Image Metadata SEO: How to Get Credit for Your Images in Google

Google reads who made an image, who to credit and where its license lives from two places: IPTC photo metadata inside the file and ImageObject structured data on the page. Add a license URL and your images become eligible for the Licensable badge in Google Images. We added this to every image on this site, and learned that your host can quietly strip half of it.

What image metadata does Google read?

Google uses image metadata to show who is behind an image and how it may be used, both in Google Images. It can read that information from the file itself (IPTC photo metadata, stored as IPTC or XMP) or from structured data on the page. These are the fields that count:

With structured data, the only required property is contentUrl, plus at least one of creator, creditText, copyrightNotice or license. Google uses contentUrl to match the metadata to the right image, so it has to be the actual image URL.

How do you get the Licensable badge?

The Licensable badge in Google Images links to your license and tells people how the image may be used. Google’s rule is short: with IPTC you must fill in the Web Statement of Rights, and with structured data you must include license. Both point to a page that explains the terms.

That page doesn’t have to be a paid licensing shop. Ours, Image use & credit, says our images are free to share with a credit link back to the source article. That’s a license too, and it turns every copied image into a potential backlink.

IPTC or structured data — which should you use?

Use both if you can. They do different jobs:

• Structured data is what Google trusts first. The documentation says: if IPTC and structured data conflict, “Google will use the structured data information.” It also lives on your page, so nobody can strip it.
• IPTC metadata travels with the file. When someone downloads your image and uploads it elsewhere, the creator and license URL can go with it — if nothing strips them on the way.

On WordPress with Yoast SEO, the article’s main image is already output as an ImageObject. We added creator, creditText, copyrightNotice, license and acquireLicensePage to it with a small filter on Yoast’s schema graph, so every existing and future image gets the same fields.

What happened when we added it to 42 images

On 26 September 2026 we rebuilt the hero image of every article on this site: a visible footer with the article’s own URL, the author and the year, plus metadata written with ExifTool into IPTC, XMP and EXIF. Then we downloaded the live files the way a browser gets them. Our host serves images through a CDN that re-encodes them, often as WebP, and that re-encoding decided what survived.

In other words, the optimizer kept exactly the rights information and threw away the rest. That’s good enough for crediting, but it means you can’t rely on the file alone for descriptions or keywords. Those belong in the alt text, the caption and the page itself.

Two smaller lessons. First, the © sign got garbled in the IPTC Copyright Notice after re-encoding, so we now write “Copyright 2026 IndexUrePoint” there and keep the © only in XMP. Second, IPTC fields have hard length limits: Credit Line and Source allow just 32 characters, so “IndexUrePoint.com” fits and a full URL doesn’t.

How do you check your own images?

Don’t trust the file on your computer; check the file your visitors and Googlebot receive. With ExifTool installed:

curl -s -o live.jpg https://example.com/wp-content/uploads/image.jpg
curl -s -o live.webp -H "Accept: image/webp" https://example.com/wp-content/uploads/image.jpg
exiftool -G1 -s -Creator -Rights -WebStatement -Credit -CopyrightNotice live.jpg live.webp

The second line asks for WebP, like Chrome does, which is often where metadata disappears. Then:

• Test the page in Google’s Rich Results Test to confirm the ImageObject has license and contentUrl.
• In Search Console, look under Enhancements for an Image metadata report. On our site it appeared within days of adding the metadata. You’ll need your pages crawled first, so check they’re indexed.

What about AI-generated images?

Google reads the IPTC Digital Source Type field and supports four values for AI involvement, including trainedAlgorithmicMedia (made by a generative model) and compositeWithTrainedAlgorithmicMedia (a composite that includes AI-generated material). The base artwork of our illustrations comes from an AI image model and every word and number on them is added by us, so we label them as composites.

Our host stripped that field, though. So we also say it in plain words on our image use page. If labelling AI images honestly matters to you, don’t rely on the file field alone.

Should you put your URL on the image itself?

Metadata is invisible and fragile; pixels aren’t. Screenshots, social platforms and AI tools routinely drop metadata, but a small footer with your domain and the article’s URL survives all of them. That’s why every image here now carries its own address at the bottom.

Image-led searches are growing enough that Google now reports them separately: Search Console’s new multimodal filter shows traffic from Lens and Circle to Search.

Keep it modest: a thin strip, not a watermark across the subject. Google’s image guidelines favor relevant, high-quality images, and a logo splashed over the content does the opposite. For social sharing, make a separate 1200×630 version so platforms don’t crop your footer off.

FAQ