Guide · SEO & Web Analytics

How to Find Broken Links with Screaming Frog

Build a repair list that shows each failed destination and the page linking to it. Here's how to find broken links with Screaming Frog using its free desktop edition. Have a public website you manage ready; allow time for the crawl, then a few minutes to inspect and export the results.

A magnifying glass traces a broken link back to its source page.
On this page

1. Start a crawl of the website #

Install the current SEO Spider from Screaming Frog's official page. The free edition checks up to 500 URLs. This walkthrough uses the basic public-site crawl, so you don't need the paid JavaScript renderer or scheduled-crawl features.

Enter the site's address in Enter URL to spider and click Start, as shown in the official broken-link tutorial. The results should begin filling with discovered URLs.

Use the public version of the site you want to inspect. A staging address and a live address can contain different content. Write down which one you scanned so the person making edits doesn't repair the wrong copy.

2. Wait for the crawl and check its coverage #

Let the crawl finish before treating the report as a complete repair list. Check whether it reached the free allowance or stopped after finding only a small part of the website. A scan that ended early can still reveal errors, but it doesn't cover the pages it never reached.

The general user guide explains the crawler and its report workflow. For this task, focus on whether the expected main sections appear in the results.

If you manage a blog, look for a few older posts as well as the homepage. If the crawl shows only the homepage, pause the repair workflow and inspect discovery. An empty error list is little comfort when the scanner hasn't reached the content.

3. Filter the failed responses #

Open Response Codes and choose Client Error (4XX). The table should show URLs returning client-error responses. Focus first on confirmed 404 destinations, then inspect the other status codes separately.

The official tutorial shows this exact filter. It also explains that the crawl overview on the right can open the same results. Use whichever route makes the list easiest for you to review.

A 403 means access was refused, and a 429 indicates request throttling. Don't replace a reference just because those responses appear beside the missing pages. Open the destination and investigate why the crawler received that response. If it's your website, the hosting logs can help distinguish a missing resource from a blocked request.

4. Trace each destination to its source page #

Select a failed URL in the upper table, then open Inlinks in the lower pane. From is the source page containing the link; To is the destination. The report also includes link text that helps locate the reference.

You should now have the information needed to find the edit. Open the From page in your CMS or source files and identify the link with the reported text. If several pages point to the same missing destination, decide whether one shared menu or template creates those links.

For an external resource, choose a replacement that answers the same question. For an internal page you deliberately moved, update the source to the current destination. If the resource has no replacement, remove or rewrite the reference so readers understand what remains available.

5. Export a source-and-destination repair list #

Choose Bulk Export → Response Codes → Client Error (4XX) Inlinks. Save the report where the person doing the edits can use it. This is the source-page export shown in the official tutorial; exporting only the failed destination list loses the place where each repair belongs.

Keep the original status and source information when adding task columns. A simple work list can add Replacement, Owner and Done beside the crawl data. Write the chosen replacement explicitly so the editor doesn't have to make the same decision again.

You should have a report connecting each failed URL to its sources. If the error is reached through a redirect, inspect that path before editing. The original content link and the last redirect destination can be different addresses.

6. Make the corrections and check them again #

Edit the content or shared template, save the change and repeat the relevant check. Confirm that the source page now points to the intended destination and that the destination opens successfully.

Read the surrounding sentence after changing an external reference. A new page may need different link text or a revised explanation. A working destination is only a good repair when it still supports what the paragraph says.

Keep the final repair list with the scan date and website address. That makes it easier to distinguish completed edits from new findings on the next scan. For a small site, a short repeatable process is enough; you don't need to turn every repair into a large SEO project.

Known problems from crawler discussions #

The crawler receives 403 responses

In a June 2026 thread, a free-edition user reported refused requests across client sites. A hosting representative offered to inspect the affected domains. Follow that route for your own site: check the host or CDN logs and ask the administrator about a narrowly permitted audit. A browser working normally doesn't prove the crawler request was accepted.

A May 2025 React-site discussion describes navigation missing even with JavaScript rendering enabled. Replies advised checking for real anchor links with href attributes. Screaming Frog's official JavaScript guide supports that distinction. Ask the developer to inspect the navigation markup. The free edition doesn't include the paid renderer.

Troubleshooting table #

ProblemCause to investigateNext action
Crawl stops at 500 URLsFree crawl allowanceTreat coverage as partial or choose a larger allowance
Only a few pages appearDiscovery or access restrictionsCheck expected sections and the site's navigation
A browser opens a reported 403Crawler request refusedInspect hosting/CDN logs for your audit request
Important navigation isn't discoveredScripted controls without anchor linksAsk the developer to inspect href links
Export has no source pagesDestination list exportedUse Client Error (4XX) Inlinks export
Many pages contain one bad linkShared template or navigationCorrect the shared source and repeat the check

Variations for smaller checks #

Check a prepared list of URLs

Choose Mode → List, then supply the URLs you want inspected. The configuration guide and official broken-link tutorial describe list mode. It's a different scan scope from discovering pages across a website, so keep the supplied list with the report.

Hand the report to a content editor

Group the repair tasks by From page so the editor can finish several changes in one visit. Keep link text beside the destination, especially on pages containing repeated references. Use a short explanation when the replacement changes the meaning of the paragraph.

Use a browser checker for an occasional scan

Dr. Link Check and Dead Link Checker are alternatives when a desktop installation isn't convenient. Check their free limits before expecting a full-site report. The free broken-link tool list compares the allowances and reporting styles.

  • The report names the site and scan date.
  • Crawl coverage and any allowance stop are recorded.
  • Each confirmed missing destination has its source page.
  • Access errors are inspected separately from 404 repairs.
  • Replacement links match the surrounding content.
  • Changed links have been checked again.

Next steps #

See Screaming Frog vs lychee if your content is built from Markdown or HTML files and you want checks before upload. For the desktop crawler's product details, use the Screaming Frog SEO Spider page.

Featured illustration: WebIndexly, created with OpenAI image generation.

D
Donna PickardSenior Technical Writer

Donna Pickard is a Senior Technical Writer dedicated to making enterprise software and SaaS products easy to navigate. With 7+ years of experience delivering everything from deep-dive API guides to intuitive online help systems, she knows how to make technical content click. Donna currently lives…