# The ONLY Hermes Agent Tutorial You Need 2026: How To Use Hermes AI Agent ## Метаданные - **Канал:** Charlie Chang - **YouTube:** https://www.youtube.com/watch?v=r5FrGDTCYSg - **Дата:** 02.05.2026 - **Длительность:** 18:28 - **Просмотры:** 10,858 - **Источник:** https://ekstraktznaniy.ru/video/51248 ## Описание Get your Hermes Agent running 24/7 on Hostinger VPS (easy 1-click install): https://bit.ly/hermes-agent-vps ^Our code CCM10 is already applied for 10% off! (only for our audience) Get the full Hermes Agent setup commands + sample prompts in my free AI Community: https://www.skool.com/ai-os/the-ultimate-guide-to-hermes-agent?p=8f07bbb7 Most people are still sleeping on AI agents right now, and if you’ve been thinking about trying one, Hermes Agent is one of the MOST powerful setups you can build today. It’s not just another chatbot, it’s something that actually works for you in the background. In this video, we teach you how to use Hermes Agent, step by step. Hermes Agent is basically an AI assistant that lives on a server, runs 24/7, and remembers everything you tell it. The more you use it, the smarter it gets. You can text it from Telegram on your phone or computer, and instead of just answering questions, it actually goes and does things for you (browse the web, create videos, ma ## Транскрипт ### Introduction [] Today, I'm going to show you how to set up and use a Hermes agent like a pro. If you're not familiar with Hermes agent yet, it's an AI agent that lives on a server and remembers everything you tell it. And the more you use it, the better it gets. It is also completely open source, and after setting it up, it's extremely easy to use. In this video, I'll take you through exactly how to set it up on a VPS, deploy it, connect it to your Telegram so you can use it on your phone or computer, and show you some of the most popular tools within the agent. I'll also be showing you how to set up your agent so it can automate your AI UGC pipeline for your business or app. So, follow along with the video, and by the end, you'll have a Hermes agent that can run your personal life and your business. Let's get started. So, first ### Setting up Hostinger VPS [0:40] thing we want to do is click the link in the description for Hostinger. This is the best, cheapest, and most efficient way to run your AI agents. And instead of paying $800 for a Mac Mini, you can just run it and deploy it in the cloud. So, after clicking the link in the description, you'll be greeted with this page, which is Hostinger's one-click Hermes agent deployment. And all you want to do is simply click deploy right here. And after clicking deploy, you'll be greeted with this page, which is basically the checkout page for Hostinger, and you can either choose 24 or 12 months. I typically choose 12 months so I can have this set up for a year without having to worry, and it only costs $8 a month. Now, if you clicked our link, you'll be able to see that you have a 10% discount off of your plan. And once you've selected the plan that you want, all you have to do is click continue. Now, if you don't have a Hostinger account yet, it'll redirect you to this sign-up page so you can just sign in right here, create your account. And after completing the checkout after we sign up for Hostinger, this is the page we'll be greeted with. And this is our admin and password page for our Hermes agent. So, Hostinger automatically fills in Hermes and a password for you, but you can change this anytime you want. And all we need to do now is just click deploy. So, right now, Hostinger is setting up our VPS, and this is basically a dedicated server for our agent to run on and live on. Now, this server runs 24 hours a day and 7 days a week, so you always have access to your agent wherever you go. So, after we set up our VPS, we'll be greeted by this little survey, so we can just select my first time using a VPS, personal reasons, and AI machine learning. We'll save that. And now, what we could see in our Hostinger dashboard that our project is being deployed. Now, a couple things to pay attention to, we have our Docker manager right here, which is where our projects will be managed when it comes to our agents. Now that our project is deployed, pay attention to this thing right here. We'll need this later. This is our four-letter code for our agent. Now, in order to get our agent connected to our ### Connecting to Telegram (BotFather setup) [2:30] phones, computers, and anything else that you want in order to text it and access it, we want to go here to the terminal. We'll click terminal. And now, we have our terminal open for our VPS. So, this is pretty much giving you access to manage the server that the agent is running on. And I know this may look kind of scary, but it's a lot simpler than it looks. Now, before we mess with anything here, we can just leave this tab open. And what we want to do is go back to our project manager, go to our Hermes agent with the code, and click manage right here. Now, this shows you that the container where the agent lives is running. And what we want to do is scroll down all the way to the bottom in order to set up the environment variables. Now, what are environment variables? These are keys to different platforms that the agent will have access to in order to do certain tasks for you. So, for example, if I want to set up Telegram for messaging in order to message the agent back and forth, what I want to do is click plus on the environment, and then we'll add the name of the key here and the value of the code that we get for the agent. So, let's quickly set up all of the environment variables that we want. So ### Setting up environment variables [3:39] we'll be using Telegram for messaging, OpenRouter for the AI models that run the agent, and file. ai for the video models. So, let's start off with Telegram. What we want to do is open our Telegram app. Then, once we're in our Telegram app, we want to click right here on contacts and look up BotFather. Now, this is Telegram's official bot creation service. So, what we want to do is just click on BotFather, and as you can see, my chat is already open. And what you want to do now is type /newbot. Press enter. Then, it will message you and ask you what you want to name your bot. So, after typing in new bot, what we want to do is choose a name for our bot. So, I'm going to say Omri Hermes agent. We'll click enter, and that's the name of our bot. And now, we have to pick a username for our bot. Obviously, as it says right here, it must end in the word bot because they don't want people having robots that can impersonate other humans. So, we'll call this Omri Hermes agent bot. And there we go. Now, our bot is live on Telegram. And as you can see right here, we have our API key for the bot. Now, please do not share your API keys with anyone because that means they could plug into your bot and use it instead of you, and you do not want that. But let's copy our API key right here, go back to Hostinger, and paste it into here. So, after we typed in our key right here in the value space, we'll go to the name and paste in telegram_bot_token. Once you have that done, click save and deploy, and your Telegram will be connected to your Hermes agent. Now, obviously, in order to use an AI agent, you need some sort of AI model to do all the work. Now, in order to do that, we'll be using OpenRouter. So, you can either search up OpenRouter or use the link in the description, but this is what it looks like. So, OpenRouter is basically a place to get API keys for all the different AI models, such as Claude, ChatGPT, Gemini, and anything else that you can possibly think of. So, what we want to do is sign up for OpenRouter, and if you already have an account, just click sign in ### Setting up Open Router API key [5:37] right here. And once you're logged in, what you want to do is click on your credits tab right here. I use it a lot, so I added $25, but you don't have to add $25. You can just add $5 if you want to use some of the cheaper models. But once you add money to your account by clicking this add credits right here, what we want to do is go to our API keys button over here. And then, we'll click create right here on the right side. We'll name this Hermes agent V2. You can name it whatever you want, just something that you remember. And you can add a credit limit right here, which is basically a safeguard so your agent doesn't go rogue and overspend all the tokens that is on your account. So, we can do $10 credit limit and resets every month. Expiration, we'll do 1 year. Now, we'll create our key, and we have this value right here. What we want to do now is copy it and go back to our Hostinger dashboard. And now that we are in the Hostinger dashboard, we'll add another environment, paste in the value right here, and write openrouter_api_key. [snorts] Then, we'll click save and deploy. Now, really quickly, let's add another API key, which will give your agent a huge set of new abilities, which is basically allowing it to create any video or image that you want it to, and that will be through file. ai. So, you can look up file. ai or just click the link in the description. So, after we reach file. ai, we can click log in. So, you can either sign up with your email, GitHub, or any other service. And after you sign up for file. ai, this is your dashboard. You can explore different ### Setting up FAL API key [7:09] models, generate within file, and use it for a lot of different things. But what we'll be using it for is the API key. So, what you want to do in order to get your API key is click on this right here, which is your account, and go to API keys. Then, we'll click add key, type Hermes agent, and then we'll create a key. Then, we will copy the key into our environment. Same process again. So, once you paste it in your value, you want to type in file_key and press deploy. And after it's been deployed, we want to go back to our terminal. And I know this might look scary, but I'll run you through everything you need to know. So, we have a document down in the description that has all the commands that you need to run. You just need to copy-paste it. Super simple. So, now that we have our terminal and our document, what we want to do is copy this command right here, paste it into here. And now, you see that there's a quadruple X right here. What we want to do is just erase those, go back to our ### Terminal commands and entering the container [8:05] project, and copy our four letters from our agent. By clicking enter, that means we are working in the project file that the agent is in. And what we want to do now is enter the container. So, what we want to do is copy this right here and paste it onto the terminal. So, after typing in this right here, the docker compose exec -it Hermes agent /bin/bash, you'll be working in the Hermes agent. Now, in order to check the status of our Hermes agent, we can just type in Hermes status. And then, we'll see everything that is currently configured for our Hermes agent. So, we can see that our OpenRouter is working. Telegram is working, and we can also see that our file API key is connected. Now, what we want to do from here is go on and choose the model that is going to run our Hermes agent. So, we'll copy that, paste it, Hermes model. And now, we have all of these options. Now, you can scroll up and down with your keyboard. But since we set up ### Checking agent status [9:04] OpenRouter, we'll click enter on OpenRouter. And now, it'll give us a list of all the models that are available. Now, I like both Anthropic and OpenAI models, but since this is an AI agent that is going to be consuming a lot of tokens, I will just be choosing Claude Sonnet 4. 6. And now, our model is set to Claude Sonnet 4. 6. Now, after we set up the model to Claude Sonnet 4. 6, what we want to do is start the gateway ### Choosing your AI model [9:27] so the agent can talk to us via Telegram. So, what we want to do is write the command Hermes gateway start, and click enter. Now, if you do get an error when typing that in, you can just type in Hermes gateway, and it will start running your gateway. So, after your gateway has started, what you want to do is go back to Telegram, and then click on this right here, which is your new Telegram bot. And after clicking on it, you can type in /start, and will get this message right here. Now, this is how you pair your Telegram bot to the Hermes agent. So, once you've copied it, you can paste it into your terminal and ### Starting the gateway and pairing Telegram [10:03] press enter. And now you can see they have approved me in order to use the bot. Now, if you are running into issues, all you have to do is click X onto your terminal, go back, open up a new terminal, and then redo the first two and three steps. Once you do those steps, you can put in any command that you need, including the one from Telegram. And now that it is set up on Telegram, I'm going to say hello. As you can see, there is no home channel set, so we'll do set home, and then we'll type hello again. And there you go, your agent is set up. So, this was the hardest part of the video. Everything from here, super easy. Now, if you're unfamiliar with it, what you want to do is basically say, "Hey, I'm new and I'm running you, Hermes agent, in a hosting or VPS in a Ubuntu container. " Ubuntu container is a thing that's running on the server. You don't really need to know what that means. Just know that it's running in a safe container within the server on Hostinger. Please let me know of your abilities, skills, and anything that I can do using your services. And there we go, it gives you a fully loaded message with everything that it can do for you. So, it has access to this server, it can do web browsing, it can code things, it can ### Testing your agent and exploring abilities [11:10] schedule different tasks. So, the schedule tasks feature is extremely useful. You can have it send you an update of your email inbox every 9:00 a. m. on every day of the week. It can spawn in sub agents and analyze different images and things like that. It can also create you different apps. Let's say you want to create an app for your kid if they want to learn some vocabulary words. You can create it using this and let your kid use it. It can also manage your Google workspace, your media, your smart home if you have Philips Hue, so it can manage your lighting in your house. You can just message it and say, "Change my light to orange. " and it can change the color of your lighting in your house to orange. Now, in order to use it and create a AI UGC skill for it to make UGC videos for your company or brand or whatever it is you're working on, we can start right here. So, first, let's make sure that the agent has access to file correctly. Do you have access to the file API now? We'll click approve right here. Cool. So, now it has access to the file API. So, what we want to do now is start creating our UGC pipeline and skills. So, let's go to file and find the video model that we want to use. So, C dance is the best one out there right now. So, what we want to do is click on C dance image to video. And then what we'll do is copy paste this link right here and send it to our Hermes agent. So, with ### Building a custom UGC video skill [12:30] the link, we'll send an example prompt. You can grab it down below to the Hermes agent for a UGC video. So, the way this works is you send the agent a picture of your product and it'll generate a UGC video based on that image. So, what we want to say is now that we have the link and the example prompt. So, here's the final message I will be sending to my Hermes agent. We have our link right here. We have our example prompt right here. And the final prompt is, "I want you to create and add a skill for yourself. " So, adding a skill basically means that the agent is creating a skill for itself that it can save and replicate and do again and again when you request for it in order to create UGC videos based on images. So, I should be able to send you an image of a product and your output should be a UGC video using C dance 2. 0. My input would be the image of the product and your job would be to run an API call to file with a prompt generated based on the image as well as the image itself. Now, let's send that to our agent and see what it does. So, as you can see, it has access to a browser, so it's navigating through the browser, it's taking pictures of the browser, it's reading different things on the page, and it's learning how it can create this skill for us. Now, this isn't limited to just creating UGC videos. The whole purpose of me showing you this is that you're able to replicate this and do it for anything in your life. So, for example, if you want to give your agent a skill that reads your email inbox, all you have to do is again, connect it to your email inbox just like we connected it to file. Then, as you saw, we give it a specific set of instructions of what it should do within the skill. So, give it instructions on how you want it to read your email inbox, summarize things, how you want it to organize things. And once you send the instructions to your agent, it'll be able to process it, understand it, and figure out how to create a skill that does what you need it to do. Let me know once the skill is ready. So, cool. So, as you can see, the agent said the skill is ready. So, what we want to do now is we'll send it an image of a product. Cool. So, as you can see, we sent it the picture of an iPhone and it's viewing the skill, UGC video generator, vision analyzer. It analyzes the image of the iPhone and now it's running the UGC skill. So, as you can see, even when the agent runs into problems, it can figure things out on its own. So, it had an issue with uploading the image, it just converted it to a different file type, and now it's asking for my approval. So, this is the prompt it came up with and let's allow it and see what it does. So, as you can see right now, it says job is in the queue, which means it's sent the video to file and now it's creating the UGC video for us. Now, the good thing about this type of agent is that if it has an issue and you want it to do one thing over the other, you just tell it to remember how to do things and it will remember it. So, as you can see, it generated our UGC videos. Bro, I am not even kidding. I still daily drive this thing. The screen is insane. Face ID is instant and the photos it takes look better than phones twice the price. I will never give this up. — So, it figured out that this was an older phone. Actually, you can see right here, iPhone 10 in silver, and it ### Testing UGC skill with product images [15:26] tailored the prompt to say that the user still uses the phone even though it's super old. Anyways, you get the point. The whole point of this is to teach you how to create these skills and learn how to use these agents so you can automate a lot of stuff in your life and your business. Now, as you can see, there's some issues with the UGC video. Obviously, it might not get it the first try. It's still AI, as you know. But, the good thing with this is if you tell the agent to make sure that one thing doesn't happen or something specific does happen. If you want the video to be longer, shorter, you want a different character, you just tell it and it will remember. So, now that we have our UGC skill creator, we can take in any product, any app, anything that you're selling, regardless of what it is, you can just upload an image and get a UGC video. So, let's try it out with another product. So, we have a Prime Hydration drink. And now the agent will see that it's an image [snorts] and will make a UGC video off of it. Now, even though we're using Telegram on our desktop, the best thing about this, and that's also why you should use the Hostinger VPS, is that it runs 24/7. So, let's say you're in bed at night or on a vacation and you don't have your computer or your laptop, you can just download the Telegram mobile app and message your bot. So, as you can see, we have our chat on our desktop and we also have the chat right here. And we'll scroll up. We have our UGC video right here. You can just download it on your phone, post it on Instagram, YouTube, whatever platform you want to, and it is ready to go. Yeah, here we go, the agent gave us the UGC video for the Prime bottle. So, let's see what it looks like. Bro, after every single workout, this is the only thing I reach for, Prime Tropical Punch. The flavor is actually insane and it hits different when you are actually thirsty. No cap, this replaced everything else in my fridge. But, yeah, that's actually kind of crazy. If you followed it along, you completely automated your UGC pipeline for whatever product you're selling. And yeah, any issues you have with the agent, the output, or anything else, you just tell the agent and it writes down files in order to improve itself over time. Now that you have this, if you want to, you can ask the agent how to connect it to different social medias and have it post on different accounts for you. And again, in order to do that ### Using your agent from mobile [17:34] all you have to do is ask the agent how. So, anyways, I hope you got some value from this video. And again, the best way to learn is by following along. So, I encourage you, if you haven't already, set up your VPS using the link down below and follow along step by step as I do everything in the video. There is a ton to cover with the Hermes agent. It has the ability to do a ton of different things as well. But, it's a really exciting AI agent and I hope that you do get a lot of value out of it. So, if you enjoyed this video, please give it a like, subscribe, and comment down below what you'll be using your Hermes agent for. We do a ton of videos about business, AI, finance, and more. Thank you so much for your time and I will see you next time.