Workflow with remote scanning and dropbox export ?


I just got a Fujitsu scansnap 1500M and enjoying its ease of use of scanning in invoices and letters. Now i’m looking for a good way of managing all the incoming scans and was wondering if I could get devonthink to work for me.

Let me outline my current scanning setup:

The scanner I’ve put next to my central server so both me and my wife can scan in incoming mail without setting up my laptop.

I’ve set up scansnap to simply scan into a dropbox folder (named “incoming”)

I’ve created a script to run when PDF’s showsup in “incoming” to automatically OCR the files.
(I do this so the main scanning goes faster and we don’t have to wait for the scanning to finish)

With this I got a pretty simply scanning setup that allows both me and my wife to scan and read the mail; now I’m looking for how to make it easier to organize instead of having everything in the “incoming” folder.

Thus, is there a way to have DevonThink

  1. automatically read in the pdf’s in “incoming”

  2. tag them with “incoming” (or somehow otherwise identify them as “unread” ?)

  3. automatically classify the PDF’s it is sure it can recognize and put into appropriate folders

  4. Keep PDF’s available in exported form on a dropbox folder layout instead of only in the devonthink directory layout with internal keynames ?

With that in place I can simply scan in the files, have devonthink automatic classify and by keeping the PDFs in separate exported “human” readable form on dropbox I avoid any “locking” issue and I don’t have to have devonthink installed on other machines.


See Folder for Scans as a Watched Folder

If you’re handy with Applescript, you can modify the folder action script mentioned above to add tags and set the unread flag (though DT should normally flag imports as unread)

This one is tricky. Depends on lots of factors. The short answer is, this is scriptable (use the “classify” command in the script), but your results may vary and you might want to consider doing this step manually.

I don’t know what “internal keynames” means. However, with folder actions, what I’m guessing you want (keep the file in Dropbox and in DT) is achievable. BUT —Dropbox does not play well with folder actions. So be careful out there. :unamused:

[size=85](My feeble mind tells me there’s a thread in this forum somewhere in the past six months regarding a law office or some other professional group trying to do what you’re looking to do - but I can’t locate it right now. Maybe sjk - the master searcher - can find it.)[/size]

okey - lets see if I can explain my self better :wink:

The structure i’m planning to layout (at least that is what the classifier seem to be pretty good at recognizing) is something like this:


PDF’s will start by being in the /incoming folder and then I manually or via the classify operation place the files in those folders.

Then I want that structure to be exported to a Dropbox folder.

The attempts I’ve made required me to Export every file every time (does not scale); and DevonThink seems to only arrange files into its own virtual filesystem (i.e. me saying /incoming/fileA.pdf goes to /car/garage/bills/fileA.pdf does not change the location on disk).

Any suggestions on how to keep folders inside DevonThink “live” on the disk ? Either directly or via some incremental export ?

You want DT to classify and sort documents into the right groups, then have that group structure mirrored in Dropbox?

The only way to have DT’s group structure mirror a folder structure in Finder is to index the Finder structure, keep your data in Finder, and synchronize with DT when the Finder folders or documents change. DT doesn’t do “two-way” sync. (I wrote a script to do this in a limited manner, but it won’t serve your purpose.) Your experience with Export not doing incremental updates is the way it’s designed now.

Two-way sync, and incremental export would be feature requests for the feature managers at DTech to decide. In the mean time, in your use case does Dropbox have to be up-to-date at all times?

A shame ;( I thought DevonThink wouldn’t keep my files “hostage” into its own structure.

Seems like the only way devonthink let you structure your data the way you want is if you let it keep it in its own database/directory structure. The “Index” feature seems very limited since every change I seem to do to those files results in devonthink to copy/duplicate the file.

With respect to the question about if the dir layout have to be instant available then no it doesn’t but I would like it to be available “sooner rather than later” to avoid having to battle with possible multiple devonthink instances.


Here are the things I try/do:

Index file and convert to searchable pdf - a copy is created

Open a pdf, split document - a copy is created