Improve pdf storage and reference management? BibDesk + DEVONthink + Bookends?

I want a better way to store PDF documents that I reference in DEVONthink. I use the PDFs in two ways.

  • As a reference for technical information, such as details for a LaTeX package, programming language reference, etc.
  • As a reference for a citation where I need the details for inclusion in a bibliography.

Only a subset of the PDFs show up as research papers in the typical places, and only a handful have DOIs (digital object identifiers).


I use Calibre for books. It works well for books and pdfs but I end up having to manually type in metadata for PDFs that are not available from a popular source like Amazon.


Bookends does a reasonable job of finding details for popular research papers, but appears to be incapable of reading author, title, etc. directly from the document.


Zotero has a “retrieve metadata” which I suspect works for popular research papers?

Other reference management software

Keep trying other and keep thinking BibDesk + DEVONthink must be the answer.

Monthly service fee tools

I have zero interest in tools that charge a monthly/periodic fee for use. I do not want a ransom tied to my workflow.


When I import a PDF, DEVONthink typically reads the metadata from the PDF. This is great but it does not work well as a source for a bibliography.


There must be a combination that could work well for both storage and bibliography.
I really like BibDesk for building a bibliography. I see threads here for BibDesk and DEVONthink. This seems close? BibDesk does not search metadata for me when that is possible (for popular research papers). So Bookends + BibDesk + DEVONthink? If so how?

Bookends does a great job of reading metadata from many PDFs; it also is very capable of looking up additional metadata by itself online if it has a doi to start with. It is also scriptable which makes it easy to transfer data to/from DT3 or other apps. But scripting often is not necessary because it is highly customizable in terms of its citation formats, including RIS; you are likely to be able to determine an output format for text or CSV output compatible with most other software with which you want Bookends to interface.

I’ve used BibDesk much longer than I have DT (15+ years vs. 3 or so) - it was one of the first apps I downloaded when I switched from Windows XP, and I still use it daily (even though I don’t actually write in LaTeX). Its AppleScript support is top notch. A BD record can link to any URL or file, so it’s easy to integrate with DT.

But yeah, it’s not great at scraping PDFs and finding citations - I work around this with AppleScripts to search Google Scholar etc., and this: