DevonThink with BibTeX/LaTeX - how do you use it?

As I’m no longer mastering the chaos, I’m trying to store all relevant
information (at least which I have on my PowerBook) in DevonNote:
Interesting websites, blogs, forum posts, personal notes, etc.

I’ve been reading a lot about the powerful AI features of the Devon
apps and I would like to use it also for my dissertation (I’m currently reviewing literature and starting writing at this time), so it
would be interesting to be able to have all the pdf articles in my
database and use the “classify” or “see also” features of DevonThink.

I want to write my thesis with LaTeX and will use BibTeX (with JabRef)
for the reference management.
I read that there are some difficulties
synchronizing BibTeX files with DevonThink.

[color=blue]Can anybody here tell me how (good) that really works?
Currently I’ve got several BibTex files with far more than 1000
entries and for many of those entries the corresponding pdf files.
Nearly each day I find a paper I’ve overlooked up to now and add it to
my database and I’m also adding notes to BibTeX entries in JabRef
while reading papers (especially, as I don’t have a mac at work but a
So in fact, my BibTeX files are evolving and devlopping each day and
I’ve read in several discussions here in the forums, that DevonThink can not synchronize
the entries in its database with a bibtex file.

[color=blue]Has that been improved during the last updates?
How exactly does DevonThink Pro handle the BibTeX entries and how/
where does it store the information?
(If it would be a plain text file on the hard disk, maybe
synchronizing could be done directly between the .bib and this text
[color=blue]Does DevonThink also import hyperlinks to local pdf files in the
bibtex field “pdf” (which jabref uses)?

Could there be a way to do the sync via AppleScript?

2nd question: [color=blue]can DevonThink deal with LaTeX markup?
I was thinking about making extended notes about read papers (with
formulas, etc.) in LaTeX or LyX… I suppose DevonThink can not really
read the LaTeX files, but it could import the created pdfs?!

That was a very long comment - thanks to everyone patient enough to read everything or at least have a look at it,

I’m looking forward to your comments


Hi Martin, sorry but I am not an expert on scripts or jabref, but I had a similar problem at the beginning of my dissertation until I had to realize that DT is useful for information management but not (yet) for advanced citation management - features like generating bibliographies and most importantly, cite-while-you-write do not exist and are hard to simulate.

However, if you let every program do what it can do best, you could use both:
Currently, I have a similar setting for dissertation writing, although I use bookends instead of jabref. My workaround:

  • bookends stores only bibliographic information, DT the rest

  • all notes, PDFs etc… go into DTPro (this might involve copying your old notes as well)

  • in jabref (or bookends), assign a unique reference number to every document. In Bookends, I set up a special input field for that purpose and I also note this number on papers itself, if I have a hardcopy. (In my system, first filed paper of today would be 070207-a etc…)

I use this number for in-text citation as well as for any note in DevonThink, that refers to a BibTex entry. I found DTs wiki-links feature very useful with these numbers: Once you have these numbers as file names, you can refer to the documents quickly.
(I recently put a photo of my document windows on flickr, here’s the link.

This might look complicated, but this way I can benefit from DTs search functions, and store all related information in one place. The bibliographic software is just a command-tab away…

I hope this workaround helps a bit although I could not answer your scripts and latex questions. All the best for your dissertation project!


Hi Mark,

thank you for your comment.
I agree with you that DevonThink does not seem to be suited for citing (in the academic way) but for connecting the information.

I think I’ll try a similar approach as you describe:
In BibTeX, a “citekey” is used for inserting a citation of a source in the paper you write, so this is the unique identifier for each paper. I also write it on hardcopies, the only drawback is the risk of having duplicates when using several BibTeX databases in parallel.

I have to try out DevonThink and will give the hyperlinks also a try.
However, I can’t do all the comments in Devonthink alone, as I’m working a lot on my PC at work which is not able to run DevonThink (at least I don’t know yet how I could run OS X on it at the same time with Windows XP).

So my idea was to write comments also containing the citekey (= unique identifier) to Documents in LateX (->PDF) or plain text and to index them with DT.

This way I would be quite independent of editing documents with DT (= Mac only) but could use the intelligent search features of DT while writing or searching for information (at least when I have my Powerbook).

Could there be anything wrong or problematic with this approach?


the pc-mac-gap remains a problem for me, too. When I work in the office, I have access to a bulky PC, at home, my Macbook. Therefore I seldom visit the office whenever I have some writing to do…

If you can tell jabref to store all notes in a particular folder on your PC at work, it should be easy to sync your notes with your powerbook at home, using a USB-stick, for example. Regular backups might be a good idea anyway…

If you do not have a special folder for your notes, it becomes tricky:
I had a quick look on jabref’s user manual and it should be possible to build a custom export filter.
For example, can you tell jabref to export only author-title plus the relevant notes, plus a special delimeter character in between? This way, you’d have a huge file with the titles plus your notes, which you can split automatically into separate files. kalisphoenix was so kind to write a script for this: It can be found here: “split file in multiple files”. After running this script on your export file, you should have multiple files in DTPro containing the bibliographic info with your notes.

As for the PDFs that you store with jabref; if they are not in a special folder, they should be easy to locate with the windows search function (search for *.pdf or was it?). If you dump them into the database, you’d have some sorting to do, but even if they are all in one folder, DT can find them with “see also”.

Good luck,

by the way, if you store your notes in plain text or .rtf on your windows pc, you can index your notes using desktop-search-applications on your PC like google desktop (I heard from my colleagues that there are other programs, which might be better)

Although these are not as intelligent as DT and more like a spotlight-replacement, these apps can be useful if you want to find your notes on a particular topic, when your powerbook is not near.


Hi Mark,
thanks again for your comments!
The notes I take at this time in JabRef are store with the bibliographic data in the bibtex field “review” as plain text.

If I index my bibtex-File (the literature database) with DT Pro, I should also have access to those plain text notes.

As it would be very interesting to note also mathematical formulas and maybe integrate pictures in my notes, I’ll try to take my notes with LaTeX or LyX, then create a pdf document for each paper reviewed and should be able to index those notes also with DT Pro.
(And - as you say, the bibtex file as well as LaTeX files or PDF notes can be read on my PC too, the only thing I’ll be missing is the AI features of DT and all notes taken exclusively in DT, but I think I’ll try to avoid that and write all notes in separate files outside the DT database.


Hi, I have some of the same questions. I am reading for a dissertation in urban studies.

The way I work is that I keep a group for each reading, named as follows:
‘Peterson, The birth of city planning in the United States, 1840-1917 (2003).’
The folders contain notes and long quotes as rich text, sometimes PDFs (esp. for articles), book reviews, other bits. Each folder also has a plain text document named ‘bibtex’ with a full bibtex formatted entry. The readings are organized into themes and subtheme groups.

What I have been doing so far when working on a bibliography is searching for ‘bibtex’, which brings up all the files in my database named bibtex – I then can easily grab the bibtex code for various readings and assemble project specific bibtex files (for papers, chapters, etc).

I am using DevonThink personal. I know that pro and office (as of several months ago) have capacity to handle bibtex within a structured table but I don’t really see how I would benefit from this given my system. I have been wondering if there’s anything new with bibtex handling in the new edition of pro. What I am really after is a way to attach individual records within a structured document to different groups somehow…

Given my system, I suppose one thing to do would be to write an apple script that extracted the content of the all the files named ‘bibtex’ and wrote them to a giant bibtex file. I don’t know how hard this would be – I haven’t messed around because I haven’t upgraded to pro (and don’t have applescript access).

I am definitely interested in streamlining this process further and would appreciate any feedback… I agree that devonthink is not built for this but it’s only a completely surmountable step or two away from being a killer app in this regard for me.