Hi everyone,
I'm trying to determine whether a recent change in Confluence Cloud could be affecting how attachments/images are accessed by third-party tools.
We use MadCap Flare to import release notes from Confluence Cloud. The page content imports successfully, but all images fail to download. The import log reports errors similar to:
Unable to download resource: image-xxxxxx.png
The pages themselves are imported correctly, but no image files are downloaded. The target image folder remains completely empty.
Some additional details:
Confluence Cloud
Managed Atlassian account using SSO
Authentication performed with an Atlassian API token
Multiple users can reproduce the issue
The same process worked previously
The issue was observed recently after upgrading our documentation tool (MadCap Flare 2026)
At the moment I'm trying to understand whether:
Confluence Cloud has recently changed how attachments/images are served or authenticated
There are known limitations for API-token-based access to attachment downloads
Other users have seen similar issues with third-party integrations that import Confluence content and attachments
Has anyone encountered something similar, or is there a recommended way to verify whether attachment download URLs are still accessible through API-token authentication?
Any insights would be appreciated.
Yes, Confluence Cloud has recently changed how attachments and images are authenticated, causing third-party integrations to fail with 401 Unauthorized errors.
Atlassian officially deprecated and removed the legacy path (/wiki/download/attachments/...) platform-wide on Confluence Cloud. Many third-party tools, including MadCap Flare's importer, rely on reading the _links.download (v1) or downloadLink (v2) metadata paths provided by Confluence. However, those metadata paths point directly to the now-removed legacy endpoint, which blocks API token authentication and results in completely empty image folders.
don't forget to click the Accept Solution button
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Thank you, that's very helpful, and it would certainly explain the behaviour we're seeing.
In our case, the page content imports correctly, but all image downloads fail, and the target image folder remains empty. We also reproduced the issue across multiple users and machines, so a platform-level change seems plausible.
Have you come across any workarounds for this?
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello @Giulia Carnovale
Based on the answer from Atlassian, fix is on the way, and they provided a workaround for that.
More details here https://jira.atlassian.com/browse/CONFCLOUD-84433
Best,
Arek 🤠
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.