Kitefold
Now on the Atlassian Marketplace

Link Health & Cleanup for Confluence

Find every broken link, orphan page and unused attachment across all your spaces, see which pages link to each page, and clean up safely. Made for Confluence admins and anyone who looks after documentation, especially after a move from Data Center.

Illustration with made-up data

What it finds

Confluence can't tell you which pages link to a page, which links are broken, or which files nobody uses. Link Health keeps an index of every link on your site, inside Atlassian, and turns it into reports and a byline on every page.

Broken links

Links to pages, blog posts, files or spaces that don't exist or are in the trash, with the page the link is on and what it points to.

Old server links

Enter your old Confluence Data Center or Server address and see every link that still points to it. These stopped working when you moved to Cloud.

Orphan pages

Pages that no other page or blog post links to. Archive the ones nobody needs, or mark the rest as not orphans.

Unused attachments

Files that no page shows or links to, across every space, with their sizes and the largest first. Move them to the trash in a few clicks.

Stale pages

Pages nobody has edited for 1 to 24 months, oldest first, straight from Confluence search.

Incoming links on every page

“Linked from 12 pages” under each page title. Click it to see the pages, grouped by space, or add the Incoming links macro to a page.

Careful by design

An app that reads every space has to be careful about who sees what, and about what it changes.

  • Shows only what you can see. Every list and every “Linked from” count is checked against Confluence's permissions as the person viewing it. Pages in spaces or pages you can't open are left out and counted, never named.
  • Stores links, never page text. The index holds page and file ids, titles, link targets, file names and sizes. It's kept in Atlassian's storage for the app and deleted when the app is uninstalled.
  • Clean-up you can undo. Archiving pages and moving files to the trash happen as you, so Confluence's permissions apply, and both can be undone. Nothing is ever deleted permanently.
  • Nothing leaves Atlassian. The app runs on Atlassian Forge and makes no calls to outside services. No analytics and no AI.

Read the privacy policy

Pricing

Free for sites with up to 10 users. Larger sites pay per user through the Atlassian Marketplace, billed by Atlassian with your other apps. Every plan includes all the reports, the byline and the clean-up actions, and you can try it free first.

Requirements

  • Confluence Cloud (the app is not available for Data Center)
  • A Confluence admin to install it and use the reports
  • Tested on sites with up to 100,000 pages (larger sites)
Help

Help and FAQ

New to cleaning up Confluence? Our guides cover what Confluence can do on its own: broken links, which pages link to a page, orphan pages, unused attachments and old links after a move to Cloud. See also what's new.

1. Getting started

  1. A Confluence admin installs Link Health & Cleanup from the Atlassian Marketplace (in Confluence: Apps → Explore more apps).
  2. The app starts building its index straight away. On most sites this takes a few minutes; large sites can take several hours, because Atlassian limits how much apps can read from Confluence each hour.
  3. In Confluence, open Settings (the cog, top right) and choose Link Health & Cleanup in the menu. The first time, Confluence asks you to allow access.
  4. While the index is being built, the page shows how far it has got. The reports appear by themselves when it's finished.

Tip: if you moved from Data Center or Server, enter your old address on the Old server tab (for example wiki.example.com) to find links that still point to it.

2. The reports

The admin page has a tab for each report. Every list can be filtered by space and exported as CSV.

TabWhat it lists
OverviewThe totals for each report, then each space's numbers. Click a column heading to put the spaces with the most first.
Broken linksLinks to pages, blog posts, spaces or files that don't exist or are in the trash. Links to archived pages aren't counted as broken. Links to a missing file usually don't show on the page at all; edit the page to find and remove them.
Old serverLinks to the old address you enter (subdomains are included). Separate several addresses with commas.
OrphansPages that no other page or blog post links to. Space homepages aren't included, and the page tree doesn't count as a link, so an orphan may still be easy to find. Mark pages that are meant to be reached another way as Not an orphan.
Unused filesFiles that no page or blog post shows or links to, largest first, with their sizes. A page's header image counts as used. Files on a page with the Attachments macro count as used too, unless you press List them too.
Stale pagesPages nobody has edited for the chosen time (1, 3, 6, 12 or 24 months), oldest first. This list comes straight from Confluence search each time, as you.
SettingsSpaces to leave out, and the index: when it was last built, and Build index to rebuild it now.

3. Incoming links on each page

  • Under every page title, the byline says Linked from 12 pages (or No incoming links). Click it to see those pages, grouped by space. When Confluence is very busy the byline may briefly say just Incoming links; the list in it still works.
  • To show the same list on a page, edit it, type / and add the Incoming links macro.
  • The count only includes pages you're allowed to open, so two people can see different numbers.
  • When a page changes, its links are updated within a few minutes.

4. Cleaning up

  • Archive pages: on the Orphans tab, tick up to 50 pages and press Archive selected. They move to their space's archive, where anyone with permission can restore them.
  • Move files to the trash: on the Unused files tab, tick up to 50 files and press Move selected to trash. A space admin can restore them from the space's trash until it's emptied.
  • Both ask you to confirm first, then show what happened to each item. They run as you, so Confluence refuses anything you don't have permission to change, and the list says why.
  • The app never deletes anything permanently.

5. Settings and the index

  • Spaces to leave out: choose spaces you don't want to clean up, such as templates, Atlassian's sample content or old archives. From the next index they aren't read at all: nothing in them is listed, links into them show as “can't check” rather than broken, and links from them don't count.
  • How fresh the reports are: page edits reach the reports within about 15 minutes (on sites over 100,000 pages, within about an hour). The whole site is re-indexed overnight when pages have changed, and at least once a week (once a week on sites over 20,000 pages), which also picks up deleted pages and new files.
  • Build index rebuilds the index now, for example after leaving spaces out or adding them back.

6. Large sites

Link Health is tested on sites with up to 100,000 pages and blog posts. Sites with up to 300,000 work too; the first index can take a day or two because Atlassian limits how much apps can read from Confluence each hour, and the admin page shows how long is left. Above 300,000, the app asks you to leave some spaces out first. If you need it for a larger site, email support@kitefold.app.

7. Common questions

Why does it ask me to “Allow access”?

This is normal and happens once per person. Apps built on Atlassian Forge must ask before they read Confluence on someone's behalf. Link Health needs it so it can check each list and each “Linked from” count as you, which is what keeps pages you can't open out of view. Press Allow access. Until you do, the byline shows just Incoming links, without a number.

What does the index store? Does it copy our pages?

No page text is ever stored. To work out the reports, the index keeps, for each page and blog post: its id, title, space, status and short-link code, and the targets of its links (including link addresses). For files it keeps the id, file name, size and the page it's on. It's kept in Atlassian's storage for the app, inside your Atlassian cloud, and deleted when the app is uninstalled. See the privacy policy.

Why are some links marked “can't check”?

They point into a space the app isn't allowed to read, or one you've left out. The app can't tell whether the page exists, so it doesn't call the link broken. To include a restricted space, give the app permission to view it in the space's permissions.

A page is listed as an orphan, but people can find it

Only links count. Pages reached through the page tree, or listed by macros that build their list automatically (such as Children display or Content by label), have no link pointing to them. Mark them as Not an orphan, or leave out spaces full of them.

I restored a file from the trash, but it's not back on Unused files

Confluence doesn't tell apps when a file is restored, so it reappears after the next full index (overnight, or within a week on sites over 20,000 pages; or press Build index on the Settings tab).

Does it check links to other websites?

No. Checking outside websites would mean sending requests outside Atlassian, which the app never does. It checks links within your Confluence site and links to your old server address.

The app says the licence isn't active

The app needs an active licence on your site (it's free for sites with up to 10 users). A Confluence admin can check or renew it under Apps → Manage apps. The index is kept while the licence is inactive, and rebuilt once it's active again.

What happens if we uninstall the app?

Atlassian deletes what the app has stored on your site, its settings (the old server address, the spaces you left out, the pages you marked Not an orphan and the other choices on its tabs) and its link index, under its retention rules for Forge apps. Nothing in Confluence itself changes. If you install the app again within 21 days, Atlassian can restore the stored data, settings included; after that it starts fresh, builds a new index straight away, and you'll need to enter your settings again. Either way the index is brought up to date on its next run. Updates from us never touch your settings.

8. Contact support

Email support@kitefold.app. We reply within one working day (Monday to Friday, 9:00 to 17:00 UK time), and the same working day for urgent problems: see support hours and response times. It helps to include your Confluence site address, the tab or page you were on, and what you expected and what happened, with a screenshot if you can.

Please don't send passwords or API tokens. We'll never ask for them.