API to enable plugins like Zotero

Would be a relief, if you establish an API for Scrivener to enable developers to write plugins for it.
I’m using Zotero for my academic writing and depend on it. It offers a word/open office plugin, which works very well.
I would like to move my academic writing to Scrivener as well, as every aspect is better for writing than with Word, but the lack of the Zotero plugin holds me back using your fantastic software.

Zotero would love to create a plugin if you would offer an API for them.
Furthermore, Grammarly or Writing Aid would be on the ship too, I suppose.

It would be fantastic to use Zotero and other plugins right inside Scrivener.
If this would be possible, I could jump entirely over!

Sorry, but there are no plans for any kind of API. We are a very small team - I’m the sole developer of the Mac version - and implementing, documenting and supporting an API just isn’t practical, I’m afraid.

All the best,

That’s really sad.
Scrivener would be a fantatic tool for academics, but without a proper support for a bibliography tool like Zotero it is not a good option.
The work around with Zoteros RTF scan feature is a “pain in the ass”.
Saldy I’m stuck with word/open office again for my academic wiriting, would have to been too great to switch to Scrivener…

Speaking as an academic, I can say that Scrivener is a fantastic tool to work and write with. I have written two monographs using it, and am working on a third. If you search the forums, you will find many threads containing workflows for using Scrivener with various citations managers including, I believe, Zotero. They’re worth exploring, if you can find the time.

To second what Kinsey said, there are plenty of ways to work around the lack of native citation support in Scrivener. Remember it’s the ideas that will get you published, not the citations. You can always sort that out later.

I have tested these workarounds, and they are all ducktape approaches, which are a pain in the ass and often don’t work as expected. I don’t want to lose time by fiddling around with some workarounds to get them to work, what should be an easy and one click plugin approach!

… and to second that, you have to read all the articles, carefully, before citing, and a normal human being can only read a limited number of articles and remember them well enough to really be able to discuss them and one’s own results. So a bit of manual handling of references shouldn’t be what stops your research.

Yes of course. Could also write with a typewriter, because writing too much a day is not good either.
That’s ridiculous.
Several other forum threads are requesting these plugins for reference management, so there are more that do not like to use painful workarounds or to do the whole thing manually.
The manuscript app (manuscriptsapp.com) will become open source, and it supports some reference managers out of the box.
I think the new Scrinever is a great approach, but I don’t like to fiddle around with something, which should be easy going and I do not want to do things manually, which can be automated and have already been automated within other software.

You could not get me to use Word for an extensive academic manuscript if you paid me. People have literally done that: I had a client who was so insistent that I use Word that they bought me a copy. I still wrote the manuscript in Scrivener.

Your workflow is your workflow and your business. But if you’re willing to inflict Word on yourself for the sake of fully integrated citation management, well, let’s just say that you and I approach the writing process very differently.


I don’t like to use Word. That’s the reason I am searching for alternatives. But I don’t want to trade one point for another.
Word is a pain for writing long manuscripts, stories, in fact, everything but a letter.
Thought I have found a solution with Scrivener, but not without this reference support, so I’m moving on.
Love to use Scrivener for my non-academic writing of course.
We will see, where I will end up…

Things are not always that easy to accomplish, technically, with limited resources.

I tested the Manuscripts app in the beginning and really liked the basic idea. I followed the user forum until the developers closed it last year. They worked hard to get the app to work with Papers and other reference handlers. The problem was that simple things like copy-paste didn’t work and there were also other serious bugs in the app, which caused users to post serious complaints on the user forum. After a number of such posts the forum was closed down. And now the developers are making it all open source to enable others to continue with the development. Well, who knows, maybe something good will emerge in the end.

If I was to write a text book alone with several hundred references I would probably appreciate an automated system, but for more normal science papers a certain degree of manual handling of references is not that big a problem.

Sometimes creating a system that does things automatically takes longer than doing the things, but creating the system is of course more fun… :slight_smile:

Yes, that’s true. Finding the perfect system is a never-ending story.
Hopefully, the open sourcing of the Manuscript app will be a happy end. Will test the current release, to see how it works now.
I will see, which software I will use at the end.

Pandoc is definitely not a ductape approach, and generates a fully referenced bibliography automatically from Scrivener.

I write grant applications with hundreds of refs (and scientific papers with 50-100 refs), collaborating with multiple others (including Endnote users) and I really really do not understand what the problem is with using temporary citations. At least for me with Bookends, I can select them and instantly go back to the source in my database or trigger a Pubmed search, easily change output style with one line of metadata etc.

Not to be ornery, but following the DFW method of retyping my entire manuscript (even though I could have just copied and pasted) in the process of turning my PhD into publications was actually incredibly valuable, & a process I recommend to all my grad students. It forced you to reconsider every word and leads to better writing (and less back & forth with editors/reviewers) imho.

I’m a programmer, computer should help humans not make our life a misery. I can understand that they don’t have the resources to support an API (but I think it’s a bad strategy).

But common

. Did you notice the acceleration in science and the society in general? You know what? Because of automation and computer calculation (a big part). Next time, try to do some research without internet and electricity.

I agree that sometimes things are gimmicks and there is a tendency to over-engineering . But asking a citation manager and an editor to play together nicely without having to start a terminal and install pandoc seem reasonable to me.

The answer “it works for me” doesn’t help. If you want to help, write a nice tutorial instead.

Say that, I tried the RTF scan and various methods: they break. If you are careful, you can make it work (kind of), but they do break. Computer programs should not break.

Yes, but it still takes a researcher to read what is published and to use that as a basis for new research, and reading can’t be automated. And getting the references into the manuscript for a scientific article isn’t the hurdle that stops people from publishing.

That’s not the question. Nobody has time for a boring task that can be automatized.

Actually there is a tone of research to automatize reading, classifying papers and try to find good evidence automatically. But that’s another subject. Of course I’m not saying human are out of the equation, I’m just giving another example.

What make me crazy is this “you can do it by hand, it’s not a bid deal, it only takes a couple of hours” attitude. It’s why so many peoples use a computer like a typewriter, because they “just do it by hand”.

… and I am a scientist who hence write scientific articles for a living, using Scrivener and using Papers 3 reference software. I can understand if a programmer sees an opportunity to fix something, but as it stands L&L have clearly stated that fixing this is out of the scope so it won’t be done, so why keep arguing? Especially as there is a whole bunch of scientists in here who occassionally comment on similar requests in the same way that I did. Fixing the references and the reference list is most likely the smallest problem.
One of the reasons why it is becoming even less of a problem is that several of the major publishing houses who produce the scientific journals have realized that fixing the reference list can be automated by themselves as long as the author uses a consistent structure.
So, for PhD students to spend time on this is bad use of their time. They should focus on other things.

I am the author and maintainer of citeproc-js, the citation formatter that runs in Zotero, Mendeley, and numerous other tools, both on the desktop and in the cloud. I am also responsible for the Zotero ODF Scan plugin, and I heartily agree that it is a pain in the ass—but the best we can do at present. I have regular contact—both in my day job as academic supervisor, and on the Web in connection with software support for my “hobby”—with writers using Scrivener for large, heavily cited projects. Relying on post-processing to finalize citations is a bit like base-jumping for authors. The learning curve is steep, the conversion process is a significant source of stress, and I catch a certain amount of the fallout from that.

I would prevail on the Scrivener team not to give up entirely on the prospect of some form of integration with Zotero (and by extension, with my own Juris-M variant that supports legal and multilingual referencing). The Zotero integration layer for LibreOffice and Word goes to great lengths to produce a responsive “just-works” WYSIWYG feel in the target editor. Simpler approaches are possible, and may be appropriate for Scrivener.

Speaking for the subset of users that crave both the flexible and scalable approach to document organization offered by Scrivener and the accuracy and convenience of reliable reference management support, I hope that the team will at least keep this possibility in the roadmap, and spare a few cycles to consider possible approaches to what is a very real problem for many.

The whole Zotero thing has been done to death on a number of threads, and from what I can see consistently stated as outside os scope and resources by Keith and team.

Endless demanding, posturing and holding one’s breath until you go blue is unlikely to change that.