Academic non-fiction workflow? (dealing with citations)

Dear Sirs and Madams,

I’ve been reading blogs and forum posts about citation management for the past week or so and many are quite old so I figured I’d post this.

My current project is likely going to be at least 300 pages in A5 with about 5-10 citations per page and possibly an end of chapter bibliography or end of book bibliography. My boss will want a .docx using Word header formatting so they can change formatting on the fly.

I’m new to Scrivener and I wonder about citation formatting. I’m used to using citation management software to keep things solid. I see a lot of mentions of Zotero on this forum, which I’m familiar with. From what I’ve seen, Scrivener compiles and in doing so changes all the formatting to presets that I can set… but what about the citations? I don’t want to have to go through all citations just to italicise and small caps individual things.

I’m mostly interested in finding out other people’s workflows, more specifically where the citation management comes in.

Is the Zotero ODF plugin scannable cite the best way to go?

Kind Regards,



I haven’t used Zotero before so not sure how it works.

I’ve written a couple of academic papers with Scrivener.

I managed to sort out the citation problem you mention by using Papers, which is the only one I found to communicate with Scrivener. It adds a text-tag in scrivener and when you export it to word it formats the bibliography automatically. Fixing the italics, etc has more to do with customising Papers style rather than Scrivener’s part.

In Scrivener, however, you will not be able to see the full reference list at the end of the document. What Papers does is to add something like {ID:444:2015, p.3} and then in Word it translates it to (John:2015, p.3) and adds the full reference.

Hope this helps.


I routinely use Scrivener for academic non-fiction writing that requires a reference manager. I do use Sente. As far as I remember, Scrivener and Sente do work well together. However, I haven’t really used Sente with Scrivener in a while, because in the last few years I got the habit to write whatever I have to write in Scrivener, then I compile it for Word, and then I enter the refs in Word with Sente

The problem is that the OP is on Windows. Sente is Mac only, as is Bookends, my preferred bibliography manager, and Papers too, I presume. On Windows, I have no idea what is available apart from Zotero and Endnote.

There is a thread somewhere in these forums about new approaches to working with Zotero and Scrivener, so a search for “Zotero” should bring it up. But basically, I would think that the safest way would be to get Zotero temporary citations into Scrivener and then scan the compiled file with Zotero, or perhaps there are hooks in Word, or one of OpenOffice/LibreOffice to do the scanning within there.

If Zotero is anything like Bookends, it should have the script to format the citations and bibliography in any of the major formats.

Endnote, can certainly do it all, but I have no experience of that and comments on Endnote are extremely mixed.

Mr X

I have used Scrivener and Bookends (which I highly recommend), for dozens of scientific papers and grant applications. They are fully integrated. With a keystroke you insert citations from Bookends into Scrivener. You then compile the Scrivener project as an RTF (or, if you must, MSWord, or many other formats) and format, with Bookends, both the in-text citations and the list of references cited, using one of the dozens of bibliographic templates provided, or painlessly create your own.

Papers has version for Windows and iPads, too.

Basically, what you all describe is that all these apps enter a tag in Scrivener and then you need to do a formatting in Word or another word processor.

This is not what I would call “integration”. Integration means what Endnote, Refworks and other apps do in Word. You enter the citation from within Word and you see the in-text and end-of-text reference.

It would be great if Scrivener had this feature. I am aware of the concept of content separation from the format but I feel that Scrivener needs to move forward. It’s a great app for writing and IMHO real integration is a must.

Good thing you don’t get to make that determination. In my experience, the more a program tries to integrate with various third party apps, the more bloated and unreliable it gets. I like Scrivener as it is.

I downloaded Bookends and played with it for a little while. But I couldn’t figure out how to import PDFs and get the biblio info, which must be a feature of Bookends, since Sente and ReadCube and other refs manager have it. How do you do it?

:laughing: :laughing:

It’s not me…it’s how needs and software evolve.

marcoiac: you just drag a PDF into bookends, if you drag it over an existing ref it will attach it by default though you can change that; drag it elsewhere and it will bring up the autocomplete, where it extracts the DOI or title and automatically does a Pubmed search, you select the correct ref, it renames the PDF, and completes the Ref. All very smooth and simple. The user manual is also very comprehensive.

j_kour: all my writing is non-fiction will lots of citations. I don’t need or want Bookends, Endnote or others to “integrate”. I want this to happen once and only once the writing is finished. CWYW doesn’t offer any benefits to me, only distractions (and all sorts of buggy incompatibilites regarding Endnote at least collaborating between windows and Mac).

Don Leo: Well Endnote+Word is the Gorilla in the room for Windows, but if i wrote on Windows, I’d probably use Zotero and LibreOffice for my final output. One point i do want to mention is if you want Headings automatically from Scrivener, the cleanest solution IMO is to use Multimarkdown (or Pandoc if you stick with Word), this compiles automatically into a well-formed outlined document, with the scrivener structure complete. No need for fiddling with replacing styles to create a structured document. Pandoc would also compile your references but I’ve no experience of this (using bookends on OS X), so can’t comment how robust it is, but Pandoc is very solid overall so I think it should work well…

I had a quick look at Pandoc. I can see what it does but how do you use it with scrivener? Do you check the title checkboxes in the “Formatting” section before you compile? And then you you export to Mulitmarkdown only?

Thanks in advance for the tip.

j_kour: yes exactly. I follow a standard MMD workflow, with all my text, block quotes, lists equations and so forth in MMD formatting (I do use Richtext presets to make my work look clearer, but it is all lost on compile). I compile to .MD only, and I have a folder action or script that runs both MMD to generate the ODT; and pandoc for Word with the following options:

pandoc --from markdown_mmd+tex_math_dollars+pipe_tables+implicit_figures+subscript+superscript-all_symbols_escapable -o MyDoc.docx

This makes sure Pandoc ingests the MMD formatting as closely as possible and doesn’t mangle my bookends citations. For each compile I thus get an ODT (from MMD) and a DOCX (from Pandoc) automatically. For citations in Pandoc you add in extra command-line flags to point to your bibliography file and your CSL style, it seems very nice and streamlined though I haven’t tried it. I am seriously thinking of switching to Pandoc more completely though as it is far more flexible, and has a more expressive (though complicated) syntax. As for my citations, as I love Bookends and its Formats manager is perfect for my needs, I need to spend more time with CSL styles to know if I should switch to the bibtex style temporary citations.

The next version of Scrivener is going to be making quite a few changes to hopefully make this type of workflow more flexible, though the details are unclear.

marcoiac: The developer of Bookends posted a number of short, informative video tutorials here: In brief, you can simply drag-and-drop a PDF into Bookends and, as long as the PDF has a DOI, Bookends will import the metadata for the PDF. The video tutorials, and the manual, describe several other ways to do this.

Thanks. I thought about that but since I was playing with it in full screen I couldn’t drag and drop and didn’t see a menu option to add the pdf to the library. Maybe I missed that one too. Anyway, I have been using Sente for ages. I looked into Bookends a while ago, and now that you can download it and use it at no cost (with a small library), I may try to see if it’s worth the switch. It’s the time investment that is the most costly aspect of it. But Bookends does look good, for sure

What’s the big advantage of Bookends versus Papers?

When a book ends there’s no more papers.

Thanks nontroppo. That’s really very helpful. I’ve play a bit with MMD and PanDoc and see how it goes.

Bookends is a rock solid application, with an active and highly responsive developer and a very helpful online community. The developer typically responds within minutes or hours to every query, and is very open to (and actively solicits) suggestions for improvements. My experience with Papers can be characterized as buggy and frustrating. I read that Sente, like Papers, suffers from customer support issues.

More importantly, Bookends offers super flexibility in workflow and in formatting references.

Bookends has other excellent features, such as flexible identification of duplicates, various ways to group references, a useful tag cloud feature, Finder-aware tagging of PDFs, notes that link to marked-up PDFs, Spotlight search of PDFs, RegEx searches.

A “highly responsive developer” is not always a proof of a good software. The relatively new “Manuscripts” app that was released last year (and which has been discussed elsewhere in the forum) has “highly responsive developers” that answer queries almost instantly, in an attempt to calm irritated users who have discovered that version 1 of the app is more like version 1 of the gamma-version (not even ready for beta testing)

My question was about the apps, not the developers. Are there any specific things that Bookends can do and Papers 3 can’t? None of the things you mentioned are specific for neither Bookends nor Papers.

My workflow is very simple:

  • I search for new articles using “Search” in Papers, and add new interesting articles to my pdf-collection (and I do this on both my MBP and my iPad Pro), and also add them to my “collections” in Papers, i.e. different topics
  • I read and annotate articles, mainly on my iPad Pro
  • I copy reference and annotations to Scapple, if I am in the process of drawing the basic logic of the subject I am working on
  • I write scientific articles (in Scrivener) and enter either citekeys or complete references to relevant articles in the process
  • I export the finished manuscript from Scrivener to Word, translate the citekeys and create the reference list, if that part wasn’t done during the writing process

How could Bookends simplify this process or make it more efficient?