Broken Link Checker for Confluence: how to find and fix broken links in Confluence Cloud
Video transcript. Recorded with sample pages in a development installation on 2 October 2026.
Video on YouTube: https://www.youtube.com/watch?v=RPm9Ql1QpME

0:00 A broken link in Confluence
Here is our Support Handbook, and this is the page new agents use on day one: Onboarding guide for new agents. It has a first-week checklist, a tools table, and a link to the support tools repository. Click it, and we get a 404. The repository moved, and the same old link is copied on other pages too. By the end, you’ll know how to find broken links in Confluence, fix this repository link everywhere, and verify the result.

I’ll use Broken Link Checker for Confluence, a Confluence broken link checker for space admins and knowledge-base owners.

0:36 Run a space scan
In Confluence’s left sidebar, expand Apps and select Broken Link Checker for Confluence. Choose Support Handbook beside Scan space. You need space administrator permission to scan. Then select Scan space.

Public HTTPS links and supported internal Confluence page links are checked automatically. When you see Scan complete, select Refresh results and read the coverage message. A finished scan can still be incomplete if it hit a limit or could not read some content. A scan checks at most one thousand published pages and ten thousand links.

1:14 Read Broken and Unverified results
Open All links and choose Broken. Broken means a public external destination returned HTTP 404 or 410. Our support tools repository returns 404 on three pages: Onboarding guide for new agents, New starter checklist, and Customer escalation process. Release notes archive returns 410. These are confirmed failures, so prioritize them first.

Review Unverified separately. It means the check was inconclusive. Here, atlassian.com timed out, and fresh results can appear as Not checked. The filter also includes Access restricted and HTTP link. Those need manual review. Excluded means a space rule prevented the check.

Available means the destination succeeded, or the internal Confluence target is accessible. Link removed means a later check confirmed that the URL no longer appears on the source page. Fix Broken results first, then review Unverified by reason.

2:19 Owners and work states
Each issue follows the current Confluence page owner. My queue shows issues on your pages, and ownership changes move issues to the new owner after refresh. Mark in progress records active work. Ignore removes an exception from the default queue, Reopen brings it back, and Recheck queues another check.

The default Needs action filter hides resolved and ignored issues. Use All work states when checking completed work. Owners shows open, in-progress, ignored, and resolved issues by owner, plus recent scans.

2:55 Replace a URL across pages
Now let’s fix the migration. In All links, select the three support-tools rows and choose Replace links. Under Match, choose Exact URL. We have one complete old repository address and one complete new address, so Exact URL is the right match.

Base URL and path is for a wider domain or path move. For example, if an old docs base moves to a new help base, a setup link below it keeps the remaining path, query, and fragment. The match respects a path boundary: docs matches docs slash setup, but not docs-old. Enter base URLs without a query or fragment.

Back to our repository. Check Current URL, enter the HTTPS destination in New URL, and open the new destination yourself first. Select Preview changes. Review every old and new destination and each source page. We can see the same change on all three pages. Untick anything you do not want to update. Include no more than twenty pages.

When the preview is correct, select Save changes to three pages. Keep the panel open until every page has a result. Each successful save publishes a new Confluence page version, while visible link text stays the same.

4:18 Verify the fix
Now verify it. Close the panel, choose All work states and All link results, and refresh after the queued check completes. Before you trust a search, use Load more results. Our three pages now point to the new address, and the result is Available. Load more also found a fourth page with the old link: IT support runbook. For one page, use Replace link in its row. Preview, save, and it is fixed.

Return to Onboarding guide for new agents and click the support tools repository link. This time it opens the repository. Then open Confluence version history. The page has a normal new version labeled Broken Link Checker: replace reviewed links. Version history can restore an earlier version if needed.

5:06 Schedules, exclusions and limits
For ongoing maintenance, open Settings. Under Automatic checks, choose Off, Daily, or Weekly, then Save settings. Scheduled checks start within approximately an hour after they become due, subject to Forge availability and quotas. Only one space scan runs at a time; Stop scan can end a stalled or unwanted run.

You can exclude destinations too. Domain exclusions skip an exact hostname. For one exact destination, select rows and choose Exclude URLs. Remove a rule in Settings and recheck affected pages when you want them included again.

Some cases remain manual. Plain-text URLs and third-party macro internals are not scanned. URL fragments and content meaning need manual review. HTTP links need a manual check, and an HTTPS link that redirects to another protocol appears as Redirect not checked.

Before exporting, remember that filters apply to results loaded so far. Use Load more results while it is available. Then use Export loaded results, or Export CSV once everything is loaded, to export the currently filtered, loaded rows.

6:21 Recap
The workflow is: scan, check broken links, separate confirmed failures from uncertain results, route work by owner, preview replacements, save, and recheck. Broken Link Checker for Confluence is free for up to ten Confluence users on Atlassian Marketplace. The app and user guide are linked below.