I Switched Back to Obsidian (because of AI)

Okay, guys. So, I have to admit Okay, guys. So, I have to admit something. I just switched my main something. I just switched my main something. I just switched my main note-taking and project management tool note-taking and project management tool note-taking and project management tool from Notion back to Obsidian, which I from Notion back to Obsidian, which I from Notion back to Obsidian, which I know is funny because not long ago I know is funny because not long ago I know is funny because not long ago I made a whole video about why I moved made a whole video about why I moved made a whole video about why I moved away from Obsidian and switched to away from Obsidian and switched to away from Obsidian and switched to Notion. So, you might ask why the heck Notion. So, you might ask why the heck Notion. So, you might ask why the heck did I switch back? Now, the reason is did I switch back? Now, the reason is did I switch back? Now, the reason is that my workflow changed quite a lot that my workflow changed quite a lot that my workflow changed quite a lot recently, especially when I started to recently, especially when I started to recently, especially when I started to use more terminal based AI agents, local use more terminal based AI agents, local use more terminal based AI agents, local tools, and so on. And when you then try tools, and so on. And when you then try tools, and so on. And when you then try to use tools like notion but also other to use tools like notion but also other to use tools like notion but also other developer focused tools like confluence developer focused tools like confluence developer focused tools like confluence or clickup air table and so on they all or clickup air table and so on they all or clickup air table and so on they all work great out of the box. But at the work great out of the box. But at the work great out of the box. But at the moment you want to bring your own AI moment you want to bring your own AI moment you want to bring your own AI agents, your existing local tools or agents, your existing local tools or agents, your existing local tools or your custom workflows in. This is where your custom workflows in. This is where your custom workflows in. This is where things get complicated because these things get complicated because these things get complicated because these bigger note-taking and project bigger note-taking and project bigger note-taking and project management platforms are mostly all management platforms are mostly all management platforms are mostly all cloud-based. And increasingly they also cloud-based. And increasingly they also cloud-based. And increasingly they also want you to use their own AI assistants, want you to use their own AI assistants, want you to use their own AI assistants, of course, with their own subscriptions of course, with their own subscriptions of course, with their own subscriptions and in their own closed ecosystem, and in their own closed ecosystem, and in their own closed ecosystem, something like Apple tries to do all the something like Apple tries to do all the something like Apple tries to do all the time. And sure, you can technically time. And sure, you can technically time. And sure, you can technically sometimes connect this with MCP servers, sometimes connect this with MCP servers, sometimes connect this with MCP servers, APIs, and CLI apps, but in my opinion, APIs, and CLI apps, but in my opinion, APIs, and CLI apps, but in my opinion, it is still clunky, relies on a cloud it is still clunky, relies on a cloud it is still clunky, relies on a cloud connection, and sometimes it's really connection, and sometimes it's really connection, and sometimes it's really slow and not portable. So that's why I slow and not portable. So that's why I slow and not portable. So that's why I have decided to quit using those kind of have decided to quit using those kind of have decided to quit using those kind of tools and move my workflows more to tools and move my workflows more to tools and move my workflows more to local tools, more ropen tools. And in local tools, more ropen tools. And in local tools, more ropen tools. And in that way, Obsidian just fits pretty that way, Obsidian just fits pretty that way, Obsidian just fits pretty well. It just uses simple markdown files

well. It just uses simple markdown files well. It just uses simple markdown files because here's the thing. If your notes because here's the thing. If your notes because here's the thing. If your notes are just markdown files on your are just markdown files on your are just markdown files on your computer, there are super easy to work computer, there are super easy to work computer, there are super easy to work with by both humans and also machines. with by both humans and also machines. with by both humans and also machines. So you can easily just use it with any So you can easily just use it with any So you can easily just use it with any coding agent like pi, codecs, cloud code coding agent like pi, codecs, cloud code coding agent like pi, codecs, cloud code or any other local LLM supporting agent or any other local LLM supporting agent or any other local LLM supporting agent in your terminal and you can still work in your terminal and you can still work in your terminal and you can still work with your entire knowledgebased system. with your entire knowledgebased system. with your entire knowledgebased system. Simple and efficient. So that's the main Simple and efficient. So that's the main Simple and efficient. So that's the main reason why Obsidian became interesting reason why Obsidian became interesting reason why Obsidian became interesting for me. Again, not because it is the for me. Again, not because it is the for me. Again, not because it is the better productivity app, but as a better productivity app, but as a better productivity app, but as a modular interface on top of markdown and modular interface on top of markdown and modular interface on top of markdown and together with my own PI agent plug-in together with my own PI agent plug-in together with my own PI agent plug-in that I have written, or honestly, I that I have written, or honestly, I that I have written, or honestly, I should better say voded in Obsidian, should better say voded in Obsidian, should better say voded in Obsidian, this became the perfect productivity this became the perfect productivity this became the perfect productivity tool for me again. So, in this video, tool for me again. So, in this video, tool for me again. So, in this video, I'd like to share this with you. I'm I'd like to share this with you. I'm I'd like to share this with you. I'm going to show you the new PI agent going to show you the new PI agent going to show you the new PI agent Obsidian plugin that I have created and Obsidian plugin that I have created and Obsidian plugin that I have created and how you can use this to level up your how you can use this to level up your how you can use this to level up your productivity system as well without productivity system as well without productivity system as well without having to use any vendor login at all. having to use any vendor login at all. having to use any vendor login at all. So you can enjoy complete freedom with So you can enjoy complete freedom with So you can enjoy complete freedom with any open source tools or your existing any open source tools or your existing any open source tools or your existing AI subscriptions or local LLMs that you AI subscriptions or local LLMs that you AI subscriptions or local LLMs that you want to use. Once you start using agents want to use. Once you start using agents want to use. Once you start using agents and tools like PI, the next question is and tools like PI, the next question is and tools like PI, the next question is where does the agent actually get where does the agent actually get where does the agent actually get current external information from?

current external information from? current external information from? because my obsidian vault is local because my obsidian vault is local because my obsidian vault is local markdown, but sometimes I want Pi to markdown, but sometimes I want Pi to markdown, but sometimes I want Pi to search the web or check Google results, search the web or check Google results, search the web or check Google results, pull structured search data and so on. pull structured search data and so on. pull structured search data and so on. And SER API, the sponsor of today's And SER API, the sponsor of today's And SER API, the sponsor of today's video, is a search results API for video, is a search results API for video, is a search results API for developers. So it gives you clean developers. So it gives you clean developers. So it gives you clean structured results from Google, YouTube, structured results from Google, YouTube, structured results from Google, YouTube, Bing and many many more services so that Bing and many many more services so that Bing and many many more services so that all of your apps, dashboards, all of your apps, dashboards, all of your apps, dashboards, automations or AI agents can easily work automations or AI agents can easily work automations or AI agents can easily work with real search data instead of using a with real search data instead of using a with real search data instead of using a fragile browser scraper. For example, fragile browser scraper. For example, fragile browser scraper. For example, I've just created a small skill with the I've just created a small skill with the I've just created a small skill with the SER API interface so that my agent can SER API interface so that my agent can SER API interface so that my agent can use it directly from any workflow. That use it directly from any workflow. That use it directly from any workflow. That means I can ask Pi to search Google for means I can ask Pi to search Google for means I can ask Pi to search Google for a topic. Maybe compare sources for my a topic. Maybe compare sources for my a topic. Maybe compare sources for my researching notes or pull YouTube search researching notes or pull YouTube search researching notes or pull YouTube search results to see which video topics I have results to see which video topics I have results to see which video topics I have already covered or what other creators already covered or what other creators already covered or what other creators are covering and then even collect are covering and then even collect are covering and then even collect results for deeper technical research results for deeper technical research results for deeper technical research from the web. It is really super cool from the web. It is really super cool from the web. It is really super cool and the best is I don't have to build and the best is I don't have to build and the best is I don't have to build and maintain all the scraping logic and maintain all the scraping logic and maintain all the scraping logic myself because SER API handles captures myself because SER API handles captures myself because SER API handles captures proxy rotation and changing layouts proxy rotation and changing layouts proxy rotation and changing layouts completely in the background. So I can completely in the background. So I can completely in the background. So I can really just focus on the workflow I really just focus on the workflow I really just focus on the workflow I actually want to build. It is also super actually want to build. It is also super actually want to build. It is also super fast. So the results come back in around fast. So the results come back in around fast. So the results come back in around just one or two seconds. And the cool just one or two seconds. And the cool just one or two seconds. And the cool thing is the pricing is quite flexible.

thing is the pricing is quite flexible. thing is the pricing is quite flexible. There's even a free plan to get started. There's even a free plan to get started. There's even a free plan to get started. So, in my opinion, SER API is the So, in my opinion, SER API is the So, in my opinion, SER API is the perfect expansion for AI workflows to perfect expansion for AI workflows to perfect expansion for AI workflows to search reliable information from search reliable information from search reliable information from qualified sources. And if you want to qualified sources. And if you want to qualified sources. And if you want to try it out yourself, just use the link try it out yourself, just use the link try it out yourself, just use the link in the description box down below or in the description box down below or in the description box down below or scan the QR code on the screen to get scan the QR code on the screen to get scan the QR code on the screen to get started with 250 free credits. started with 250 free credits. started with 250 free credits. All right, guys. So now before we start All right, guys. So now before we start All right, guys. So now before we start looking into some specific AI workloads looking into some specific AI workloads looking into some specific AI workloads and how exactly I'm using the PI and how exactly I'm using the PI and how exactly I'm using the PI Obsidian plug-in, I want to start from Obsidian plug-in, I want to start from Obsidian plug-in, I want to start from the beginning and show you some of the the beginning and show you some of the the beginning and show you some of the changes that I have made to my changes that I have made to my changes that I have made to my productivity system. I know I probably productivity system. I know I probably productivity system. I know I probably keep changing it every year. So it might keep changing it every year. So it might keep changing it every year. So it might look a bit different compared to my look a bit different compared to my look a bit different compared to my previous videos about Obsidian or even previous videos about Obsidian or even previous videos about Obsidian or even notion, but I think this new structure notion, but I think this new structure notion, but I think this new structure allows me to better work with my home allows me to better work with my home allows me to better work with my home lab and have a better documentation and lab and have a better documentation and lab and have a better documentation and operation instruction nodes for my AI operation instruction nodes for my AI operation instruction nodes for my AI agents or for myself if I'm managing my agents or for myself if I'm managing my agents or for myself if I'm managing my home lab servers. So here for example, home lab servers. So here for example, home lab servers. So here for example, this is the main directory where all the this is the main directory where all the this is the main directory where all the information for my home lab it is information for my home lab it is information for my home lab it is written in simple markdown files. So written in simple markdown files. So written in simple markdown files. So nothing complicated. I'm still using nothing complicated. I'm still using nothing complicated. I'm still using some of the plugins that I've shown you some of the plugins that I've shown you some of the plugins that I've shown you in earlier videos such as the data view in earlier videos such as the data view in earlier videos such as the data view that allows you to better query your that allows you to better query your that allows you to better query your data across the vault and adjust custom data across the vault and adjust custom data across the vault and adjust custom JavaScripts to better customize the look JavaScripts to better customize the look JavaScripts to better customize the look and feel of your notes. So, of course, and feel of your notes. So, of course, and feel of your notes. So, of course, um AI helps a lot with creating these um AI helps a lot with creating these um AI helps a lot with creating these instructions as well. I might show you a

instructions as well. I might show you a instructions as well. I might show you a few things later on, but here you can few things later on, but here you can few things later on, but here you can see the overview of all of my homelet see the overview of all of my homelet see the overview of all of my homelet nodes. for example, all of the services nodes. for example, all of the services nodes. for example, all of the services that uh I'm managing in my home lab. And that uh I'm managing in my home lab. And that uh I'm managing in my home lab. And then I'm using the new Obsidian base then I'm using the new Obsidian base then I'm using the new Obsidian base tables. So these allow you to add tables. So these allow you to add tables. So these allow you to add different views for tables layouts. You different views for tables layouts. You different views for tables layouts. You can even add canban and calendar. When can even add canban and calendar. When can even add canban and calendar. When you're using the right community you're using the right community you're using the right community plugins, probably I might do another plugins, probably I might do another plugins, probably I might do another video on community plugins updated at video on community plugins updated at video on community plugins updated at some point. If you want to see that, some point. If you want to see that, some point. If you want to see that, please leave me a comment because there please leave me a comment because there please leave me a comment because there are a couple of new community plugins in are a couple of new community plugins in are a couple of new community plugins in a new marketplace format. It's pretty a new marketplace format. It's pretty a new marketplace format. It's pretty cool. And I'm also using this for cool. And I'm also using this for cool. And I'm also using this for developing my own PI agent plug-in as developing my own PI agent plug-in as developing my own PI agent plug-in as well. But again back to my home lab. So well. But again back to my home lab. So well. But again back to my home lab. So here you can see I'm managing all the here you can see I'm managing all the here you can see I'm managing all the services like my authentic uh services like my authentic uh services like my authentic uh repository. Um then checkmk, cloudflare, repository. Um then checkmk, cloudflare, repository. Um then checkmk, cloudflare, cockpit. Of course, I'm also managing cockpit. Of course, I'm also managing cockpit. Of course, I'm also managing front meta sections here that you can front meta sections here that you can front meta sections here that you can use in the tables to filter or better use in the tables to filter or better use in the tables to filter or better visualize things like the environment if visualize things like the environment if visualize things like the environment if it's a testing or production it's a testing or production it's a testing or production environment, the platform, the git environment, the platform, the git environment, the platform, the git repository, then how I'm managing repository, then how I'm managing repository, then how I'm managing secrets in GitLab or locally or any secrets in GitLab or locally or any secrets in GitLab or locally or any other secret uh management platforms other secret uh management platforms other secret uh management platforms priority. I've not really filled in priority. I've not really filled in priority. I've not really filled in this. For example, DNS probably should this. For example, DNS probably should this. For example, DNS probably should become critical because I need DNS for become critical because I need DNS for become critical because I need DNS for almost everything in my home lab. Then almost everything in my home lab. Then almost everything in my home lab. Then of course the status with the of course the status with the of course the status with the description and so on. All the fields in description and so on. All the fields in description and so on. All the fields in the table are basically coming from the table are basically coming from the table are basically coming from these front meta sections that you these front meta sections that you these front meta sections that you define in the free dashes and then you define in the free dashes and then you define in the free dashes and then you can easily just add text in here. It's can easily just add text in here. It's can easily just add text in here. It's super simple and easy. Again, this is

super simple and easy. Again, this is super simple and easy. Again, this is really just plain markdown file and I'm really just plain markdown file and I'm really just plain markdown file and I'm using some community plugins to add using some community plugins to add using some community plugins to add visualizations to these front meta visualizations to these front meta visualizations to these front meta sections or to query certain sections or to query certain sections or to query certain information. I try to add a template information. I try to add a template information. I try to add a template structure for all these things where I structure for all these things where I structure for all these things where I have defined the specific structure how have defined the specific structure how have defined the specific structure how these documents should look like. So these documents should look like. So these documents should look like. So when an AI agent needs to modify a when an AI agent needs to modify a when an AI agent needs to modify a service documentation or when I say hey service documentation or when I say hey service documentation or when I say hey can you create a new service that the AI can you create a new service that the AI can you create a new service that the AI agent has the specific template in the agent has the specific template in the agent has the specific template in the correct format that I want it to be. The correct format that I want it to be. The correct format that I want it to be. The same I'm also managing for same I'm also managing for same I'm also managing for infrastructure. So here you can see all infrastructure. So here you can see all infrastructure. So here you can see all of my physical and virtual of my physical and virtual of my physical and virtual infrastructure machines such as my infrastructure machines such as my infrastructure machines such as my firewall or my Kubernetes clusters, the firewall or my Kubernetes clusters, the firewall or my Kubernetes clusters, the storage server, backup server, um the storage server, backup server, um the storage server, backup server, um the free Proxmox nodes and my virtual free Proxmox nodes and my virtual free Proxmox nodes and my virtual servers and my soft switch. So there's servers and my soft switch. So there's servers and my soft switch. So there's also the IP addresses that I'm tracking also the IP addresses that I'm tracking also the IP addresses that I'm tracking here. The platform, whether this is a VM here. The platform, whether this is a VM here. The platform, whether this is a VM or an appliance or a bare metal, or an appliance or a bare metal, or an appliance or a bare metal, Kubernetes, whatever it is, Kubernetes, whatever it is, Kubernetes, whatever it is, infrastructure type also uh critical. I infrastructure type also uh critical. I infrastructure type also uh critical. I probably should also put the uh Proxmox probably should also put the uh Proxmox probably should also put the uh Proxmox nodes and raise the priority to I'd say nodes and raise the priority to I'd say nodes and raise the priority to I'd say high. It's it's not really not high. It's it's not really not high. It's it's not really not everything is super critical, but these everything is super critical, but these everything is super critical, but these three nodes are quite important for my three nodes are quite important for my three nodes are quite important for my lab. Also, IPAM is quite interesting. So lab. Also, IPAM is quite interesting. So lab. Also, IPAM is quite interesting. So here I've documented all my networks, my here I've documented all my networks, my here I've documented all my networks, my VLANs, all the IP addresses for all of VLANs, all the IP addresses for all of VLANs, all the IP addresses for all of my servers so that when I'm searching my my servers so that when I'm searching my my servers so that when I'm searching my home lab for IP addresses, I have a good home lab for IP addresses, I have a good home lab for IP addresses, I have a good understanding of where things are understanding of where things are understanding of where things are located and what IP addresses are used.

located and what IP addresses are used. located and what IP addresses are used. But yeah, I don't think I need to show But yeah, I don't think I need to show But yeah, I don't think I need to show you each and everything. Let me show you you each and everything. Let me show you you each and everything. Let me show you a few examples how AI can help me a few examples how AI can help me a few examples how AI can help me managing this. I've just opened the managing this. I've just opened the managing this. I've just opened the directory where my Obsidian vault is. So directory where my Obsidian vault is. So directory where my Obsidian vault is. So where all the markdown files are simply where all the markdown files are simply where all the markdown files are simply stored. So for example, when I do an LS stored. So for example, when I do an LS stored. So for example, when I do an LS home lab, you can see these are just the home lab, you can see these are just the home lab, you can see these are just the plain markdown files and directories in plain markdown files and directories in plain markdown files and directories in the vault. And when I now open the PI the vault. And when I now open the PI the vault. And when I now open the PI agent in this vault, the agent now has agent in this vault, the agent now has agent in this vault, the agent now has direct access to all the markdown files. direct access to all the markdown files. direct access to all the markdown files. And by the way, if you have not watched And by the way, if you have not watched And by the way, if you have not watched my PI agent video, I will link you this my PI agent video, I will link you this my PI agent video, I will link you this in the description. So that tells you in the description. So that tells you in the description. So that tells you how to use this tool. It's a super slim how to use this tool. It's a super slim how to use this tool. It's a super slim and tiny and fast AI agent that runs and tiny and fast AI agent that runs and tiny and fast AI agent that runs directly in your terminal. And this does directly in your terminal. And this does directly in your terminal. And this does not really use any complicated MCP not really use any complicated MCP not really use any complicated MCP servers or a bunch of different tools. servers or a bunch of different tools. servers or a bunch of different tools. It really has some core utilities like It really has some core utilities like It really has some core utilities like reading writing files running ls grap reading writing files running ls grap reading writing files running ls grap and executing uh shell commands and and executing uh shell commands and and executing uh shell commands and stuff like that. And it does not pop up stuff like that. And it does not pop up stuff like that. And it does not pop up any permission questions or something any permission questions or something any permission questions or something like this which can be at some point a like this which can be at some point a like this which can be at some point a bit destructive. So you have to be aware bit destructive. So you have to be aware bit destructive. So you have to be aware of if you say hey please destroy my of if you say hey please destroy my of if you say hey please destroy my entire infrastructure it probably will entire infrastructure it probably will entire infrastructure it probably will do this without asking for another do this without asking for another do this without asking for another confirmation. So be aware of this. But confirmation. So be aware of this. But confirmation. So be aware of this. But at the same time, this is also super at the same time, this is also super at the same time, this is also super super fast and it can work with your super fast and it can work with your super fast and it can work with your existing uh codeex or chat GPD plus and existing uh codeex or chat GPD plus and existing uh codeex or chat GPD plus and pro subscription, GitHub copilot pro subscription, GitHub copilot pro subscription, GitHub copilot subscription. It also supports local subscription. It also supports local subscription. It also supports local LLMs if you want. When you're using LLMs if you want. When you're using LLMs if you want. When you're using local LLMs processed at your machine, local LLMs processed at your machine, local LLMs processed at your machine, there's literally no connection to any there's literally no connection to any there's literally no connection to any external service at all. So you can

external service at all. So you can external service at all. So you can really keep your entire structure, your really keep your entire structure, your really keep your entire structure, your notes, your documents under your notes, your documents under your notes, your documents under your control, which is pretty cool. So for control, which is pretty cool. So for control, which is pretty cool. So for example um I can easily attach any files example um I can easily attach any files example um I can easily attach any files as context for example home lab DNS and as context for example home lab DNS and as context for example home lab DNS and I say describe how to manage DNS. If you I say describe how to manage DNS. If you I say describe how to manage DNS. If you would ask an AI agent like chat GPT would ask an AI agent like chat GPT would ask an AI agent like chat GPT without any context this question it without any context this question it without any context this question it would give you a general answer but now would give you a general answer but now would give you a general answer but now because I attached my local because I attached my local because I attached my local documentation it comes back with useful documentation it comes back with useful documentation it comes back with useful information. This is a way how you can information. This is a way how you can information. This is a way how you can easily work with markdown files and AI easily work with markdown files and AI easily work with markdown files and AI agents even in your terminal. And for agents even in your terminal. And for agents even in your terminal. And for example, I can also do the same with example, I can also do the same with example, I can also do the same with home lab IPAMS. So when I want to know home lab IPAMS. So when I want to know home lab IPAMS. So when I want to know what's the next free IP address in my what's the next free IP address in my what's the next free IP address in my home lab for production VMs. So let's home lab for production VMs. So let's home lab for production VMs. So let's see what this comes up with. It see what this comes up with. It see what this comes up with. It immediately finds the IPAM file here. immediately finds the IPAM file here. immediately finds the IPAM file here. You can see it even uses simple inline You can see it even uses simple inline You can see it even uses simple inline Python scripts to do some data Python scripts to do some data Python scripts to do some data transformation. You can see how fast transformation. You can see how fast transformation. You can see how fast this is because it really is just an this is because it really is just an this is because it really is just an inline simple script. And now it comes inline simple script. And now it comes inline simple script. And now it comes back with useful information. The lowest back with useful information. The lowest back with useful information. The lowest currently unassigned IP address is currently unassigned IP address is currently unassigned IP address is 1020.7.

1020.7. 1020.7. So yeah, this is still free, but it So yeah, this is still free, but it So yeah, this is still free, but it appears to be the next free documented appears to be the next free documented appears to be the next free documented IP. If you prefer continuing newer VM IP. If you prefer continuing newer VM IP. If you prefer continuing newer VM ranges, then use 10202. ranges, then use 10202. ranges, then use 10202. So really cool. And this is how fast I So really cool. And this is how fast I So really cool. And this is how fast I can get this information. Of course, I can get this information. Of course, I can get this information. Of course, I can also go further down this rabbit can also go further down this rabbit can also go further down this rabbit hole and say, "Okay, uh please can you hole and say, "Okay, uh please can you hole and say, "Okay, uh please can you um reserve this for a new server that um reserve this for a new server that um reserve this for a new server that I'm installing or whatnot." And this is I'm installing or whatnot." And this is I'm installing or whatnot." And this is the way how I keep my documentation up the way how I keep my documentation up the way how I keep my documentation up to date. And this is much better than to date. And this is much better than to date. And this is much better than using notion with their own AI using notion with their own AI using notion with their own AI subscriptions or whatever. I can use the subscriptions or whatever. I can use the subscriptions or whatever. I can use the tools that I'm already using for tools that I'm already using for tools that I'm already using for managing my infrastructure or for managing my infrastructure or for managing my infrastructure or for developing code or for any other daily developing code or for any other daily developing code or for any other daily workflow. That is the power of markdown workflow. That is the power of markdown workflow. That is the power of markdown files. So I absolutely love this. But files. So I absolutely love this. But files. So I absolutely love this. But this is only half of the story because this is only half of the story because this is only half of the story because I've also created something that is I've also created something that is I've also created something that is called skills and prompts. So a skill is called skills and prompts. So a skill is called skills and prompts. So a skill is basically a small reusable instruction basically a small reusable instruction basically a small reusable instruction package for agents. In PI, for example, package for agents. In PI, for example, package for agents. In PI, for example, this is usually stored in a a skill this is usually stored in a a skill this is usually stored in a a skill directory with a simple skill.md directory with a simple skill.md directory with a simple skill.md markdown file. I've also created a sim markdown file. I've also created a sim markdown file. I've also created a sim link from my Obsidian vault to the PI link from my Obsidian vault to the PI link from my Obsidian vault to the PI agent skill directory so that I keep my agent skill directory so that I keep my agent skill directory so that I keep my prompts and my skills. managed in prompts and my skills. managed in prompts and my skills. managed in Obsidian. So I can easily update all the Obsidian. So I can easily update all the Obsidian. So I can easily update all the skills that I've written. PI agent has skills that I've written. PI agent has skills that I've written. PI agent has direct access to this no matter if it's direct access to this no matter if it's direct access to this no matter if it's running in Obsidian via my plug-in or in running in Obsidian via my plug-in or in running in Obsidian via my plug-in or in the terminal. So this is pretty cool. I the terminal. So this is pretty cool. I the terminal. So this is pretty cool. I just started Pi without skills because just started Pi without skills because just started Pi without skills because otherwise it would use it automatically.

otherwise it would use it automatically. otherwise it would use it automatically. But uh when I just run pi with skills, But uh when I just run pi with skills, But uh when I just run pi with skills, it immediately knows how to connect to it immediately knows how to connect to it immediately knows how to connect to my Apple mail, my Apple reminders, my my Apple mail, my Apple reminders, my my Apple mail, my Apple reminders, my boiler plate CLI. And here is a skill boiler plate CLI. And here is a skill boiler plate CLI. And here is a skill that I have specifically written for my that I have specifically written for my that I have specifically written for my home lab. So when you go in this skill home lab. So when you go in this skill home lab. So when you go in this skill file, you can see that it starts with file, you can see that it starts with file, you can see that it starts with two front meta sections, a name and the two front meta sections, a name and the two front meta sections, a name and the description is used by the agent to find description is used by the agent to find description is used by the agent to find out whether to use this skill. So out whether to use this skill. So out whether to use this skill. So sometimes it can find it out by itself sometimes it can find it out by itself sometimes it can find it out by itself depending on the description and if that depending on the description and if that depending on the description and if that probably has something to do with the probably has something to do with the probably has something to do with the task that you want it to do. I will task that you want it to do. I will task that you want it to do. I will later demonstrate how exactly that later demonstrate how exactly that later demonstrate how exactly that works. But here you can see uh I've works. But here you can see uh I've works. But here you can see uh I've described uh what is my home lab um how described uh what is my home lab um how described uh what is my home lab um how I'm managing things uh that I'm using a I'm managing things uh that I'm using a I'm managing things uh that I'm using a hypervisor behind firewall I'm using hypervisor behind firewall I'm using hypervisor behind firewall I'm using infrastructure code I'm using CI/CD infrastructure code I'm using CI/CD infrastructure code I'm using CI/CD repositories internal DNS proxies and repositories internal DNS proxies and repositories internal DNS proxies and monitoring logging so I try to again monitoring logging so I try to again monitoring logging so I try to again document as much as possible in a clean document as much as possible in a clean document as much as possible in a clean and specific format so these are very and specific format so these are very and specific format so these are very precise instructions for the AI agent so precise instructions for the AI agent so precise instructions for the AI agent so that it knows where is the obsidian that it knows where is the obsidian that it knows where is the obsidian vault where I'm managing my home How can vault where I'm managing my home How can vault where I'm managing my home How can I access this? So that it knows about I access this? So that it knows about I access this? So that it knows about the different sections and can use the the different sections and can use the the different sections and can use the templates. For example, project templates. For example, project templates. For example, project management is happening there and I'm management is happening there and I'm management is happening there and I'm using it for this and that and this is a using it for this and that and this is a using it for this and that and this is a template. Here's here are also rules and template. Here's here are also rules and template. Here's here are also rules and guidelines like do not duplicate service guidelines like do not duplicate service guidelines like do not duplicate service specific deployments or resources in the specific deployments or resources in the specific deployments or resources in the details in the skill. Keep those details details in the skill. Keep those details details in the skill. Keep those details in the service node or for operational

in the service node or for operational in the service node or for operational tasks. Start with irrelevant root tasks. Start with irrelevant root tasks. Start with irrelevant root documentation. Then inspect the owning documentation. Then inspect the owning documentation. Then inspect the owning repository if needed. Preferred repository if needed. Preferred repository if needed. Preferred documented infrastructure as code documented infrastructure as code documented infrastructure as code workflows over ad hoc runtime changes. workflows over ad hoc runtime changes. workflows over ad hoc runtime changes. So that means that the AI agent when So that means that the AI agent when So that means that the AI agent when reading this knows that I want most of reading this knows that I want most of reading this knows that I want most of my infrastructure and resources to be my infrastructure and resources to be my infrastructure and resources to be managed in IA such as Terraform, open managed in IA such as Terraform, open managed in IA such as Terraform, open tofu and so on. most of my important tofu and so on. most of my important tofu and so on. most of my important repositories like GitLab DNS super repositories like GitLab DNS super repositories like GitLab DNS super important proxfana important proxfana important proxfana Prometheus traffic and so on and yeah Prometheus traffic and so on and yeah Prometheus traffic and so on and yeah repositories may but do not necessarily repositories may but do not necessarily repositories may but do not necessarily have to contain the following have to contain the following have to contain the following subdirectories. that tells the AI agent subdirectories. that tells the AI agent subdirectories. that tells the AI agent that whenever I need Anible playbooks that whenever I need Anible playbooks that whenever I need Anible playbooks for a service to manage this in its own for a service to manage this in its own for a service to manage this in its own directory. Okay, so let's go back to my directory. Okay, so let's go back to my directory. Okay, so let's go back to my terminal. Let me show you what happens. terminal. Let me show you what happens. terminal. Let me show you what happens. For example, when I say slashkill For example, when I say slashkill For example, when I say slashkill homelab and say, I'd love to create a homelab and say, I'd love to create a homelab and say, I'd love to create a new VM in my home lab. Show me an new VM in my home lab. Show me an new VM in my home lab. Show me an onboarding plan plus checklist. And onboarding plan plus checklist. And onboarding plan plus checklist. And let's see what comes back with. Of let's see what comes back with. Of let's see what comes back with. Of course, I have not written a specific course, I have not written a specific course, I have not written a specific onboarding guide, so the result might onboarding guide, so the result might onboarding guide, so the result might vary from time to time, but you could vary from time to time, but you could vary from time to time, but you could technically do this. If you want a technically do this. If you want a technically do this. If you want a specific format for an onboarding guide, specific format for an onboarding guide, specific format for an onboarding guide, just write a template and define that in just write a template and define that in just write a template and define that in the skill. But here you can see that the the skill. But here you can see that the the skill. But here you can see that the AI agent by reading the HomeLab skill AI agent by reading the HomeLab skill AI agent by reading the HomeLab skill knows exactly where to find the relevant knows exactly where to find the relevant knows exactly where to find the relevant information. So, it reads the Proxmox information. So, it reads the Proxmox information. So, it reads the Proxmox documentation, then it knows how I'm documentation, then it knows how I'm documentation, then it knows how I'm naming things. For example, server pro

naming things. For example, server pro naming things. For example, server pro two, app test one, kubernetes pro one, two, app test one, kubernetes pro one, two, app test one, kubernetes pro one, worker, etc. It knows where my proxmox worker, etc. It knows where my proxmox worker, etc. It knows where my proxmox nodes are. So, pick a proxmbox node nodes are. So, pick a proxmbox node nodes are. So, pick a proxmbox node where to run this capacity planning where to run this capacity planning where to run this capacity planning IPAM. So, it knows choose a VLAN subset IPAM. So, it knows choose a VLAN subset IPAM. So, it knows choose a VLAN subset from the IPAM documentation, a from the IPAM documentation, a from the IPAM documentation, a slack.home.ccrave.de slack.home.ccrave.de slack.home.ccrave.de provisioning approach, monitoring, provisioning approach, monitoring, provisioning approach, monitoring, backup restore documentation. So it backup restore documentation. So it backup restore documentation. So it knows that at the end of the day it knows that at the end of the day it knows that at the end of the day it needs to create a documentation under needs to create a documentation under needs to create a documentation under infrastructure slack.mmd with the infrastructure slack.mmd with the infrastructure slack.mmd with the specific front meta section so that I specific front meta section so that I specific front meta section so that I immediately can see that in the table immediately can see that in the table immediately can see that in the table here when I say hey please create this. here when I say hey please create this. here when I say hey please create this. I'm not doing this right now but these I'm not doing this right now but these I'm not doing this right now but these are all the things that it knows from are all the things that it knows from are all the things that it knows from the skill and I can know go on and say the skill and I can know go on and say the skill and I can know go on and say um how and where would you manage th um how and where would you manage th um how and where would you manage th this VM to give me a better this VM to give me a better this VM to give me a better recommendation how exactly to do this. recommendation how exactly to do this. recommendation how exactly to do this. So here I would primarily manage this So here I would primarily manage this So here I would primarily manage this through the homelab prox repository with through the homelab prox repository with through the homelab prox repository with documentation in the obsidian vault. documentation in the obsidian vault. documentation in the obsidian vault. Here it describes the life cycle VM Here it describes the life cycle VM Here it describes the life cycle VM creation documentation. So these are the creation documentation. So these are the creation documentation. So these are the front meta fields the IP address front meta fields the IP address front meta fields the IP address monitoring that should be included such monitoring that should be included such monitoring that should be included such as Prometheus and Graphana. I can also as Prometheus and Graphana. I can also as Prometheus and Graphana. I can also say show me the IA files that you would say show me the IA files that you would say show me the IA files that you would create for me and ask if you need create for me and ask if you need create for me and ask if you need additional information. So this now additional information. So this now additional information. So this now should come hopefully come back with should come hopefully come back with should come hopefully come back with Terraform files and yeah so that's Terraform files and yeah so that's Terraform files and yeah so that's exactly what's happening. It uses exactly what's happening. It uses exactly what's happening. It uses Proxmox VMQ Proxmox VMQ Proxmox VMQ so the Proxmox terraform provider that

so the Proxmox terraform provider that so the Proxmox terraform provider that allows me to manage a new VM in allows me to manage a new VM in allows me to manage a new VM in infrastructure is code. It also found infrastructure is code. It also found infrastructure is code. It also found the Ubuntu template because that's my the Ubuntu template because that's my the Ubuntu template because that's my preferred operating system. here the preferred operating system. here the preferred operating system. here the description um memory and so on a DNS description um memory and so on a DNS description um memory and so on a DNS record so it knows to automatically record so it knows to automatically record so it knows to automatically create a DNS record and how to create create a DNS record and how to create create a DNS record and how to create this and netbox resource if needed. this and netbox resource if needed. this and netbox resource if needed. Yeah, because I'm managing some of my Yeah, because I'm managing some of my Yeah, because I'm managing some of my IPAM documentation and machine documents IPAM documentation and machine documents IPAM documentation and machine documents in Netbox add obsibilian infrastructure in Netbox add obsibilian infrastructure in Netbox add obsibilian infrastructure nodes information I need from you. So I nodes information I need from you. So I nodes information I need from you. So I need to tell it the VM slack the need to tell it the VM slack the need to tell it the VM slack the description environment and so on. Of description environment and so on. Of description environment and so on. Of course, I could now go on and ask, "Hey, course, I could now go on and ask, "Hey, course, I could now go on and ask, "Hey, give me some uh template for this or give me some uh template for this or give me some uh template for this or what would you add as a name or IP what would you add as a name or IP what would you add as a name or IP address? What is the next free IP address? What is the next free IP address? What is the next free IP address? Pick on your own or whatever." address? Pick on your own or whatever." address? Pick on your own or whatever." So, this is really different from a So, this is really different from a So, this is really different from a generic AI task like how do I generally generic AI task like how do I generally generic AI task like how do I generally create a virtual machine? Then it would create a virtual machine? Then it would create a virtual machine? Then it would come back with Microsoft HyperV, VMware, come back with Microsoft HyperV, VMware, come back with Microsoft HyperV, VMware, whatever. And of course, I've just used whatever. And of course, I've just used whatever. And of course, I've just used chat to demonstrate this. But I could chat to demonstrate this. But I could chat to demonstrate this. But I could easily just go on and say create this. easily just go on and say create this. easily just go on and say create this. Then the Pi agent would connect to my Then the Pi agent would connect to my Then the Pi agent would connect to my Git repository. It would use the Git repository. It would use the Git repository. It would use the credentials that I've added to the credentials that I've added to the credentials that I've added to the system and it could easily create a new system and it could easily create a new system and it could easily create a new git repo deploy the machine. I've git repo deploy the machine. I've git repo deploy the machine. I've recently asked it to create some recently asked it to create some recently asked it to create some services like I don't know Arcane for services like I don't know Arcane for services like I don't know Arcane for example is a docker manager that you example is a docker manager that you example is a docker manager that you will see in one of my future videos and will see in one of my future videos and will see in one of my future videos and I just said hey please can you create an I just said hey please can you create an I just said hey please can you create an arcane production environment in my home arcane production environment in my home arcane production environment in my home lab and it literally had done it the way lab and it literally had done it the way lab and it literally had done it the way I want. So this is really absolutely I want. So this is really absolutely I want. So this is really absolutely amazing. One tip that I also want to

amazing. One tip that I also want to amazing. One tip that I also want to give you with this, there's one give you with this, there's one give you with this, there's one challenge that markdown files have where challenge that markdown files have where challenge that markdown files have where systems like notion still have systems like notion still have systems like notion still have advantages and this is when you start to advantages and this is when you start to advantages and this is when you start to manage entities or items in front manage entities or items in front manage entities or items in front matter. For example, here you can see I matter. For example, here you can see I matter. For example, here you can see I have many many different services that have many many different services that have many many different services that can be filtered by environment, can be filtered by environment, can be filtered by environment, platform, repository and so on. But platform, repository and so on. But platform, repository and so on. But markdown technically it needs to go markdown technically it needs to go markdown technically it needs to go through all the files, read these files through all the files, read these files through all the files, read these files and find the information you're looking and find the information you're looking and find the information you're looking for. And I just found out there's a for. And I just found out there's a for. And I just found out there's a super useful tool that helps with super useful tool that helps with super useful tool that helps with querying markdown by front matter. This querying markdown by front matter. This querying markdown by front matter. This is called MDBase query. MDBase query is is called MDBase query. MDBase query is is called MDBase query. MDBase query is absolutely amazing. It's a simple tool absolutely amazing. It's a simple tool absolutely amazing. It's a simple tool that is an obsidian basis compatible that is an obsidian basis compatible that is an obsidian basis compatible query engine for markdown vault. So you query engine for markdown vault. So you query engine for markdown vault. So you can easily just install and run this can easily just install and run this can easily just install and run this small CLI program in the terminal and small CLI program in the terminal and small CLI program in the terminal and then use this to query for specific then use this to query for specific then use this to query for specific front meta sections, filter by date, front meta sections, filter by date, front meta sections, filter by date, filter by names and output different filter by names and output different filter by names and output different fields in formats. Pretty similar to fields in formats. Pretty similar to fields in formats. Pretty similar to something like SQL queries. For example, something like SQL queries. For example, something like SQL queries. For example, when I want to query my home lab when I want to query my home lab when I want to query my home lab projects and filter it by type is equal projects and filter it by type is equal projects and filter it by type is equal project and repository platform is project and repository platform is project and repository platform is GitHub. So that it shows me all the GitHub. So that it shows me all the GitHub. So that it shows me all the projects with these specific values and projects with these specific values and projects with these specific values and it's selecting the file path. It's it's selecting the file path. It's it's selecting the file path. It's selecting the status and the repository selecting the status and the repository selecting the status and the repository and the output format is markdown and and the output format is markdown and and the output format is markdown and limit by 8 results. So then I quickly limit by 8 results. So then I quickly limit by 8 results. So then I quickly get back this table. You can also add get back this table. You can also add get back this table. You can also add different outputs for example CSV here.

different outputs for example CSV here. different outputs for example CSV here. So this is really amazing and what I So this is really amazing and what I So this is really amazing and what I have done is when we go back to obsidian have done is when we go back to obsidian have done is when we go back to obsidian and open my obsidian skill. So here I and open my obsidian skill. So here I and open my obsidian skill. So here I have described how to work with walls, have described how to work with walls, have described how to work with walls, how the markdown files are structured, how the markdown files are structured, how the markdown files are structured, how obsidian links are structured. So how obsidian links are structured. So how obsidian links are structured. So the back links and how this is working the back links and how this is working the back links and how this is working and about querying. So here I've added and about querying. So here I've added and about querying. So here I've added an instructions to say use MDB query for an instructions to say use MDB query for an instructions to say use MDB query for searching, filtering and reporting on searching, filtering and reporting on searching, filtering and reporting on nodes by front matter properties instead nodes by front matter properties instead nodes by front matter properties instead of using find or rip grap etc. Here's an of using find or rip grap etc. Here's an of using find or rip grap etc. Here's an example. Here are some useful filter example. Here are some useful filter example. Here are some useful filter examples such as publish dates is empty examples such as publish dates is empty examples such as publish dates is empty functions and stuff like that. And this functions and stuff like that. And this functions and stuff like that. And this way my agent can also query information way my agent can also query information way my agent can also query information from my wallet in a super easy and from my wallet in a super easy and from my wallet in a super easy and simple way. So, for example, when I go simple way. So, for example, when I go simple way. So, for example, when I go back to pi, I probably just have to back to pi, I probably just have to back to pi, I probably just have to raise the thinking level a bit because raise the thinking level a bit because raise the thinking level a bit because sometimes the AI agent with low thinking sometimes the AI agent with low thinking sometimes the AI agent with low thinking mode does not use multiple skills at the mode does not use multiple skills at the mode does not use multiple skills at the same time. But when you increase the same time. But when you increase the same time. But when you increase the thinking level, you can also say thinking level, you can also say thinking level, you can also say questions like, for example, list all my questions like, for example, list all my questions like, for example, list all my home lab services where I'm managing home lab services where I'm managing home lab services where I'm managing secrets locally. So, let's see what secrets locally. So, let's see what secrets locally. So, let's see what happens. It's a pretty general question, happens. It's a pretty general question, happens. It's a pretty general question, but because the agent has access to the but because the agent has access to the but because the agent has access to the different skills, it knows, okay, for different skills, it knows, okay, for different skills, it knows, okay, for inspecting HomeLab services that is inspecting HomeLab services that is inspecting HomeLab services that is stored in an obsidian vault, I probably stored in an obsidian vault, I probably stored in an obsidian vault, I probably should use both of these skills. So, you should use both of these skills. So, you should use both of these skills. So, you can see I've not really specified the can see I've not really specified the can see I've not really specified the skills. It just knows that by the

skills. It just knows that by the skills. It just knows that by the descriptions descriptions descriptions and then because it uses the obsidian and then because it uses the obsidian and then because it uses the obsidian skill, it knows that it needs to use an skill, it knows that it needs to use an skill, it knows that it needs to use an MDB query and query the front meta MDB query and query the front meta MDB query and query the front meta section specifically for my secret section specifically for my secret section specifically for my secret management local because I have defined management local because I have defined management local because I have defined that in the template and in the skill that in the template and in the skill that in the template and in the skill and then you can see it quickly comes and then you can see it quickly comes and then you can see it quickly comes back with the result here and then I can back with the result here and then I can back with the result here and then I can go on and say hey can you transform the go on and say hey can you transform the go on and say hey can you transform the secrets add gitlab secrets for these secrets add gitlab secrets for these secrets add gitlab secrets for these repository repository repository or whatever. So, this is really nice. By or whatever. So, this is really nice. By or whatever. So, this is really nice. By the way, there's one thing that I want the way, there's one thing that I want the way, there's one thing that I want to say. If you are familiar with my to say. If you are familiar with my to say. If you are familiar with my boiler plates project and my template boiler plates project and my template boiler plates project and my template library, I will also put you a link to library, I will also put you a link to library, I will also put you a link to this in the description below. But here this in the description below. But here this in the description below. But here is the library where I'm managing all is the library where I'm managing all is the library where I'm managing all these templates. I'm planning to add a these templates. I'm planning to add a these templates. I'm planning to add a new static format so that we can share new static format so that we can share new static format so that we can share templates for something like skills for templates for something like skills for templates for something like skills for example because I thought when I have example because I thought when I have example because I thought when I have taken the time to write all this stuff taken the time to write all this stuff taken the time to write all this stuff maybe I should share it with you. For maybe I should share it with you. For maybe I should share it with you. For example, the obsidian skill is something example, the obsidian skill is something example, the obsidian skill is something pretty general and we can improve this pretty general and we can improve this pretty general and we can improve this as a community. Um so why not just add as a community. Um so why not just add as a community. Um so why not just add some boiler plates for these skills here some boiler plates for these skills here some boiler plates for these skills here might be useful. So I'll start that might be useful. So I'll start that might be useful. So I'll start that somewhere in the next week. So you need somewhere in the next week. So you need somewhere in the next week. So you need to give me some time to prepare it. But to give me some time to prepare it. But to give me some time to prepare it. But I think we could benefit a lot from this I think we could benefit a lot from this I think we could benefit a lot from this as a community. Again, you do not have as a community. Again, you do not have as a community. Again, you do not have to actually use Obsidian. You could also to actually use Obsidian. You could also to actually use Obsidian. You could also use this with markdown files in your use this with markdown files in your use this with markdown files in your terminal agent and so on. But yeah, terminal agent and so on. But yeah, terminal agent and so on. But yeah, let's let's come to the next point.

let's let's come to the next point. let's let's come to the next point. Currently, I've just shown you how to Currently, I've just shown you how to Currently, I've just shown you how to use PI agent in the terminal. But I also use PI agent in the terminal. But I also use PI agent in the terminal. But I also had the thought, well, when PI agent is had the thought, well, when PI agent is had the thought, well, when PI agent is so amazing, I'm using this for almost so amazing, I'm using this for almost so amazing, I'm using this for almost everything. Could I not just integrate everything. Could I not just integrate everything. Could I not just integrate the PI agent in the Obsidian app itself? the PI agent in the Obsidian app itself? the PI agent in the Obsidian app itself? So, what I have created is a community So, what I have created is a community So, what I have created is a community plug-in called PI agent. Of course, I plug-in called PI agent. Of course, I plug-in called PI agent. Of course, I will link you this in the description will link you this in the description will link you this in the description and this is a plugin that you can easily and this is a plugin that you can easily and this is a plugin that you can easily add to your Obsidian and then you can add to your Obsidian and then you can add to your Obsidian and then you can use the PI agent CLI to work with this. use the PI agent CLI to work with this. use the PI agent CLI to work with this. There are some requirements. You need to There are some requirements. You need to There are some requirements. You need to install PI agent, the PI coding agent install PI agent, the PI coding agent install PI agent, the PI coding agent locally on your device. Of course, I've locally on your device. Of course, I've locally on your device. Of course, I've also made this open source so you can also made this open source so you can also made this open source so you can easily find this project on GitHub. easily find this project on GitHub. easily find this project on GitHub. However, I also want to say this and be However, I also want to say this and be However, I also want to say this and be completely transparent with you. I've completely transparent with you. I've completely transparent with you. I've not touched a single line of code in not touched a single line of code in not touched a single line of code in this repository. It is 100% vip coded. this repository. It is 100% vip coded. this repository. It is 100% vip coded. I'm honest with you guys. But I'm honest with you guys. But I'm honest with you guys. But surprisingly, it works pretty well. So, surprisingly, it works pretty well. So, surprisingly, it works pretty well. So, um I was not running into major issues. um I was not running into major issues. um I was not running into major issues. Of course, a couple of things here and Of course, a couple of things here and Of course, a couple of things here and there, but I was really surprised how there, but I was really surprised how there, but I was really surprised how well the PI agent could write code for well the PI agent could write code for well the PI agent could write code for this and create the plugin. It's it's this and create the plugin. It's it's this and create the plugin. It's it's absolutely amazing. So once you install absolutely amazing. So once you install absolutely amazing. So once you install the plugin you have access to it. You the plugin you have access to it. You the plugin you have access to it. You should see that icon here on the left should see that icon here on the left should see that icon here on the left side open PI agent and then you have a side open PI agent and then you have a side open PI agent and then you have a chat window and in this chat window for chat window and in this chat window for chat window and in this chat window for example here you can switch the model.

example here you can switch the model. example here you can switch the model. When you connect PI and log in in your When you connect PI and log in in your When you connect PI and log in in your terminal the Obsidian PI agent can query terminal the Obsidian PI agent can query terminal the Obsidian PI agent can query all the models that you have added. And all the models that you have added. And all the models that you have added. And I have created three different modes I have created three different modes I have created three different modes here. Um let's quickly go to the here. Um let's quickly go to the here. Um let's quickly go to the settings and I can show you this better. settings and I can show you this better. settings and I can show you this better. So here in the settings there is a So here in the settings there is a So here in the settings there is a default thinking level and here there default thinking level and here there default thinking level and here there are four different tool modes. So these are four different tool modes. So these are four different tool modes. So these settings basically just enable tool settings basically just enable tool settings basically just enable tool calls in the PI agent. Chat has no CLI calls in the PI agent. Chat has no CLI calls in the PI agent. Chat has no CLI tools. Review has read search list only. tools. Review has read search list only. tools. Review has read search list only. Then editing edit write but no shell Then editing edit write but no shell Then editing edit write but no shell commands in full agent can even execute commands in full agent can even execute commands in full agent can even execute shell commands in the background. You shell commands in the background. You shell commands in the background. You can also add custom instructions here. can also add custom instructions here. can also add custom instructions here. You can change the pi executable path. You can change the pi executable path. You can change the pi executable path. And then there's also a skill directory. And then there's also a skill directory. And then there's also a skill directory. So you can include default PI skills or So you can include default PI skills or So you can include default PI skills or specify a skills folder. I probably will specify a skills folder. I probably will specify a skills folder. I probably will also add prompts to this. Currently also add prompts to this. Currently also add prompts to this. Currently prompts are missing. But um yeah, I will prompts are missing. But um yeah, I will prompts are missing. But um yeah, I will probably release a new version with probably release a new version with probably release a new version with prompts that should not be too difficult prompts that should not be too difficult prompts that should not be too difficult to add. The cool thing is that it to add. The cool thing is that it to add. The cool thing is that it automatically adds the current file as automatically adds the current file as automatically adds the current file as context. You can see it here. So context. You can see it here. So context. You can see it here. So whatever file you have opened will whatever file you have opened will whatever file you have opened will automatically be sent with the query. So automatically be sent with the query. So automatically be sent with the query. So then you can see the agent is starting then you can see the agent is starting then you can see the agent is starting to think. Currently we have no to think. Currently we have no to think. Currently we have no visibility what exactly the agent is visibility what exactly the agent is visibility what exactly the agent is thinking or what it's doing in the thinking or what it's doing in the thinking or what it's doing in the background. I probably need to update background. I probably need to update background. I probably need to update this at at some point. But here you can this at at some point. But here you can this at at some point. But here you can see the summary is being created and see the summary is being created and see the summary is being created and once it is finished it will once it is finished it will once it is finished it will automatically be converted to markdown automatically be converted to markdown automatically be converted to markdown rendered files. You can also cue rendered files. You can also cue rendered files. You can also cue messages here. For example, summarize messages here. For example, summarize messages here. For example, summarize your changes. When the agent is still

your changes. When the agent is still your changes. When the agent is still running and you send a new message, it running and you send a new message, it running and you send a new message, it will be cute here and then execute it will be cute here and then execute it will be cute here and then execute it automatically when the agent is done. Of automatically when the agent is done. Of automatically when the agent is done. Of course, you can also cancel this action course, you can also cancel this action course, you can also cancel this action here. Remove cute message or cancel the here. Remove cute message or cancel the here. Remove cute message or cancel the entire processing. And I'm also entire processing. And I'm also entire processing. And I'm also supporting threats here. So you can fork supporting threats here. So you can fork supporting threats here. So you can fork the message, you can create a new chat the message, you can create a new chat the message, you can create a new chat and you can change the name of the chat. and you can change the name of the chat. and you can change the name of the chat. So for example, demonstration. So here So for example, demonstration. So here So for example, demonstration. So here for example, you can see all the the for example, you can see all the the for example, you can see all the the chat history and you can see what is chat history and you can see what is chat history and you can see what is currently running. It also supports currently running. It also supports currently running. It also supports multi- threading. So you can easily just multi- threading. So you can easily just multi- threading. So you can easily just let one thread be processed and then add let one thread be processed and then add let one thread be processed and then add a second one and work in that in a second one and work in that in a second one and work in that in parallel. You can also add some threads parallel. You can also add some threads parallel. You can also add some threads to favorites. So then they are added on to favorites. So then they are added on to favorites. So then they are added on top of this. I mostly use this when I'm top of this. I mostly use this when I'm top of this. I mostly use this when I'm researching YouTube videos and topics. researching YouTube videos and topics. researching YouTube videos and topics. So um that is pretty useful because I'm So um that is pretty useful because I'm So um that is pretty useful because I'm not only managing my home app here in not only managing my home app here in not only managing my home app here in Obsidian. I'm also managing my YouTube Obsidian. I'm also managing my YouTube Obsidian. I'm also managing my YouTube videos and all other projects for work videos and all other projects for work videos and all other projects for work and private stuff. And then the PI agent and private stuff. And then the PI agent and private stuff. And then the PI agent is super helpful with this. Again, if is super helpful with this. Again, if is super helpful with this. Again, if you want to test this project, go to the you want to test this project, go to the you want to test this project, go to the Obsidian marketplace, add it to Obsidian marketplace, add it to Obsidian marketplace, add it to Obsidian. Feel free to contribute fork Obsidian. Feel free to contribute fork Obsidian. Feel free to contribute fork this project or help improving this with this project or help improving this with this project or help improving this with buck reports and stuff like this. So buck reports and stuff like this. So buck reports and stuff like this. So that's pretty cool project. And again, that's pretty cool project. And again, that's pretty cool project. And again, it's completely viodated. There may be it's completely viodated. There may be it's completely viodated. There may be the one or another issue. So use it with the one or another issue. So use it with the one or another issue. So use it with caution, but um I think it could be caution, but um I think it could be caution, but um I think it could be super valuable for us. So yeah, these super valuable for us. So yeah, these super valuable for us. So yeah, these are basically all of the reasons why I are basically all of the reasons why I are basically all of the reasons why I switched back to Obsidian. As you can switched back to Obsidian. As you can switched back to Obsidian. As you can see, uh I have much more control of the see, uh I have much more control of the see, uh I have much more control of the data, how it's presented, what AI data, how it's presented, what AI data, how it's presented, what AI workloads I can use in my documentation

workloads I can use in my documentation workloads I can use in my documentation and note-taking systems. And it is super and note-taking systems. And it is super and note-taking systems. And it is super fast. And of course, this is so much fast. And of course, this is so much fast. And of course, this is so much different when you're using a system different when you're using a system different when you're using a system like notion or Confluence. I still do like notion or Confluence. I still do like notion or Confluence. I still do not think that these systems are bad. I not think that these systems are bad. I not think that these systems are bad. I want to be clear here. Yeah. So, notion want to be clear here. Yeah. So, notion want to be clear here. Yeah. So, notion confluence and other systems are confluence and other systems are confluence and other systems are probably much more polished when it probably much more polished when it probably much more polished when it comes to this all-in-one workplace. It comes to this all-in-one workplace. It comes to this all-in-one workplace. It has much better databases. So, you don't has much better databases. So, you don't has much better databases. So, you don't need to fiddle around with CLI tools need to fiddle around with CLI tools need to fiddle around with CLI tools like MDBase query or stuff like that. like MDBase query or stuff like that. like MDBase query or stuff like that. That all works much better in these That all works much better in these That all works much better in these bigger platforms. And for a lot of bigger platforms. And for a lot of bigger platforms. And for a lot of people, that might probably be the people, that might probably be the people, that might probably be the better choice. But for my workflow, if I better choice. But for my workflow, if I better choice. But for my workflow, if I want to have full control over the data, want to have full control over the data, want to have full control over the data, over the tools, over the AI over the tools, over the AI over the tools, over the AI subscriptions that I want to use for my subscriptions that I want to use for my subscriptions that I want to use for my documentation and project management documentation and project management documentation and project management systems, I think simple markdown files systems, I think simple markdown files systems, I think simple markdown files are just the best and obsidian works are just the best and obsidian works are just the best and obsidian works great for me. But you can certainly also great for me. But you can certainly also great for me. But you can certainly also think about using any other markdown think about using any other markdown think about using any other markdown based tool like Joplin or Lock. I know based tool like Joplin or Lock. I know based tool like Joplin or Lock. I know some of you guys prefer these kind of some of you guys prefer these kind of some of you guys prefer these kind of tools. Not a problem at all. just run tools. Not a problem at all. just run tools. Not a problem at all. just run the PI agent in your terminal and then the PI agent in your terminal and then the PI agent in your terminal and then use whatever you like as an interface use whatever you like as an interface use whatever you like as an interface for these markdown files. It actually for these markdown files. It actually for these markdown files. It actually does not really make so much difference does not really make so much difference does not really make so much difference when you use these simple and portable when you use these simple and portable when you use these simple and portable tools. But yeah, that's just my opinion tools. But yeah, that's just my opinion tools. But yeah, that's just my opinion about it. No, I'd like to hear yours.

about it. No, I'd like to hear yours. about it. No, I'd like to hear yours. What do you think about this? Hopefully, What do you think about this? Hopefully, What do you think about this? Hopefully, it inspired you to think about your it inspired you to think about your it inspired you to think about your project management and how to work with project management and how to work with project management and how to work with AI tools in your home lab. Please let me AI tools in your home lab. Please let me AI tools in your home lab. Please let me know um if that was interesting and what know um if that was interesting and what know um if that was interesting and what else you want to see. And of course, else you want to see. And of course, else you want to see. And of course, thank you so much for watching. A big thank you so much for watching. A big thank you so much for watching. A big thanks goes out to all of my supporters thanks goes out to all of my supporters thanks goes out to all of my supporters on Patreon and on YouTube. You guys are on Patreon and on YouTube. You guys are on Patreon and on YouTube. You guys are really amazing and make all these free really amazing and make all these free really amazing and make all these free tutorials and videos possible. And of tutorials and videos possible. And of tutorials and videos possible. And of course, I'm going to catch you in the course, I'm going to catch you in the course, I'm going to catch you in the next one. Take care. Bye-bye.