Broken and Redirected Image Audit
Diagnose image 3xx, 4xx, and 5xx responses, trace failed assets to source pages, and repair the component or media reference.
Use Success (2xx) as the healthy comparison, Redirects (3xx) for moved image sources, Client Errors (4xx) and Server Errors (5xx) for failures, and Broken for the combined unsuccessful set.
Investigate the complete image request
- Open the source URL and confirm status, content type, and final destination.
- Inspect every Source Page and determine how the URL was generated.
- Check whether the failure affects
src, a responsive candidate, or both. - Group findings by CDN path, upload process, template, and filename pattern.
- Compare browser requests with Screpy's crawl response.
- Check authorization, hotlink protection, firewall, and case-sensitive paths.
A 404 often means a file was deleted or moved without updating content. A 403 can indicate external hotlink protection. A 5xx pattern can reveal image transformation or CDN failures. A redirect may be harmless for old external references but avoidable in your own markup.
<!-- Old source -->
<img src="/uploads/old/hero.jpg" alt="Product overview" />
<!-- Current source -->
<img src="/media/hero.webp" alt="Product overview" />Choose the correct repair
Restore the file when the reference is still valid, update source markup to the current URL, preserve a redirect for legitimate old public usage, or remove the image when it no longer serves the page. Update every responsive candidate and metadata reference, not only the fallback src.
Do not replace an informative missing image with an unrelated placeholder merely to return 2xx. Correct content matters as much as transport success.
Example decision
A shared product component requests a deleted thumbnail path while the full-size source remains available. Update the media transformation mapping and verify every product state. Do not edit hundreds of product records individually when one component creates the URL.
Verify the fix
Load pages with the browser network panel, test image URLs directly, resize across responsive candidates, and recrawl. Confirm the Broken count falls, redirects are intentional, and images display correctly on every affected template.
HTTP Status Codes
Understand access, rate-limit, client, and server responses.
Responsive Images
Check every srcset candidate after repairing an asset.
Broken Links
Audit failed page and resource destinations from their sources.
Find the real failure mode
A failed image may return 404, 403, 429, or 5xx; it may also fail because of DNS, TLS, hotlink protection, or an invalid URL. Open the resource directly and compare what the crawler receives with what a normal browser receives.
Group failures by host, path pattern, and source template. One missing CDN variant can affect every product card, while a single editorial typo affects only one page. External image failures require a decision: replace the asset, host a permitted copy, or remove the dependency.
Expected changes after repair
A successful fix should reduce Broken and the relevant error-code filter while increasing Success (2xx). If the image now redirects, Broken may disappear but Redirects (3xx) will rise. Update first-party markup to the final image URL when practical.
If the total image count falls, confirm whether files were intentionally removed or whether affected pages stopped being crawled. Also verify that lazy-loaded images remain discoverable in rendered markup.
Example investigation
Suppose /media/product-800.webp returns 404 on 300 pages. Check whether the image transformation service changed naming rules, whether the original exists, and whether the shared card component requests a nonexistent size. Repairing the generator or component is safer than manually changing 300 pages.
HTTP response classes are summarized in MDN's status-code reference. After deployment, recrawl affected templates, open several images at their natural URLs, and check that dimensions, format, and alt behavior were not accidentally changed.
Frequently asked questions
Why does an image work in my browser but fail in the crawl?
Your browser may use cookies, cached files, a different location, or a slower request rate. Compare response headers and security rules, and ensure the crawler can access the resource without a logged-in session.
Should a missing image URL redirect?
Redirect when the same asset moved and the new file is a true equivalent. For first-party markup, also update the source to the final URL. Return an error when no replacement exists.
Related guidance
Return to the Images overview to compare broken resources with redirects and source types. Google's image SEO best practices explain how crawlable image URLs, page context, and supported markup affect image discovery.
Image Dimensions and Layout Stability
Audit missing width and height attributes, reserve image space, and verify layout stability without treating every finding as a Core Web Vitals failure.
Responsive Images and Srcset
Audit missing and present srcset usage and deliver appropriate image candidates with sizes, picture, and a reliable fallback source.