I stumbled into this solution out of desperation and wanted to document what worked for me here in case the same thing happened to someone else. I realized I’d accidentally deleted half of the text on a document, which made it even more complicated than if I’d just deleted the whole file. So it wasn’t in the trash. I went to my backups, and horrifyingly discovered that all of my saved backups were from AFTER this mistake. I was trying to get in the habit of saving more frequently, and it bit me hard this time. (I’ve since increased the cap to keep 25).
Dropbox saved me, and I hadn’t intentionally made any settings to have it do this. Here’s how to look at the full version history of a specific document. I am on a Mac; YMMV on Windows.
Navigate to where your file is saved (mine is Dropbox > Apps > Scrivener
Right-click > Show Package Contents
Files > Data (I viewed it as a list and sorted by date modified, which helped me easily spot the latest)
Open the “content.rtf” files until you find the document in question
When you have the right file, right-click and look down to the Dropbox options. View version history.
I had multiple versions of every individual minute. I went version by version and found the point the text disappeared, and then was able to pick the very last version of my text before I deleted it. Progress saved. I’m so relieved. If you’re in the same boat, hopefully this can be helpful.
I’m glad this worked for you! Losing work is so scary and heart-breaking.
As a “public service announcement”, it’s worth saying that Dropbox, iCloud, OneDrive, etc., are cloud-syncing solutions and should not be considered to be a backup plan. Scrivener’s internal backup routine is helpful, as can be its “snapshot” feature – but the “snapshot” feature is really a form of version control (for undoing rewriting), and the internal backup routine is limited to a certain number of copies (after which it will overwrite old versions), and Murphy’s Law is that it will miss the critical window of the past version that you are desperately searching for.
That said, it is definitely worth having a true backup solution. On a Mac, TimeMachine makes this simple, and restoring an old version is as simple as using the Finder. Personally, I use Arq, which can schedule versioned backups to myriad locations. On my computer, it runs every night and saves a backup (over WiFi) to my NAS server, with daily/weekly/monthly backups (they get thinned automatically over time) extending back over a year. This might be all you need, but there are other good practices to consider.[1][2]
Automation tools like Hazel and Keyboard Maestro can also be used to define a backup routine for files based on some kind of trigger. For example, you might set Hazel to monitor for changes in your Scrivener project, and if it has been changed today, then it would copy it to another folder (your backup folder) with the date appended, e.g., my_project_20250205.scriv. I wouldn’t use this on every single file (although it would be possible) – but if I was dealing with a PhD dissertation or a novel, then this would be my extra backup for sure!
Whole-drive backup programs like CarbonCopyCloner and the venerable SuperDuper are arguably less necessary nowadays than they used to be. Their strength has always been to make “bootable backups” if your entire computer crashes. I’m content to have my projects (and other documents/files) backed up, but in the event of a major computer crash, I’m content to re-install all the applications etc. from scratch. ↩︎
Many people recommend having one local backup (on the same computer), one backup on a different hard drive, and one off-site backup (on a cloud drive or server, or through a service like Backblaze). ↩︎