usdigitalmarketingindex.com SEO
When should an image sitemap include assets hosted on a separate CDN?
Updated November 15, 2026
When images that matter are served from a different hostname and might not be found through the pages alone, and when the conditions for listing them can be met. Google’s documentation permits image URLs on another domain in a sitemap, with a condition: “In some cases, the image URL may not be on the same domain as your main site. This is fine, as long as you verify both domains in Search Console” (Google’s image sitemap documentation). It adds a second requirement, that robots.txt must not disallow crawling of content you want indexed. So the decision has two parts. First, is an image sitemap useful for this site at all? Second, if the images sit on a CDN host, are that host’s verification and crawl access in order? An image sitemap listing CDN URLs that crawlers are blocked from fetching accomplishes nothing.
External image URLs in sitemap entries
An image sitemap attaches image entries to the page they appear on. The page URL belongs to the main site. The image URL can be anywhere.
A realistic CDN-hosted example, with placeholder domains:
<?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://www.example.com/products/thermostat-t200/</loc>
<image:image>
<image:loc>https://cdn.example-media.net/products/thermostat-t200-front.jpg</image:loc>
</image:image>
<image:image>
<image:loc>https://cdn.example-media.net/products/thermostat-t200-wall-plate.jpg</image:loc>
</image:image>
</url>
</urlset>
Points from the documentation:
| Point | Detail |
|---|---|
| Structure | Image tags sit inside the page's url entry; they can be added to an existing sitemap or kept in a separate one, and the documentation calls either approach "equally fine" |
| Required tags | The image container tag and the image location tag |
| Limit | Up to 1,000 image entries per page entry |
| Removed tags | Caption, geo location, title, and license tags were removed from the documentation; do not add them |
| Purpose | To tell Google about images "we might not otherwise find (such as images your site reaches with JavaScript code)" |
When the image sitemap earns its place:
| Situation | Include images in a sitemap? |
|---|---|
| Images loaded by JavaScript, galleries, or carousels not present in the served HTML | Yes |
| Images important to the business: products, portfolio, original photography | Yes |
| Images in plain <!–INLINECODE0–> elements on crawlable pages | Optional; they can be found through the pages |
| Decorative images, icons, and interface graphics | No |
| Images on pages that are not indexable | No |
Hosting on a CDN does not by itself make a sitemap necessary. An image in an ordinary img element is discoverable through its page wherever the file lives. The CDN matters because it adds conditions.
Verification and crawl access
Two things must be true of the CDN hostname.
Verification. The documentation says the arrangement is fine “as long as you verify both domains in Search Console”, and, for a content delivery network, to “make sure that the hosting site is verified in Search Console”.
| CDN setup | Verification route |
|---|---|
| A subdomain of your own domain, such as images.example.com, pointed at the CDN | Covered by a Domain property for example.com, or verify the subdomain as its own property |
| A separate domain you own, such as example-media.net | Verify that domain |
| A hostname owned by the CDN provider, shared with other customers | You cannot verify a host you do not control; ask the provider about a custom hostname |
The first setup is the simplest to manage: a hostname under your own domain, served by the CDN.
Crawl access. The same passage says to “make sure that your robots.txt file doesn’t disallow the crawling of any content you want indexed”. Each hostname has its own robots.txt.
| Check | On the CDN host |
|---|---|
| robots.txt exists and loads | Fetch it at the root of the CDN hostname |
| It does not disallow the image paths | Read the rules for Googlebot and Googlebot-Image |
| Images return 200 to a crawler | Request with a Googlebot user agent |
| No hotlink protection blocks crawlers | Request without a referrer, and with the main site as referrer |
| No signed or expiring URLs in the sitemap | URLs must be stable |
| HTTPS | Certificates valid |
A CDN can be configured with a default robots.txt that blocks everything, or with hotlink rules that refuse requests lacking the site’s referrer. Either makes the images unfetchable, sitemap or no sitemap.
Image inventory coverage
Decide which images to list, and generate the list from data.
| Step | Detail |
|---|---|
| Source | The CMS or product database, which knows each page's images |
| Select | Content images on indexable pages |
| One URL per image | The canonical, full-size or standard variant, not every resized version |
| Stable URLs | No session tokens, signatures, or cache-busting parameters that change |
| Match the page | Each image listed under the page that displays it |
| Match the markup | The same image URL that appears in the page's HTML |
| Keep current | Regenerated when images are added or removed |
The sixth row, on matching the markup, matters most for a CDN. If the page’s HTML references one URL for an image and the sitemap lists a different URL for the same file, for instance a different size or hostname, the two do not corroborate each other. Use one form.
A coverage check:
| Measure | Compare |
|---|---|
| Images in the sitemap | Against content images found by crawling the pages |
| Pages with image entries | Against indexable pages that have images |
| Images in the sitemap that return errors | Should be none |
| Images on pages and absent from the sitemap | Decide whether they should be listed |
Sitemap testing
| Test | Method | Pass condition |
|---|---|---|
| Well-formed XML with the image namespace | Validate the file | Valid |
| Page URLs | Request each | 200, indexable, canonical |
| Image URLs | Request each, or a sample | 200, image content type |
| Crawler access to images | Fetch with a Googlebot user agent; test against the CDN's robots.txt | Allowed |
| Verification | Search Console property list | CDN hostname verified or covered |
| Submission | Sitemaps report in Search Console | Read successfully; no errors |
| Agreement with pages | Compare sitemap image URLs with <!–INLINECODE2–> values in page HTML | Same URLs |
| Stability | Fetch the sitemap on two days | Image URLs unchanged for unchanged images |
| Outcome | Performance report with search type set to image, by page | Impressions for pages whose images were listed |
Read the outcome with care. A sitemap is a way of telling Google about URLs. It does not cause indexing, and an image listed in a sitemap is assessed like any other.
Decision summary:
| Condition | Action |
|---|---|
| Important images, on a CDN hostname under your domain, crawlable | List them |
| Important images on a separate domain you own | Verify the domain, check its robots.txt, then list them |
| Images on a provider-owned shared hostname | Arrange a custom hostname first |
| CDN blocks crawlers | Fix access first; the sitemap changes nothing until then |
| All images in plain HTML on crawlable pages, low importance | A sitemap is optional |
| Expiring or signed image URLs | Do not list; provide stable URLs |
Changing which host serves the images is a migration, with redirects and mapping of old image addresses, and that work is described in the guide to how image URLs can retain search visibility during a media library migration.
The rule: list CDN-hosted images when they matter and might be missed, after confirming the CDN hostname is verified in Search Console and open to crawling.
Add your comment to this answer