Au revoir Bookends, bonjour Zotero

The new Zotero app on the iPad isn’t bad for taking notes and since I switched my file syncing for Zotero to WebDAV (making use of my Fastmail space), it’s been rock solid and easy to access on my iPad with no further need for additional plugins or software. I can then export the notes to markdown on the Mac to us in Obsidian.

I’m still on Zotero and since the recent update with the integrated pdf reader/annotator I’m sure it was the right decision.

For anyone who was sitting on the fence w Zotero 5, I’d recommend checking out Zotero 6 it also has an iOS app that works flawlessly for me annotating pdfs.

Note, though, that zotero 6 saves annotations to its database rather than the pdf file. But you can save the annotations to the pdf by “exporting” them there. To add new annotations, you’ll have to go back to database mode and then re-export. That’s probably the biggest issue and I thought it might be a dealbreaker, but so far I’m happy the way it works since I didn’t see any need to use other pdf editors.


Zotero all the way, My breakthrough was using Highlights as my PDF annotation tool on the iPad and I personally save the un-annotated and annotated PDFs in my Zotero system.


Perhaps my use case is unique. I have to chosen to manage about a half dozen separate citation libraries to work across as many different individual projects and team members. Everyone else uses Windows.

I am staying with Bookends. It offers the best all-around support for AppleScript, multiple libraries with their own attachments sub-folders (no other app appears to offer this), iPad annotation toolset (the Highlights app on the iPad is cumbersome by comparison), and integration with DevonThink.

I’ve dropped Papers. It is professional and very well-designed. I could use it for its ability to search for comparable articles. But this is not my primary need, and its limitations in other areas leave me stranded.

I use Mendeley solely to share PDFs with members of my research team.

I just completed a proof of concept test for how my workflow will run going forward on my projects. A team member marked three journal articles in Mendeley for me to review. I exported the PDFs from Mendeley to an > inbox folder at the Finder level. This folder was set as a Watch folder for a certain library in Bookends. When I open that library in Bookends, it imported the PDFs, completed all the meta-content, and moved the references to a Static Group called > to review. I synced only that category to Bookends on my iPad. I pulled up each PDF, annotated for highlights such as #why, #how, #result, and so on. I synced back to Bookends on my computer. I exported the annotated PDFs into a Finder folder > to Mendeley. I replaced the PDFs in Mendeley with the newly annotated files. The most difficult step in this was the last one. Mendeley is not well designed to swap out PDFs with any degree of ease. The biggest frustration in all of this is that Mendeley does not recognize any form of annotation or note from external apps. I tried annotating in Acrobat … Mendeley does not recognize it. Yes, you can see the annotation on the PDF. But you cannot see it in the notes list.

Otherwise, I’ve tried but not taken to Zotero. I get lost trying to set it up properly for citation management. I hear all the great things about Zotfile or the annotation options on the new iPad version. But … setting up Zotero and understanding how to get it to work always seems to cause me more friction compared to working with Bookends. I keep Zotero if only as an option to replace Mendeley for sharing in case one or the other of my team member prefers.



Another possibility is to set up the Zotfile plugin to automatically store files in cloud storage (so the database will just have the links).

Then PDFs can be annotated in any app with access to the cloud storage.

Notes can still be kept in Zotero itself.

That’s what I used to do (linked files) because I didn’t want Zotero manage my pdfs for me (Because I would still open them on various devices with a pdf tool to annotate them). I did this even though that meant that some Zotero features didn’t work because they were reliant on Zotero internal reference ID for each item.

But with version 6 I see no more need for this. Since the internal annotation works fine for me - and even works across devices - I am now letting Zotero handle my pdfs. This means I can no longer use cloudstorage like dropbox, onedrive etc to store and sync my pdfs (because Zotero doesn’t support these) but it does support webdav and because I have plenty of Webdav storage available, that’s where my pdfs now reside. But chances are, that in a year or so, I may have forgotten where they actually are, because from a user perspective they are just “in Zotero”…

I haven’t used it extensively yet, and I usually find some quibbles/bugs, but so far I haven’t.

Yes, you can do that if you want (and that’s what I’ve done for many, many years), but unless you have specific reasons why you want to store your annotations in the pdf file, there is no need for this any more.

1 Like

Yeah, WebDAV storage is what I didn’t have — but I do have 2TB of iCloud storage. :slight_smile:

I’m curious — which Zotero features rely on the internal reference ID? My reference manager needs are fairly simple, so I haven’t encountered those features so far as I know.

Good question. I think one thing was links to zotero items from outside zotero. Like linking to a reference while writing notes in a note taking app like obsidian.

Edit: here is a specific app that doesn’t work with linked attachments: Zowie

1 Like

Got it, thanks! 20 characters…

Tell you the truth I am now back to pencil and paper and a kind of ‘lab book’ system for references: all I really need. I used both Zotero and Bookends in the past and still do for formatting sometimes, LaTeX too which is my real favorite. I was interested in the comments regarding Legal citation here, I feel few lawyers use any of these?
I came to the conclusion with Bookends, which I like and I like the developer who was very helpful to me once, that one is expected really to design one’s own formats, that seemed to me to be the bottom line, unlike, for example EndNote which has built in nearly every citation format I know about when I looked. This is a kind of question out of sheer interest. What is your feeling on this?

1 Like

I wanted to use Bookends on the iPad to read/annotate PDFs, but discovered that tags (not labels) do not stick when you « resync » with MacOS… Anybody experienced this? Or found a workaround?

I index the Bookends attachments folder in DT, and use tags extensively to classify the PDFs & search in DT. So that’s a dealbreaker for me… :frowning:
Well, for the iOS app only. I still use the MacOS app, but I annotate in other apps on the iPad…

Personally, I don’t need much from this type of software. I read all my PDFs on macOS in Skim, for its inline annotations, which allow me to tag text (using a few tricks I developed long ago). For extensive note-taking, I use BBEdit (I hook the PDF to my BBEdit).

I’ve not written much academic stuff since switching to Bookends. What I mainly want that doesn’t seem to be in Bookends, as @ryanjamurphy metioned, is excellent ability to get the full reference from a PDF. So I’m having a look at Zotero given that Zotero 6 is out and seems to be getting good reviews. I’m also considering (finally) switching to LaTeX after all these years.

I recall @ryanjamurphy published a method to get the best of both worlds (zotero and bookends). But I would prefer avoiding going through that loop time and again if I can. From his comments on Zotero Version 6 Released - Software - MPU Talk, I don’t see an indication that he’s made the jump to Zotero.

The key difference with your workflow and mine is that I read all my PDFs on my iPad in either Bookends or PDFExpert (for documents that are not bibliography journals).

Admittedly this can be a hit-or-miss in Bookends. I too would like to see some improvements in this front. The most frustrating limitation is the inability to fill the abstract field even with the DOI.

The decision may be whether you do not mind working with the format that Zotero uses to store the PDF and to store annotations to the PDF. I prefer the approach in all other citation managers where the PDF is not “hidden” in a subfolder. I prefer the standard where annotations are stored within the PDF not stored external. Since you mentioned Skim, you are already familiar with the processing needed using apps that store annotations external to the PDF.

I’d consider this a highly beneficial step when you work in the hard physical sciences or engineering fields. It beats Word for writing any long-form reports, especially when they include mathematical expressions and figures.


1 Like

I have been thoroughly enjoying Bookends’s scriptability. I have many complaints with it, but I’m finding I can script my way around all of them.

(I still use that Zotero → Bookends automation several times a day!)


Thank you, @ryanjamurphy . I’ll stick with Bookends and try your automation. Much appreciated.

1 Like

last time I checked, it didn’t have AppleScript automation to get title and address of selection.

I’m in the process of moving my Bookends library wholly into DEVONthink. It’s experimental but I think it’ll work.

The process is to create a markdown file in DEVONthink for each reference item and then add all the Bookends metadata for that item as custom metadata. The markdown file itself contains any notes. Attachments are simply related as item links, having been already indexed.

I use Pandoc to process my citations, so that side is already taken care of. One benefit of moving everything into DEVONthink is that Pandoc uses CSL as a standard, which has a different and much wider range of metadata types. So I just have a bunch of custom metadata fields set up in DEVONthink, named and structured as per CSL. It’s much neater.

Translating the one data structure into the other is pretty easy, if time-consuming to set up. The challenge will be getting new references into DEVONthink without passing through Bookends or Zotero first. Getting book metadata from the Google Books API is straightforward, as is getting academic article metadata from DOIs via Crossref. I also plan to look into this at some point: GitHub - zotero/translation-server: A Node.js-based server to run Zotero translators


Another lawyer here. I use Zotero, but I don’t use it for giving me Bluebook citations in briefs. There is a Bluebook citation style, but it’s not very accurate in my opinion. I’ve toyed with the idea of creating my own, but it’s too ambitious a project for the moment. I do use Zotero for managing my library of authorities and precedent, and my annotations of them. It’s a gem of a tool.

P.S. Late to this party, I know. But this thread just showed up for me. Maybe because of the recent comment activity.


I’m a European, rather than an American, lawyer, so I typically use Oscola rather than Bluebook. From what I’ve seen though, there is no automated Bluebook solution. It’s just too idiosyncratic.

Oscola has a few problem areas (EU law and cases for example), but it seems much easier to automate.

[Insert rant about us having to deal with citation formats that can’t be fully automated 30 years after word processors became ubiquitous…]

1 Like

This has been working very well for me. Thank you @ryanjamurphy !

1 Like