Snapshot Everything

I’d love a command that will take a snapshot of every document all at once when I finish a draft.

I find when I go back to edit, I forget to snapshot before I make changes. It would be nice to say “click–here’s a snapshot of this draft” and go in and change everything.

I do make a copy of the full draft, too, but I love the snapshot so much and it would be great to just snapshot the whole thing at once.


That’s not possible, but you can configure Scrivener to do a full backup on each manual save. (Tools -> Options -> Backups in the Windows version; Scrivener -> Preferences -> Backups in the Mac version)

You can also do a full manual backup at any time, using the File -> Backup -> Backup Now command.

One problem with snapshots is that they are contained within the Scrivener project. I helped one user who was creating both a snapshot and a full backup on every manual save. Over time, his project had become extremely large, and the time to create the backup had become long enough to drag Scrivener to a halt. Especially if your manuscript has a lot of images or is otherwise large to begin with, overly aggressive snapshot-taking can really blow up the project size very quickly.

Backups, in contrast, are stored completely separately. This avoids performance issues and also makes them more robust in case the original project is damaged.


You can though select multiple documents in the binder and take a snapshot of them all at once. The key is that you have to select the specific documents; selecting a parent container alone won’t also snapshot all its subitems.

If your draft docs are scattered into a varying folder hierarchy, you might be able to use search to easily forward all the docs you need to select for that snapshot shoot. IF you use keywords or labels in such a way that you could do a keyword/label search that would turn up every single doc that is part of your draft (and nothing else), you are in business, since you can then from the search-results binder select all the things you mean to snapshot.


When viewing the created snapshot later, does one have to again select all the same specific documents?

Snapshots are saved per document, accessible in the inspector for the single focussed document; that is, you can only view snapshots for one document at a time. Creating the snapshots in one go via the multiple selection just makes a regular snapshot for each document in the collection, the same as if you had created them singly. So you wouldn’t have to recreate the identical selection in order to view the snapshots; you’d select the one document within that whose snapshots you wanted to view.

Ah, I see, so it’s not one large snapshot of all the selected documents, it’s just a bunch of individual snapshots. Cool.

One of my favourite features of Scrivener is that, when I do a manual save, it will also take a snapshot of every document that has been changed since the last manual save. I probably manually save once, maybe twice per writing session (a hangover from MS Office use). My understanding is that this decreases the utility of my backups (as I have settings to keep only the most recent 5); is that the case? My presumption, though I’ve never tested it, is that Time Machine has me covered for older backups just in case I ever need them.
As for project bloat, I don’t have any images in my documents and keep all of my research in a separate programme, so I presume I shouldn’t be in any danger there?

I also would really like a “Titled Snapshot All” command without having to “expand all”, then “select all”, then “Take Titled Snapshot of Selected Documents”. Lazy mode needs more work!!! 8)

The best way could be to make a “special snapshot collection”, which contains all the documents at a specific date, and those documents will not be affected by any thing change in the blinder’s documents.

Then, this technic will offer the possibility to compare an entire document between the document in the blinder or with the “snapshot collection”

You should be fine with this arrangement if your Time Machine backups are connected and backed up to every day. I’d make sure the date is part of the backup file name, and that it’s zip-compressed; it just makes managing and retrieving backups that much more organized.

As long as your backups aren’t taking very long, then using the snapshot-on-manual-save feature is perfectly fine. I think the issue is in the contrast between making snapshots of 1 or 2 documents on average every writing session, and creating snapshots of every document in your binder once or twice per writing session; you end up with a lot more duplicates of your document in the later case, whereas the former case isn’t nearly as voluminous.

  • [size=85]Snapshots are copies of the entire document, not just the changes to the previous document, if I understand them correctly.[/size]

Because each snapshot of all your manuscript’s documents duplicates every document (internally, as far as I understand it), doing a “snapshot all” frequently can really bloat your project. It’s not exponential growth (as with the ‘start with a penny, double your money every day for a month’ thought experiment goes), but it’s still pretty substantial over the period most people require to write and edit a book.

But if you want to do it, and want to make the process a little less tedious, then try this:

Using the project search tool (in the toolbar), modify its settings to restrict the search to your manuscript folder (or expand it beyond if you prefer). Also modify the settings to only search in “Text”, rather than synopsis or other metadata which are not included in a snapshot. Then search for a single space character. That will give you every document with any writing in it whatsoever, unless you’re doing something quite avant-garde, like stringing every word in your book together, or writing one word per chapter…

Now save that search result as a collection.

From this point forward, all documents that match your search criteria will appear in that collection whenever you visit it, so just click on one document in the results, select all (CTRL-a or CMD-a), and take a named snapshot. So you get a process that requires two clicks (one on the collection, one on any document in the result), a keyboard shortcut (CTRL-a or CMD-a, depending on platform), and then (potentially) a click in the toolbar, if you add the titled snapshot icon there, or another keyboard shortcut, if you prefer.

It’s much easier in practice than it seems in writing… give it a try, and I think you’ll be pleased with the efficiency that the saved search adds to the process.

Robert: yes the use of a collection saves one step, great thanks. 8)

In relation to the “bloat”, I keep lots of PDFs and graphics in my binder, so extra RTF files will be small fry in relation!!!

I think that is not right. Kinsey’s description of the pref option is correct, it /is/ only snapshotting the changed docs (not the whole lot) with each manual save. Scriv > Prefs > General > “Take snapshots of changed text documents on manual save”

Yes, as gr pointed out, the snapshots are only of the edited documents. Which isn’t to say they might not still build up over time!

Also, for this bit:

This is only true if you also have the preference set to create a backup each time you manually save. The snapshots are created within the project; they aren’t related to your full project backups, of which you’re keeping the most recent five. If you’re only set to create a backup on close, for instance, it doesn’t matter how many times you save within a session.

I meant to assuage Kinsey’s concern about using the existing snapshot-on-manual-saves feature, but I just managed to muddy the waters. Sorry!

Clarification: The existing alternative function of manual saves should be just fine when used sparingly (once or twice per session). Especially if the files you’re working on only contain text.

No worries gr & Robert, I’m perfectly clear. And I have the preferences set to back up on each manual save MM, thanks!

I didn’t read all posts so this may be a repeat comment.

click in the binder to select a single doc
hit the (expand all) button on tool bar
type (Command A) to select all documents in the binder
hit the take snap shot button tool bar.

is this process too difficult?

You recently posted a wish list request asking for the ability to limit the number of snapshots. The conflict between that request and this post should be obvious.

If you would like to preserve the current project state, taking a full backup is a better choice, as it will capture everything, not just the changed files.


I was only submitting a solution and posing a small inquiry about my proposed solution.
There is no correlation between this post and my wish list ideas, however thanks for noticing.