DEVONthink ‚ should I buy it?

Very nice! How often are you doing the Database Archive to ZIP?

I’ve used DT for ~15 years and find it to be an essential piece of software. I’m also an academic; I have a cell biology lab and also teach at an R1 university. I use DT as a data repository - kind of an electronic lab notebook. It accepts all types of files and I collect my notes alongside data in a variety of formats and relevant primary literature as PDFs. I have never had an issue with data corruption.

I’ve done the experiment several times over the years where I moved by documents to the file system (by simply dragging them out to replicate the directory structure) to determine if I really need DT, and I always return after a couple of days.

The responses above lay out many of DT’s capabilities, so I’ll add to those with a few features that I have come to depend on:

  1. Every document in the database is assigned a unique URL link that can be inserted into other apps to allow for interconnectedness.

  2. The OCR and PDF annotation, while not as full-featured as some apps, are pretty great. I typically use Zotero to collect manuscripts and index the PDF files in the Zotero directory in DT. This allows me to use the advanced search capabilities in DT and provides a nice platform to read and highlight using aniPad/Apple pencil that all syncs back to the original PDF file. DT can extract highlighted notations into a text or markdown file. I can also replicate DT files to appropriate project files. DT can also internally linnk files similar to a wiki, which can be very useful.

  3. By far my favorite feature, however, is that I can edit my notes (sometimes text or .md files but mostly RTF) in the editor window and switch between them just by selecting them in the browser pane. The combined file browser/editor is similar to other apps (e.g. IA writer, Drafts) and it allows quick switching between files. without tying to the app (as in the case of the examples I provided). This is what I actually missed most when I tried to step away from DT - the integration of file editing and organization across file types.

  4. DT has a deep tagging system and a parallel “labeling” feature that can be defined by the user but also adds a nice visual cue by color. For example, all active projects in my database has a label of “Active” and shows as green in the file browser while “Completed” are blue, “waiting for” yellow, etc.

You can certainly replicate many of these features in the file system, but when integrated into DT, the whole is greater than the sum of its parts.


Have you had any file corruption issues over the last 15 years?

The second part of this thread How do I move 15 years of email to a safe and accessible archive? Should I even bother? - Tech Support - MPU Talk turned into a discussion of some pros and cons of DEVONthink that a prospective user may wish to consider.


No. I see the occasional error message that a file can’t sync but it usually resolves itself. I do check/repair databases occasionally, but just to be proactive not to fix an issue.

1 Like

Besides some complaining about the “Open-Database-Warning”, I could not see a real valuable discussion on the second part of the linked thread?

But when you depend on a software developer to take into account that they advertise their system as being suitable for archival of critical data, and you FOLLOW ALL THE RECOMMENDED PROCEDURES, to safeguard your data for future usein years or decades in the future if I needed to do my own file by file verification I would have xpeted at least a sentence or two saying that was important.

DT was a total disaster in the face of a critical data loss issue. They ignored, poo poo’ed and otherwise blamed me for their own SW shortcomings. I left afer that. They can be fine for 99% of the users but idf you are in teh1% that gets screwed you will not trust them.

Contrast that situation with the known bug in Obsidian about a very rare potential for a data loss. If you are using sync, make a change in one system and don’t allow at least 2-3 seconds before closing the vault on one systes in some cases you can end up wth a zero length file. As soon as the problem was discovered the devlopers investigated, verified a problem, came up with a workaround and are continuing to investigate soutions. No finger pointing at the "bad users’but a genuaine desire to make sure it didn’t fail and lots of help on how to recover lost data.

Nothing even close to that has ever come from DEVONThink. and I reiterate, I was a very loyal user for over a decade. They left e hanging and I cannot forgive or forget. And I REALLY don’t want anyone else to expericne the disaster their problems cause me personally.

So I will continue to raise the caution flag about DEVONThink until there is a clear, complete and open discussion of their failings and they take responsibility for the hash they made of some long time user data.


On the contrary, I’d say it provides yet another example on what happens when DT (former) users say anything negative. Just as happened in this thread. :wink:


Hey, I will take your criticism to heart. I’m sorry for poking fun at you when you were unhappy with Devon’s support and their forum last year. I should’ve just let karlnyhus complain without saying anything today, too.

1 Like

Firstly, yes you should buy it!

Secondly, you’re not missing anything, except perhaps the versatility of this application.

My uses of DEVONthink over the last few years have been:

  1. Managing a service unit in a university
  2. eLearning course provider
  3. Studying for a MSc
  4. File management for all personal documents

It helped me with my dissertation by (automatically) gathering data from the web and helping to synthesise huge amounts of data. I still use a separate reference manager (currently Zotero) but I index my reference library so that I can include journals in all searches.

I’m not sure what you mean by ‘connected’ as there are many ways to do this (copy and paste item links, wiki-links, tagging, automation features etc.). It’s the type of application that will grow and shrink to meet your needs without getting in the way.

DEVONthink has been rock solid for me and is an essential macOS app. The number of times it has saved my bacon (ok tofu, I’m vegetarian).

:pray: DT

Can’t recommend it enough.


You could say anything negative, if you want, about anything, if there is a base for that.
If you complain about something, and this is just because you did not understood the function and mistake a feature for a bug, than it shouldn’t be criticized if someone is correcting you!

And if you Link within a thread to an other one, where you specially mention

and it in reality turns out, that this second part mostly contains your complaint about a feature, that you are simply not understood, and think it would be a bug, than I think it is legitime to mention also, that this discussion is not as interesting as you had mentioned before.

I am myself not happy with everything on DT, there are some functions (e.g. DT is turning nowadays a “/” within a Tag into a command that is separating the Tag into two, while we need this “/” in Germany as a part of official file numbers), I would be very happy if I could turn them of, or if they would be more useful.
And I had a few discussions myself on the DT-Forum with some Fanboys, and part of the crew, who were be slightly “unhappy”.
But it is not helpful at all, if you do not want to let any discussion going on, if someone is thinking something is a bug, while it is a feature!

And after all, (unfortunately) DT is the only(!) real/useful DMS on the Mac…

As much as I like Devonthink it’s easy for it to become a data dump where you can only find stuff via search.

As years go by this grows and Devonthink works well so you don’t really need to sort it. The challenge is that we really should be thinking about our data and storing it appropriately either with meta data or structure and remove cruft.

At some point you will need to sort it so if you buy Devonthink I would encourage you to still distill your data and create a knowledge base rather than a data dump.


Do you figured out the technical reason behind your data loss?

OK - got it. Looks very uefull!

1 Like

Wow - What a thread!!!

Thanks everybody for your many pros and the one contra.

Concerning the contra:

As always with software, one should not trust it completely to keep data safe - there is nothing like a solid backup strategy:

  1. keep backups onsite and offsite
  2. Keep old backups following a multiple generation model (e.g. keep yearly backups forever, the monthly of the last year, weekly of the last month, daily of the last week)
  3. verify your backups and data for corruption

Most of us (including myself) do 1 (e.g. time machine and backblaze in my case), many 2 (partly - I ave the one year version history for backblaze), and, only a few follow 3.

The problem with 3 is that, that it is a complex task which is a complex task, which likely differs between file formats.

Corruptions can occur through software bugs, race conditions, bitrot, etc. But I think an approach as I will outline could make the chance of data loss much less likey, in regards to DT. I assume, that the most likely culprit for the data loss in DT was something along the synching (maybe soime race conditions? No idea.)

  1. disable automatic sync with iCloud
  2. create a smart rule (as mentioned by others) which checks for empty files, abort if there are any
  3. verify and repair databases
  4. only if all checks result in no errors or warnings, do the sync.

It should be easily possible to implement such a rule and execute it either manually or on quit.
Or - could one leave “automatic sync” on, and only sync if these checks pass?

I agree, that this should be in the software itself, but there is nothing like software without bugs - and if they are very rare and only in specific cases, they are nearly impossible to fix.

So: I will very likely go the DT route and if I loose my data and complai here, please direct me to y post….

Thanks a lot.


Googling around, brought an old friend of mine (from Linux times) to mind: rsnapshot. Based on the rock-solid rsync (as Carbon Copy Cloner is as well) , it makes it easy to implement three generation backups.

… Looks at his 4 Seats.

I am going to say yes. I started with YoJimbo 13 years ago, moved to Evernote 8 years ago, now Devonthink 2 years ago. I think it is too powerful and versatile to ever be replaced.


And here is a manual to install and run rsnapshot on a Mac. It is from end 2017, but the only thing which likely has changed, is the implementation and options of the launchd Agent.

To me, DevonThink is essential and irreplaceable!!!

Two hints;

  1. Buy David Sparks DevonThink Field guide I had been using DT for years when the guide came out and I still learned a ton. To this day, when I want to try something new, I go back to David’s fieldguide and watch a couple of specific videos.

  2. I bought a 1tb SD card card that fits flush to the surface of the card slot from Amazon
    and I have Carbon Copy Cloner back up by DT data files twice a day. Once at lunch time and once at 6 pm. Since these are incremental backup they only take a few minutes.

BTW, this is not my only backup but it’s always current.

Also, the reason I purchased the 1tb is that I use a rolling five-day backup. This way I have five days of slack to become aware of a corrupt file or some other malady. Each Monday I then rewrite last Monday’s backup.

  1. Yes - I will invest in thge field guide - no question about that.

  2. This depends a lot on the settings for CCC - if you have unlimited Safety_Net enabled, then you can go back. If not, it would not help in the case of the undetected corruption discussed above. But I like that idea a lot - a second (in addition to Backblaze) backup on the go!

1 Like