How to Add Captions and Transcripts to Images and Video

This check flags pages where images and videos sit on the page without text that explains them, such as captions, transcripts or a nearby description. By the end of this guide, every meaningful image on your key pages will have alt text and a visible caption, and every video will have a transcript, a captions track and structured data.

By Updated ⏱ 30–90 minutes, depending on how many pages have media🎯 Intermediate● Low impactGEO / AI
Illustration: Context for images and video fix guide
GEO / AI · FIX GUIDE Context for images and video IMPACTEFFORT TIME30–90 minutes, depending on how many pages have media LEVELIntermediate

Why it matters

AI answer engines and search crawlers don't watch videos or interpret images the way people do. They rely on the text around the media. A video with no transcript is effectively empty to them.

A chart with no caption can't be quoted or cited. Captions and transcripts give these systems text they can index, summarize and attribute to you. They also help users who are deaf, skimming, or browsing with media blocked, and they make the page eligible for video and image rich results.

✕BeforeA product demo page with a bare<video> embed and three screenshotsthat have alt="" and no captions or✓AfterEach screenshot sits in a <figure>with descriptive alt text and a<figcaption>. The video has a visible

What you need

Edit access to your CMS or page templates
A crawler such as Screaming Frog SEO Spider or Sitebulb to list pages with images and video
A transcription tool (YouTube auto-captions, Descript, Otter or Whisper) for video audio
Google's Rich Results Test for checking structured data

Step-by-step fix

1

List the pages that contain images or video

Crawl your site and export every page that contains content images, <video> tags or YouTube/Vimeo iframes. The check scores the share of pages that pass, so start with high-traffic templates such as blog posts, product pages and guides. Ignore decorative images like icons, dividers and background patterns.

Those don't need captions.

text
Screaming Frog > Configuration > Custom > Extraction:
Name: video   XPath: count(//video | //iframe[contains(@src,'youtube') or contains(@src,'vimeo')])
Name: figcaption   XPath: count(//figcaption)
Name: img_no_alt   XPath: count(//img[not(@alt) or @alt=''])
Pro tip: Sort the export so pages with a video count above 0 and a figcaption count of 0 come first. Those are usually the quickest wins.
2

Wrap meaningful images in figure and figcaption

Any image that carries information (screenshots, charts, diagrams, product photos, team photos) should sit inside a <figure> with a visible <figcaption>. The caption should say what the image shows and why it matters. Include specifics like numbers, names or dates rather than a generic label.

html
<figure>
  <img src="/img/checkout-funnel-2026.png" alt="Bar chart of Example Co checkout steps: 62% reach payment, 41% complete purchase" width="1200" height="675">
  <figcaption>Example Co checkout funnel, Q1 2026. Most drop-off happens between shipping and payment.</figcaption>
</figure>
Pro tip: A caption is read far more often than body text. Put your key takeaway there, not just a title.
3

Write alt text that complements the caption

Alt text describes what the image shows for people who can't see it. The caption explains its meaning for everyone. Don't paste the same sentence into both, because screen readers will read it twice.

Keep alt text under about 125 characters, and use alt="" only for purely decorative images.

html
<!-- Informative -->
<img src="/img/team.jpg" alt="Five Example Co engineers reviewing a whiteboard sprint plan">
<!-- Decorative -->
<img src="/img/divider.svg" alt="">
4

Publish a visible transcript for every video

Generate a transcript, then edit it for accuracy. Fix product names and add speaker labels. Place it on the same page as the video, in the HTML, not inside a PDF or behind a script-loaded widget.

A collapsible <details> element keeps the page tidy, and the text stays in the source for crawlers.

html
<figure>
  <iframe src="https://www.youtube-nocookie.com/embed/VIDEO_ID" title="Example Co dashboard setup in 4 minutes" loading="lazy" allowfullscreen></iframe>
  <figcaption>How to connect your store to the Example Co dashboard (4:12).</figcaption>
</figure>
<details>
  <summary>Read the full transcript</summary>
  <p><strong>Maya (Example Co):</strong> In this video we'll connect your store in three steps...</p>
</details>
Pro tip: Add a 2–3 sentence summary above the video as well. AI answers often quote the summary rather than the full transcript.
5

Add a captions track to self-hosted video

For <video> files you host yourself, add a WebVTT captions file using the <track> element. For YouTube or Vimeo, upload the corrected transcript in the platform's subtitle settings:

  • YouTube: YouTube Studio > Subtitles > Add language > Upload file
  • Vimeo: Video settings > Distribution > Subtitles
html
<video controls preload="metadata" poster="/video/setup-poster.jpg">
  <source src="/video/setup.mp4" type="video/mp4">
  <track kind="captions" src="/video/setup.en.vtt" srclang="en" label="English" default>
</video>
Pro tip: Serve the .vtt file with the content type text/vtt, or some browsers will ignore it.
6

Describe the video with VideoObject structured data

Add JSON-LD that names the video, describes it and includes the transcript. The transcript, description, thumbnailUrl and uploadDate properties give search engines and AI systems a machine-readable summary.

html
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "VideoObject",
  "name": "Example Co dashboard setup in 4 minutes",
  "description": "Step-by-step walkthrough of connecting a store to the Example Co dashboard.",
  "thumbnailUrl": "https://example.com/video/setup-poster.jpg",
  "uploadDate": "2026-01-15",
  "duration": "PT4M12S",
  "contentUrl": "https://example.com/video/setup.mp4",
  "transcript": "In this video we'll connect your store in three steps..."
}
</script>
7

Re-crawl and confirm the pages now pass

Re-run your crawler with the same custom extractions and confirm that pages with media now report figcaptions and no missing alt text. Test a few video pages in Google's Rich Results Test to confirm VideoObject is detected without errors. Then re-run the audit to see the share of passing pages go up.

text
curl -s https://example.com/guides/dashboard-setup | grep -cE '<figcaption|<track|"transcript"'
Pro tip: Check the rendered HTML as well as the source. Some page builders inject captions only through JavaScript.

Platform-specific notes

WordPress (block editor)

Click an Image or Video block and use the 'Add caption' button in the block toolbar. This outputs <figure> and <figcaption> automatically. Set alt text in Block settings > Alternative text. For transcripts, add a Details block below the video. Yoast SEO and Rank Math output basic VideoObject markup, but you may need a custom HTML block to add the transcript property.

Shopify

Add product image alt text in Admin > Products > [product] > click the media > Add alt text. Themes don't show captions by default. Add a metafield (for example custom.media_caption) and render it in sections/main-product.liquid inside a <figcaption>. Put video transcripts in a Collapsible content block in the theme editor.

Webflow

Select an image and set alt text in the Settings panel (gear icon) > Alt text. In Rich Text and CMS rich-text fields, click the image and choose the caption icon to add a <figcaption>. Paste VideoObject JSON-LD into Page settings > Custom code > Before </body>, or bind CMS fields inside an Embed element.

Wix

In the Wix Editor, click an image > Settings > 'What's in the image?' to set alt text. Wix Blog images support captions through the image toolbar. For video, add the transcript in a collapsible text box or Wix Blog section. Wix adds some video structured data on its own but doesn't include a transcript.

Next.js

next/image renders only an <img>, so wrap it in your own <figure> and <figcaption> component and make alt a required prop. Render transcripts on the server (in Server Components or getStaticProps output) so they appear in the initial HTML, and inject JSON-LD with a <script type="application/ld+json"> in the page component.

Check that it worked

✅ How to verify

  • Re-run the audit and confirm the 'Context for images and video' check shows a higher share of passing pages.
  • View the page source (Ctrl+U) and confirm that <figcaption>, the transcript text and any <track> element are present in the HTML.
  • Paste a video page URL into Google's Rich Results Test and confirm VideoObject is detected with no errors.
  • Turn on a screen reader such as NVDA or VoiceOver and check that images are announced with useful alt text and are not read twice.

⚠️ Common mistakes

  • Copying the alt text into the caption word for word, so screen readers repeat it and the caption adds nothing new.
  • Hosting the transcript in a PDF, a separate page or a script-loaded widget that crawlers never see.
  • Publishing raw auto-generated captions with misheard product names and no punctuation.
  • Captioning decorative images such as icons and dividers while leaving charts and screenshots unexplained.

Is this still an issue on your site? Run a free audit — ZORO checks this and 294 other SEO, AI-visibility, accessibility and security signals, with fixes for your platform.

Check my site free

FAQ

Do AI search engines like ChatGPT and Google AI Overviews read video transcripts?

Yes. They rely heavily on the text on the page. A transcript in the HTML, plus a short summary, is often the only way the content of your video can be cited in an AI answer.

What's the difference between alt text and a caption?

Alt text is hidden by default and describes the image for people who can't see it. A caption is visible to everyone and explains the image's meaning or context. Informative images benefit from both.

Is YouTube's auto-generated transcript good enough?

It's a starting point. Edit it to fix names, technical terms and punctuation before you publish it on your page or upload it as subtitles. Inaccurate transcripts can spread errors into AI answers.

Does every image need a caption?

No. Caption images that carry information, such as charts, screenshots, diagrams and product or people photos. Decorative images should have alt="" and no caption.