As a Historian, i’m trying to use Obsidian as a Knowledge Discovery theory experiment.
Working with historical documents (scanned, mostly) needs a lot of pre-processing. So, before creating notes, i’m trying to OCR those documents to find Named Entities, and that I decided to turn into single .md files. Each person, place, idea, scientific concept, I turned into a .md so the Graph can be populated and interpreted, to find new narratives using the Journey plugin maybe, or any other possibilities. About articles, books and note-taking, I just go with the flow. I read, if I identify an important entity, I turn it into a .md, after that I try to populate some information about it, to enrich the whole graph view. So, the process of knowledge and narrative discovery is basically a loop.

Hope you liked my way of working with Obsidian as a Scholar. Thanks for this post.

1 Like

Hello again,

I try and do as much in Obsidian as possible. It’s also where I manage my daily tasks, meeting notes, and where I develop lectures and tutorial content for my teaching too. The final version of my work depends on what I’m writing. If it’s a journal article/chapter, I’ll do the final edit in Word. If it’s a Tweet, I typically use the Twitter website or Typefully.

Really the key to my workflow is getting the best bits of what I read into my vault as efficiently as possible and in a way that makes them easily retrievable when I need them. I don’t write my own summaries of what I read before putting the highlights into my vault because this takes a long time. I need to read multiple papers each day in my job (if possible), so spending a whole day or longer on one reference is a poor use of time when I may not even end up using that reference in my own writing. I do summarise and recontextualise what others have written into my own words, but not until I know I want to use their specific points in my outputs.

I’ve used Obsidian for 12 months now, and in that time, I’ve read and processed 235 articles/book chapters. My outputs have increased considerably too, which I put down to the system. In the past, I found it a real chore to read any academic articles. In fact, I’d only ever do so as I was writing because I needed to find a quote to support what I was already arguing. This obviously limited my writing to my own ideas, whereas now, I’m reading, learning, and thinking every day, I’m reading widely in my own field and into other fields, and all these practices are improving my writing.

There are many PKM workflows you can try with a tool like Obsidian, and different workflows suit different people and needs. If anyone wants to increase the amount of reading they do and have more to write about when they need it, the workflow I’ve developed seems like a good way to go.

Good luck!


Shhhh. Don’t reveal our secrets…

More seriously: how long did it take you to kick this habit? I feel like I’m constantly playing catch up with the deadlines/projects I already have, so I don’t have the freedom for this kind of “freedom.” But maybe that’s just PhD life (plus being a new parent) and I’ll be able to shift gears in the future…


Haha! Your joke is both funny and sad at the same time.

I’ve been an academic since 2014 (PhD conferred in 2015), and things only really changed for me when I started using Obsidian around a year ago. The key wasn’t really using Obsidian, it was in changing my mindset from ‘publish or perish’ to a sense of just wanting to learn and grapple with the ideas in my field that I find most interesting/pressing. I had to be able to trust that focusing almost exclusively on thinking and learning (as serviced by reading and writing), all the publications, grant applications, and other academic activies would be addressed as byproducts. In my mind, my identity as an academic is now to be a prolific reader, writer, thinker, and learner, and all the other stuff is just the fruit of these core practices. Obsidian (and similar tools) help with this since you know that you only need to read a paper/book/whatever once. It’s funny to admit this since I really wasn’t motivated to read in the past, but now I literally go to bed asking myself if I can fit in reading a little more of the latest paper before I go to sleep. As I sit there watching a movie with my wife (after the kids are asleep), I’m also reading a paper on the side, chuckling occasionally so she thinks I’m fully engaged in the movie. If I didn’t know that all the good bits of what I read are stored safely and easily accessed in my Obsidian vault for all eternity, I wouldn’t be motivated to read, since I know I’d just forget it all in a heartbeat. It’d be too much work. But by largely forgetting about the outputs you’re working on, and focusing on just learning about your field, it changes not only how much you get done but who you are as an academic/PhD student. Or, it has for me at least!

A book that helped me realise this stuff is ‘Essentialism: The disciplined pursuit of less’ by Greg McKeown. It’s not really about academia at all, but it helped me trim away the stuff in my life that was taking up my time in unhelpful ways. When there’s not as much ‘fun’ or ‘distracting’ stuff to do, your genuine interest in your field and the knowledge that whatever you read is ‘safe’ in your vault will prompt you to pick up the next paper. In a nutshell, this mindset to be a prolific reader, writer, thinker, and learner, coupled with Obsidian, has allowed me to feel happy about being an academic for the first time. I’m not stressed at all - I’m just focused on learning. This is a pretty positive step for someone who literally burnt out in the first few years as an academic.

Good luck with the PhD and the much more important task of being a parent :+1:


What a wonderfully helpful and kind response. Thank you! You just pushed me over the edge, in a good way. :slight_smile:

I’m going to work to finish my book project in Obsidian and then export the near finished product to Word or Scrivener for final editing.

Thanks again for taking time out of your busy schedule to respond in such a thorough and helpful way.

Blessings to you.


I also enjoyed this post - thanks for your comments on this thread (as another academic using Obsidian, I find it very encouraging to know that others are finding it as useful as I am). If you do have a moment, I’d be very interested to hear how you gain an overview of your notes (as I understand your workflow, you don’t have ‘atomic zettals’). This is where I have found Obsidian quite useful - having markdown files with relatively long file names (say Professor Smith made this comment about this text) - and then using graph view/the side bar to replicate the experience one would have with spreading notecards across a desktop. Do you mainly use your tags when searching through your notes and constructing outputs? (Thanks again! - And, just to say, it is wonderful to hear from an academic really enjoying their work for once!)

1 Like

Hi Ryan,

Thanks for your reply. Hmm, great question. I’d say I don’t really use atomic notes, as such. I have my highlighted notes, which are long notes with highlights from books with tags added to each quote to suit the quote’s topic (e.g., writing, assessment, phonics, etc.) and rhetorical purpose (e.g., definition, example, solution, etc.), and then I trust the search functionality to find each quote when I need it. I initially followed the How to take smart notes approach, with fleeting, literature and permanent notes, but once I could search for a specific quote so easily, I figured it was adding an extra step to take the quote and put it in an atomic note. I also initially had an index note with a similar set up to what you explained (descriptions of the quote/note), which let me see quickly what I had on different topics, but I haven’t missed it in the months that I’ve stopped using it and not having to write the descriptions has saved a lot of time. In a way, my current approach has the potential to limit my vault from being a ‘conversation partner,’ since I now have to know what I’m looking for before searching for it. But I usually know what I need next, so it hasn’t really been an issue. In a way, the use of tags and searching allows each quote to operate a bit like an atomic note, even though it’s just there, in the long highlight note with all the others.

I would be incredibly happy if someone developed a plugin that allowed the tag system to work like it does in Roam (filtering the tag list based on what tag you select to only show other tags that are used on the same lines/blocks - the more tags you select, the shorter the list of tags becomes). I’d probably pay money for a feature like that, since it ‘would’ allow for conversations with my vault, revealing connections I never knew were there.

Anyway, yes, I try and get multiple papers into my vault each day, filled with highlights and tags, then, when I’m ready to write about a topic, I create what I call an ‘ideas note’ where I dump and reorder all the direct quotes I need for an output, then I create what I call a ‘crucible note’ which is where I summarise and recontextualise the quotes in the associated ideas note into my own words and mix in original thoughts. So, all the novel thinking and connecting still happens, but not until I’m actually ready to write an output. Initially I was summarising and putting quotes into my own words after highlighting every source, which was very time consuming and frankly unsustainable since I may not even need it for an output. It took so long to process a single article, so while it did help me to think in new and interesting ways, it didn’t really keep up with my workflow as an academic. My current approach is much more streamlined and I’m sure wouldn’t appeal to many people using Obsidian because it doesn’t really rely on many of Obsidian’s key features. But it’s helped me to become a more prolific reader, writer, thinker and learner, so that’s good enough for me.

All the very best with everything. I’d love to hear about your approach to this PKM work. How brilliant that you’ve found this while working on your PhD. I feel like all the reading I did for my PhD back in 2021-2014 is largely forgotten and wasted now. Oh well!



I do things similarly to @Dpthomas87. (I’m in philosophy, but read a lot of old texts, medieval manuscripts etc., so lots of pre-processing there too.) I keep my references organized in BibDesk (which I can really, really recommend if you’re on the Mac), including my pdfs.
When I read and highlight a pdf article or book, I export the highlights to markdown, which goes into Obsidian. I name my lit notes by their citekey, and use tags pretty much the way Damon described it.
If I come across something that I think is very important, or just very well put, I tend to create a separate note for it (and link it to the main lit note). All these “snippets” get their yaml field with the author and the parent note, so that I can parse them later.

When I write, I have an “ideas” note as well, where I collect all kinds of things that come to mind about the project, as well as link to all the lit that may be relevant.
Nowadays I do my first draft writing in Obsidian and then export to latex for final editing and compiling.

Good luck with the PhD!


Amen and a thousand upvotes. I left Roam for the Obsidian beta at least 18 months ago and this is the main feature I still pine for…I love writing in Obsidian but searching for and resurfacing information is still a significant concern. This adds a LOT of friction to my workflow. Using search in Obsidian is high friction as well (as it was in Roam). For my use cases this (ease/certainty of resurfacing) is by far the biggest remaining issue with Obsidian.


Thanks for all this - absolutely fascinating.

Might experiment a bit with my workflow (basically at the moment my Kindle highlights are copied into Obsidian atomic note by atomic with tags which I then use on the graph view to get an overview of a topic and start building an outline).

Experiencing a lot of friction creating a new note for even quite small pieces of information (which I already probably have from other scholars in the vault anyway) and adding similar set of tags again and again (though like you I find rhetorical purpose tags exceptionally useful). Something nice too about having the atomic note title to force you to get to the point and think ‘how will I actually use this in future’. But it is very time consuming, especially when a text is very dense and complex (just been going through an edition of a primary source where I was making multiple atomic notes even from just a single footnote). I’ve seen a new plugin which may help provide best of both worlds - going through one document of highlights, adding tags and headings, and then running the plugin to split the highlights into those atomic notes.

Glad to hear how well everything is going though - it is always a pleasure to read and hear from people who are becoming more productive (and, crucially, enjoying their work at same time).

Take care,


1 Like

@Dpthomas87 not sure if you are aware, but there is a discord role to make these sort of requests: @dev for hire. Hope that helps!

1 Like

Hi and thanks for the interesting conversation. After I bit of struggling I’m currently using the following process and convention, if it’s of interest to anyone. I’m a researcher in Sociology.

  1. I read a piece of writing and highlight, take margin notes and sometimes take handwritten notes
  2. Once I formed some thoughts of the writing, I start writing a summary of the work. I don’t like any process that imports all highlights because it is too much information to be useful later. I use my highlights, which I use plenty, to quickly go back through the argumentative structure of the book or essay I had been reading. My summary includes some quotes, which I find particularly useful and/or might want to use in my own writing later on.
  3. Once my text summary is finished, I go through my text summary (which lives in /summaries) and identify arguments and thoughts I think are worth storing and create the actual notes for them my /notes folder.
  4. I use bibtex/jabref to manage references.

Notes have the following strucutre:

  1. A metadata header that looks like this:

cre: 210127
upd: 210519
aut: Bateson, Haake
top: Spiel, Zeichen, Bedeutungsambiguität, Paradox, Pragmatismus, metakommunikativer Marker, inhaltsebene vs. metaebene
ref: bateson2000theory, haake1984paradoxical

The goal is to be as platform agnostic as possible. I currently don’t use hashtags, although hashtags might turn out to be the better solution. The metadata header gives me the option to use the search function similar to hashtags, since they just produce identifiable strings. As such I can, for instance, search for “top:*topicname” and find it.

  1. I curate a note which basically serves as an index. The note has an alphabetical list of topics. The notes that reference a topic are linked behind it. The index system can scale once I have more notes by making the topics themselves pages that are lists of notes that reference the topic.

This way I end up with the following ressources:

  • Highlights and margin notes in the work i have read to guide rereading
  • An abstract summary of the text for reference and to look for quotes or thoughts
  • A box of arguments on various topics that are searchable by metadata

I wrote this some time ago in a different post but I think its worth reiterating: What I think is getting lost about Zettel is that the value of a Zettelkasten type system is in being forced to continously reread your prior notes. The value lies not only in storage and recall but in technologically forced engagement. Only this type of reengagement with the notes provides actual value in my opinion. The two challenges im trying to solve are (1) being platform agnostic and scalable (2) being very specific of what I store (reducing complexity). In the end, as always, the best note is the one you actually use. If theres too much information (or too little actual information) you will get lost, even if you have a super good reference system.

addendum: i also sometimes draw mindmaps in goodreads on my ipad. I just add these images as part of my text summaries.
addnedum2: I’m currently thinking of adding another step to create flash cards with mnemodyne of things I really want to remember. My brain used to be a lot better at recalling stuff and it’s really useful if you can be specific in discussions you really care about.


@bmosbacker – I’m a few months behind in sorting out an academic workflow for my law and policy work. And apparently asking similar questions! Also, currently, taking MacSparky’s new DEVONthink course and playing with Bookends and Obsidian When you have some time to spare, would you mind sharing some information about your academic workflow? Kind regards.

1 Like

Thanks so much. This kind of uncomplicated workflow is key for us in academia.

This resonated a whole lot with me, thank you! As a recently minted PhD I’ve reflected about my identity as an academic. What’s often forgotten in the busyness of academia is exactly what you are getting at here, and what will ultimately make a difference.

The approach described here essentially captures both the Hamming question and how Feynman kept a running list of interesting problems and tested new ideas against them. Thank you!