Storage & data · Amazon S3 source
What your pipeline writes to the bucket shows on the page as it is now
With the Amazon S3 source the build artefacts, nightly exports, generated reports and diagrams that automation puts in object storage render on the Confluence page straight from there. Your readers see the object as it is at the moment they open the page, and nobody downloads and re-attaches it.
You use the object storage you already run, AWS or not
S3 is a protocol as much as a product. The connection speaks to AWS S3, MinIO on your own hardware, Wasabi, Backblaze B2, DigitalOcean Spaces, Cloudflare R2 or any other S3-compatible store through a custom endpoint, with the region and path-style addressing to match. If your infrastructure team already runs object storage, it is almost certainly on that list - so you change nothing.
Your key stays with the admin
An administrator adds the connection once: access key, secret, endpoint, region. The secret is encrypted before it is stored and is never shown again, not even in the admin screen. Everyone else picks the connection by the name it was given and browses to an object in one of its buckets - so publishing from storage never means circulating credentials, and revoking access is one change in one place rather than an audit of every page.
You scope a connection before anyone points at it
A connection carries its own list of buckets, so it reaches only the ones it was set up for, and it can be restricted to chosen Confluence groups. The connection your data team uses does not have to be the one the whole instance can browse, and a page author cannot reach past what the connection was given.
What the other apps leave you re-attaching
None of the 90 competing apps in the homepage comparison lists S3 as a source. Some accept a URL, which reaches an object only once it has been exposed as a public link; the rest leave a person on your team downloading the file and re-attaching it to a page whenever it changes. That is the loop this closes for you: 28 of the 31 macros here read from a bucket directly, so the page is never a copy.
Amazon S3 is one of 13 places Macro Pack reads from
One app, one licence, one connection your admin sets up - and 12 more places a page can pull content from, whether that is a page attachment, a Git repository, an S3 bucket or your database.
See every source compared