apps together. Second, the shared cache turns expensive profile lookups into free sub-second hits. Third, when you upload a file for Whisper, always check durationSeconds against creditsCharged. Finally, summarize gives you the AI digest plus the full transcript in the same JSON payload.",
    "transcriptSegments": [
      {
        "text": "Hey everyone, welcome back to the channel. Today we are breaking down structured data APIs for social video.",
        "start": 0,
        "duration": 8.2,
        "end": 8.2,
        "timestamp": "00:00"
      },
      {
        "text": "First, one API key across platforms beats stitching OAuth apps together.",
        "start": 8.2,
        "duration": 5.1,
        "end": 13.3,
        "timestamp": "00:08"
      },
      {
        "text": "Second, the shared cache turns expensive profile lookups into free sub-second hits.",
        "start": 13.3,
        "duration": 5.4,
        "end": 18.7,
        "timestamp": "00:13"
      },
      {
        "text": "Third, when you upload a file for Whisper, always check durationSeconds against creditsCharged.",
        "start": 18.7,
        "duration": 6,
        "end": 24.7,
        "timestamp": "00:18"
      },
      {
        "text": "Finally, summarize gives you the AI digest plus the full transcript in the same JSON payload.",
        "start": 24.7,
        "duration": 5.5,
        "end": 30.2,
        "timestamp": "00:24"
      }
    ],
    "wordCount": 70,
    "segments": 5,
    "language": "english",
    "durationSeconds": 30.2,
    "duration": 30.2,
    "creditsCharged": 2,
    "noSpeech": false
  }
}