Hermes Agent is the greatest AI tool ever made. Here's how to set it up

Hermes Agent is the number one AI tool you should be learning right now. It is a super powerful AI agent that is basically your 24/7 AI employee. The problem is most people have absolutely no idea how to use it or get any sort of value out of it. In this video, I will show you how to use this incredible tool. I will cover literally everything from installation to get it up and running and choosing the right model, choosing the right computer to use cases, everything. If you stick with me until the end of this video, you will be a master of the most powerful AI agent on planet Earth. Now, let's lock in and get into it.

If you haven't used or mastered Hermes Agent yet, this is about to be the best 60 so minutes of your entire life because it is going to change everything for you. This is Hermes Agent. What I'm going to show you in a second. First is what it's all about, everything you need to know about it. Then I'm going to show you how to install it, use cases, all that. This video is built to be jumped around if appropriate. So, if you don't need to see installation, if you already know what Hermes Agent is, you can look down below and find different sections. So, feel free to jump around however you'd like. But let's get into it.

First, let's first talk about what Hermes agent is for those that are new or for those that just need a refresher. After that, we'll talk about Hermes versus OpenClaw, which one you should be using. We'll talk about which models you should be using. We'll talk about should you use local models, should you put this on a Mac Mini. We'll talk about where you should install this, how to use their dashboard that I promise you 99% of people are using wrong, their built-in automation and canband board, how to take advantage of their self-improvement. We'll talk about which skills you should be using, use cases, security, all that. So, jump around for that.

But first, let's start off with what Hermes agent is. It is a 24/7 AI agent. That means it can work for you around the clock, completely autonomously, completely proactively figure out tasks and activities you should do to advance you, to make you more money, to get you closer to your goals. It is a custom employee for you. And I truly believe it's the most powerful AI technology out there right now. It learns everything about you and the way you work. It learns your goals, what activities you do every day, what tasks you do every day, how you make money, what you're trying to achieve. It learns everything about you and saves it into its ultra powerful memory. The most powerful memory in AI agents. It's accessible from everywhere. So, you don't have to install any sort of Hermes app. You can just have this integrated into your Telegram, your Discord, your iMessage, whatever you want. So, it lives where you live. It can do work autonomously. So in the middle of the night while you're sleeping and go and build an entire SAS application and launch it for you. And here's the part that separates Hermes agent from everyone else. It is self-improving. Every time you send it a message, every time you send it a prompt, it will do that task and it will learn, okay, what went well? What didn't go well? Next time I do this, how do I do it better? And actually learns and edits its own skills. So the more you use it, the more powerful it gets.

The number one question I get though is Hermes or Open Claw. Well, I got some answers for you. For those that have been watching this channel for a while, you know I was really, really big into Open Claw. I still am, but they have run into many issues that I'm hoping and I'm sure they will solve. I am good friends with the team. I know they will improve things. But at the moment, I am recommending Hermes Agent for several different reasons. One, it's lighter weight. Open Claw has gotten and feels very very bloated because it feels bloated. It's not quite as performant. It's not quite as quick. Hermes just feels a lot snappier and a lot quicker. But here's the big reason. Here's the big reason why I think you need to go with a Hermes at the moment and why I think you're watching the right video is because Open Claw's biggest issue the last couple months is every update has broken it. Every time I do OpenClaw update, I can be like 90% sure I'm about to rip my hair out. I don't know if they're not testing it. I don't know what the deal is, but every time I upgrade OpenClaw, it breaks. And for me, someone who lives with AI, who's always using AI agents, it's a little bit quicker for me to fix it, right? But for those who don't use AI agents a ton, it is a pain in the woo to have to fix their agent every single time it breaks. And so that reason right there, the more reliable updates I think is the biggest reason uh to be using Hermes at the moment.

They also have a bunch of other really cool things. They have the new portal which is they actually have their own AI model services built into Hermes. So if you want to save money, you can use their built-in models. It's built to be multi-agent. So it's really easy to spin up multiple agents at once, have them do a ton of work, have them work autonomously, which is amazing. And their social media team is just great. So, if you go on X, you have recommendations, you have feedback, you tweet it, you can be about like 90% positive that their official account's going to reply to you within like 10 minutes. They have a really, really great social media team.

Now, before we get into installation, I want to answer another question I get quite often, which is, which should you be using, Hermes Agent, Claude Code, Codeex? A lot of videos on my channel are about vibe coding, using Claude Code, using codecs. Still amazing tools, still tools I use every single day. But there are times where you should be using Hermes agent and when you should be using Claude Code and Codeex. Hermes agent you want to think as like your general chief of staff, your general employee that can do any kind of work or task for you. Day-to-day tasks, doing research, building documents, building spreadsheets, being your administrator for your computer, moving around files, cleaning up files, bouncing ideas, being your business advisor, right? That's where your Hermes agent is going to be. Think of it as your general purpose business advisor, employee. It's also great at building prototypes. Don't get me wrong, Hermes, OpenClaw, they're great at vibe coding. You can use them for vibe coding. I primarily use them though to build prototypes. So if I'm out on the go, I'm at the gym, I get a great idea, I pull out Telegram, I hit Hermes up, I say, "Hey, build this for me." I come back, I have a full prototype. It's also great at building its own tooling, right? So if I'm building like, you know, extensions to the kanban board. If I'm building a mission control, I'm using Hermes agent to build the tooling for itself because it's most familiar with how it works. If I'm doing things where I want the process to continuously improve, be better every time I do it, I like to use Hermes agent because again, the real strength of this agent overall is the self-improvement. If I just want quick changes in an app, if I want to change a button color real quick, this or that, I'll say Hermes, go quickly do it because I can just spin it up real quick. You just want to think of it as like your own general purpose employee.

When you use Claude code and codecs, this is when you're building large-scale complex applications. Right now I'm building Henry intelligence machines about to change the world. Link down below. I'm using a combo of codeex and clawed code for that just because they have so many different features purpose-built for vibe coding for getting complex applications built. So if you're going deep into a vibe coding session, you have a trillion dollar idea for a business. I still recommend using claw code or codeex for that. That's your focus locked in work. You're going to have open your two windows side by side codeex cloud code. You're gonna have 20 different agents going. You're sitting down. You're locked in. You have codeex cla code going and end to end testing. There's tons of really good testing features with Claude code and codec. So I still prefer again for that more complex coding session. You go with cloud code or codec.

Now that we covered some of the basics of what Hermes agent is, let's get into actually installing it. So, we're going to cover installing it, which models you should choose, first things you should be doing, use cases, and a whole lot more. If you learned anything at all so far, make sure to leave a like down below, subscribe, and turn on notifications. Also, make sure to join the Vibe Coding Academy. I do live boot camps covering Hermes and all the other AI tools every single Friday. Link for the Vibe Coding Academy down below. It's the number one AI community on planet Earth. I promise it'll be the best decision you ever make.

Okay, so if you go to Hermes agent website, I'll put the link for it down below. There is a command here, a curl command. You simply click copy on. From there, you open up your terminal and what you're going to do is you're going to paste that in. Because I already have Hermes installed, I'm going to do a slightly different command, so bear with me here. Once you hit enter on the command, you're going to go into setup mode. Couple options here. First thing off the rip, if you already have OpenClaw installed, it's going to give you the option to actually import all your OpenClaw stuff, all the memories, everything involved with that. Feel free to do that if you want. That's one option. If you're doing this clean, if you want to set up a clean agent, I actually recommend this for most people. I think using OpenClaw and Hermes side by side is a great idea. I don't think there's any need to have them all have exactly the same memories and skills. So, even if you have OpenClaw installed, it might be a good idea just to say no there. What we're going to do is we're going to do quick setup. We're just going to configure the things we want to configure.

So the first option we're going to get is the model. Let's discuss the different model options for Hermes. You have basically three tiers of options, right? Expensive, moderate, and cheap. Obviously, the more money you spend, the better results you're going to get. So on the expensive side is going to be anthropic. Claude is just the best model for these types of AI agents. Unfortunately, Anthropic cut off use of OOTH for this, meaning you can't use your existing Anthropic subscription. So, you're going to have to do API key. With an API key, you're basically paying for every single thing you do. This is going to be more expensive than your $200 a month for the subscription or your $200 a month for Chat GBT subscription. So you probably, if you have good use, probably spend a few hundred a month. It's going to be the most expensive, but you're going to get the best use out of it.

The good news is on the moderate tier is Open AI codecs. You get a good amount of usage with your Open AI subscription inside Hermes. So even if you're on the $20 a month subscription, you get decent usage. Couple pieces of good news there. Chat GBT inside these agents, OpenClaw and Hermes, absolutely sucked up until recently. Chat GBT 54 and earlier literally completely useless in OpenClaw and Hermes. Chad GBT 5.5 though, it's very good, right? Is it as good as Anthropic? Is it as good as Claude? No. But now it's actually usable. Now it's actually solid and good. So, if you're on a tighter budget, or even if you just want to save a couple dollars a month, you can go ahead with your Chat GBT subscription here, still get pretty good usage, still get good functionality out of it, get a ton of usage. You're going to get way more usage than Anthropic out of this, and you'll be good to go. So, there's nothing wrong, in my opinion, with using your Chad GPT subscription here. With Claude, it's just going to do tasks a little bit better, and it's going to have a little bit warmer of a personality, more fun to interact with.

On the cheaper side, you have many different options. If you're already paying for XAI, you can use your XAI OOTH here. You can also go with the new portal. This is actually the creators of Hermes Agents uh AI subscription. They kind of curate different models. It's only like $20 a month. You're getting cheaper Chinese models uh powering your agent, which honestly is not going to be quite as good, but it's usable, right? Miniax, Kimmy, very usable models, very cheap models. The personality is just going to feel very robotic and it might not be that great when it comes to actually executing on tasks. I would at the very least recommend the moderate option which is Chad GBT. You're going to get, you know, good functionality out of it. It's going to have a decent personality. So, choose your model from there. Whether it's ooth or API, it'll instruct you on what to do from there. If you got to get your API key, if you need to log in, if you're doing chat GPT, proceed with that.

Now, next is going to be what messaging platform to use. And you're going to have a few options there. You're going to be able to choose between Telegram, WhatsApp, Discord, iMessage. Let me tell you this, the best way to communicate with your agent is without a doubt going to be Telegram. Telegram is the best way to do it. Telegram is actively working to make their platform like a platform you communicate with agents with. They're adding tons of functionality. They're allowing you to do a bunch of different things like topics, allowing your agents to actually talk to each other. There are so many cool things in Telegram they're adding. They're really the only plat messaging platform on planet Earth where they're trying to make it like AI agent friendly. You could do Discord. There's some cool things with Discord there. I have a video on using Discord with your agents, but as a main communication channel, I don't think anything comes close to Telegram. You need to be doing Telegram. It's also free, so that's pretty cool. Telegram is the way to go here.

You go through those couple things, you'll be all set. You'll be good to go. You'll have it all set up. When you set up Telegram, it's going to give you a couple steps on actually connecting it to Telegram. It's really simple. You don't need me to put in the video. Really easy to do. You're just copying a token from Telegram into your Hermes and you'll be good to go. Then you're all set up. You can start chatting with your agent.

So, let's talk about first things to do. This is your own AI employee, right? This is your own personal assistant. If you were to right now go and hire a personal assistant, they come over and it's day one and they're ready to do work for you. What's the first thing you would do with that person you just hired? Well, you'd probably first tell them about yourself and your business so they're familiar with you, right? That is what we're going to first do with Hermes. Hermes has the best memory in the game. Hermes remembers everything about you. It's constantly self-improving. So, let's do this. Let's tell Hermes all about us.

So, here's what I did, and you should do this, too, alongside with me. I am Alex Finn. I'm an entrepreneur and content creator. I have a YouTube with 200k subs and a Twitter 450K. So, I told all about myself, what I've achieved up to this point. I built a social media tool called Creator Buddy and I'm building a new tool called Henry Intelligent Machines, which will allow anyone to build their own business. So talking about what I'm working on and then I talk about my goals and ambitions. My goal is to make Henry Launch super successful and increase my subscribers. So talk about your goals and ambitions as well and what you want to achieve. That's super important because when we get Hermes to start doing things proactively for you in a second, it will be able to look at your goals and ambitions and figure out tasks to do proactively for you. So make sure you cover those three things. Who you are, what you're working on, and your goals and ambitions. Hit enter on that. That will be saved into your Hermes memories and it will know you so much better.

And let's do this before I start going through the user interface. So the amazing dashboard they built out, the kanban board, all of that. Let's set up a task. Let's set up a scheduled repeatable task Hermes can do for you every single night. So there's this thing in Hermes called cron jobs. Basically what they are just scheduled tasks that run autonomously. We're going to set up a schedule task. Everything we do, everything we set up is just plain English into Hermes. It's just plain English into the text box. You talk to it like a normal human being. It's the best part about AI. So, what we're going to do is we're going to talk to like a normal human being, and we're going to tell it to do something super specific and helpful for us every single night that brings us closer to our goals.

So, let me put this prompt in here. Feel free again to pause, follow alongside with me. "I want you to schedule a task for yourself to do every single night at 2 a.m. It should be a micro app, UI, or system that helps us get closer to my personal goals and ambitions. It should save us time, make us more productive, or just be generally useful. Make sure this is scheduled for every single night at 2 a.m. Make sure it surprises me and makes me happy." Boom. Putting that down below. What this is going to do is make your Hermes agent proactive. It's going to have it schedule a task for every single night where it's going to look at memories, systems, goals, ambitions, and put something together, build something out, vibe code something for you that will get you closer to your goals and ambitions. Something that will just be generally helpful and delightful. You hit enter on that. It's going to schedule the task.

Let's hit enter. Here we go. And GPT means this is my Hermes agent powered by GPT. I have a lot of different Hermes agents, a lot of different open claws. If you want like a video on multi-agent approach, feel free to let me know down below as well. And as you can see here, this is one of the cool parts about Hermes. You can see all its tool calls. So, it's scheduled a cron job called create. It's scheduled. It's going to happen every night at 2 a.m. The next run is tonight at 2 a.m. And it's going to deliver it here inside of Telegram. I set it up to create a small tangible thing every night, a micro app. It's instructed by us towards things that save time, support your goals. Boom. Just like that, Hermes is getting to work for us. Our employee is getting to work. This is how you talk to your employees, right? You talk in deliverables. This is what I want. This is when I want it. That's how it goes. It's talking to us just like an employee. So, that's amazing.

Now, I have a whole bunch of other tips, tricks, use cases. We're going to go through that in a second. I first want to take you through the Hermes dashboard. I want to take you through all the really cool UI stuff they built out. This is something OpenClaw, no one else is really doing. So, let me show you that now. And again, if you need to, you're familiar with different things, feel free to skip down to different chapters down below. So, if you go into your terminal and you say Hermes dashboard and hit enter, it is going to open up a browser with your Hermes dashboard inside of it. This is what it looks like here. Tons of things going on. I want to kind of walk you through this real quick.

If you go to models, you can very easily swap out your different models, try new models. Very, very easy to do. You don't need to go through terminal for all of this. As you can see, my model for some of my agents is claopus, others is GPT5. If you go into your cron section, you can see some of your scheduled cron. So, you should be able to see the one you just scheduled where it's going to do something proactively for you. If you want, you can create one manually here. This is a good way to create crown jobs. If like you want to make sure maybe it's a little bit more complex and you want to make sure it builds it the correct way, going into the UI is a good way to do it.

You also have skills. So, this is really cool. We're going to dive more into this in a second when we go back into Telegram. But again, the strength of Hermes is the self-improving skills. Every single thing you have it do, it reviews what skills it used and how to improve the skills based on what you asked it to do. So, it's constantly self-improving. So, you can come in here and see a lot of these skills it has. You can go in, browse, turn it on. I can even play Pokemon and Minecraft, which is unbelievable. And so, it learns things as it goes. You can see all those skills here. And make sure the right skills are turned on.

We have plugins here. These are like extra powerful abilities your Hermes agent can have. Make sure you come on here and you can turn on any ones you need. It has some things like browser use, browser base, fire crawl which allow it to crawl the internet really well. It already has built-in tools around using Chrome and using your computer computer use, but you can add on if you already have API keys around any of these, add them on here. I highly recommend enabling the image gen from OpenAI allows it to generate images. The OpenAI image generation model is the best in the game by far. So I would turn that on so it can generate its own images as well. That makes it super good if like you're building games with your agent. I use the OpenAI image gen model to build assets for games. So if you go to your Hermes say, "Hey, build this game for me." It'll be able to go in and actually generate all the assets for the game it's building to if you're a nerd like me.

Profiles. This is a good one if you want to set up a multi-agent approach. A profile is basically just a different Hermes agent. As you can see here, I am using something like seven or eight different Hermes agents at once. So you can go in here if you want to do multi-agent approach. Just click create, set up a second Hermes agent. It'll be booted up and ready to go. This is cool. You can see the skills for each. So you can see which ones I'm using more. My default, which is just Hermes, has 150 skills, which means I've been using it more than the others. So this is a good way to set up your multi-agents config. There's a whole bunch of options here. You can go through this if you want. Text to speech. I'm not a big text to speech guy.

But where I want to take you next is kanban. This is a really cool screen in here. This is probably where you'll spend most of your time in the dashboard is the kanban board. This is super powerful and they just shipped it recently. This is basically an automated way to multitask, right? So coming in your kanban board, it has the different statuses of tasks. Triage, to-do, ready, in progress, blocked, done. What's really cool here is if you put tasks into triage, what will happen automatically is your Hermes agent will take that task, split it into subtasks, move it to to-do, and then automatically assign it to different sub agents to-do. So, if I for instance go in here and I say script me a master class video for Hermes, and I hit create, it's going to go in there. And what's going to happen in a second is the Hermes agent will actually take that, split it into steps and assign it to agents to automatically do things for me. This is really cool and this is what I highly recommend doing. I write down all my to-dos on this piece of paper every single morning. What I do is after I write down my to-dos, I go in here to this kanban board and I take any tasks that an AI agent can do for me and I just put it into triage, right? I'll just dump in like five or six tasks into triage. They go in there. I walk away. I do something else. I make breakfast. I make a pop-tart. Whatever. By the time I come back, and you saw it happen live, the tasks move to to-do. Agents start grabbing it and they start getting things done for me. By the time my breakfast is over, I already got half my to-do list done. My AI agent has already tackled half my day for me. That's how you should be operating as well. Start every morning. Kind of triage yourself with a list of tasks. You get them into your Hermes kanban board and by the time you're done your breakfast, you already got, you know, half the things you need to get done today done already. So, that's a really cool way to use this kanban board. This kanban board is super powerful. You probably can find 20 other different things to do and how to use this. But, that's the way I like to use just using in the morning, getting my tasks in. By the time the day is halfway over, I'm almost done everything I needed to do, which is really amazing.

Now that we went through the dashboard, let's go through a couple different use cases here. I'll show you how it works, how it self-improves, how it uses different skills, things like that. I'm putting out over the next few days a Hermes like seven use case video. If you want to catch that when I drop that, feel free to turn on notifications down below, too. It's going to cover in-depth beginner, medium, advanced use case. But let's go over a few of those here.

So, I'm going to show you three simple use cases any of you guys could take advantage of and get some value out of. The first one is a daily tutor. If you're anything like me, you watch a ton of educational videos on YouTube. I'm watching them all the time. I mean, you're watching one right now. And one thing I love to do is give those videos to my Hermes agent, have them get the transcripts, learn everything that's in the video, and then be my tutor every morning. So, it'll give me some fun facts from the videos I watched recently. Maybe even quiz me and ask questions. Watch how we do this.

So, I got this video I put out on my second channel link down below on how to make $2 million AI in 2026. Say I'm a viewer. I watch the video. I find it super helpful. Let's take this link. What I'm going to do is I'm going to give it to my Hermes agent. And I do this with basically any video that has a ton of content that I find helpful. I say, "Hey, check this video out. Read the transcript. Then every morning remind me of a concept or two from the video, then quiz me, and then I give the link to the video." I hit enter. And then, as you can see here, it's using its built-in YouTube content skills. It's going in. It's pulling down the video. It's reading the transcript. It's pulling out all the important facts and information from it. And now it's going to schedule a cron job, which again, as we talked about earlier, just a scheduled task that happens at specific times. And it's going to send me a message every morning saying, "Hey, here's a couple things to remind you from that video." You should be doing this with any video or educational video you watch that has important, interesting information in it. That's how you reinforce these facts in your brain. Got it? There's a share your journey playbook. It's extracted 10 important concepts from it and it's scheduled the QR for every morning at 8 a.m. and I'm going to get my first lesson at 8 a.m. tomorrow. This is how I'm learning more and more with Hermes and improving my life. Let's get into the next use case.

So, you should be setting this up. Everyone watching should be setting this one up. The next one is actually maybe the most powerful one I use, which is my computer administrator. If you put every device you got, your iPhone, your computers, your Mac minis, whatever you got, your iPad, all in the same kind of private network, your Hermes agent can move around any of them and get any content you need. Here's what I mean by that. A lot of the times I'll be out and about and I'll be like, "Oh man, I wish I had that document from my computer. Oh man, I wish I had that presentation from my iPad. I can't access it because I didn't put it on Google Drive or anything." Now with Hermes agent, Hermes can act as the computer administrator, basically like my device administrator, and get me any file I need from any of my computers at a moment notice. This is how you do it. You go on tailscale.com. For those who don't know, tail scale is an amazing tool. It's completely free. It basically gives you your own private network on all your devices. Not sponsored in any way, by the way here. I just really like their product. You put it on all your devices and what's going to happen is now your Hermes can move across all of your devices and get any documents, files, anything at all you need and move them around your devices when you need them.

So, for instance, I'm on my laptop. I need an agent.md file from my Mac Studio. I just go on my Telegram and say, "Hey, Hermes, get me the agent.m MD rules file from my Mac Studio and drop it off on my MacBook Pro." It goes, it does that. It can move around any of my devices, no matter where I am in the world, and get me the files I need. This is also great if you're doing development, too. So, if you're a Vibe Coder or using Codeex or Claw Code, this allows you to actually test and load up local hosts from any of your devices. You don't need to be on the device the local host is running. So that's really helpful as well. Now Hermes is my device/administrator across all of my devices which is really awesome.

The next use case I'll go over is an amazing one and that is a session recall use case. What does that mean? Hermes has a feature that OpenClaw doesn't have. No other AI tool really has which is really incredible session recall. Every session you have with your Hermes agent, every conversation you have, everything you've ever said to your Hermes agent is stored in a log, a session log. And that means at any point you can go and ask your Hermes to recall anything you've ever talked about. So if you want to say, "Hey, tell me what we talked about last Thursday at 8 a.m." It can do that. If you want to say, "Hey, tell me what we talked about 3 months ago at 2 a.m." It will be able to do that. No other AI agent has this type of memory.

So what's the use case of this? What's a good use case here? Well, we can take advantage of this incredible memory. We can say, remind me of every Hermes use case we've talked about a month ago. This is incredible memory. Now, at any point, you can go, hey, remind me that thing we talked about a month ago. Remind me that thing we talked about a week ago. What's every YouTube video I dropped in here last week? What's every YouTube video I dropped in here three months ago? What's every link I gave you two months ago? You can quickly recall anything you've ever talked about, any idea you've idea. What's every business use case we've discussed last month? You know, I am constantly dropping in business use cases to my Hermes agent going, "Oh, is this real? Can we turn this into a product?" And now I can quickly get it to recall all of those things. You can see here, look at all these use cases we've talked about. Really amazing stuff here that I can just I'll never forget anything now. Everything it remembers. It's the ultimate second brain for your life. As long as you're running everything through your Hermes agent, you'll literally never forget anything. So, this is a really amazing use case of Hermes agent as well, being able to remember everything you've ever talked about.

So, let's go into the next section, which is very particular to Hermes agent, which is the memory and skills functionality inside Hermes and what makes it so special and what makes it better than every other AI agent out there. First of all, just the basics on how those things work. All your memories and all your skills live in markdown files on your computer. And this is really important. When you have memories on Chad GPT or Claude, they live in the cloud. You don't know what they are. You don't know what's saved. You don't know what those AIs know about you. With Hermes, it's different. They all live in markdown files on your computer. So everything it remembers about you, every skill it learns, you can audit and look at and change because it's just on like a text file on your computer.

But here's the big one. Here's what makes Hermes agent so unique is the self-improvement. Hermes built into it every step of the way. Self-improving. So whenever it does anything, every task you give it, it figures out along the way what was the best way to do that thing and updates its skills and memories for that best path. What do I mean by that? Say you tell it to go and read the transcript from a YouTube video and take out all the best practices from it. It'll go, it'll search for how to find transcripts in YouTube videos and then it'll grab the transcript and it'll find all the important points from that video. Once that task is done, it'll figure out and review how it did that task and then save it to memory how it did it. So next time you give it that task to do, it can just go, "Oh, I did this before and this is exactly how I did it." And it'll do it that way again. Now, if you have him do it and then maybe it does something wrong or you say, "Hey, actually, can you do it this way?" It'll remember that, save it to memory, and again, self-improve. So basically every time you have your Hermes agent do something, do a task, learn something, go and do something, it reviews what it did, saves it as a skill in memory, and it use that every time it does that task again. So it is constantly self-improving. The more you use Hermes agent, the better it gets, the more custom it gets to you, which is what makes it so powerful.

And third, this is a big one, their memory and skills are constantly evolving. Every update in Hermes Agent adds new built-in skills. A big one from the last update was computer use. I now can have Hermes Agent go and control things on my computer. If you want to see all the tools it's adding, all you need to do is go in, say Hermes tools, hit enter. It's going to go in. You'll be able to say configure and you can see a lot of the tools it's adding here. They're constantly adding. If you haven't done Hermes tools yet, make sure you do that. And you can turn these on. You want to make sure you have on browser automation as well as computer use which you see at the bottom. That way if you're ever away from your computer, you can go in and say, "Hey, Hermes, do this on my computer. Build this report. Build this slideshow." Whatever it is, and it'll go into it.

And I know an objection I'll get there is, "Hey, Chad GBT can already build presentations. Claude can already build presentations." The reason why you want to be doing those things with Hermes agent is because Hermes knows everything about you. It has the best memory of any AI there. So, if you say something like, "Hey, build me this presentation." It'll be able to do it better than Claude and Chad GBT. One, because it's self-improving and it learns every step of the way the way you need to do things and the way you want it to be done. And two, it knows all the context around you. So, it knows what your preferences are. So, these tools are super super valuable.

Another new one they added is this video generation one. I love this one. Image generation. Video generation. If you plug your Hermes into Gro's OOTH or Chad GBT's OOTH, you can now use those tools to generate assets. Let me show you using XAI Gro. "Imagine generate me an AI video of a basketball player in a blue jersey wearing the number two flopping to the ground." I'm going to hit enter on that. And because it just learned this new skill of being able to generate AI videos, it's going to be able to go in and create that video for me. So, it's constantly learning these new skills and improving and getting better. There's no other AI out there that can go from moving files around on your computer to getting transcripts to YouTube to generating videos. Hermes can do it all and is constantly improving, expanding its functionality. All right, looks like the video is done. Let's see what we got here. Oh man. Yep, that number two guy in the blue jersey is certainly taking a dive there. Wow, that's a big flopper. This is not based on any sort of a realistic person, any resemblance to anyone in the NBA, purely coincidental. So, pretty good.

The next part about Hermes agent, which is really awesome, is mission control. Mission control is a custom interface for your AI agent. This allows you to build out any tool you need for your AI agent to get good work done. Everyone should have a mission control for their AI agent. One, because it's just super helpful because you can build out custom tooling for your AI agents. Two, because it's just fun. This is a fun thing to do. You're allowed to have fun with AI. You're allowed to have fun with AI agents. Not every single thing you do needs to be some sort of minmax hyper optimization to ship more code. You're allowed to just have fun with your AI agent and the mission control is a really good way to do that. There's a lot of custom tooling I built out here. One thing is this content pipeline where I can put in content ideas and then Hermes will go in and actually add scripts, create thumbnails, things like that. All I need to do is put a new idea in here and then Hermes takes it and moves it along the content pipeline. I have a memory page. This is basically like a memory wiki where I can go in and see all the memories my agent has. So I can go in, read everything we've talked about from the last month, day by day, see our long-term memory because Hermes memory is so freaking powerful. Basically everything we say and do is recorded. So sometimes I just like to almost like a journal or a diary go back and see things we worked on in the past. I had now have this really nice user interface where I can see those things. I have a docs page. This allows me to see every single artifact and document we've ever created. Makes it really easy to see things. Usually when you have your agent build out different documents, it goes into like a random folder on your computer. Now I can just say, "Hey, put in our docs and our mission control." And I can quickly easily read it like this. I have a kind of an animated 2D office view where I can see all my agents walking around doing things. All my open claws, all my Hermes always working. I can come in here and see exactly what they're doing at any given time and many more things here.

All you need to do to build your mission control is literally go to your Hermes agent and say, "Hey, I want a mission control. Build out a custom mission control where we can later add tools in that we can use to improve our workflows." You don't need to know how to code. There's nothing super crazy or complex here. Every time I say, "Hey, build someone in the audience goes, "Oh, why don't you go into more detail?" There's no more detail you need to know. Literally, anything you need, anything you want from your agent, just say, "Hey, do this for me." That's how smart the agent is. You want a new tool, go to your agent, say, "Hey, build a new tool." Whatever you want. So, to build a mission control, literally go to your Hermes agent and say, "Hey, build out a mission control for me." Then once you have that built, say, "Hey, okay, I want this tool now. I want this tool now. I want that tool now." And it'll go and build it out for you. You should have your own custom dashboard face just like this where you can play around, build any tools you need and have some fun. Mission control super critical concept you need to have in your Hermes agent.

Next part we'll talk about is fixing issues with Hermes agent. This hasn't been as much of an issue with Hermes agent as it was with OpenClaw. OpenClaw would crash every time I'd update it. Doesn't happen quite as much with Hermes Agent, but you still should be prepared with how to fix things inside Hermes agent. And the way to do this is very simple. You're basically just going to open up Hermes agent in Claude Code or Codec. So you go into Cloud Code, you go into Codeex. Inside either one of those tools, you open up the folder that your Hermes agent is in. This is typically in the root of your computer. You'll see like a Hermes folder. You open that up in claw code or codeex, then you say, "Hey, I have this problem with Hermes. Please fix it." And your Claude Code or Codeex will go and fix that issue. That literally solves 100% of the issues I've ever had with Hermes. To be quite honest with you, Claude Code and Codeex are great at debugging these agents. So, any issue you have, just make sure you open up in Claude Code and Codeex and it will go and fix those issues for you.

Next thing I want to talk about is security. This is a really big request for me. Security. Here's the thing. When it comes to security with these agents with OpenClaw and Hermes, and this is going to be a really controversial thing I'm about to say. This is going to get a hundred replies in the comment section going, "Oh, you're such an idiot. You're so dangerous. You're going to hurt people. Shut up." Here's the thing. When it comes to security with Hermes and OpenClaw, the security concerns are vastly overrated. Yeah, I said it. There I go. The security concerns with OpenClaw and Hermes are vastly overrated. My question to you is this. If you go on the internet, if you watch YouTube, if you're on Twitter, you probably know a lot of people that are using Hermes and OpenClaw. How many of those people have had a major catastrophic security event? Truly, do you know anybody who has had a major catastrophic security event with OpenClaw or Hermes agent? My guess is the answer you're going to give me is no. Now, does that mean security events are it's impossible to have with Hermes or it's absolutely perfect when it comes to security? No, it doesn't mean that at all. I'm not saying it's completely invincible. But what I'm going to tell you is this. Hermes agent and OpenClaw only do what you tell it to do. So, if you say, "Hey, Hermes, go and create me a presentation on LLMs." It's not going to go and then go into your private photos folder and then leak all your nudes to the internet. That's not how it works. It's not like you're going to go say, "Hey, Hermes, check the score of the NBA game last night." And then it's going to go and leak all your text messages on Twitter. That's not how any of this works. These agents only do exactly what you tell it to do. So, the only way it's going to have a major catastrophic security event is if you go and you tell it to do something extremely stupid, if you go and tell it to delete all your text message and all your emails, it's going to delete all your text message and all your emails, but nothing else. If you tell it to go and drop you a document in your folder, it's going to drop you a document in the folder. It's not going to do anything else other than that one thing. So, what that means is you need to have some accountability, right? You need to think before you give it a prompt, is this going to cause it to do something really stupid? Am I asking it to do something really stupid? It's not like you can say, "Hey, code me an app." And then it's going to text your ex-girlfriend. That's not how this works. So, just think before you go, is this thing I'm asking you to do bad in nature? If not, then you're good to go. You have nothing to worry about.

But for some reason, these naysayers and doomers on the internet say, "Oh, Hermes, it's the most dangerous thing ever. Don't put it on your computer. Don't put it on your main computer. No, put it on your main computer. I have it on my main computer. I have it on every computer I own. I didn't isolate it. I don't think you need to isolate it. I think going and creating a new iCloud account just for your Hermes agent is actually a quite stupid and timewasteful thing to do. I don't think there's any point in doing it. As long as you just have some personal accountability and think about what you're telling your agent to do, I don't think there's any risk at all. And I know that's going to piss people off. I know the purists online who've been going around telling everyone, "Oh, this is a major security threat." That's going to make them go to the replies and cry like little babies, but I don't care. Until someone on planet Earth can tell me someone who's had a bad security event with one of these agents, I don't really care. Just use good judgment. Be an adult. Think about what you're doing and security won't be a concern. I don't think you need to make an iCloud account. I don't think you need to make a separate Google account. I think you're adding radical amounts of complexity for no reason. I think you should install it on your main computer. I don't think you need to have a VPS, which for the record, the VPS's are scams. Don't put on VPS's. Every AI YouTuber on planet Earth is getting paid by Hostinger to shill you a VPS. I promise you all those AI YouTubers who are shilling hosting jer literally zero of them are actually using hosting jer themselves. So don't listen to them. Just install Hermes on your main computer. I don't care. You're not going to have security issues if you practice some accountability and good judgment. That's security. You don't need a Mac Mini for this. You don't even need to buy like a super expensive Mac Studio like I did. All you need is whatever computer you own. Could be an old dusty laptop in your closet. Be whatever you're using right now. You can install it on there. There's no concern. There's no issues all of a sudden expose all your text messages. Relax.

That is the basics of Hermes agent. If you watch this video, you know how to set it up, what to do, how to run it, what you should be doing with it. I make tons and tons of content on AI agents like Hermes, like OpenClaw, like Claude Code, like Codeex. So, if you learned anything at all, like, subscribe, turn on notifications. I do live live live boot camps every single Friday in the number one AI community on planet Earth, the Vibe Coding Academy. Link down below for that. Make sure to join. We have thousands of people in there that absolutely love it, that are learning a ton. If you want to stay on the cutting edge, join it. Best decision you'll ever make. I am so deeply and eternally grateful that you would spend all this time with me to watch these videos and learn from me. Is an honor of a lifetime. I can somehow build this platform and make you educational AI videos. I'm so grateful.