Transcription
Obsidian is an application for writing notes that does it a bit differently. It's an application that emphasizes the connection between note and note, and this means we can write a certain note and then take any word in that note and link it to another note, which eventually develops and develops, and over time we get a graph of notes linked to other notes, to other ideas, to more information that enters the system on the go, as we document everything that comes to mind. As this network grows, it becomes more complex, it contains more connections, it slowly reveals more and more patterns of thought, patterns of behavior that, let's say, were difficult to identify at a quick glance. I can catch myself suddenly one day looking at this graph and saying, "Wow, almost every brilliant idea I had this year happened while I was running on the treadmill, and I wasn't even listening to music." And I tell myself, "Okay, I need to try running less with headphones. The weather is nice now too, maybe I'll go for a run in the forest." Then Claude Code bursts into the room, looks at me with a strange look I'm not used to. And I say to her, "What's your story? Why are you looking at me like that? Speak." And she says to me, "Listen, before you go to the forest, I saw your Obsidian, and I think we need to talk." And I say to her, "What's the problem with going for a run in the forest? What could I miss here? What do you see that I don't see?" And she says to me, "Notice that every time you start thinking about this project you planned half a year ago, you notice, you escape to a run in the forest. And now it's also without headphones, but listen, you're working on yourself. Maybe if you take one small step forward on the project, you'll also feel less stressed, and you'll also see that you'll definitely have an hour free for a run, and you'll also sleep much better that day, which is something you're lacking." Notice that with this connection of Claude Code to Obsidian, we get a very great power. We have smarter eyes to look at the graph and smarter, faster hands to write into this graph. And what's really fun here is that the connection is very, very simple because all we need to do is run Claude Code on the folder that Obsidian uses to store our notes. And then Claude Code will run freely from there, and we'll also improve it later, you'll see. And you ask me, "Rani, okay, but what are these notes? It sounds like something very, very complex." So the answer is no. The answer is that we're talking about Markdown files here, files in MD format. For those who don't know, files that support all sorts of symbols that give text a graphical form. For example, these two hashtags before a line turn the line into a heading. These two asterisks surrounding a specific word make it a bold word. And then a program that supports Markdown and Markdown rendering knows how to read these symbols and display something more readable. And of course, it continues with diagrams, tables, flowcharts, and so on. And what Obsidian adds to this is the ability to link MD files or link one MD file to its friend, and thus create such a network, such a graph that organizes our ideas. Then the space each idea occupies also gets its emphasis in the graph. Notice how here, as I write, connections are created on the graph. On the right, I'm saying that after a long day of meetings, I noticed my stress levels usually rise. As I was writing this, you can see a link was created to the "meetings" point and to the "stress" point. And all I do to create such a link is to surround one word or a few words with two square brackets on each side. Then, if what I surrounded is already in the graph, it immediately gets a link to what I'm currently writing. And if what I surrounded is not in the graph, a new point is created, a new note is also created. For example, notice that "yoga" is here because, of course, I wrote it, but I never wrote anything in a document called "yoga." So we see the point in gray, and it's a new document for all intents and purposes. If I go in and write, I can link yoga to the rest of the points in my graph. Now we say, "Cool, but in today's era, we're not going to write this document manually. That's not going to happen. Rani, forget it." So, of course, we'll talk to Claude Code in English of three units, and we'll want Claude Code to do the work for us. And by the fact that we tell her she's writing into Obsidian, she'll already know how to use these connections and thus preserve our information in that associative graph structure. And of course, beyond writing the information, reading the information and searching for all sorts of patterns will also be something Claude Code will do much faster, much better in our place. So we need two things: we need Obsidian, and we need Claude Code. If you don't have Claude Code, you go to the official Claude Code website. I've put a link in the video description, and you use this command you see here. In this case, I'm on Windows, so this is the command I'm using. But if yours is different, if yours is otherwise, that's perfectly fine. I'm copying it to my terminal. And the installation, of course, begins. To install Obsidian, we go to the website. Notice it's a nice website, Obs-idian, which is a cool pun. We click here on "Get Obsidian for Windows," double-click the installation file, and we start the simplest installation in the world. They ask me if I want to install Obsidian for all users on my computer or just for me. I choose "Only for me," continue here with "Next," then click "Install," wait a few seconds, and then I reach the opening screen. Obsidian tells me, "Listen, choose the folder where you want to store all your files, all your notes, what's called your vault." So I choose here "Vault," give it a name like "Code Cave Vault," and I choose the location of the folder to be "users/run a," and we reach our workspace. There are various buttons around, nothing too scary. I'm sure you've dealt with scarier things in your lives, but generally, what we see here is all our notes on the left. Usually, something is open in the middle with the note we're currently working on. And the coolest thing here is the graph view, which if you accidentally closed it out of excitement, you can open it by clicking the button with the graph icon in the top left corner. We have a note here called "Welcome." Notice this heading, of course, "Welcome Note," a sample like this, regular text. Notice "This is your new world." And the note called "Welcome" is linked to a note called "Create a Link." But because we haven't created one yet, it's gray like the "yoga" note we saw earlier. Now, you don't need me to show you how to write text and surround it with square brackets. You didn't come here for that. What would be cool to see, what would be cool to do, is to now run Claude Code on this folder, talk to it, and see how this graph changes on the go. So I open Obsidian here on the left with a vault that contains a few notes and a few connections between them that I created beforehand, and on the right, Claude Code on the same vault. Notice that I'm here on "Code Cave Vault" both in Claude Code up here and in Obsidian, you can see it down here on the left. And I tell Claude Code that I had a great day today. I slept enough, learned advanced topics in language models, and even managed to survive the day on just one coffee, which is very, very surprising. I felt less stressed overall, and it would be cool if you could document all of this in my Obsidian. Oh, and also create the file called "learning.md" which, judging by its name, will contain all the relevant information for learning in general. She tells me, "Okay, Rani, let me first find your vault." Of course, because she's running on it, she finds it after a few seconds. She also sees specifically the folder here in the top left corner called "Journal," which is essentially a folder that functions, according to its name, as a journal where we will actually enter all sorts of documentation that we identify during the day. She understands the vault structure after a few seconds. She sees that the journal folder is empty. She also understands this style of creating links with these square brackets, which is, of course, no big deal for her. And after a little while, she writes the file "March 15, 2026.md," Sunday, which actually describes how the day went in MD format. Notice also the links that were created here, like "sleep" and "coffee." And as you saw, the graph also started to move a bit because connections were added during the writing itself. The file "learning.md" that I requested was also created here. And she summarizes and tells me that she created two files in the vault, and in both, they use the links that Obsidian supports. They are also called "wikilinks," by the way, worth noting. And thus, they are integrated with the rest of the graph. I can also continue and tell her more specifically that I learned about "deep learning," and I learned about "sparse attention," and about "mixture of experts." And these are the topics I want you to document in Obsidian, please. And naturally, now that we have "learning.md," connect them to "learning.md" as well. And she does all of this without any problem at all. She creates more files for us like "mixture of experts," "sparse attention." Notice how automatically they are linked to our journal here in the corner from March 15th because I'm talking to her today, so she knows how to make this connection to our journal. And notice also how she updates the file for that day. She removes this line that spoke generally about general learning. The one you see marked in red, and she adds the green. She adds the line that details it further and talks about learning "sparse attention" and "mixture of experts." And here we get a summary of two files created and an update of two existing files. She tells me that they are all linked to each other. The two new files are linked to each other and also linked together to "learning.md." And that our journal closes everything and links this whole story together to that day. And so I can continue to add things like, let's say, I had a lot of motivation today and I could manage my day perfectly, and document that for me too. And so, slowly, more and more bridges are created. And here, for example, you can see "time management" which has started to be linked to many points in the graph. Then comes the moment when I ask Claude Code, and I ask her to look at my graph, not just to write to it, but also to look at it and tell me, "Tell me, do you see anything I can do that will improve my sleep?" And she tells me that from my vault, it can be seen that it would be a very, very good idea to build an evening routine, some kind of evening routine before sleep. Put the phone aside because social media before sleep ruins sleep quality. Darken the room, cool it down, and make it quiet, and other tips. And continue doing what I've been doing until today, or rather, today, because it reduced my stress, that one coffee, one coffee instead of two, and don't compensate for any lost sleep hours by staying up late. That's not a good idea. She also recommends hydration, drinking enough throughout the day because it helps with everything in general. And of course, these are tips that we all know. We haven't discovered anything new here, but I also asked a very general question. Think about it, if I continued and documented more and more things over, say, a whole year, I don't know, or a few years, Claude Code would not only have a lot of information and a very large graph, but this whole story would also be documented on a timeline as a journal, which could reveal and close patterns in various periods of my life. And then see what I'm repeating, which habits I need to give up, which habits I need to keep or add. And also things that are not related to my personal development. Various things that can be more technical, more specific, on various topics in the graph. So I go along with her, I say to her, "Listen, come on, let's add a sleep routine for the evening." And she really does it. She adds to my journal this whole section we see here in green, which contains seven steps, seven things I need to do tonight before sleep. So this is the basic use of Obsidian, but you can imagine that it doesn't end here, and it's true, it's just starting here, because we can add plugins to Obsidian and thus expand its capabilities. The number of plugins, as you can guess, is infinite, but I will ask Claude Code to give me at least the prominent plugins so that I can install a few and show an example of what I can add to Obsidian. And I specifically ask her for plugins related to productivity, to improve my productivity, and for her to also give me the option to choose what to install. A story that takes a few seconds here is summarized in a table that shows me some leading plugins. Notice, for example, the "Data View" plugin here, a plugin that allows us to query our vault, our graph, like we query. We have the plugin called "Templates" here, which allows us to create note templates. And then I can tell Claude Code to rely on a specific template when she writes something for me. I have "Calendar," which gives me a calendar widget linked to my daily notes and weekly notes. And so through this calendar, I can jump to notes. So if I have this plugin, I can tell Claude Code to link notes to my calendar, and thus I can schedule tasks or various things related to my calendar and generally get a more organized view of things when I jump to Obsidian and look at it. If we go down here, we have "Auto Note Mover," which automatically organizes our notes into folders according to tags or according to various rules we've added. So I tell Claude that I want "Data View," I want "Calendar," and this "Auto Note Mover" also sounds excellent, so I want it. And everything is installed here after half a minute, a minute at most. A summary shows us all the new plugins. What we need to do now is close and reopen Obsidian, restart Obsidian so that, naturally, all the new capabilities become active. And I do that, of course. And afterwards, I ask Claude Code, "Let's use Data View, it's actually interesting to me. I want to use Data View to create a query on Obsidian regarding, let's say, stress management and coffee consumption." And she does it. She creates a "dashboard.md" for me, which has six queries. On the left, in a table, it shows me the query. On the right, what the query does. And as written below, if I open the note in Obsidian, if I open "dashboard.md" here, Data View will update all the tables live, and as I add more things to the journal, the dashboard will update automatically. This sounds amazing to me, a great feature, because this automatic update, in some cases, depending on my query, is something that Data View does now, and it's not something Claude Code needs to worry about. So we'll also guess what, we're saving a few more tokens along the way here. On the graph, you can see the new dashboard, which is called "Stress and Coffee Dashboard." And if we click on it, you can see its textual representation. You can say it looks like this. Notice there are quite a few links to specific points in the graph, as you can see. And because my graph is not really organized, notice on the left I have many files. I would like to organize them into folders, and for that, we installed "Note Mover." So I ask Claude Code to organize this graph for me, at least the notes themselves, not just the graph itself. She shows me this table that summarizes the order that will happen here, and she also asks me if I want to start these changes according to the table, and I, of course, agree. I tell her, "Looks good," and thus give her the understanding that she needs to start. And of course, Anthropic asks me at this exact moment what I think about Claude Code's performance so far. So I give them quick feedback while all of this is running, because why not give credit? So notice how slowly all of this starts to fold, step by step. You can see how on the left, the left, this long list we had from the beginning starts to get a structure, a folder structure that organizes all our notes as it should be, so we have some order for our eyes, and also ultimately, if we have order, then Claude Code will also have the same order, and that will obviously significantly ease her work. And while all of this is happening, you can also see that our graph is being shaped, taking on a new form. It's very nice to see it organizing itself according to the order we agreed on in the table a few seconds ago. And I think not only that, notice also how powerful this plugin is and how smartly Claude Code uses it. She tells us that all the notes have moved, but let me define the mover now so that new notes you add later will automatically go to a specific place in the vault where they should go. And she therefore edits a file located at this path. Notice this configuration file for "Auto Note Mover," called "data.json" at this path. Here is the beginning of the file. Naturally, if we want to see the entire edit, we can also do "Ctrl+O" here, as written here, and then expand another 60 lines that are written in the configuration, and we get a summary of this whole process. And you can see here that notes whose name is in this date format will automatically go to our journal, to our journal. And if we also add a tag to a certain note, it will automatically move according to the topic, because that's very logical to organize notes. And to add a tag, we just need to do a hashtag and the topic name somewhere in the notes, and of course, all of this will be automatically organized. What's very nice to see and important to pay attention to here is that now it's not just us who benefit from the order and organization in our daily lives, but also Claude Code benefits from this additional memory that we've connected to her. Because now all these connections here on the graph are something that Obsidian already does, and she doesn't have to strain so much when she formulates an answer. She will have to search our graph, but because this graph is built this way, in this graph search, we already have existing connections, so it's very convenient for her to go through specific paths that are suitable for constructing an answer to what we requested. As this order is maintained over time, the structure of the graph becomes much, much more solid. And as we document and document our lives on Obsidian, statistically, we can hit all sorts of recurring patterns in recurring behaviors and thus get quality feedback, perhaps even discover things about ourselves that we never paid attention to. This, of course, won't happen in a week or two, maybe not even in a month, but over time, and as the connections strengthen here, we will absolutely be able to pinpoint completely new things. And when we discover them, perhaps we can also improve the way we input new information into the graph. So the accuracy here can even potentially grow exponentially. In my last video, I talk about the integration of Claude Code with Google Colab notebooks, and notice that there's nothing preventing the connection to Obsidian now as well. And it seems that the monster we created in that video has now gained new superpowers. If you stayed until now and watched, I thank you for stopping by, clicking on my video, and listening to what I have to say. And if you want to see my videos at least a day before they are released to the public, if you want me to answer you in the comments before everyone else, you are welcome to join the "Branga." Below this video, you will see a button called "Join," you click on it, and from there you will see what the "Branga" is. And even if you decide not to join, that's perfectly fine too. I will still answer you in the comments, and I will still be here on YouTube. You will still see me on YouTube as usual because I don't intend to go anywhere. Until next time. Alright, bye.