Video Schema & AEO: Help AI and Google Understand Your Videos

Video is becoming an increasingly popular content format, but most AI can't watch videos — they read text. VideoObject schema is how you translate video content into language AI can understand and cite. Without schema plus a transcript, even your best video is invisible to AI.

How AI processes video

AI (ChatGPT, Gemini, Perplexity) doesn't stream video. When asked a question with video as the answer, they look for:

1. Text surrounding the video: page title, article description, alt text, caption
2. VideoObject schema: name, description, thumbnailUrl — structured text
3. Embedded transcript: actual video content in text form
4. YouTube auto-transcript: if video is on YouTube, AI may read the auto-generated transcript

Priority #3 — transcript embedded in your page — is the best way to ensure AI understands and cites your video content.

Complete VideoObject schema

{
  "@context": "https://schema.org",
  "@type": "VideoObject",
  "name": "How to Create HowTo Schema in 5 Minutes — JSON-LD Guide for Next.js",
  "description": "HowTo schema helps AI and Google understand your tutorial as step-by-step. This video guides you through proper JSON-LD HowTo schema syntax, from simple steps to advanced HowToSection. Result: rich snippets and AI citation for technical guides.",
  "thumbnailUrl": "https://aeosaigon.com/images/howto-schema-thumbnail.jpg",
  "uploadDate": "2026-01-15T08:00:00+07:00",
  "duration": "PT7M45S",
  "embedUrl": "https://www.youtube.com/embed/VIDEO_ID",
  "contentUrl": "https://www.youtube.com/watch?v=VIDEO_ID",
  "author": {
    "@type": "Organization",
    "name": "AEO Saigon"
  },
  "publisher": {
    "@type": "Organization",
    "name": "AEO Saigon",
    "logo": {
      "@type": "ImageObject",
      "url": "https://aeosaigon.com/logo.png"
    }
  },
  "hasPart": [
    {
      "@type": "Clip",
      "name": "What is HowTo schema?",
      "startOffset": 0,
      "endOffset": 60,
      "url": "https://www.youtube.com/watch?v=VIDEO_ID&t=0s"
    },
    {
      "@type": "Clip",
      "name": "Basic JSON-LD syntax",
      "startOffset": 61,
      "endOffset": 200,
      "url": "https://www.youtube.com/watch?v=VIDEO_ID&t=61s"
    },
    {
      "@type": "Clip",
      "name": "HowToStep vs HowToSection",
      "startOffset": 201,
      "endOffset": 380,
      "url": "https://www.youtube.com/watch?v=VIDEO_ID&t=201s"
    }
  ]
}

Duration format (ISO 8601)

Video duration uses the PT#H#M#S format:

Transcript: how AI actually reads your video

The transcript is the most important part but the least implemented. How to add it:

Option 1: Collapsible section (recommended)

<details>
  <summary>View full video transcript</summary>
  <div class="transcript">
    <p><strong>[00:00]</strong> What is HowTo schema and why does AI need it...</p>
    <p><strong>[01:05]</strong> Starting with the basic JSON-LD structure...</p>
    <!-- full transcript -->
  </div>
</details>

Option 2: Auto-transcript from YouTube YouTube automatically generates transcripts for videos. Export from YouTube Studio → translate and edit → add to your web page.

Option 3: Use automated services Whisper (OpenAI), AssemblyAI, or Rev.com for high-accuracy transcripts.

Google Key Moments — special display in Search

When you declare hasPart (Clip schema) with timestamps, Google may display Key Moments in search results — a list of timestamps with segment names, allowing users to jump directly to the part they need.

Requirements:

• Video must have Clip schema with startOffset, endOffset
• Video must be publicly accessible (not private)
• Web page must have VideoObject schema and embed that video
• At least 3 clips declared

Key Moments is one of the few rich snippets that AI search engines also reference when answering "which step..." type questions.

Optimize your YouTube channel for AEO

Beyond schema on your website, optimize the YouTube channel itself:

Video titles as questions or answer-first:

• "How to add HowTo schema for Next.js App Router (2026)" — good
• "Video #12 — Schema tutorial" — bad

Video description — the first 250 characters matter most:

HowTo schema helps tutorial content get cited by AI. In this video:
→ 00:00 What is HowTo schema
→ 01:05 Complete JSON-LD syntax
→ 03:20 Implementation in Next.js App Router
→ 05:40 Testing with Rich Results Test

Chapters in YouTube (equivalent to Clip schema): Add timestamps in the YouTube description → YouTube auto-creates chapters → supports Google Key Moments.

VideoObject schema + transcript is the essential pair for video content in the AI search era. See HowTo schema and FAQ schema to combine with video content, or use AEO checker to check your current page schema.

Frequently asked questions