{
  "title": "Improve Media #2",
  "video_id": "iBhpIpokajE",
  "published": "2024-02-19T13:45:46+0000",
  "complexity": "Normal",
  "old_url": "https://workflows-of-ai.com/workflow/improve-media-2",
  "description": "Transcribe audio media with Deepgram and describe image media with GPT-4 Vision, so Drupal fills summaries and search words for you.",
  "subtitle": "Get transcripts, summaries, search words and preview images filled in for audio and image media without any editorial work.",
  "summary": "Audio media gets a Deepgram transcription, and OpenAI builds a long summary, a short summary, search words and a DALL-E preview image from it. Image media skips the text step: GPT-4 Vision in advanced mode, which needs the Token module, describes the image and suggests search words from it. Video support is planned for a later Deepgram version.",
  "categories": [
    "Drupal AI"
  ],
  "intro_html": "<p>In the last video, called <a href=\"/blog/improve-media-1\">\"Elevate Drupal Media module document field with AI elements\"</a>, I showed you how to improve the metadata for a PDF or Word document.</p><p>This is a follow-up video where we do the same for audio and images.</p><p>The idea is basically that the end user does not have to do much or any editorial work to get the media entities well described and indexable by Search API.</p><p>This makes it possible to show improved previews and makes it easy for both the editors and the end users to find the correct results.</p>",
  "intro_added": false,
  "chapters": [
    "0:00 Following up on document media",
    "0:30 Installing the Deepgram plugin",
    "2:29 Transcribing audio with Deepgram",
    "4:07 Summaries and search words for audio",
    "5:52 A preview image with DALL-E",
    "7:12 Planning the image media fields",
    "8:14 Advanced mode with GPT-4 Vision",
    "10:20 Search words from the image",
    "12:29 Checking the results",
    "13:38 Video support coming later"
  ],
  "corrections": [
    [
      "deep CR",
      "Deepgram"
    ],
    [
      "deep cram",
      "Deepgram"
    ],
    [
      "decam",
      "Deepgram"
    ],
    [
      "deep gr",
      "Deepgram"
    ],
    [
      "deep G",
      "Deepgram"
    ],
    [
      "Deep module",
      "Deepgram module"
    ],
    [
      "dream Studio",
      "DreamStudio"
    ],
    [
      "stable diffusion",
      "Stable Diffusion"
    ],
    [
      "open AI",
      "OpenAI"
    ],
    [
      "gp4",
      "GPT-4"
    ],
    [
      "GPT for Turbo",
      "GPT-4 Turbo"
    ],
    [
      "f fmeg",
      "FFmpeg"
    ],
    [
      "a inator",
      "AI Interpolator"
    ]
  ],
  "ai_note": "The video and the introduction are Marcus's own, first published on workflows-of-ai.com. AI wrote the subtitle, the summary and the video chapters from the transcript, and corrected misheard names in YouTube's automatic captions."
}