Free Google Discover Technical Audit Tool: Check Image, Crawl and Preview Issues

I’ve built a free technical audit tool to help identify some of the issues that can affect Google Discover visibility.

It is also useful for checking technical signals that matter for Google News and Top Stories as well as Discover, especially around image eligibility, crawler access, structured data and how Google is able to preview a page.

You can try it here: Free Google Discover Technical Audit Tool

This started as a set of checks I used when troubleshooting specific site problems for publishers and other businesses that rely heavily on search visibility.

One recurring issue was image handling. In some cases, images were not appearing properly in Google News. In others, Discover performance had dropped and one of the underlying problems was that Google could not use the right images in the way the site expected.

Sometimes the issues were connected to CDN changes. Sometimes they followed a migration. Sometimes they were caused by caching rules, structured data, crawler access, image transformations or template changes.

The awkward thing about these problems is that the page can look fine in a browser.

An editor opens the article and sees the headline, image, byline and page layout they expect. But Google may be seeing something slightly different. Or the page may be technically accessible, but not providing Google with the image, metadata or preview permissions it needs for the best presentation in Discover, Google News or Top Stories.

I found myself checking the same things manually, then later with a set of small Python scripts, so I decided to pull them together into one tool.

What the Google Discover audit tool checks

The tool audits a single URL and looks for observable technical signals that can affect how Google crawls, understands and previews a page.

It checks things like:

• crawler access
• simulated Googlebot vs normal browser responses
• rendered or stealth browser access where needed
• image size and accessibility
• whether max-image-preview:large is set
• structured data
• schema image consistency
• canonical and robots signals
• basic byline, date and publisher signals
• headline and content presentation signals
• caching and delivery issues that can appear after CDN or development changes

The aim is to surface the kinds of things that are easy to miss in normal editorial workflows.

For example, a page may have an article image, but the image exposed in og:image may be too small for Discover. A URL may appear to request a 1200px image, but the delivered image response may still be smaller. A JSON-LD image field may point to a page URL rather than an image file. A CDN may send different cache-control headers to normal users and simulated crawlers. A site may serve the article to a real browser but return a challenge or incomplete page to automated requests.

None of these issues automatically explain performance. But they are worth knowing about.

What this Google Discover tool is not

This is not a full SEO audit.

It is not a qualitative assessment of a website or article.

It will not tell you whether your journalism is good, whether your site has topical authority, or whether your content deserves to perform well in Discover.

It also will not guarantee that a page will appear in Google Discover, Google News or Top Stories.

That matters because Discover performance is shaped by far more than technical setup. Story selection, timing, image choice, headline strength, topical authority, site reputation, user interest and editorial quality all matter.

A technically clean page can still perform badly.

A technically imperfect page can still appear in Google News.

The purpose of this tool is narrower. It is there to help identify technical risks that may be getting in the way, or to rule some of them out more quickly.

Why image checks matter for Google Discover

A lot of Google Discover troubleshooting eventually comes back to images.

Discover is a highly visual surface. If Google cannot access a suitable image, or if the page does not allow large image previews, the page may be less likely to get the presentation you want.

That does not mean an image issue is always the cause of a Discover decline. It often is not.

But image problems are common after migrations, CMS changes and CDN changes. They are also easy to overlook because the visible image on the page is not always the same image Google is being asked to use in metadata or structured data.

The tool checks whether images are accessible, large enough, representative and consistent across signals such as og:image, schema image fields and visible article images. It also checks whether the page allows large previews using max-image-preview:large.

That last one is important. Missing max-image-preview:large is not an indexing blocker, but for publishers who care about Discover and image-led Google surfaces, it is a technical issue worth fixing.

Why crawler access checks matter

The tool compares a normal browser-style request with a simulated Googlebot request. In some cases, it also uses a rendered or stealth browser-style fetch to understand whether the page is accessible in a normal browsing context.

This is useful because some issues only show up when requests are treated differently.

A site might serve normal users a full article but give automated requests a challenge page. A CDN or WAF might treat simulated crawlers differently. A page might be accessible in a rendered browser but incomplete in the raw HTML response. A template might expose different metadata depending on user agent or rendering path.

The tool tries to identify those differences without overclaiming.

It is important to be clear about this: the tool does not make requests from Google’s verified Googlebot IP ranges. A simulated Googlebot block does not prove that real Googlebot is blocked.

If the report shows access problems, the next step should be verification in Google Search Console URL Inspection, server logs, CDN logs or WAF logs.

The tool can show that something looks suspicious from the outside. It cannot replace proper log-level diagnosis.

How it can help with Google News and Top Stories

Although the tool is primarily useful for Google Discover troubleshooting, many of the same checks are also relevant to Google News and Top Stories.

For Google News, technical issues around crawl access, article metadata, structured data, canonicals, publication dates, publisher signals and image handling can all affect how easily Google understands and processes a page.

That does not mean the tool can tell you whether a page deserves to be included in Google News. It cannot.

But it can help identify whether there are obvious technical issues that may be making it harder for Google to crawl, interpret or preview the page properly.

Why I’ve tried not to make it alarmist

One of the things I dislike about many free audit tools is that they are designed to scare people.

They find a long list of issues, give everything a dramatic score, and leave people feeling like their site is broken.

I have tried to avoid that.

Some findings are marked as informational. Some are caveated. Some reports are marked as incomplete if the tool cannot retrieve a reliable version of the article. Simulated Googlebot issues are treated carefully because they do not prove real Googlebot behaviour.

That does not mean the tool is soft. If a page has no suitable image, blocks image access, misses max-image-preview:large, serves incomplete HTML, or exposes inconsistent metadata, it should say so clearly.

But it should also explain what the issue does and does not prove.

Limitations

There are plenty of limitations.

The tool checks one URL at a time. It does not look at wider site architecture, internal linking, topic clusters, Search Console performance, Google News traffic, Discover trends, server logs, analytics or competitor performance.

It does not assess the quality of an article.

It does not know whether a site has earned authority on a topic.

It does not know whether a headline is genuinely compelling, whether an image is editorially strong, or whether a story is likely to match user interest in Discover.

It also will not work cleanly on every site. Some sites block non-Google bots or anything that looks automated. In those cases, the report may still provide some useful signals, but it may be incomplete.

This is not a universal website audit tool. It is mainly designed for sites where Google Discover, Google News and Top Stories matter.

Who it is for

The tool is most useful for publishers, editorial teams, audience teams, SEOs and developers who need a quick technical check on a specific URL.

It may be useful when:

• a story is not appearing as expected in Google News
• Google Discover performance has dropped after technical changes
• a site has recently changed CMS, CDN, templates or image handling
• images are not appearing properly in Google surfaces
• you want to compare what a normal browser sees with a simulated crawler request
• you want to rule out obvious technical issues before spending time on deeper editorial analysis

It is not a substitute for a proper investigation. But it can help identify where to look next.

How to use the Google Discover audit tool

Using the tool is simple.

Paste in the URL of the article or page you want to check, then run the audit.

The tool will fetch the page in a few different ways, including a normal browser-style request and a simulated Googlebot request. Where needed, it may also use a rendered or stealth browser-style fetch to understand whether the page is accessible in a normal browsing context.

Once the audit has finished, it produces a report showing the main issues found, grouped by severity.

The score is there to give you a quick sense of whether the page has obvious technical risks, but it should not be treated as a performance prediction. A high score does not mean the page will perform in Google Discover. A low score does not mean the content is poor. It means the tool has found technical issues worth reviewing.

The most useful parts of the report are usually:

• image eligibility
• max-image-preview status
• crawler access
• normal browser vs simulated Googlebot differences
• structured data and image metadata
• canonical and robots signals
• article byline, date and publisher signals
• caching or delivery differences

If the report finds high-priority image issues, start there. For Google Discover, image size, accessibility and preview permissions are often among the most practical technical checks to fix.

If the report says the audit is incomplete, that usually means the tool could not retrieve reliable article HTML. This can happen when a site blocks automated requests, returns a challenge page, requires JavaScript, or treats non-Google bots differently. In that situation, the report may still be useful, but you should verify what real Googlebot sees using Google Search Console URL Inspection, server logs or CDN/WAF logs.

If the report finds only low or informational issues, that is still useful. It may mean there are no obvious technical blockers on that URL, and that the next stage of investigation should focus more on content quality, headline and image strength, topical authority, internal linking, timing, competition and wider Search Console or Discover performance data.

The tool is best used on specific URLs rather than homepages or section pages. It is designed mainly for article-style pages where Google Discover, Google News or Top Stories visibility matters.

Try the free Google Discover technical audit tool

You can try the tool here: Free Google Discover Technical Audit Tool

I have tested it with well over 100 URLs across more than 50 domains, mainly sites where Google Discover, Google News, Top Stories and other search surfaces matter.

That is not exhaustive. There will be edge cases. Some advice may need refining. Some sites will block the tool. Some findings will need manual interpretation.

If something looks wrong, please let me know. I have added a contact form to the tool.

If you disagree with the advice, I would like to know that too.

I would much rather improve the tool than pretend it is perfect.