**Julian Goldie** (0:00)
Today, I'm going to show you the Agent Operating System that basically automates everything for us. So this is running with Hermes and Claude as an Agent Operating System. And basically, Hermes is the most powerful AI agent you can run right now, but only if you build these three layers around it. If you skip them, your Hermes is basically just a Smart Chat window. But the good news is even if you've already set up Hermes already, adding these three layers, I'm going to run you through today, turns it into a 24-7 AI employee that knows your business and never forgets things. So it works 24-7, it gets smarter every single day and the whole thing can be built as I'm going to show you today with custom workflows as well. So by the end of this guide, you're going to have the full blueprint, the brain, the mission control and the memory system. And the third layer is really the one that changes everything. Most people skip it, but it's exactly why that agent doesn't work properly as a stack. So the problem here, right? And this is why it's so important to understand, is like a chat window is not an operating system, right? And the pattern I see over and over again is like someone's using Claude or they're using Hermes agent, for example. They install an agent, it's impressive for a day, and then reality hits, right? Every session starts from zero, every job you need needs you present, type in, you close your laptop, everything stops a week later. You just sort of abandon the agent that you're using. Now, if you compare that to machine that you're using this system on, right? Then you have a full system of all your agents orchestrated together, and you have different layers plugged in. So an agent alone forgets everything, it needs you to be there to babysit it, it stops when the laptop closes. The Agent OS system remembers everything about you, your brand, your business, your goals. It's got one dashboard, and jobs just keep running whilst you sleep, right? And the great thing about this is like, if you're using an Agent OS like you see right here, we can visualize our memory galaxy, so these are all our memories plugged in together. And so all of our agents, so for example, like Claude, Hermes, whatever agent we plug into this stack, has our context ready to go inside there. Now also, this is way more useful than, for example, like going over to Buzz, and then going over to Claude, and then opening up Chat Cheapity or Codex in another tab, which is super messy and time consuming and very fragmented. Instead, you want to plug them all into one system where they all work together. Then also, the great thing about this is, it doesn't matter what model comes out, like GPT 5.6 could release something new, like they actually did today. No problem. We can plug it into Codex right here. Claude Fable 6 comes out. We don't need to panic. We can just plug it straight into Claude setup right here. So everything is controlled, and you're not really worried about the models, so trying to play catch up or get shiny object syndrome. You've got everything ready to go inside one system that just improves every day no matter what. Now also, what I like about this is, we can build our own custom workflows in this. So for example, we have a video agent over here that can create amazing videos automatically with our video setup as you can see.
And so this has, for example, a voice service setup. It has, for example, a fully edited video. This is an AI avatar video that's plugged together. And all we need to do is type in a prompt, and it's ready to go. And so anything that you work on day to day, you can build a custom workflow like this. I've looked at all the tasks that I spent time on and then automated it using this full system so that everything is ready to go in one place. And also you can do really interesting stuff with this. So for example, let's say we're using Hermes. Well, we can have a custom made voice agent ready to go. It speaks back whenever I want, listens to me whenever I want, as you can see right here. And then also it can control my browser. It can control my computers, listen in the background. We can see our full conversation history. We can see all the stuff we've built with it directly over here, as you can see. These are all previous builds. We can even, for example, go over to the Hermes Oracle. This is running 24-7 on autopilot, so no need to babysit it. And you can see over here that it finds the latest news, for example, like Grok 4.6 dropping in seven days. And then we can draft social media content. We can publish content about that straight to a WordPress website. We can see the direct link to the actual post that just dropped 24 hours ago, right? And so it's automating all of the things that I used to do manually myself. For example, like monitoring competitors as well. It pulls that data in. As you can see, monitors are competitors, gives us content ideas based on that, and then we can go from there. So it's a really powerful system. Now, you might also say, okay, I'm not technical.
12 more minutes of transcript below
Try it now — copy, paste, done:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090
Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/YOUR_EPISODE_ID