How to Put Product Specs in HTML Text, Not Images
Many product and service pages show dimensions, materials, ratings or pricing tiers only as a JPG spec sheet, an infographic or a downloadable PDF. This guide shows you how to find those pages, rebuild the specs as real HTML text and tables, and confirm that crawlers can read them.

Why it matters
AI assistants and search engines quote what they can parse from the page's HTML. Text baked into an image depends on OCR, which is unreliable and often skipped. PDFs get crawled separately, so their facts are not tied to your product page.
", a page with a readable spec table can be cited, and an image-only page usually is not. HTML specs also help screen-reader users, work in Google's product rich results and are easier to translate.
What you need
Step-by-step fix
List pages where specs live only in images or PDFs
Start with the URLs flagged in your audit. pdf files or use images with names or alt text such as 'spec', 'datasheet', 'dimensions' or 'size-chart'. Then open each template and check whether the same facts also appear as selectable text.
If you can't highlight a number with your mouse, crawlers usually can't read it either.
Screaming Frog: Configuration > Custom > Search Contains: datasheet|spec-sheet|size-chart|\.pdf Apply to: product and service URL paths
Extract the spec data into a structured source
Take the specs from the original source, such as a PIM, ERP or supplier spreadsheet, rather than retyping them from images. If the PDF is the only source, copy the text out of it. For scanned PDFs, run OCR (for example, Adobe Acrobat's Scan & OCR tool) and proofread every number.
Use the same field names on every product, such as Weight, Dimensions, Material, Power and Warranty, and always include units.
Render specs as a semantic HTML table
Add a spec table to the product or service template. Use <th scope="row"> for the labels so screen readers and parsers can match each label to its value. Add a <caption> or a heading that names the product.
Keep units next to the values, and use real characters such as × and ° instead of images.
<section id="specifications">
<h2>Example Co X200 specifications</h2>
<table>
<caption>Technical specifications for the Example Co X200</caption>
<tbody>
<tr><th scope="row">Weight</th><td>1.2 kg</td></tr>
<tr><th scope="row">Dimensions (W × H × D)</th><td>30 × 20 × 10 cm</td></tr>
<tr><th scope="row">Material</th><td>Anodised aluminium</td></tr>
<tr><th scope="row">Battery life</th><td>Up to 12 hours</td></tr>
<tr><th scope="row">Warranty</th><td>2 years</td></tr>
</tbody>
</table>
</section>Make sure specs are in the initial HTML
Tabs and accordions are fine, as long as the content is already in the page source and only hidden with CSS. Do not load specs through a click-triggered fetch or an iframe. Many AI crawlers do not run JavaScript, so content that appears only after scripts run may never be seen.
<details open> <summary>Specifications</summary> <!-- spec table here, present in server HTML --> </details>
Mirror key specs in Product structured data
Add the same values to your Product JSON-LD. Use standard properties such as weight, width, height, depth and material where they exist. Put everything else in additionalProperty as PropertyValue entries.
The values must match the visible table exactly.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Example Co X200",
"brand": {"@type": "Brand", "name": "Example Co"},
"material": "Anodised aluminium",
"weight": {"@type": "QuantitativeValue", "value": 1.2, "unitCode": "KGM"},
"width": {"@type": "QuantitativeValue", "value": 30, "unitCode": "CMT"},
"additionalProperty": [
{"@type": "PropertyValue", "name": "Battery life", "value": "Up to 12 hours"},
{"@type": "PropertyValue", "name": "Warranty", "value": "2 years"}
]
}
</script>Keep images and PDFs as extras, not the only source
You don't need to delete infographics or datasheets. Keep them, but add descriptive alt text to the images, such as alt="X200 dimensions diagram: 30 × 20 × 10 cm". Label PDF links clearly, for example 'Download full datasheet (PDF, 240 KB)'.
The HTML table should hold every fact a buyer would compare.
<p><a href="/files/x200-datasheet.pdf">Download the full X200 datasheet (PDF, 240 KB)</a></p>
Verify the result and re-run the audit
Fetch the raw HTML of a fixed page and confirm the spec values appear in it. Test the structured data in the Rich Results Test. Then re-run the audit and confirm the flagged pages have dropped off.
Spot-check one page per template type.
curl -s https://example.com/products/x200 | grep -iE "Weight|Dimensions|1\.2 kg"
Platform-specific notes
WordPress / WooCommerce
Go to Products > Edit product > Product data > Attributes. Add each spec, tick 'Visible on the product page' and click Save attributes. WooCommerce then shows them as an HTML table in the 'Additional information' tab. For non-product pages, use the block editor's Table block, not an Image block. SEO plugins such as Yoast WooCommerce SEO or Rank Math can map attributes into Product schema.
Shopify
Go to Settings > Custom data > Products and create metafields such as specs.weight and specs.material. Fill them in on each product. In Online Store > Themes > Customize, add a block to the product template and connect it to the metafield with the 'Connect dynamic source' icon. Alternatively, loop over the metafields in Liquid inside a <table>, for example {{ product.metafields.specs.weight.value }}.
Wix
In Wix Stores, open a product and use 'Additional Info Sections' (Product Info Sections) to add a Specifications section as text. You can also add a Table element from Add Elements in the Editor. Avoid uploading spec sheets as Image or Gallery elements.
Webflow
Add fields for each spec (Plain text or Number) to the Products or CMS Collection. Bind them to a table built from Div blocks or a Collection List on the template page. Webflow renders CMS content server-side, so the values appear in the HTML.
Next.js / headless
Render the spec table in a Server Component, or with SSG/SSR, so it is in the initial HTML response. Don't fetch specs client-side inside useEffect, and don't place them in a modal that only mounts on click. Check the result with curl, not in the browser.
Check that it worked
✅ How to verify
- Run curl or use 'View page source' (not DevTools Elements) and confirm the spec values appear as text inside a <table> or a list.
- Paste the URL into Google's Rich Results Test and confirm Product is detected with your weight, material and additionalProperty values.
- Use Search Console > URL Inspection > View crawled page and search the crawled HTML for a spec value.
- Re-run the audit and confirm the 'Specs in HTML, not images' check no longer lists the fixed URLs.
⚠️ Common mistakes
- Moving specs into an image's alt attribute only. Alt text helps, but it is not visible page content and cannot hold a full spec sheet.
- Loading the spec tab through JavaScript after a click, so crawlers that don't run scripts see an empty container.
- Letting the HTML table, the schema and the PDF disagree. Mismatched values erode trust and can disqualify rich results.
- Pasting specs as one long paragraph without labels or units, which makes individual values hard to extract and quote.
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 freeFAQ
Can Google read text inside images?
Google can extract some text from images, but it is not reliable for ranking or for answers. Many AI crawlers do not run OCR on page images at all. Put any fact you want indexed or cited in HTML text.
Are PDFs bad for SEO?
No. Google indexes PDFs, but as separate documents. Their content is not credited to your product page, and the PDF may rank in place of the page. Keep the PDF as a download, and put the key specs on the HTML page as well.
Should specs be in a table or a list?
Use a table for label–value pairs and for comparisons across models, with <th> headers. A description list (<dl>, <dt>, <dd>) also works for simple single-product specs. Either is far better than an image.
Do hidden tabs or accordions hurt how specs are indexed?
Not if the content is in the initial HTML and only hidden visually. Google indexes such content normally. Problems start only when the content is fetched by JavaScript after a user interaction.
How to Fix Semantic HTML Structure for AI and Search Parsers
How to Fix Missing Pricing Information for AI Crawlers
How to Fix Question-Style Headings for AI Search
How to Fix Wrong Facts About Your Brand in AI Answers
How to Add a TL;DR or Key Takeaways Box to Long Content
How to Build Review Platform Presence That AI Assistants Cite