[Video Tour] (by Beau Haan) Zettelkasten in Roam. Using block references in bulk and keeping all fleeting notes

In fact, the functionality I mentioned is there: you may right click a references concept and click show content to add the backs of concepts. Don’t know how i missed it.
Here’s how it looks.

So the only thing missing are portal backlinks, and it’s all good.

1 Like

bottom portal on the right in your pic, are you creating a portal to a reference of “This is a concept” on purpose? not just a portal to the direct concept?

I thought the point here is to

  • make each Zettel a card with extra structure (literature + fleeting slots + further nested children)
  • make portal/reference to each Zettel (eg “This is a concept”) from anywhere?

Is that the pattern you are talking about? (sorry I didn’t watch the 1 hour video and am not familiar with the terms, also I have never spent time reading a book on how to take notes)

If I’m rereading correctly, It was on purpose, but just for completionism’s sake. I was testing the ability ito show reference content anywhere, since at the time, only references and tags had visible backlink counters, so you would have to use references if you wanted to multi-add the same zettel to related ones and have a way to view all those related zettels from it. May as well test how references in portals with shown content looked. Since then, portal counters have been added, so you can just use portals to put related zettels in the related section of the target zettel.

Mind you, I’ve never done that, preferring to delete fleeing notes and use references for zettels (adding titles to them via concept :: a la evergreen notes), so the video is still your best source on what the results would be, much like “How to Take Smart Notes” is your best source on Zettelkasten in general.

Just to make sure I don’t get the wrong book, it’s “How to Take Smart Notes” by Sönke Ahrens? the guy who said he’s using Roam because it’s “most innovative and lean” ? Tools

What about the Evergreen notes by Andy Matuschak ? Similarities and differences between evergreen note-writing and Zettelkasten

The best place to get all the information you need is

I’m currently trying to implement Haan’s method into Remnote as a student and I’m having trouble with some things.

  1. When moving the permanent note I created to the permanent notes page I make an alias for the tag (concept acting as a tag) “permanent note” or do I create a completely new concept descriptor that acts as my keyword
  2. How would I use portals in Remnote to replicate the inline block view that Haan has? How do I make it show the other notes that I referenced in my relevants notes section under that topic
  3. Whenever I try to copy to reference the entire Remnote flashcard it only shows the title of the page along with the the concept. But I want it to show the concept and the copy of the flashcard?
  4. Does having multiple portals (in the hundreds) of portals on one page slow it down?
  5. Finally are there any current Remnote themes that has a better visual differentiations between references and bolded concepts.

Thank you in advance :slight_smile:

No idea what you’re trying to accomplish with the tags and aliases stuff, there are no queries yet, so multiple tags are pretty useless. You can use any single rem or emoji to serve as a single tag for permanents, you’ll still have to use the outliner headers and portals/references to organise them.

You’ll have to help me out with a screenshot or timecode, given the age of the video (and the age of our discussion with Hannes above, though RemNote portals only grew visually thinner and functionally remain the same). Do have a look at this article for a more recent practical text guide that should be trivial to translate to RemNote.

Assuming you mean it only shows the front and you want both front and back, right click the reference and “show content”. See images.

Portals in general aren’t nearly as slow as they used to be, but I’m not sure about hundreds. If you’re using a single tag for permanents, you can just use that tag’s automatic portal, which will chug a bit, but display them eventually.

AFAIK there are no current RemNote themes that aren’t dark versions of the default, though if you ask in the CSS forum for the change you want I’m sure someone can help you out. Check out this snippet for the concepts bold selector

  1. Thanks. I realized that what I needed was to create new concepts to act as the keywords that Haan highlights in his permanent page
  2. Thanks for the article. But I still can’t figure out a way to do the inline view thing. Is there any alternative to portals other than CSS or something? Maybe another reason is because I don’t have too many notes referenced to my permanent notes page yet.
  3. I’m still having issues with this because when I tried to right click like you said the closest option I found was “show full name” instead of “show full content”. So it sort of gave me a breadcrumb trail instead of the descriptor framework that I was looking for

only shows “full name”

I want what comes up in that little preview to show up together with my concept. When I click on show full name what comes up is

breadcrumb instead of the descriptor framework

I would also like to point out that I’m not trying to reference the original concept from the stub. I want it to link directly back to it’s place in the permanent notes.
4. Finally how would I use the date reference (!!) visible from both my permanent notes and initial dumping ground. I don’t want to make the date the parent block like Haan does. Would the date be guaranteed to show as long as I do all my notes from the daily documents page?

Thanks in advance!

Again, Roam has queries, RemNote does not. Whatever multiple keywords you assign to a permanent note, as long as you are also searching for the permanent note tag, it will show you all the permanents. I presume this is what led you to attempt to manually store all of them in the same page, but that’s a road to nowhere: you will need to create keynotes per topic at some point and if you use regular flashcards, you’ll also want regular headers for them. Alas, creating a keyword index is not viable at the moment due to the above mentioned query limitations.

The only way to include rem children as you want is to use a portal (which will also lead back directly to the permanents notes page, should you choose to continue down that route) and unfold it. The example I gave was using flashcard functionality (concept) to create a front that serves as a title for the note, while its expanded contents are the flashcard back. Titling notes is Matuschak’s Evergreen notes feature rather than Zettelkasten, but I found it integrates easily enough (though you can use aliases for references instead, of course). See the awesome list for links to comparison between Evergreens and Zettelkasten.

If you leave the permanent note in the daily page and only ever portal or reference it elsewhere, yes. However, if you portal, you can’t indent under it and if you reference, you can’t see its children. Do keep in mind that every rem, regardless of where it was made, has its creation date in its /metadata, so you can just move them out of daily documents. Personally, I use a single inbox rem that has no date or a header for the closest topic or a potential parent/child to draft new permanents (using fleeting notes and literature notes from a source in a separate pane), then move and portal them as needed.

More broadly, don’t be tempted into thinking that you can use only connections between individual permanents to organise everything (possibly this is what goes on in the video?..). You will need keynotes with references and/or outliner with portals, so might as well get in the habit of looking for macro similarities between permanents. Hopefully, queries will come soon enough to allow for using tags more meaningfully, but for now get comfortable creating headers and portalling a note under multiple relevant topics.

Hey quick addition. The only way I can see to have the concept and descriptor show when referenced is to reference them individually and just put them into another card. But I’m afraid that will put duplicates in my knowledge base. How does that work?

As a point of terminology, “Concept” refers specifically to a bold flashcard that may be practised front to back or back to front, “Descriptor” to an italic flashcard that may be practised front to back or descriptor to parent rem. Regular rem are just parents and children (grandchildren etc.). Some regular children may be prefixed with a reference to rem with a ~ prefix and a highlight colour that is called a Universal Descriptor (and may be turned into a Descriptor flashcard, but doesn’t have to be). This situation is a result of some historical naming conventions and usage assumptions, so understandably easy to be confused by.

If you need a parent and a child without any extra text, just use a portal and manually hide (ctrl+alt+h) extraneous children. If you’re adding extra text, it should be meaningful enough to distinguish from just bare references. In either case, backlinks are created, so you’ll have the opportunity to manually clean up any duplicates later.

Personally, I use a rule of three: if I’m referencing three previous notes in the same paragraph more than once, there’s probably a way to use them to create a single new permanent that more precisely describes what I mean and reference that instead.

1 Like

Thanks. I think that portals will be really helpful. But I want to know if there is a way to get rid of the stubs page. Like is it possible to reference something so that it takes me to the place where I originally set up the reference instead of to the automatically generated top level rem of the link. Because I already made a page to organize all my class notes normally with headers and I want to link that you my permanent notes page. So I want whatever references I have on the 1st page to link to the 2nd and vice versa. Thanks :slight_smile:

“Stub” is an automatic tag applied to rem that has been created inside a reference or portal and is therefore placed at the top level with no parent.

Any reference should already take you to the original rem, including a portal - just click on the rem’s bullet inside the portal.

Trying to mix regular notes with atomic is dicey - only go down this route if your linear note headers are reasonably source-agnostic, i.e. can be expanded easily and folded meaningfully. Ideally, you’d separate all the ideas from regular notes into individual literature notes, then either turn into permanents, store under permanents as citations or expansions, or create an outliner structure that can support them. In any case, do settle on a way to delineate permanent notes from all other (an emoji is useful due to being a unique tag and immediately visible).

1 Like

Thank you both. I am completely new to RemNote and Smart Notes but I see the potential for learning and would like to incorporate into how I extract knowledge from books.


Have you found a way to replicate Beau Haans Roamkasten into RemNote? I read your thread where you are trying to wrap your head around how to replicate it in RemNote.
Like I said I am new to RemNote and I still need to learn all it’s possibilities and it’s nomenclature. I can follow how Beau Haan is doing his notes up to the Permanent note. But his linking, tagging and crossreferencing with his Zettelkasten is not easy to comprehend.

I also don’t really understand what you wrote about Boxed References and aliases.


I have been trying to implement Beau Haan’s method too and have “finally” made some good progress (or so I believe). Like Harald, I too am confused by:

If only there were a way to combine a block reference with a concept reference (hint hint [Implemented] Rem Reference Aliases).

Can somebody please provide some insights? Examples are always very helpful. Thank you.

Here is a general outline of what I have done. I’m just starting out so this is a work in progress and any comments are very much appreciated:

  • Create an index page containing all my keywords. These are concepts so retain their uniqueness, e.g.,

  • My Relevant Notes (RN) are usually contained in a page linked to a PDF I edit. So for example, if I am reading a book (R- reference) and highlighting sections, I will add the RN to the page L- reference:

  • The RN must then be indexed:

  • Once I write my Literature Notes (LN), I can then create the Permanent Note (PN). I must first add this to my Structure Notes (SN). I consider the SN, at least for the time being, as my Zettelkasten which contains a concrete outline of the notes as they are created, thereby forcing me to think about context throughout these steps (which is so important!):

Most of this is just a translation from the Haan video, however, my solution using concepts as indexes took me a while to figure out and I’m not sure if this the best solution (whatever that may mean). I only have a few PNs at this very early stage so I don’t have links included to other PNs in the RNs. However, so far so good and as I suggest, I’m always open to being enlightened by more experienced folk and users, especially on the use of aliases in this context.


I’ll just go ahead and delete that line about aliases, since at the time they did not exist, and I did not realise concepts had this functionality:

Now, you can either manually create short aliases for long references of blocks of text or explicitly title the block of text a la Evergreen notes and reference that concept title only (expanding if required as in the post linked above).

I’m afraid I cannot provide any more specific advice about replicating Haan’s setup, as I consider van Doorn’s article both superior as a practical guide and closer to my own workflow:

Thanks for your reply @UMNiK, much appreciated. I too find van Doorn’s configuration inspiring, however, I was wondering if you (or others) could please clarify a few elements which I’m finding difficult to comprehend?

  1. In the following image there are two parts: 1. Relevant Notes (RN); and 2. Index tag


How do you (or van Doorn) materialise this index tag? For example, is it a RemNote concept, a tag (i.e., #Some tag), a link (i.e., [[]]), or other?

Or: Could this index tag simply be a link to the keyword in Keyword Index Page? (van Doorn suggests to collect the keywords after adding the key/index tag in the relevant note. If this is the sequence then the keyword would not exist in the Keyword Index page, therefore ruling out this possibility. (Edit: Perhaps another way to say this is: Is van Doorn suggesting the keyword (or index) created in the RN is a top level rem (without a parent) which is then added as a link to the Keyword Index Page? If so, then it must exist as a link in this Keyword Page. I am wondering is it significant that the keyword remains top level and why so? Does it matter?)

  1. When van Doorn suggests to “collect and transfer all your keywords to a keyword index page”, does he mean to add the link to the created index tag (in the relevent note in #1 above)? For example, if the index tag is a RemNote tag, then in the Keyword Index page I will add a link to this tag?

  2. The image below is taken from van Doorn’s Figure 14. Example of concept handle (Popper’s Revenge) with title, relevant notes, and reference note filled out (Credit: Maarten van Doorn) https://roambrain.com/wp-content/uploads/MVDFig14-342x600.png I am confused by the same text appearing at the top level of this note and for the atomic declarative phrase for the Permanent Note. Why is this so? Does the Permanent Note link to this top level rem and therefore back on itself (i.e., the Permanent Note resides within this rem hierarchy) within this top level rem or does this create a separate page (i.e., a separate rem with the text “Collective conflict over what’s …”?


  1. In Figure 30. Structure Note partial example (Credit: Maarten van Doorn) https://roambrain.com/wp-content/uploads/MVDFig30-445x600.png, what is the form and significance of the preceding tags/keywords/links I have higlighted in the graphic?

Because I don’t have a familiarity with Roam, I find it difficult to translate these operators into RemNote. For example, is this two separate tags? Is it a single keyword? What is going on here?:


Many thanks in advance!

I honestly think it’s pointless to get this bogged down in the technicalities of the apps rather than the actual usage of the slip-box, especially when trying to map things one-to-one across apps that aren’t exact clones of each other (though Roam and RemNote are relatively close). Nonetheless, I will give my opinion on what they do, why they do it, and how to do it in RemNote. Needless to say, this is all extrapolation and speculation, and based on hazy memories and hasty re-skimming at that, so feel free to mentally prefix every following sentence with “I believe” or “If I recall correctly”.

Both Haan and van Doorn assume that most permanent notes will be stored at the top level or else in Roam’s daily note, i.e., the location of the actual permanent note within the system does not matter. In order to get to the notes, they both add a keyword to their title, which in RemNote could be done with either a tag or reference, since they would both be shown and link to they same rem. The practical differences are that references are more explicitly visible (tags are liable to get rolled up into a useless number icon if applying multiple), while tags allow for easier filtering if there are multiple tagged notes in the same document (filter menu rather than search portal query). In order to replicate Luhmann’s setup, the keywords themselves are stored in a simple list in a separate location, the index, so that is where all the keyword backlinks will lead. Keynotes are populated with references to permanent notes that are indented using the outliner. Alas, you may not indent under portals in RemNote, so references are your only option there.

Haan decides to keep all his fleeting and literature notes along with the permanent ones as well as to bring in other related permanents close by. Sadly, there is no way to remove notes from search in RemNote, but you can manipulate the search prioritisation somewhat by making your permanent notes into concepts (either by using the title as the front or just raw ctr+alt+c) and leaving the others as plain rem. For related permanents you can either use references, which may be shortened using aliases, or portals, which allow easy access to the children.

Van Doorn functionally adds Universal Descriptors in order to sort his connections between notes/related notes. From what little I know of Roam’s internals, the reason to use a reference prefixed with # rather than a pure [[ is to be able to style just that tag with Custom CSS. Obviously, in RemNote you can just colour the original Universal Descriptor using the highlighter, or, for more advanced appearance changes, tag it with a rem that has text that would apply the snippet (see detailed example).

(the following is my own advice rather than trying to decipher other people’s, which may or may not be of interest)
The main difference (for Zettelkasten purposes, so excluding flashcards) until recently between Roam and RemNote was the existence of queries, thus limiting what could be done with tags, and, by extension, the usability of having hundreds of notes at the top level with no easy ways to grab subsets of them (except for keynotes). For my own workflow this resulted in having to use portals and the outliner to store permanent notes that apply to multiple topics directly under them (along with keynotes and universal descriptors) rather than building a centralised keyword index. I may re-evaluate and shift to a tag approach now that queries are finally in, but there is something nice in being able to easily mix hierarchical flashcards with looser interconnected permanents (though also hampered by the inability to indent under portals). Of course, the different note tiers do have to be clearly distinguished - I prefer using an emoji and making permanents into titled concepts to use them directly as clauses in new notes.

Again, all these details are ultimately up to you and have little bearing on the actual usage: as long as the notes are usable by themselves while also being connected and able to be surveyed, use whichever feature you find most convenient. Maybe you love the outliner, so you go light on keynotes of references, preferring to arrange the notes themselves. Maybe you’re a tag maniac, so you have a few permanent queries that pick only permanent notes with a specific field/certainty threshold keyword. Maybe you enjoy a graph, so you keep a few entrypoint keynotes and open their local graphs to get to all other notes. In every case, the goals are to enable yourself to easily produce a new note if one is obviously missing, quickly and accurately connect two notes that contain connected ideas, and see the rough shape of the arguments without having to reread every word.

Apologies if that isn’t specific or concrete enough, but I truly don’t think you’ll have a better guide than your own usage - after all, it will always ultimately be your own words you’re writing, connecting, or surveying.

1 Like

:slight_smile: I couldn’t agree more. Think I was at the end of my tether yesterday so frustratingly posted my queries. Once again thanks for the considered reply and I’ll take my time to ponder it.

For anybody else seeking insights into the practical how of implementing a zettlekasten, I found this source invaluable (in addition to those mentioned above), Roamkasten - a practical how to guide to optimize Zettelkasten in Roam Research. However, as UMNiK has suggested, the proof is in the pudding.

1 Like

From a light skim, seems like it’s more or less a text version of Haan’s tour, though feel free to open a separate topic if you feel it’s warranted (though it’s a bit old by now). In my experience, tags (or Universal Descriptors) with raw references are not enough for a strong connection - some words are required to specify the nature of the link to be able to easily understand it later. The counter-argument would be that that locks in a particular line of thought rather than leaving a plain link, but since the reference is still there, I find it more valuable to be able to see the exact reason it was made and add or revise it if needed.

I assume you’re aware of the materials in the awesome list, but it’s definitely worth checking out Matuschak’s second video, as it contains making flashcards alongside permanent notes, which is quite relevant to RemNote.