Exporting Scrivener projects to Devonthink

I know that some (many?) use Scrivener and Devonthink in parallel. Your own thoughts in Scrivener, everything else in DT. That wouldn’t work for me. Probably because my own thoughts would fit on a beer mat.

Anyway, exporting folders and documents (or a whole project) from Scrivener to DT is very easy. But how do you make sure that keywords, labels and notes are also included? If I understand correctly, this is not possible. Or is it? Does anyone have experience with this?

I use both Scrivener and DEVONthink a lot, and in cooperation with each other. I don’t perceive a benefit to connect Scrivener’s files into DEVONthink via DEVONthink’s index features. Or using Scrivener’s external folders (which I know nothing about).

Why do you want to do this and for what benefit? What’s your purpose?

I agree that trying to duplicate metadata between both systems is probably too much overhead with not a lot of benefit. Something to consider is, instead of duplicating data into both systems, use linking instead. Both Scrivener and Devonthink have an external URL protocol that can be used to load a database/project and then a specific item within that. Both also have fields and mechanisms for storing and triggering URLs.

Overall it would make more sense to me to refer to a Scrivener resource via such a link from a Devonthink note that might speak of it, and vice versa, than to replicate the data and then use their respective internal linking to refer to it.

1 Like

I’m going to second @rms. What are you trying to achieve?

Thanks rms and AmberV and kewms :smiley:

I was obviously unclear … probably because I don’t want to do what I feel forced to do. To leave Scrivener. Not today, not tomorrow, but I’m getting ready to.

Scrivener is for writers. But I’m not a writer. I’ve always deliberately ignored that, and thus also suppressed the fact that this app wasn’t developed for me. I write a lot, but I don’t compile. What my texts look like doesn’t matter. Whoever pays me is only interested in the content of the texts. To get to this “content”, I talk to people, collect documents/information of all kinds, evaluate them, comment on them and draw conclusions.

I’ve been doing this with Scrivener for maybe 10 years. Organizing my work with Scrivener works perfectly. All my processes are optimized with and for Scrivener. With Keyboard Maestro and BetterTouchTool I have adapted Scrivener to my needs in such a way that it has almost become a different app. In short: I’m really attached to Scrivener :heart_eyes: But my material is growing and growing. I have tried many tricks to compensate for Scrivener’s weaknesses when searching through large projects. So far, I’ve succeeded well. But I’m slowly reaching a limit and have to think about how to continue.

DT is the best alternative. Even if it actually has disadvantages for me. I don’t like the interface and a lot of things don’t work the way I want them to. The only advantage: I can search through many documents quickly because DT can search specifically in certain fields. Text only in the text, keywords only in the keywords and labels only in the labels. And all at the same time, on one line. Unfortunately, Scrivener can’t do exactly that. And probably never will be able to. That’s ok, because most Scrivener users don’t need that.

But it seems that I am trapped in Scrivener. How am I supposed to import the metadata I need for my work into another app? If I can’t do that, I don’t even have to think about another app. And I might even like that. :innocent:

I suppose one immediate question is: does Devonthink have any standard or easy way of importing multi-field data? Does it have a format or convention (like CSV) that would make it so importing such a file not only establishes its content and title, but would set metadata values to the item?

If it does have a file format, or structured text file approach for doing that, then it is likely Scrivener can create it. That example demonstrates a more complex interchange format, using the JSON syntax, but sometimes complex examples can illustrate the spectrum of capabilities you have. If for example CSV is something you could use, well, if Scrivener can make that, it can surely formulate data onto a line with some commas or tabs in between them.

Whatever the case, getting trapped into Scrivener is nearly impossible, unless one imposes that upon themselves as a constraint.

Another approach: the Export feature does have a metadata checkbox, which creates a .txt file alongside the main content file(s). This uses a consistent format that AppleScript (which DT supports) could work upon.

Another alternative might be to split the project into more manageable parts, possibly along topical lines. For any given query, you can use a system-wide search tool to get you to the right project, then Scrivener’s own tools to find specific items. You could augment this by indexing (not importing) your projects in DevonThink.

Edit: FWIW, what you’re doing sure sounds like writing to me.

1 Like

I thought that Scrivener projects (packages) could not be indexed. Can you please explain how to do this?

Somehow yes, that’s probably why I chose Scrivener :slightly_smiling_face:

I just tested, and apparently indexing a Scrivener project shows the project’s QuickLook preview in DevonThink. So it’s searchable, but only at the project level, not down to the individual component documents or metadata.

(Apparently my memory of having done this myself was mistaken.)

Yes, you will be able in DEVONthink to index the Scrivener project files. But they be in a file structure managed by Scrivener and may not mean much. Dunno. Try it. Anything can be indexed into DEVONthink–just is it a good idea or not. If you index be sure to read about indexing in the DEVONthink Manual.

That being said, I wonder if it increases the risk of corrupting the Scrivener project files if you inadvertently make changes to files or folders inside of DEVONthink where you have indexed them. Something to consider.

From what you say above, if it were me …

  • Keep using Scrivener as your authoring/writing tool. Yes, can write with DEVONthink tools, but nothing as powerful as Scrivener.

  • as your writing deliverables are text, pehaps now for each of your Scrivener projects you have collected over the years, compile them into individual *.txt files and then import those final versions into DEVONthink.

  • You’ve probably collected a lot of files that are in Scrivener Reseach folder. For those Scrivener projects that are old and no longer active as writing projects, drag and drop those research files into DEVONthink (thus importing them) … perhaps a named group for these being the same as the Scrivener project.

Me, I’d not try to actively link the files from Scrivener into DEVONthink by indexing. Possible? Yes. Useful? Not sure.

My guess would be no, as DT opens the project in Scrivener by default. And the whole point of DT’s indexing function is that the material remains under control of the source application.

But this was just a quick check. As always, confirm that you have good backups before trying anything potentially damaging.

I am travelling sans laptop and cannot experiment. But my understanding is that Scrivener files are RFT and when indexed in DEVONthink, then DEVONthink will use the system defined app or default DEVONthink editor for RTF files which may or may not be Scrivener. And the folder structure, expressed as groups in DEVONthink, will be something not to touch else corrupt Scrivener. Maybe I am overthinking, and I would like to dig into it but cannot.

Again, I just tested this. DT does not show the internal Scrivener project structure, it just shows the QuickLook preview.

Edit: Now, another alternative would be to use the Sync with External Folder feature in Scrivener, and have DT index the External Folder. I don’t have time to test at the moment, but I would expect that to behave the same way that other External Folder manipulations do. In particular, the synced folder is a flat list that does not preserve Scrivener’s internal hierarchy.

(I’d expect External Folder Sync to be on the slow side if the source project is already big enough to have performance issues.)

1 Like

I’ve stored .scriv projects in a DT library before without any mishaps. It would be a problem on their end if it couldn’t handle packages well (Scrivener is incidentally programmed to handle packaged formats dragged into the Binder). Both present the item in its “native” format, like Finder would, and display the Quick Look data in their respective previews. Neither digs into the contents or handle internal files directly. When you open in external editor, it uses the associated software to open it, and changes go directly into the container package.

(Disclaimer: I haven’t tried this in ages, but unless they messed up something, it should still work. Test with the Tutorial.scriv!)

Many thanks to all. :pray: I will take a closer look at everything. At first glance, I think there is always a functionality missing that is present in Scrivener.

Is there any chance that Scrivener 4 has a better search engine than Scrivener 3? Or maybe the new L&L app? I don’t mean the search with prefixes. I can do that myself. It would be enough for me if many documents could be searched faster.

Not quite what you are looking for but when I want to run complex searches for text in Projects I dump them into #LancsBox (a corpus linguistics tool from Lamcaster University in the UK). #LancsBox: Lancaster University corpus toolbox

Thank you @reepicheep Unfortunately I don’t understand how this works. How can I use it to search my Scrivener projects?

Ah. I now see. Incredible. I didn’t expect that. What I thought would happen is completely different that what is happening. Learning for the new Month of May!