Bookends and DevonThink "side by side" with same PDF folder -- risky?

Hi guys -

This post in the Bookends forums suggests that Bookends (unlike Zotero) doesn’t require any changes to your Finder file structure. You can just point it at your existing pdf library and hit “go.” That seems to open the possibility of having both (1) DevonThink (via indexing) and also (2) Bookends “pointed” at the same underlying PDF files. This would then let me deal with bibliography stuff through Bookends and important stuff ( :grin:) through DevonThink, while keeping the same native Finder folder structure.

At first glance, it seems like it’s all upside? But my PDF library is ginormous so I want to be sure I’m thinking about it right before running my PDF library through Bookends. Am I running any risk by having two different database apps drawing off the same constantly refreshed source folders/files in the Finder? Or am I right that the two could basically play happily “side by side” in the same sandbox, so long as they were taking turns?

Any thoughts (especially on risk assessment!) would be most appreciated…



My current bibliography folder has nearly 14k items (327M words / 40 GBs), and it is indexed both by DT3 and BE. So far, so good. I do not see any problems, and I know many people who do the same thing here. The worst that could happen is renaming a file and having DT3 re-index it as a new record, but recently I am seeing that even when I rename a file in BE, DT3 is quick to pick up the change and adjust its corresponding record accordingly. (BE won’t do the same thing though, that is, automatically finding the file after it was renamed).

Hope this helps. If you need anything, just let us know.


I also index my Bookends attachments folder. It is only about 1,500 items, but no real trouble so far. I’ve had things this way off and on for over ten years, I think.


Awesome–thanks so much @bernardo_v and @mbbntu

I bet I will have a couple of setup questions, and will circle back to this thread then. But this feels like a huge load off trying to manage biblio stuff without messing up what really matters–the actual substance of what I have in DevonThink!


One advantage compared to every other citation manager AFAIK is the ability in BE to split a library into robust, stand-alone libraries. You might find some relief to your anxiety by splitting your large library, not only to provide a test case to try before you take the full plunge but also to improve the oversight across possibly disparate research areas and/or to better regulate the eventual need for archival storage.


1 Like

If you are thinking to index the folder in DT, the process is very simple. WE all are doing it. You managing your pdf files inside a parent folder which is managed by BE, and, you index it to DT.

The real challenge is if you want to import the pdf files to DT (while they still being managed by BE). And the answer given by Jon in that reply describes a method for this approach: to import the pdf files into DT, and at the same time manage them with BE. I personally, however, find this approach very difficult to to maintain in the long run.


Yeah, I definitley want to keep indexing my PDFs rather than importing, so this seems like the way to go. I guess the only downside is that I need to rebuild my PDF Library, at least in the sense of indexing all of my PDFs all over again (since to get the PDFs “into” Bookends requires that I feed them into a new folder, it turns out.) Many thanks for the thoughts!

1 Like

I thought Jon showed you how to use your existing folder pdf files here: adding Bookends to an existing PDF folder structure - Sonny Software

I don’t think re-indexing in Devonthink is necessary if you follow his suggestion.



Do you know what, I had misread his response—I thought he’d just misunderstood my point, and was confirming what I had feared. But re-reading I realize that “new library” in his response must mean the bookends database, not a literal folder. Oh this is great news! Thanks for flagging; not sure how I misread before


Following this, and going to try it!

Thanks to all who posted.


Yes, a library=database in BE terminology.
Great that you find solutions.

1 Like

I have combined Bookends and DT for years and as long as you keep all your pdfs in the one folder you define in Bookends and index that folder in DT you will have no problem. When using Bookends to rename the pdfs, DT simply index them again without you even noticing it. This way you can use Bookends for managing all your references and citations, at the same time as you use DT’s incredible search capabilities and AI to find and curate your material. I currently have 3960 references and if I select any one pdf in DT, its “See Also” function is incredibly good at finding what other pdfs that are relevant. I know it is smart text analysis, but it feels like magic :slight_smile:




awesome – thank you!

commencing Operation Bookends Overlay today : )

1 Like

A followup question for @Per @Dellu @Bernardo_V or anyone else who has indexed their Bookends library to their DT database: what’s your workflow for adding new PDFs? I’ll sketch my setup, then flag what seems like a problem.

I’ve got a “PDF Library” folder in the Finder. It’s got nested folders inside it, organized by file type:
Screen Shot 2022-05-24 at 12.57.57 PM

I’ve imported this entire PDF library into Bookends. For now, I’ve recreated the Finer folder structure as a mix of Folders and Groups in Bookends, partly to facilitate mass tagging. (In the future I probably won’t file new Bookends records in any particular Bookends folder; I’ll just tag each one with all relevant subjects and use smart groups as needed–this is one of the huge attractions of Bookends).
Screen Shot 2022-05-24 at 12.58.52 PM

I want to
(1) download new scholarly articles, often multiple articles at a single sitting (possibly using Bookends itself for the search process)
(2) use Bookends to tag the new downloads and store them in Bookends database as new records, and
(3) file the actual PDFs in the right place in my Finder folder structure (which is very important to maintain for DT search purposes, and also for accessing stuff from my iPad, which is Dropbox syned),

And it seems I hit a problem.

  • If I feed the new articles into Bookends first, then filing them in the Finder breaks the link to where they are.
  • If I file them in the Finder first, then I have to remember where they are and go item by item in different folders to make sure everything is run through Bookends, which is inefficient and highly subject to (ahem) user error

I feel like I must be missing something. What’s the best workflow for adding NEW scholarly articles to a DT indexed Finder folder structure, while also as seamlessly as possible importing them to Bookends at the same time?

Many thanks for any thoughts

I’ve always let Bookends take care of pdfs. I have a single folder for attachments, and all pdfs go into that. I don’t bother with creating sub-folders. I find that just creates a lot of work, and I don’t find that it helps me in any way. Search will allow me to find whatever I want in seconds, so I don’t consider that folder hierarchies add much, if any, utility for me.


I agree with @mbbntu . You want to attach your pdfs to your references in Bookends. Bookends can name the files automatically in whatever format you want, so they are easy to tell apart even when stored in the one folder you assign Bookends to use for them. I let Bookends have the folder on iClouds to have all references synced between desktop and iPad/iPhone (Bookends is useful for me across platforms). Having all the pdfs in the same folder also helps DT to keep all files indexed and ready for its search and AI magic. You can arrange your references in folders in Bookends if you like, but that does not mean the pdfs go in different folders.

The many ways to get pdfs into Bookends, and thus also DT through the indexing, should perhaps be better discussed in the Bookends forum. However, I like Bookends “Add/References using quick add” in which you simply paste a doi, isbn or a bunch of other identifiers and Bookends do the rest. It even have the option to add a pdf automatically if available online. Otherwise you just drag your file to the reference. I also let Hazel (another handy app) keep an eye on my downloads folder to identify files with a doi number and then to move pdfs to Bookends’ “ref import” folder after I assign the tag “ref” to the file. Voilà! Bookends does the rest. DEVONThink and Bookends work incredibly well together, and Hazel is a handy little helper until the files are either in DT or indexed in DT. After that, DEVONThink is all you need.


Incidentally, there is a handy Bookends bookmarklet that you can put in the bar of your browser, and if you are on a page (say in PubMed, or similar) for an article, you can just click the bookmarklet and it will open the page in Bookends, import the reference, and even download the PDF, if there is one available. See this page:

1 Like

wow wow wow - just tried that, that’s awesome!!

This is super helpful, thank you! It sounds like DevonThink really isn’t the thing I need to think about for the import stage; that is just a Bookends question and so long as I keep DevonThink pointed at the right place for indexing, there’s no reason to think about it.

I do need to figure out if it’s possible to continue filing in folders; I have too many other workflows dependent on a folder structure that I can navigate to find stuff (outside of Bookends or DevonThink). But that sounds like a truly pure Bookends question, so I"ll head over there now!

(at some point also need to figure out how to get Bookends records to link to the indexed file in DevonThink, and the indexed file in DevonThink to link to the bookends reference so I can jump back and forth between them. but hopefully there’ll be writeups on that over there too.)

Thanks so much guys!

1 Like