YouTube Instagram TikTok How it works FAQ

The fastest way to turn talking videos into clean, searchable text. Free, private and no account required.

About

Built by Rishav Raj with one goal: free transcripts for everyone.

rishav@yttotranscript.com

Transcribe

YouTube to text
YouTube with timestamps
Instagram Reels
TikTok videos

Learn

How it works
FAQ
YouTube transcripts guide

More

Chrome extension
Contact
Privacy policy
Terms of service

© 2026 YTtoTranscript. Made for creators & researchers.

Not affiliated with YouTube, Instagram or TikTok. Respect creators’ copyright.

The Claude Code + Obsidian Setup That Now Runs My Life — Transcript · YTtoTranscript

Home/Transcripts/YouTube

YouTube Transcript

The Claude Code + Obsidian Setup That Now Runs My Life

Chase AI · 3,488 words · 17 min read · EN-ORIG

Watch on YouTube

Share X LinkedIn WhatsApp

Below is the complete, readable transcript of The Claude Code + Obsidian Setup That Now Runs My Life by Chase AI on YouTube. Read the full text, copy any part you need, or generate a transcript for any video with our free tool.

Most of you are using Claude Code with Obsidian the wrong way because if all you're doing is opening Claude in your vault and using Obsidian as a way to read markdown, you are leaving a ton of value on the table. These tools can do way more than that and one example is this. I've turned Obsidian into the

command center for my entire Claude OS. It has a fully integrated terminal, custom observability, and all of my most used skills and automations are just one click away. So today I'm going to break down how this command center works, how you can build it for yourself, and what's going on under the hood so we can

get the most out of Claude Code and Obsidian. So this is my Obsidian Command Center and I love this thing because it kind of gives me the best of both worlds. I have the integrated terminal here so I'm able to use Claude Code the same way I always use it, which is in

the terminal, but at the same time I get all of the benefits of the Obsidian ecosystem. I have my entire vault set up over here on the left so I can very easily see all my markdown files and right here I have what is essentially a custom plugin that Claude Code built for

Obsidian. And it's completely custom. Everything I see here is related to me and what I need to see in my day-to-day. Obviously, my stuff is very content forward so that includes stuff like, hey, what are my subscriptions doing? What are my follower accounts doing across all my platforms? What's going on in my schedule that day? What are my

daily tasks? And this is just sort of like the overview. So I can go deeper, I can see, okay, let's do a deep dive into my audience. How has my YouTube been doing? What's been working? What hasn't been working? If I look at research, you know, I can see stuff like all the

Transcribe another video

Paste any YouTube, Instagram or TikTok link to get a free transcript.

Free · No sign-up · Unlimited

Add to Chrome — transcribe in one click

trending GitHubs that have been created in the last week. I can see Hacker News. I can see also sort of a pulse of the AI landscape at large. Headlines, X conversation, YouTube trending, and then content opportunities. The cool thing with this isn't necessarily what I have here, but the fact that you could change

any of this or add anything you want to with a single prompt inside of Claude Code. On top of that, I have a lot of more common automations and skills I run on a daily basis, things like deep research, doing an inbox brief, a morning report. Now, a lot of these automations and their outputs are

actually tied to what you see here visually. So, something like the research side is tied to an actual skill. So, if I click this and look at the full GitHub trending, I can see the full list of all 10 trending GitHub repos that have been created in the last 7 days that are

related to AI as well as the top five trending. And because of Obsidian, it's really easy for me to just click inside here, and I also have a web viewer. So, it also has a browser, which is nice. All on the same pane. Also, I have my actual calendar over here on the right,

and this just isn't like a special plugin with Obsidian. This is literally just a pinned browser tab, which works for me. And so, when you look at this all together, what I'm able to do is I'm able to do all my work I would normally do inside of Claude Code in the terminal. Yet, I get all this additional

sort of observability, sort of like visual reports that you naturally don't get inside of the terminal. Because don't get me wrong, I love the terminal, but it has some weaknesses, which is why everyone uses IDEs like VS Code and that sort of thing. But over time, as I've moved away from things like VS Code, I

still like to have stuff like this up kind of at eye level. So, I have found a command center pretty much like this inside of Obsidian is a great way for me to do my work. And because it's completely custom, knowing how to create something like this, I think is a huge value add, especially if you're someone

who's more of a solo operator. Plus, if you build something like this, naturally, you're going to have a more symbiotic relationship between Claude Code and Obsidian, which means we're using Claude Code in a manner where we have some sort of memory construct that allows us to get better outputs over time. So, let's talk about the process

and what you need to be thinking about if you were to create something like this for yourself. But first, a quick word from today's sponsor, me. So, last month I released my Claude Code Masterclass, and it It the number one way to go from zero to AI dev, especially if you don't come from a

technical background. My Claude Code Masterclass also includes an Agentic OS and a Codex Masterclass as well. So, whether you're trying to master Claude Code or Codex or you're trying to build something exactly like we're talking about today, you can find it inside of Chase AI Plus. It includes all the prompts to pretty much get my exact

setup, including this command center. So, if you want to get your hands on that, there'll be a link to that in the pinned comment. So, what do we need to do if we want to build an Obsidian command center that sort of sits on top of Claude Code? Well, there's two things we need to be thinking about. One

is sort of the fancy part, the observability piece. What do you actually want this to show? So, for me, I like you saw, I broke it out into three areas. I have the overview, which is where I normally sit. I can see my 5-hour token burn. I can see what's going on in my sub count. I can see my

latest upload, right? My schedule, my daily tasks, all that. Well, what do you need? What do you want to actually see? And again, mine goes deeper for audience and research. I've seen people do this a number of different ways. I have certain people I've talked to who have, you know, they're running multiple

businesses. So, each tab inside their OS relates to a different business or even like a different persona, right? Sometimes they have their business owner hat on. Sometimes they have their dev hat on, etc., etc. So, that's a question that I can't give you the answer to. But, you need to know what it is you're actually trying to

track. What metrics do you care about? What do you want to see when you're also coding or using Claude Code? That should drive your answer. Ideally, that answer is also tied to stuff you already do. So, like you saw here on the research side of stuff, when it looks at GitHub trending, Hacker News, and my morning

brief, like this morning brief, this report right here, is something I already do regularly. Like, it's an actual skill. It's kind of giving me a summary of a deeper report, which is tied to a skill. So, hopefully by now, if you've been using Claude Code a lot your business or in your daily life, you've kind of set a

series of like workflows and skills and automations. Well, instead of having these just like hidden away in some markdown file that you would occasionally use Obsidian to bring up, well, let's just put it in one place. Let's make it really easy for you to see. So, there are 1 trillion different iterations of what

that could be. All you have to do at that point is pretty much have a back and forth with Cloud Code and try to come up with some ideas. And you don't have to come up with a single one. You just like kind of need to experiment because it's not too hard to swap them

out. When it comes to the actual design, the sort of aesthetic you're going for, I think a great trick is using Cloud Design to help you out. So, have Cloud Design create a prototype as if it was creating a web app because it's going to create a plugin. This whole command center, this is a plugin. It's It's code

that Cloud Code wrote. So, go to Cloud Design, have it create some sort of project, but when you create it, have it create like a bunch of different variations. So, I already knew I kind of wanted this like black and orange type of style. But, what you can do, instead of just showing it with one single

style, have it create like three, four, five different sort of macro variants with completely different styles. And eventually, you can hopefully nail something that you like. That's my sort of tip for trying to nail some sort of visual style. Now, the second thing you want to think about for your dashboard is, do you want to set up some sort of

shortcuts for your skills and automations that you use very regularly. So, that's what you see right here. These are kind of the most common skills and automations, and if they're automations that allows me to do it on demand, that I use in my day-to-day. So, for my plan today skill, it will go out

into my calendar, grab everything. It also takes a look at my email and gives me sort of my daily tasks. My YT pipeline allows me to like actually search for specific things on YouTube. I have a weekly review, a morning brief, etc., etc. Again, totally custom. But, I found it to be super convenient having

these to be simple buttons that often generate some sort of report instead of having to do it manually inside of Claude or open up another terminal and have it run it that way. Because what this does is it just opens up a headless version of Claude. And don't worry, we'll talk about Claude-p at the end

here in terms of some of the recent Anthropic changes, but it opens up a headless version of Claude and then runs it for me. So, it doesn't like sort of have to populate whatever current active session I'm using. And that's kind of the meat and potatoes of the command center itself. Like you say, it's

completely customizable. It just what you want to see. Beyond that, and remember this is what Claude code is going to create via a plugin. Beyond that, then we're talking about actual Obsidian plugins and things in the Obsidian ecosystem itself. So, there's a couple of free wins. So, if we go down to like settings and go to

community plugins, there's a few you're probably going to want to add. So, first of all, you definitely want to add terminal. You are going to want to You can see here the command center. You can see how this has been added as a specific plugin. That's the one Claude code did. You're going to want want to add hot

reload. So, hot reload means if you make a change or have Claude code make a change, you don't have to restart Obsidian every single time. The iconize, that's what allows me to have these little like icons next to all my folders. Another one you're going to want to do if you go to core plugins is

you definitely want to make sure you have web viewer actually enabled. For whatever reason, it's not by default available. So, web viewer just means I can actually look at things. Like I click this link to a video, it actually takes me there versus pulling up, you know, my actual Chrome browser in another window. And also

remember, you can adjust all this however you want. So, like if I want to have some sort of, you know, tab up here, I can do that and on and on and on. That's That's the nice thing about Obsidian. It's just like there's so many different variations of this you can create. Now, the other

thing you have to nail if you decide to go with some sort of like Obsidian command center route is like, "Okay, what's actually going on in terms of the file structure of Obsidian and Claude code because if you want to use Claude code with Obsidian and you plan on having any level of

scale like thousands and thousands of documents, you have to have some sort of file structure that makes sense. Now a very common structure you'll see is the Karpathy file structure. Now the Karpathy file structure looks something like this where we have our vault and then within the vault we have three subfolders, the raw, the wiki, and the

outputs. So raw is where unstructured data goes. Wiki is kind of just like internal reports or internal Wikipedia pages that actually break out what we've done in the unstructured data. And then third, we have some sort of output, some sort of deliverable. So example would be, "Hey, I want you to do some research

on AI agents." All that research goes down here into the raw. Hey, we take that information about AI agents and we turn it into an article inside of our wiki. And then lastly, "Hey, I now want to create a slide deck to show someone information about AI agents." Well, that goes into the output

section. And at each level of Obsidian, every single layer deep, we add some sort of index file that tells Claude code, "Hey, at this layer of the folder, this is what exists." So for example, if I told Claude code, "Hey, can you look at stuff that we've created that has to do with AI agents?"

It would start at the vault. It would go to the wiki. It would read the master index which says, "Hey, inside of here there's AI agents, rag systems, and content creation." It would then go to AI agents. And then inside of AI agents it also had what it would have an index file, a table of contents saying, "Hey,

here's all the files inside of here." Now obviously that's overkill if we're talking about just three different things, but think of it being 3,000. But the truth is when it comes to creating a command center with Obsidian and actually having some sort of like symbiosis between Claude code and Obsidian, you don't have to do this.

It doesn't have to be raw, it doesn't have to be wiki, it doesn't have to be output. You don't have to follow Karpathy's recommended system. You just need something that makes sense for you and you need something that anyone can kind of travel through if they you said, "Hey, find this document." Because if we think big picture and we

have thousands and thousands and thousands of documents, if we don't have a system that makes sense and there's a clear sort of highway to go down to find things, well, that's going to cause Claude code to eat up a bunch of tokens and it's going to be slower and then it's like, "Okay, why don't we just use

a full-blown rag system?" So, you could potentially be put yourself in a corner where you need to add a lot of complexity to your system. So, for me, I have an archive, content, daily notes, dashboard, inbox, op, projects, raw system, and wiki. So, if you break all those out, they kind of fall into these three

similar buckets. I've just added a few to sort of like delineate some stuff. But, at the end of the day, they're still kind of an unstructured area, a structured area, and then output/deliverables area. But, just like everything else we've talked about with the Obsidian setup, it's totally custom. You just need to find out what makes

sense to you. If you can explain it and if someone can understand it, then Claude code probably can, too. And speaking of that, the other thing you're going to want to pay attention to is your claude.md. So, what should we put in our claude.md? I'm of the opinion that less is more. So, when it comes to the sort of

conventions that go into this folder, I think primarily should have to do with your vault structure. Like, this way Claude code knows exactly where to go to get things. So, my vault structure kind of breaks out what is happening at each folder. And then I kind of talk about stuff like the navigation pattern, how it should

actually move through each folder to find stuff. And then I have some information about file names and organization of Obsidian markdown. So, I think at the bare minimum, what you need to have inside your claude.md is some sort of explanation of like your file structure, how it should navigate, and then I would add a little

blurb about Obsidian best practices. Basically, like how you want it to use Obsidian, which means like have Wiki links. How you want it to embed, how you want things to be tagged. That way, we can get the most out of Claude code creating things inside Obsidian in an Obsidian type fashion. And by

doing so, you're able to get the most out of these Obsidian type systems without having to manually do anything, right? That's the symbiotic relationship here. Claude code writes the files in Obsidian format because it's in Obsidian, and it makes it really easy for you, the human being, to actually read all these things and get the most

out of it. That's sort of the idea. And you combine that with the observability you get with a command center system and integrated terminal, and I think you can be sitting really pretty with a Claude OS using this kind of setup. Now, the last thing I want to talk about was sort of cost, especially when we talk about,

"Hey, if I click this button, a skill or automation gets used, and it's in a headless format with Claude, so Claude-P." Well, for those of you who haven't been paying attention, that sort of use case for Claude code will no longer pull from your max subscription. Instead, what it's going to do is every

month they're going to throw you $200 on top of your max sub, and then headless programmatic use of Claude code will pull from that pool, that $200 pool. Issue with that is it's not using the subsidized sort of usage that you get with a max plan, instead it's going to be on API

sort of costs, which is like 10x more expensive. So, can you run into problems with that? Realistically, I don't think so. You know, in an Obsidian setup like this, I'm using these, you know, throughout the day, but I'm not hitting here sitting there spamming these, right? I'm still primarily operating inside of the terminal itself.

So, while it sucks, I don't think that's really going to be a problem. If it is a problem for you, well, then just move over to the Codex you like, to be totally honest. Having all this be a layer on top of Codex instead of a layer on top of Claude code, like there's no There's basically

no difference. You can refactor this in like 5 minutes. Right? Like, we're just changing Claude-PE to Codex headless. So, and Codex is a really good tool as well. So, it's not like you're also getting a dip in the outputs. So, stupid, annoying, yes. A huge problem that's a big blocker, no. So, that's

where I'm going to leave you guys for this one. I really enjoy using this sort of command center for my Claude OS. I think a lot of people could get some use out of it because, like I keep harping on, it's custom. So, you just make it work for you. There's no one-size-fits-all.

So, as always, let me know what you thought. Make sure to check out Chase AI Plus if you want to get your hands on the Claude code masterclass as well as the prompts to get my exact set of if this exact thing is kind of what you're going for. And besides that, I'll see you around.