PKM (Part 2): Landscape

If you look to the east, you’ll see the Outliners Forest. It is a magical place where branches and leaves can talk to each other. Trees can do that, too, using the fungal network that entangles their roots. You can hear various bird songs, but what grabs the attention is the screech of the spaced parrots.

The Outliners Forest covers part of the Markdown Valley, yet the valley stretches over a much bigger area. The soil is very fertile there. If you walk around, you can find all kinds of plants. A few knowledge streams go through the forest. The bark of some trees around the streams is chewed out by beavers who build dams further down to keep as much water area to themselves as possible before the streams flow into the river down the Markdown valley.

The river flows fast at first, but it slows down when it leaves the Markdown Valley and enters the Canvas Canyon. Intricate drawings cover the walls of the canyon. If you come closer, you’ll see that the drawings are made by the slime mould. It is a bit like the fungal networks of the Outliner Forest, but the legend has it that the slime mould is so intelligent that it can solve any problem.

White clouds cover the sky above. They resemble fluffy thought bubbles, float gently, and seem deceivingly light. Yet, everybody knows they can carry a lot and expand to cover the sky from end to end.

Weary travelers usually enter the Landscape of Personal Knowledge Management from the south, from the Lonely Elephant Savanna. It is a place where one can reflect and prepare for the wondrous adventure that awaits them in the Outliners Forest, the Markdown Valley, and the Canvas Canyon.

Hello and welcome. This is the second article in the series on Personal Knowledge Management (PKM). The previous part was about the explosion of PKM tools in the period 2020-2022. Yet, it did not mention any tool. This part will go to the other extreme – it will try to list them all. But before that, let’s have an overview of the main options and capabilities of the currently available PKM tools.

This article started with a description of the PKM tools landscape, out in nature, as the word “landscape” would suggest. Now, let’s shift the metaphor. Imagine you enter a PKM Tools shop.

At the PKM Tools Shop

The shopkeeper greets you and asks you with a smile:

How can I help you?

I’m looking for a PKM tool.

You came to the right place. We have seventy models in stock.

That’s a lot. I’ll need your assistance.

My pleasure. What would you like to know?

My first question is, where is the data stored?

Some tools store it locally and treat the user’s computer as the primary location. They call this principle local-first.

And then?

You can keep your content stored locally, but if you want to synchronize it across devices or collaborate, then you need to have it copied somewhere on the cloud.

How can I do that?

You can use a sync service from the same vendor or sync your files using another service.

What do you mean by “files”? Isn’t there a database?

Most PKM tools we offer, especially the local-first ones, work with Markdown files. The benefit is that you can access your files with another PKM tool or any text editor. And speaking of editors, we have PKM tools that work as plug-ins to editors of programming code. They don’t require coding skills to use, but still, mostly developers use them so that they don’t have to leave their main working environment when they want to take notes.

That sounds nice. But how are then the files linked?

There is always some kind of database or index, but it can just be working in the RAM of your computer.

If I have my files synced, how am I sure others cannot access them?

Most tools that we have offer end-to-end encryption. By the way, when it comes to synchronization, those that truly follow local-first principles should allow synchronization between devices just using the local WiiFi or Bluetooth without storing any data on the cloud.

And in terms of where the data is stored, what other options do you have?

There are typical cloud-based PKM tools, but you could tell in a similar way that they are not cloud-only but rather cloud-first. That’s not something you’d read on their websites, though. I’m just trying to help you with the comparison.

Thank you. This implies they can also store the data locally, right?

Exactly, well, most of them anyway.

In terms of storage, are these all the options you have?

These are the most popular options, but we also have others. For example, you can store your data in a secure vault called Solid Pod. There are other decentralized options, like this one on the shelf to the right, that use the so-called content-addressing, where your content can work in a peer-to-peer network.

Peer-to-peer? So, something like torrents, I guess?

It is similar in some ways but very different in others. BitTorrent is something where people share files in a way that you can get them by downloading different parts from different seeders. It is a way of sharing and not a way of publishing data. If all seeders are gone, you can’t get the file anymore. Content addressing of publishing, rather than just sharing. It is an alternative to location addressing used in all file systems and on the web. The identifier of a piece of content is created from the content itself and not from the path to the location or mapped to it in some way.

Does this mean I can have my content synchronized across my devices without using a centralized synchronisation service in the cloud?

Yes, indeed.

Then why are there not more tools like that?

Maybe there will be, but at the moment, that’s not the case.

I’m asking way too many questions, I guess.

Oh, not at all. In fact, I’m pleased that what you wanted to know first was about where and how the data is stored. Many people who come here ask only about the functionalities.

Well, it’s about my personal data.


From what I have read so far, all PKM tools have note-taking features.

Yes, and the trend is to have the notes linked to each other.

What does the user do to link them?

Most tools offer some simple mechanism for linking, usually with double square brackets. You type square brackets, and then you start typing the name of the page. This works like a search and brings you all matching pages. If there is no such page, whatever you put between double square brackets will create a new linked page with that title. Then, a very popular feature is showing backlinks.

You mean the pages that mention the page you look at?

Yes. Some tools offer filtering and searching over the backlinks, sorting options and creating links to pages that contain the title of the current page in the text.

It’s then like a graph of pages.

Indeed but some tools go further, and their nodes are not just the pages but the paragraphs of each page, or more generally called “blocks”. A block can be a paragraph of text, an image, a video and so on. And then, a few of the tools that users tend to classify as “outliners” can nest blocks. Like in any other editor, it’s just indentation, but usually, there are some powerful features around that. We have clients whose notes often have nested blocks to a very deep level, and they find it quite useful.

But if understood right, in the common case, when links are at page level, there is only one type of relationship, “mentions”. If that’s the case, this seems to be very limiting.

I think so, too, but for most clients, that’s sufficient for what they need.

It’s not much different with the block-based PKM tools, is it? They have only “has child”/”has parent” relation types, right?

Yes, that’s the case, but there are exceptions. We have a few PKM tools that have typed relations.

And they call them “relations”?

There is at least one tool from those we currently have where they are called exactly that. In most PKM tools, they are called “properties”. There is even one tool vendor who calls them attributes, which I find strange. We have PKM tools that don’t offer many functionalities around the properties, but with some, you can select the value type. That value can be either an object or a simple literal, where some tools allow you to specify the data type, like integer, string, date, and so on.

Do they follow some common standards for that?

Unfortunately not. With a few exceptions.

So, basically, there is no interoperability between PKM tools?

At the level of schema, not, but still, there is some interoperability between those using Markdown. They do have specifics, yet most follow similar flavours of Markdown and use double square brackets for internal wikilinks.

You mentioned that note-taking is a common feature. Can you tell me more about it?

You can write directly or use templates for specific types of notes. Or the template can come once you define the type. That’s the case only for tools that allow users to define their own schema. You pick the type, and then it shows the properties such type of note should have, depending on what the user has defined.

Are there other ways to create notes?

Oh yes, plenty, depending on the tools. Some have integrations with other tools, like calendar or annotation tools.

And what about images and voice?

It is common that a PKM tool will offer a browser plug-in for clipping and a mobile app for directly capturing pictures, recording voice notes and getting a file or a link shared from another mobile app.

If there are so many ways to write and capture things, then probably too many things get captured. How are they actually used?

It is indeed a problem. Users make a lot of notes that they never open again, but PKM tool-makers came up with a lot of features to counter this tendency.

Like what?

For notes that represent tasks, using due dates and tags helps in resurfacing. For memorising things, there are various spaced repetition systems, from the common flash-card ones to some that are way more sophisticated. And then, some tools offer ways to define an algorithm when, on certain dates, notes, say, having some tags but not having others, will appear on your daily page.

What’s a daily page?

There is a common design pattern implemented with variations by different PKM tools. Some call it a journal, and others call it daily notes pages. It’s when your home page focuses on the current day, and you can scroll down to the previous days.

Is it initially a blank note, the current day page I mean?

If there is no note tagged with that date, yes, it can be blank. However, many users have templates for their daily notes, so there is some organization.

It seems these tools are full of features. They are probably very cluttered with menus and icons.

On the contrary, most support focused work so the users can get what they need without interrupting their workflow.

But how can they bring what they need, say a timestamp?

The common design pattern is an inline menu evoked with a forward slash for popular actions and a command palette for all actions.

When I entered, I saw a few tools on display with colourful maps of linked cards.

Ah, yes, we have a few of those where notes are created on a canvas, then linked, nested, grouped and displayed in different ways. Basically, that’s the third most popular way to create connected notes, the other two being outliner and the regular document style. Some call it long-form. The text-centred PKM tools also offer canvases, and the visual tools offer a focused text experience, but for those tools you saw on display, visual note-taking is what they excel at.

And what about AI features? I guess the hype has not missed the PKM tools.

Not at all. A few tools simply allows to write a prompt directly in the tool and get the response inside the note. Others allow you to ask questions on your notes and even support your workflows.

Asking questions on my notes? But that would mean giving an AI API access to my personal notes.

Unfortunatly yes.

Can’t I run an LLM on my computer?

The contemporary LLMs require a lot of power, way beyond the limits of even the most powerful personal computers. But I’m hoping that this will change and will get PKM tools that are able to do that.

Thank you very much for all that information. There is a lot to digest. I’ll think about my needs and will come again.

You are more than welcome. All tools we have offer some trial period, and we have quite a few that are free, at least until you need some more advanced features and services.

I’ll have that in mind. Thank you once again. Have a great day!

You too.


Hopefully, this visit to the PKM Tools shop was a useful orientation. Now, you may wonder what these 70 tools are. Here’s the list in alphabetical order. Let me know if I’m missing something.

  1. acreom
  2. Affine
  3. Agenda
  4. Agora
  5. Amanote
  6. Amplenote
  7. Anytype
  9. Bear
  10. Capacities
  11. Coda
  12. Codex
  13. Cosma
  14. Craft
  15. Dendron
  16. Dokieli
  17. Drafts
  18. Dynalist
  19. Emanote
  20. Evernote
  21. Fermat Research
  22. Foam
  23. FuseBase
  24. Heptabase
  25. Hulunote
  26. Hypernotes
  27. iA Writer
  28. Joplin
  29. Kanopi
  30. KeepLink
  31. Kinopio
  32. Kosmik
  33. Lattics
  34. Lazy
  35. LinkedDataHub
  36. Logseq
  38. Milanote
  39. MindPlace
  40. myReach
  41. Napkin
  42. Notejoy
  43. nb
  44. neolace
  46. Nota
  47. NotebookLM
  48. NotePlan
  49. Notion
  50. Obsidian
  51. Organizedly
  52. Org-roam
  53. outline
  54. Quivr
  55. Reflect
  56. RemNote
  57. Roam Research
  58. Scrapbox
  59. Scrintal
  60. Silver Bullet
  61. Subcounscious
  62. Supernotes
  63. Synapsen
  64. Tana
  65. Tangent
  66. Thunknotes
  67. TiddlyWiki
  68. TrinPod
  69. Twos
  70. Unigraph
  71. UpNotе
  72. Workflowy
  73. xTiles
  74. Zettlr
  75. Zim
  76. Zoho Notebook


An obvious question is what qualifies as a PKM tool. I haven’t provided any definition or qualification criteria. If you feel the need for that, please suggest one. You may also suggest, based on that, what should be removed or added to the list.

There are tools such as Athens Research and Clover, which have been discontinued and, for that reason, removed from the list. Other tools, like Knovigator, now Treechat, powerful and useful as they may be, specialised in a direction that, although related to PKM, doesn’t seem to be generic enough – fine, here’s one criterion – to qualify as PKM tool. For similar reasons, Heyday and Rewind are not there. There are some promising tools like WikiBonsai, which will probably enter the list, just not yet.

There was a discussion of options and features and a list of PKM tools, but nothing to bring the two together. The intention was not to make a comprehensive analysis or some classification matrix. There are two reasons for that. First, this is beyond the ambition of this write-up as it requires experience or at least in-depth knowledge of all the listed tools. Way too often, we read classifications that reveal good experience with some tools and superficial knowledge of others. Second, the PKM market is evolving fast. At the current stage of development, such a classification will quickly get out of date. A tool may be classified as not having encryption, properties, or canvas, and then in the next release, it does. Yet, there is no doubt about the usefulness of such an analysis, so if there is one item to go under “future work”, it will be this one.

What I can still do instead is provide some examples. A typical local-first tool using Markdown is Obsidian. Another such tool that is also an outliner is Logseq. Dynalist, Roam and Tana are outliners too. Speaking of Tana, it’s the opposite of local-first, it is cloud-only. However, what it excels at is the flexible schema definition, including some advanced semantic features. Other examples of PKM tools providing schema definition, more specifically definitions of objects and relations types, are Capacities and Anytype. Anytype is an example of a decentralized tool using content-addressing. Nodebook, Scrintal and Heptabase are  canvas-centered tools, giving visual interaction a priority over other means of creating content. LinkedDataHub is an example of an interoperable PKM tool using entirely W3C semantic standards. Some support for such standards is also provided by Logseq. Apart from Anytype, other decentralized tools are TrinPod and Dokieli, but they are based on different protocols and technologies. They use Solid.

That was it for this part. The next one will be about the use of knowledge graphs for Personal Knowlege Management.




2 thoughts on “PKM (Part 2): Landscape

  1. Love your work and the writing style. As someone who only discovered you and the idea of PKG within the week, but original interests in augmenting LLM’s with my own PKG. Would love if you could point me in the direction of some content to explore and learn more about. It seems to me that one’s PKG could significantly improve generation quality of various use cases, IE if I ask it to design some domain specific analytical functions it sometimes uses sub-optimal logic, but if I could utilize retrieval augmented generation from my PKG, in theory this would speed up my creative endeavors. Any insights or recommendations would be greatly appreciated.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.