What goes in an image sitemap when my product photos live on a CDN?

Build a page-to-image inventory

A fictional ceramics store serves pages from its own domain and photos from a CDN. The seller lists only product page URLs in a spreadsheet and assumes the sitemap already describes every image. Build a second column for the exact approved image URL used by each page. Include relevant actual images rather than private generation previews.

Google documents image sitemap extensions and cross-domain image hosting. The ledger here prepares that implementation; it is not a completed XML sitemap and does not submit anything to Search Console.

Review the separate hosts

A successful image request in the owner’s logged-in browser does not establish anonymous crawler access. Ask the implementation team to check public delivery on the actual host. Do not weaken access controls on private material to fill a sitemap.

Include only useful approved images

For the fictional mug, map the intended public page to approved primary and secondary image URLs. A page URL belongs in loc; each image URL belongs inside its own image:image entry as image:loc. The host names below are reserved examples, not real files or a submitted sitemap.

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:image="http://www.google.com/schemas/sitemap-image/1.1">
  <url>
    <loc>https://store.example/products/mug-blue</loc>
    <image:image>
      <image:loc>https://media.example/mug-blue-front.jpg</image:loc>
    </image:image>
    <image:image>
      <image:loc>https://media.example/mug-blue-handle.jpg</image:loc>
    </image:image>
  </url>
</urlset>

Google’s image-sitemap reference permits cross-domain image hosting with both domains verified in Search Console and no relevant robots block. Its current image extension uses image:image and image:loc; image:caption, image:geo_location, image:title and image:license were removed from its documentation. Do not revive them from an old template.

Replace each example with the actual public URL. Exclude rejected or private previews. If image URLs change, rebuild the mapping from published assets. If you cannot verify the relevant third-party image host, resolve that implementation requirement with the provider rather than asserting the sitemap is compliant or weakening private access controls.

Validate and monitor the real sitemap

Have the developer validate the XML and inspect the published sitemap response. After submission, read Search Console’s actual processing status and dates. A successful sitemap is a discovery signal, not proof that every image is indexed. Keep the approved asset record so later changes can be reconciled without confusing a new creative with a broken delivery URL. Record separate outcomes: XML accepted, public image fetch succeeds, and actual search/index evidence. None proves the others. For a missing image on an indexed product page, continue with the image-discovery diagnostic.