DEVONthink + Obsidian



@IvanPsy: Exactly!

It has been almost a year since I stopped shopping for the newest flashy tools—a mental shift from “knowledge worker” to a “writer”

The PARA system I explained above just fitted to this new me. 2022 has been my most productive year so far.


I agree.
My job is not organizing information: my job is making something from them.

I also agree with you when you say information may get old, and therefore useless overtime, so why bother with storing and organizing?
I’ve been a power user of Evernote for years, I had many thousands of notes there.
Then they raised the price, the App have become more and more stuffed with features, so I started my journey to other Apps.
I was worried about losing the information at first, so I kept my Evernote account (I wasn’t able to export the notes because I had only an iPad).
The funny thing is after 3-4 years I’ve never felt the need to open Evernote to gather some info from it!

It was very enlightening.


Easy to gather ≠ someday useful :slight_smile:

Even in DEVONthink, the ease of getting things in and organizing doesn’t prove the worth of the material being organized. This is why we end up with junk drawers, attics, and garages full of things that you later look at and think, _“Wait… why did I pick up these three random screws in a parking lot?!?” :thinking::upside_down_face:


This is an excellent video, and its remarks apply as much to DevonTHINK as to any other tools. I don’t know much about Zettelkasten, LYT, or PARA, but I am a big believer in KISS.

That is precisely why I use Obsidian. It is stupid simple, but capable of depth where it is required. The interface is about as close to a blank piece of paper as can be achieved on a computer, but it automates rote tasks. It is not on the cloud, so I can work offline. It is not in a browser, so I am not tempted by distractions. It records information in natural language, just like handwriting notes, which is my comfort spot. Its interface is WYSIWYG, but formatting can be introduced through markdown without shifting from keystrokes to mouse navigation, unlike RTF. Formatting is relatively simple. And I have an approach that organizes the material with search, tags, links, and transclusion—in that order. In summary, I use Obsidian to replace the Word and Excel documents I had been storing notes and information in. For most functions DT could suffice, but it would require just a little more effort to get to that point, and expose me to more points of distraction.

Indeed, I find DT tempted me to fill out more metadata (dates, tags, citation data) than was strictly necessary for my project, getting into the weeds on material but never really thinking about it. And even with workspaces set up, I was finding myself distracted by the sheer amount of information I was recalling with searches. Similarly, the “store anything” attitude I had was pushing me to collect too much information in archives. I was both too deep and to broad at once. DT shows its full breadth and depth when you open it for the first time, and I have found I get the most use out of it by paring it down for each task. And don’t get me wrong, I’d be drowning without DT. But even with the AI, I found myself spending more time maintaining the system than necessary, before I found a good structure.

I mean I did some great work with Word and a well organized Finder.

I think I can summarize my thinking with a heuristic I see a lot in design circles, that of a balance between engagement and resistance. That is, the way a structure, proceess, or interface engages my attention and the way it resists or slows down the task at hand. I like matrices, so here is one:

_ engaging resisting
positively flowing rethinking
negatively distracting frustrating

Positive engagement is being caught up in material that is relevant to the task at hand. Negative engagement is something that draws my attention away. Positive resistance forces me to think, and ideally look at a topic from another perspective. Negative resistance means the method is not working. Adopting a process, I try to keep it as simple as possible, organizing and automating only when they save time required for both the rote task and the thinking/brainstorming that might have gone on while I did it. To steal from Marie Kondo, I ask, “does it spark thought?” I am trying to only keep methods that do.

I found this essay on the burden of maintaining a Zettelkasten persuasive.

Where I will play the part of Eris in this discussion is that I believe shallow thinking is critical to doing valuable intellectual work. But of course depth is too, and one of the reasons paper and pencil—and, I would argue Obsidian—is so effective is that the simplicity lets it do both. I use it to help my brain, not replace it.

Let me step back and explain where I am coming from. I think I have a somewhat unusual background for this user group, since I am an architect who has also worked in trade journalism and am now writing a nonfiction book. Those fields really taught me the value of working towards deliverables and also working towards depth from breadth. Finding the right design approach or the right angle on a story requires learning how to think in a shallow way effectively. That is, not in a narrow-minded and rigid way. But eventually you do have to get into rigor. Libel is bad. Building collapse is bad. So you need processes, standards, and tools that let you do both.

I think I can thank a lack of effective shallow thinking for my book. The events that I cover fell between disciplinary cracks. Specialists in urban planning, Black history, and local history had all noticed the story as early as the 1970s but none regarded it as important enough to dig deeper and connect it to the context.

This phenomenon is, to me, the inverse of “shallowness,” which I will dub “thickness.” By that I mean research that is so concerned with depth that it is buried in the mud. Within the set narrow bounds it accepts, it sounds complete, but it is blind to simple things that affect central claims and it can’t demonstrate its relevance. I think much of this stems from the (valid) disciplinary emphasis on starting from literature or theory. In one journal article or book, this focus may not be an issue, but cited through successive layers of articles, it produces papers that think the whole ocean is the marianas trench and the book needs to be about the caribbean. In fact, I have had to double check a lot of archival citations because the interpretations of primary sources are blinded.

I find DevonTHINK and Obsidian both are helpful for covering breadth and moving into depth. Generally, I use Obsidian as a workspace and DTP to organize and re-think. I use separate Obsidian vaults for different projects and one for my day to day. DevonTHINK has also been really valuable for cross-referencing information through searches and linking. What I am working on right now is a chronology of events that my book covers. I find it really hard to keep track of dates, and I am seeing patterns and connections I did not before. It is accomplishing one of the best ways to use shallow thinking: seeing issues from new perspectives.

In Obsidian, I keep notes loose until they need to be deep. I have been putting wikilinks to good use, adoping a structure similar to what @msteffens notes. Links can convey information that is more relational and substantial than tags or search. They significantly reduce the number of tags I need and improves search. I rarely use them inline. That is the interruption case Nicholas Carr calls a “violent footnote.” Instead, I try to put them between paragraphs, because each paragraph is bascially a complete thought. I also use them in templated info sections. It’s particularly useful for the note files I keep for individuals. Relationships are a very shallow form of knowledge that can be very consequential, and it is so easy to describe them quickly with wikilinks in Obsidian. And the relationships are much more flexible than genealogy tools or spreadsheets. (Graph view is neat but I haven’t found it insightful.)

As I go along, I am replicating documents in my file structure and various notes in my Obsidian vault into groups based on the subchapters in the outline that I did. Classifier, tags, and wikilinks help here. I will later use Obsidian or Scrivener to incorporate these into a detailed outline with citations, and eventually a fully developed draft. I could use Zotero, Tropy, or just aliases, but DT really removes a lot of the nonproductive friction to assembling the material I need to write.

So, this is how I am using DevonTHINK and Obsidian together. My goal is to write a book, covering the breadth of context and moving into deeper thought. It is mostly about picking the interface that provides the right balance of engagement and resistance to perform a given task. try to use them only for functions that add value. They both remember for me, but they don’t think for me. I try to be stupid simple.

PS: @dellu the idea of formatting quoted text differently from book text is a great one. Stealing that!


Hey Ivan dont you use backlinks, how do you do that in DT3? they have no backlinks as I know? do they now?

DT3 does track back links.

Open the inspector, click the “document” icon, and go to the Links tab.

That lists incoming and outgoing links. The Mentions tab lists where the document name appears without being a link.

The new node map script is pretty nice, too, a brilliant contribution by @benoit.pointet.

These are not the same links like in obsidian with the brackets. I see how I can set a link to a whole document but not how to set it from word to word. or to set the word als a new doc link ( backlink).

the mentions thing I dont understand too? I see the window but how to set and work with it. is there a video help doc sowhere ?
thank you very much

You don’t really have to do anything for mentions, just mention documents.

For instance, let’s say you have a document called John Doe.

If another document called Weird People includes something like, “I never really liked John Doe, he was always afraid I’d find out his real name.”

The Mentions tab in the Document inspector for the John Doe document will now show that Weird People mentions John Doe.

You create links in documents. DT keeps track of back (incoming) links, and you’re right - they are links to documents, not paragraphs or words within documents.

Links are created with double square brackets, e.g., [[John Doe]], or with different link text, e.g., [Anonymous](John Doe). That would create a link to the John Doe document displaying as the word Anonymous.

Devonthink allows duplicate names, so sometimes you might want to use a more rigorous link. You can right click on a document or group in the navigation list and choose “copy item link.”

That can be used outside of Devonthink, or you can paste that into a document in DT. When you paste into a Markdown document, the source view of the document will show the link as something like:

[John Doe](x-devonthink-item://71D66155-18A9-436D-B780-68DB86081BB3)

Document level linking works for my purposes. Other than that, DT does a better job for me than Obsidian. I’ve already got DT, so it’s my go-to for what others would do in Obsidian.


Obsidian isn’t inventing anything with square bracket links. That’s a WikiLink in DEVONthink. See DEVONthink’s Preferences > WikiLinks and the Documents > Document Linking section of the built-in Help and manual for more information.

1 Like

Thank you for your worthy answer.
I am at the beginning of md or Rrtf etc. to learn what can do what. and DT is so powerful but not so intuitiv and you have to know what can be done how. So ist a learning curve which I want to take now.

I try to find my way only with DT but tools like reflect or obsidian or bear2/panda or iaWriter, show so much easy options. Some of them seem easier to handle ( sorry DT) some of them have only 10% or less of the features of DT, but some see to have some “fancy new” features, which sometime see to make big sense, when you are starting with such work. But sometimes I also find out that most or many features are fancy but you or your future workflow done need them - it like the 12.000Euro spezialiized Mountainbike which you always wanted ,and you got it. but them you find out it work help you with the mountain or even worst, you find out its not what you 100% need.

BUT at the beginning: you dont know What you need. So you try to find the best suite, meaning you take all features like collect them and try to find the tool, what covers all or most. that is really tricky but good for sales…

So here I am at my journey with my workflow tools, testing my 10 year old license of DTProOffice now DT3 and try to find out what it can do and what it can’t.

so for me its hard to find out what it can’t cause things work different in DT3 than in other tools.

Maybe it would help to know what DT is missing compared to obsidian or reflect. e.g. Can DT3 do backlinks to exact words in one file to an other word in another file. or linking from word to word in same or many files?
I dont want to use obsidian but I can not find out what it can do more -to know if I need it or not. I dont want to get used to a tools which I have to swap after months cause I need a function and learn again a new tool from start. so therefore a well done deeeeeeeeep comparison as a chart or table would help, but no company will do this because, nobody wants to show another too as better. so they list is missing the feature a tool can not do but show everything what they can do. and it will miss all things the other software can do better.

and since reviews are paid there is nothing objective out there I think.

all reviews are the same and no-one tells the deep truth.
That is my hard journey now to find out myself.

Or do you folks have a feature list comparison of all in deep features - not only what DT can do but what obsidian or others cad and DT is missing there. not nice for DT but helpful for the user and a longterm win.

For starting your journey of learning DEVONthink, I recommend you go to DEVONtechnologies | Handbooks and Extras and get and read “Take Control of DEVONthink” (a “free” ebook compliments of DEVONtechnologies) and the outstanding “DEVONthink Manual”. Read the former and at minimum skim the latter so that you know what resources are available and perhaps pick up some tidbits.

1 Like

Ditto what @AndreasEbner said.

I was a dedicated user of TheBrain before I discovered DEVONthink, and I have Obsidian installed.

In my use, the single most brilliant feature in DEVONthink is that tags and groups are very close to the same thing. Basically, a tag is a group. When you apply the tag to a document, a replicant appears in the tag. De-tagging deletes the replicant that’s in the tag.

If you haven’t seen replicants, they are like symbolic links. When you replicate a document or group, it appears in two (or many more, as you wish) places in DT and its title becomes red italics to indicate the instances are replicants of each other. You can further replicate any existing replicant to create more.

Edit any one replicant, all instances reflect the update.

As you delete replicants, the document isn’t deleted until all instances are deleted. You can delete the original first, if you want. The replicants are perfectly safe.

Back to tagging, you can develop a subset of a DT database. For instance, selected class notes from a database going back several semesters could be tagged for use writing a final paper. You can set manual ordering in the tag and arrange the documents in a logical flow.

Or, you could have your groups arranged by major subject, with documents in subgroups according to source. You might later realize a taxonomy organized by country of origin and decade is needed.

Tagging lets you form additional hierarchies. View your data from different perspectives as you need.

DT is cool.


Actually, they’re closer to hard links… if hard links still worked like they’re supposed to in macOS :wink:


Hey @AndreasEbner welcome to the club! :wink:

My strategy has changed since I wrote my previous comments.
I was so deep into the rabbit holes of Zettelkasten, Obsidian (mostly its plugins) and the Second Brain myth that I was blind to maybe the only truth: I’m a knowledge worker, but knowledge is not my end goal. My very end goal is making something with this knowledge: an article, a video, a podcast…

So a reflection came to my mind: do I really need such beautiful, structured, interconnected second brain?
As I started to ask myself such question I realized I was spending more time on connecting notes than on produce something from them.

Eventually I made a test: I moved all my MD files from Obsidian to DT, and uninstalled Obsidian.
After many month I still don’t miss it…

Ok my databases are lesser and lesser similar to a Second Brain, but I produce more content.

So I don’t rely on backlinks anymore: I simply use the link to the item when I need a fast reference to a specific note.
e.g.: I work in the educational field, so I have a repository of educational activities.
When I start a new project and I find an educational activity fits the need, I grab the link to the note and insert it to the note related to the project, so that I can jump fast to the activity.

Nothing more…

Short story: I reduced the frills and see if I was able to live without them.

My next test:

I’m used to highlight a PDF and export a summary of the highlights in a md file.
It’s a Zettelkasten inheritance: I used so that I was able to create atomic notes from those files.
I was so used to them that I had a shortcut for this!

But when I started a new project, a training course about a very narrow topic, I went to DT and looked for information inside my databases.
And I realized I was extracting the information from the PDFs and their highlights inside the PDFs, and never touched the summary files.

So my next test is I don’t extract the summary anymore, and see if I can live with only the PDFs and the highlights inside of them.

Do I miss Obsidian?
Sometimes I find myself wishing to have a more structured, beautiful to see knowledge hub, but then I realize this is not my real wish: my real wish here is to have more order in my life.

I’ve discovered the minimal approach is better for me and my life.

Hope it makes sense and helps.


I’m a knowledge worker, but knowledge is not my end goal. My very end goal is making something with this knowledge: an article, a video, a podcast…

Sometimes I find myself wishing to have a more structured, beautiful to see knowledge hub, but then I realize this is not my real wish: my real wish here is to have more order in my life.

I’ve discovered the minimal approach is better for me and my life.

These are very potent and astute realizations!.

Thanks for sharing your evolution, including deprecating applications and processes you tried but came to find weren’t as useful as you thought they’d be. All users in any application would benefit from this kind of thoughtful examination :heart: :slight_smile:

This reinforces our constant perspective that while you should be open to learning new things, ultimately it is YOUR experience that matters. There isn’t one way to use DEVONthink or DEVONthink To Go, etc.. There are many ways, each of which can provide an unknown tip or insight. But in the end, you use the applications in ways that make sense to you!


oh i understand your journey too… back and forth and around…

i am a psychological counsellor and life coach with a real license (!) and work with the DAOIST Philosophy maybe you have heard of Laozi.
I combine western psychology with Daoist life philosophy to help people with stress, decisions, order in life and to find themself and have a solid self-esteem and calm life.

So I teach, I have Q&A´s, I do online courses and counselling / Coaching and always feed my work with Daoist knowledge from the DAO DE JING Verses and their comments.
So therefor I need a solid PKM, or Database, how you call it, to find attachment points to the question of the client and connect it to the wisdom of the verses in the book to show to background of the meaning.

e.g. vers 1

so is very looks like this in English ( I use German language ).

and then I do connections to other verses or tags like “cosmology” “emptiness” “self”… all that can be seen in the vet but is not directly written. so I find quickly the passages of the book where a certain topic is mentioned or can be seen/extracted indirectly. so the real text is one thing. BUT the real work is the indirect connection, the second and third level of meaning. the interpretation , the grouping of topics…all that need a really solid way of attacking it. otherwise you will find nothing then you started a bit.

This book is just 81 short verses but I covers the same topics or even more than a book like the Bible inclusive all connected books. so 1000000000tousands of pages with connections.

This is the real reason I am now a bit obsessed with the working of links, backlinks , references, tags, groups, replicants, and the difference between all of them ( which is not 100% real yet but I work on it hard :wink: ) .

Theses connections are the basis and to money and the product afterwards I will get out of it like @IvanPsy said, the reason you do all of it, is to get something out of it, and produce something new. I do online courses, videos in future and coaching with it. and what else comes in future.

Till now DT3 was like a big tank of storage to put everything in it. And NOW comes the time to connect the dots and get all in order and get something OUT of DT. I love DT but its a big tool and a lot to understand and learn, so it takes you and shakes you, it needs your attention, and it want you to get all in - > then it gives you something.

THEREFORE I need and answer if DT3 can do all what I need - and in which way - and want it can not do, ( what maybe others can do better ) and if I really need it - or there is a workaround which is not too big.

The manuals and also not always so easy to read, cause they are written by someone who has a deep knowledge of the thing already and sometime its hard for such knowing people to get back to their early days where you can not understand what the beginner dont understand. the beginner view to a product is not a minimum view of a professional DT user divided in small pieces and explained. the pro user does things a beginner won’t use or do, because he would be frightened to do complex things in the beginning or dont understand them deep enough to get the point of doing it.

BUT I will fight myself through the manual- I like the saying. RTFM - read the fux*ing manual :wink:

Maybe you folks can help me answer the question with my DAODEJING work, if that can be done good with DT. IN Obsidian the backlinks are easy to jump between the verses and connect dots and tags and everything. I hope that can be done in DT too the same way.

@BLUEFROG will DT add some more of those fancy “PKM” named features in future, like real backlinks (as I understood now the DT are not the same as in obsidian) and putting links to certain words in a file not only the whole doc. cause I need the exact word to be the end of a link…
and getting more features in writing tools like iawriter or bear… ( and pls dont ask me what I am missing in DT: this is the problem that I can not tell what DT can do - the other tools have some sort of feature list on their website, I could not find something like this about DT. - but maybe I overlooked it.

1 Like

thank you again for your knowledge and explanation. that really helps to get use cases which express the use of a tool. which can not be easy understand by just reading a manual. uses cases make live simple and understandable.

Thank you for that!!!

I definitely will have a look at replicants - I dont use them yet. I did not understand it, was too complicated. but your example helped a lot to understand it.

Yes. that is impossible. Reading is a great introduction and I believe essential. But “doing” is the only way to really learn.

Keep it simple. Do not fall into the trap of doing everything, going after the “cool and shiny” things first, etc. And hardly anything you do cannot be changed when you find a better way for you. Use it.


The answer that can be named
is not the answer.


There is no such answer, any more than you asking if Colorado is where you should live. A thousand people will give you 1000 answers. You will never know unless you actually go there and see for yourself.

As I said, DEVONthink already has what you refer to as backlinks, as well as other options.

and getting more features in writing tools like iawriter or bear… ( and pls dont ask me what I am missing in DT: this is the problem that I can not tell what DT can do - the other tools have some sort of feature list on their website, I could not find something like this about DT. - but maybe I overlooked it.

This is impossible to answer. What writing tools specifically?

1 Like