Google Discover images and max-image-preview checks
Check whether an article offers a suitable large image for Google Discover, including image size, preview directives, crop, and the public image response.
By Crawlito · 5 min read
In short
- Google recommends relevant large images at least 1200 pixels wide for Discover.
- The max-image-preview directive controls permission for a large image preview.
- A working article URL and a working image URL need independent verification.
On this page
- Identify the image a public reader receives
- Check dimensions and the downloaded bytes
- Inspect preview permission in all relevant places
- Trace a changed image through caches and metadata
- Evaluate the page and subsequent observations
- Test a URL that silently returns a thumbnail
- Frequently asked questions
- Check the article and its image
To investigate an article's image readiness for Google Discover, check the actual public image, its dimensions and relevance, and the page's image-preview directive. Google recommends large images at least 1200 pixels wide and enabling large previews with max-image-preview:large or the documented alternative. Google decides which eligible articles to display in Discover. Google’s Discover guidance
Identify the image a public reader receives
Open the article without an editor session and inspect its principal image. Record the file requested by the browser, the source references in the initial HTML, and the article's relevant image metadata. A CMS might store a large original but publish only a small thumbnail. The media-library dimensions alone cannot establish what a crawler or visitor can retrieve.
Compare the image with the article's actual subject. A generic site logo or decorative banner may fit a theme while doing little to describe this story. Choose a relevant image that helps someone understand the article. For research or technical documentation, a legible original illustration may serve the subject better than an unrelated stock photo. Preserve necessary credit and provenance in the publication's normal workflow.
Inspect the same article on a phone. A responsive image component can choose a different resource or crop at small widths. Record the candidates available in the HTML and which one the browser selected. A small viewport selection does not by itself prove that a large candidate is unavailable; test the relevant candidate addresses directly.
Check dimensions and the downloaded bytes
Fetch the intended large image anonymously and inspect the returned file. A URL ending in .jpg can redirect to a sign-in screen, return an HTML challenge, or supply a generic placeholder. Confirm status, type, actual decoded dimensions, and visual content together.
curl -sSL --max-time 30 -D hero-headers.txt -o hero-image.bin https://media.example.com/articles/field-guide-large.jpg
file hero-image.binUse an image inspector to read the dimensions and view the result. If a transformation service controls width through a query parameter, verify the public URL after that transformation. A requested width of 1600 does not guarantee that the underlying original contains enough detail or that the service returned those dimensions. Upscaling a tiny image can satisfy a superficial dimension check while producing poor quality.
Keep the source's intended composition in mind. Text labels embedded near an edge can be lost in a crop. A face or important object can also become unrecognizable when the image is used in a different aspect ratio. Check the publisher's actual image variants and choose a composition that remains useful in the supported presentation. The image context guide explains how captions and alt text support the page itself.
Inspect preview permission in all relevant places
Look for the page's robots meta tags and response headers. An explicit preview rule can be emitted by a theme, SEO extension, or infrastructure layer. The following example permits large previews under Google's documented directive.
<meta name="robots" content="max-image-preview:large">Inspect any additional googlebot directive and X-Robots-Tag header that applies. A permissive tag in one block can coexist with a restrictive rule elsewhere. Review the complete response and Google's documented handling of combined directives before deciding that the setting is effective. The robots directive guide helps locate these signals.
A page might intentionally limit previews because of rights or business policy. Confirm the intended policy with the content owner before changing it. A technical audit should describe the current setting and its consequence accurately. The goal is to implement the publisher's actual preview preference, with clear evidence that the public response expresses it.
Google's robots specification documents image-preview values and their scope. Use that reference when a plugin labels its setting differently from the emitted directive. A control called “large previews” is useful only if the published page contains the intended output. Google’s robots metadata specification
Trace a changed image through caches and metadata
After replacing an image, inspect the page HTML, image metadata, and image bytes again. A new photograph uploaded under an old filename can be cached independently from the article. The page may reference the intended URL while the CDN still returns the previous photo. Save response time and cache headers when comparing origin and public behavior.
Check any explicit image fields in article structured data and social metadata. These may be generated from a separate featured-image record rather than the first visible image. Keep them accurate for the article and avoid contradictory image identities. A content editor changing only the body image can leave the previous feature image in those fields.
The CDN freshness guide provides a way to investigate old public bytes. Use targeted cache procedures supported by your platform and verify the result from a new request. Repeatedly changing the filename without understanding the publishing source can leave stale references throughout the article and feeds.
Evaluate the page and subsequent observations
Google says content can be eligible for Discover when it is indexed and meets the relevant policies, with no special markup required. Check the article's ordinary public access and index signals alongside its image. A large image cannot compensate for an article route that returns a challenge or an unintended noindex directive.
Use a small acceptance set with one new article, one older article whose image changed, and one page intentionally using a restrictive preview policy. Confirm the expected image and directive on each. Include a mobile view and direct image fetch. This gives the editor and developer distinct checks they can reproduce after a theme update.
If Discover performance is available in Search Console, record the reporting window and which pages actually received activity. A lack of visible Discover traffic is not a diagnosis of one image setting. Compare the technical findings with the specific article and date before drawing a conclusion about presentation.
Keep the article's relevance and editorial quality central to the publishing decision. The image check establishes whether the page offers a usable public visual and the intended preview permission. Subsequent impressions and clicks are observations to measure separately as the audience encounters the publication.
Test a URL that silently returns a thumbnail
Some media services reuse one path while choosing the output size from request headers or cached transformation state. Download the purported large image twice using the site's normal public image URLs and inspect the decoded dimensions both times. Compare the result with a known large original from the same service.
If the address consistently returns a small thumbnail, trace the CMS field or transformation preset that produced it. If the size varies unexpectedly, investigate cache keys and content negotiation with the media owner. Retain the actual files from the test, since the URL text alone cannot demonstrate what dimensions were delivered. The final check should establish that the intended large candidate is available through a stable public request.
Frequently asked questions
Is uploading a 1200-pixel image enough?
Verify the public resource, relevant composition, preview permission, and the article's normal eligibility. Dimensions are one part of the documented guidance.
Do I need special Discover schema?
Google states that special tags or structured data are not required for Discover eligibility. Keep existing article markup accurate for the page it describes.
Can an old image remain after I change the featured image?
Yes, if the page, metadata, and asset caches update independently. Fetch the public article and image separately to locate the stale representation.
Check the article and its image
Run the free page check on the article to inspect its public content and directives. Then inspect the large image file independently for dimensions, relevance, and accessibility.
Filed under Structure and schema. Topics: images, google, metadata