SuperDuper! 4

SuperDuper! 4.0.6 (September 14, 2026)

SuperDuper 4.0.6 is (once again) mostly about disk images: dealing with them when they’re already open; allowing erase; scaling band size to try to keep them fast.

Oh yeah, also “final” support for Golden Gate.¹

Running-only filter

Click the n copies running pill to filter to just those jobs. If you have a lot of copy jobs for a given source, you can filter the list with a click.

Disk images

New disk images should be faster. A sparsebundle stores your backup as thousands of small “band” files, and the size of those bands is fixed when the image is created. The default size doesn’t grow with the image, so a large backup could end up with an enormous number of tiny files—an 8 TB destination could reach nearly a million of them. (Why is this the default when the size is known at create time? You’ll have to ask Apple.) SuperDuper now picks a band size that suits the image’s size, balancing the number of bands with their size.

Note This applies to new images in 4.0.6 and later. If you have a large one and it feels slow to erase, you may want to recreate it. (You can change the band size with hdiutil, but it copies the whole thing over…impractical in most cases.)

An already-mounted image now works if you’ve selected it by filename. Not much else to say about that.

Erase, then copy is available for disk images. Again, not much to say: erase away!

Note I still don’t allow OS replication to an image…there’s basically no point.

Dropping a mounted image’s volume on a job now binds the image file. That’s what you want, so SuperDuper figures it out and does it for you.

Copying a disk image into itself has a better error message. But really, don’t do that, eh?

You can use an image as a source. It gave an error before. It no longer does. Better than that, it works.

Timeouts have been adjusted to deal with large images and slow hosts. A big image on a NAS can take a really long time, so SuperDuper is now exceptionally patient.

Copying

A copy no longer (intermittently) fails on a drive that’s already mounted. macOS mounts a volume the moment it appears, and if SuperDuper asked at the same instant, the system reported a conflict that we read as “not mounted”. No longer!

CoreSpotlight mess less messy. CoreSpotlight folders are very busy on recent macOS versions, and can be in invalid states when copied. SuperDuper is now more tolerant of this kind of temporary corruption.

OS changes to the destination mid-copy are now tolerated. The OS can sometimes whack things SuperDuper is writing, while it’s writing. Those situations are now retried rather than flagged as errors.

Improved copying into cloud folders. FileProvider-managed folders, used as destinations, now tolerate provider EA changes (I’m looking at you, Dropbox).

Password Prompting

Overprompting fixed. There were situations with locked volumes where, even though SuperDuper had successfully unlocked a volume, it re-prompted for the password.

Rules

Deletion improved. When an open rule is deleted, I now tell you how many jobs reference that rule. If you elect to delete the rule, the rule editing window is closed. Any jobs that reference that rule show its name in red and cannot be run until a different rule is selected.

Microtastic speed improvement. A small but meaningful improvement was made to rule evaluation speed. You won’t notice. But isn’t it nice to know it’s 51% faster, even so? More than half a microsecond! Woo!

Integration

There’s now a third-party application, Open Backup, that can notify mobile devices (iPhone and Android) when backups run.

Golden Gate

We’ve supported Golden Gate through its development, and I have to assume we’re running the RC, since it’s coming out today, September 14th.

That means we’re almost fully compatible with Golden Gate.

Or rather, we’re fully compatible, but Apple itself has had a regression.

Drives that have been properly replicated by Apple’s own tool—the only way to do it—are not showing up as startup drive choices.

When this started happening mid-Golden Gate development, we reported it to Apple (FB24210103, Apple friends), and hope they’ll address the issue in a follow-on update.²

Note You can get the copy to boot by reducing boot security and running bless in a Terminal session in Recovery, with —setBoot, but, really, don’t do this unless you really have to.
Really.

Remember: you do not need to be able to to boot from a backup to restore from it.

And when Apple does fix the problem, boot will just start working again, with no need to update SuperDuper.

Requirements

The same as before: macOS 14 or later, including macOS 27 (Golden Gate) on Apple silicon or Intel.

SuperDuper! 4.0.5 (August 26, 2026)

SuperDuper 4.0.5 is a quick fix release to correct some incorrect status that interfered with scheduled execution.

Scheduled copies

A scheduled copy now properly connects your server, even on subsequent runs. If your destination is a share—or a disk image living on one—and it wasn’t mounted at the scheduled time, the schedule could fail its “precheck” without trying to actually mount. This also fixes a problem where a network image catch-up run…wouldn’t run.

Disk image jobs

Similarly, an image job could go dead right after a successful backup. Copy to a disk image on a NAS, watch it finish perfectly, and the job would go amber with “Can’t verify destination” and both buttons greyed out. (Quitting and restarting fixed it.)

Shortcut status

Shortcuts is told which end of the job was unreachable. Now, “Get Job Results” tells you if it’s a source or destination problem.

Requirements

The same as before: macOS 14 or later, on Apple silicon or Intel.