Agent OS: Voice Agent + Lead Machine + Free AI

So today we're going to be answering some of the latest questions that we have on the agent operating system, how to use it, how to get the most out of it, how to run it on multiple machines, for example, how you can basically automate and build anything with this powerful system that basically gets updated every single day, controls all your agents, orchestrates anything that you want, and has some of the most powerful features I've ever seen with this. So for example, we can control local agents using this system, we've got Fusion built in, we've got all of the latest updates and every day something new comes out and every day we build something into it that just makes a whole system better. We've even got a second brain visualized into a memory galaxy so that we can see and visualize our memories and then also plug in contacts instantly into any agent that we want. And the thing that I would say here is, you know, if you've seen the situation with Fable 5, if you've seen the situation with GPT 5.6, where you can't even access it yet, like, the models are getting gated, they're less available, but the systems are something that you can build and then you can hold on to, and then you can automate anything you want.

And it doesn't really matter about models anymore. So with that, we can get straight into this and start answering some of the latest questions inside the AI Profit Boardroom. This is something I do every single day. So if you want to ask me any questions, you want to see me answer them in a video tutorial like this, feel free to post inside the AI Profit Boardroom community, link in the comments description or go to the AIProfitBoardroom.com.

So first question from Victor, AgentOS multi-system. Where would you install AgentOS if you run multiple machines with multiple systems? So for example, let's say you have a MacStudio running OpenCore, you've got MacBook Air running Hermes, you've got, for example, Microsoft 365 Copilot on a work machine, you also got active subscriptions to chat GPT, right? So one thing that I would recommend is that you do actually install everything in one place locally, it will just avoid everything being fragmented.

It'll be much easier and simpler to manage. But if you do really want to like insist on just making sure that everything is in different places and all fragmented all over the place, then what I'd recommend is you use something like TailScale. TailScale can link everything together. And for example, you could have your agent operating system set up on a VPS as well using TailScale.

John Meza inside the AI Profit Boardroom community, he's actually got a full tutorial on it. And it works really nicely. If you do have a setup like this, where it's super messy, fragmented all over the place, no problem. I know, I meet all sorts of people using AI, so I know how it is.

For me, I like to keep everything in one place and keep it organized and simple. If you don't want to do that, then you can have like multiple different computers, you can have your laptop, etc. all over the place. And maybe you just can't rely on a local only setup.

So what you can do is you can use a VPS, you can secure it. And then you can use something like CloudFlare as well. So CloudFlare, Access, or for example, TailScale. Two good options for managing this.

Now, one of the best things about this community is like so many people, number one, winning with agent operating systems. And what they do a lot of the time, for example, and I'll show you a great example in a minute, is that they customize it exactly how they want it to look, right? So you might see my system, so you might be like, I don't need this. I want to customize that, I want to change UI on that.

And you can do all of that with our systems, right? So for example, if you look at Sheena, who's an absolute legend inside the community, she's always sharing like amazing stories, so shout out to her. And what she's actually done is turn her agent operating system into a self-running lead machine, which is absolutely insane when you think about it. So if you look at this, basically, and she's actually set this up like a full website, and the website looks amazing, as you can see, right?

So what this does, essentially, is it's an interactive AI demo of her AI agents that she's hosted, right? So you can get started and see all the concepts and that sort of thing. And then it generates leads to her agency, which is insane. So she's basically showcasing the full power of her agent operating system, which is, you know, built around our sort of ideas and systems, but then customized to her.

So shout out to her. And then she has basically set up a website where people can explore it. And if they're like, right, I need this for my business, or I want to get Sheena to do this for me, well, then they can go through the website once they've seen the demo, right? And so I think that's just a really, really smart way to use this, number one, not just for yourself, but also for your clients, even the design of the website, like it looks super nice, it's clean, nice branding, etc.

So amazing work there. If you want to see the full breakdown of how she did it and what she did, she's got a great post. And she's got example screenshots over here too. Pretty wild stuff.

Probably one of the best posts I've seen recently inside the AI Profit Boarding. And I think that is one of the best things about having a community like this is just you can work together. I mean, look at all the people she's tagged and shouted out here. It's like, we're all winning, we're all learning together.

A rising tide raises all ships or boats. And when we share stuff together, we all learn and we grow and we become part of a community where we're all winning together. So it's fantastic. Great to see.

All right, this is actually pretty cool. I want to come back to this website. So you can see here, it says how it works. And then you go to interactive and it shows like the whole demo of all the systems she's built based on the industry and everything.

Look at this. And then you can click on this and it's a CTA to start working with it. Amazing. So we've got another question from Rick, which is about the Hermes lead machine we've built into the agent operating system.

So that's inside the outreach section over here. And Rick says, you know, just joined a couple of days ago and he's interested in starting to generate leads. Is this something that's available right now? And like William says, it's inside the lead section of the agent OS install.

So just to demo how this works, basically what you can do here is you can type in exactly the type of leads you want to generate. And then from there is going to start generating the leads as you can see right here. So it breaks down the domain, the company, their details, their status. So actually validates like, is this a real, you know, email that you can actually reach out to.

And then also you get the notes here. So this is enriched. So it has details about the company that you can use for the outreach email. Now, once you've typed in the leads and generated a list like this, you can delete any or remove any that you don't want to reach out to.

And then from there, what you can do is you can go over to the campaigns tab and inside the campaigns tab, you can see our live campaigns over here, right? So how many people are inside the list, um, how many emails we've sent, et cetera, whether it's in draft or published or it's canceled. And you can tell the AI what the campaign is about, right? So you can tell the AI, okay, I have this blog, I want to reach out to these companies.

I want to get backlinks for this page, for example, or you can just. Allow the subject, or you can write the subjects in the email yourself manually. So it's up to you. You can get AI to write it or you can get the, or you can manually write it yourself.

Then you get a live preview of the email. And then from there you can just click on create campaign, right? So it's a full lead generation machine. And once you've actually got the campaigns live, you can see how many leads you've contacted, how many are sendable.

Um, it also skips anything that's too big. So like if the companies are too big to reach out to, no point reaching out to them. And so it actually removes them from the list unless you want them to stay in. So that's basically how the whole system works.

But yeah, it's super powerful. We basically built our own SaaS tool with that. And it was just because like the amount of people that were reaching out to me and saying, listen, Hermes is amazing. How do we use it for lead generation?

And this is how, this is the only solution I've seen right now from anyone. You know, I haven't seen anyone else build something like this. So if you want to get that lead machine, it's inside the HNO S2. So great question from Yaba here.

And basically Yaba was asking, you know, if you have multiple businesses, can you run them all inside the HNO S or do you need to create multiple dashboards? So we have quite a few running through this, for example. So we have, for example, multiple websites. We have my agency, we have our community, we have all our social media running through this system.

So you can use it for multiple different businesses at a time. No problem at all. And the way you can do that is you just have separate workflows for each business that you want to run. So, for example, if we go into the SEO section, we can do SEO and this is for our community stuff, right?

Whereas, for example, if we go to the video agent, we can generate video content and that's more for our social media. And then, for example, if we need to generate outputs for our agency, well, no problem. We can go to the Hermes section here and we can automate the outreach for our agency inside this section. So if you have multiple different businesses or different projects going on, no problem.

And the same, for example, inside the Obsidian Memory Galaxy. So inside the Obsidian Memory Galaxy, we have many different projects, many different things going on, but they're all organized in separate folders. And so you can keep everything organized, even if you're working across multiple different projects or businesses or anything. Another question we got was from Bika, which is, can the agent operating system be set up with free models using these free models?

And can you get good outputs with them? Absolutely. So I'll give you a few examples. Inside Hermes, you can link free models to Hermes.

So, for example, we have North Mini, which is a free API and also a free local model. So you can get the free API from Open Router and you can also get the free local model from Olama. And then we have that plugged into our Hermes agent. The same, for example, if you go to the local section inside there, you can build free stuff using local models.

And that works as well. And then with a combination of free APIs, OAuth, so for example, you could use Hermes with OAuth and GPT 5.5 or free local models and free APIs. You don't need to pay for any APIs. You can just use the free stuff.

So we've got a question from Nova, a couple of questions on the agentic OS. So he was asking, or they were asking, you know, I see you use HyperFrames above ReMotion. Is there a reason for it? And do all your AI avatars run via Haygen or consider local models?

So Sheena's already answered this, which is awesome. So HyperFrames is just HTML. ReMotion is React and TypeScript. So for an agentic pipeline, HTML is way more reliable.

And then for avatars, she does the same as us, which is default to Haygen. And that's pretty powerful for that sort of stuff. So I've found better, I've found better responses with HyperFrames too. I found that ReMotion was a bit more slow and didn't look as good.

And then for Haygen, yeah, that's what we use and stick to. Next one we got is from Nick. Nick was asking about agent OS in Higgs field, right? Currently in the process of creating a walkthrough AI videos.

He's got the agent OS up and running and he's wondering, you know, how could you use Higgs field with Claude? So I think there's a couple of options here. Number one is you could actually install a separate workflow and ask Claude to build it for you inside your agent OS. And basically you'd have a custom workflow just for using Higgs field, right?

Higgs field, I think they have an API and they have a MCP as well. So just like how we've got the SEO section and the video agent section, you could have a section that says Higgs field, and then you build that into your agent operating system. So that's step number one, or option number one. And then option number two is the, for example, with Claude, you could give Claude access to the documentation for the API, the MCP and the API itself for Higgs field, and then Claude could actually control Higgs field directly that way.

So there's two different options you could use for the agent OS. Either give it to Claude directly or build your own custom workflow. If you want to build your own custom workflow, it's very simple. It's just like you tell the agent that built it for you, what you want to create, how you want it to look, you tweak, you iterate it, and then eventually you get something that you like, which is what we did with pretty much every single workflow inside there.

Wow. Look at this. And so Robert is absolutely crushing it. He said a big thank you to Julian and this group.

You know, he started at CRO four years ago. Struggle was real, pretty difficult. Had some ups and downs along the way. So sorry to hear about that.

Sounds like a tough journey, but now he's at this point, which is absolutely wild. So congrats on your success. Like that's absolutely amazing, inspiring stuff. Fantastic to hear about your story as well.

And, you know, credit to you for just taking all the action and working hard and keeping on going down this journey. I really respect it. Thank you. Again, like when I see stuff like this, it blows my mind.

It's just, you know, it's inspiring to see stuff like that. It's right. It's inspiring to see people's journeys. And I think that's the great thing about having a community like the Aeroprofit Boardroom is it just brings people together where we can all share our stories, it filters out all the noise and it's a way of connecting in a much better way, which I really like and respect.

So we've got another question from Victor about, you know, how can you use Hermes kind of like a voice activated agent that you can control and build stuff with? So we actually have that inside the AgentOS as well over here. So if you go to the Hermes Javis section, you can basically speak to it. You can switch off real time on and off.

So you can basically have it listen to you. So you can see it's responded directly here. It does have the voice. I'm not going to have the audio on right now, just because I don't think it will sound very good, but you can see that it's literally listening to us in real time right here.

Which is pretty cool. And then if we scroll down, we've got the full conversation and we've got all the builds and the creations we've made. We can put it in warm mode as well. And then we can have a wake word as well.

So this means we can have a word here that actually wakes it up. So if we say like Hermes wake up, it will respond to us. As you can see right here, it's glowed up since we gave it the wake word. So it's pretty powerful stuff.

So you got a question here as well from John about, you know, how secure is OpenClaw, is it safe to use, et cetera. Here's the way that I look at it. I mean, number one, I don't really use OpenClaw that much because of those issues. Number two, I keep everything local so that it's all sandboxed away on one machine.

And number three, I think if you're not sure, don't use it. So for example, if you're not certain about OpenClaw, then don't use it. You know, I think that's always a good sign as a heads up, you know, intuitively, if you're not sure about how secure it is, don't use it in the first place until you are comfortable with it, until you understand the power of it completely. So that's the way that I would look at it.

And then when it comes to agents, if you want something that's more stable and reliable, then just stick with Claude, you know, you can use inside the agent OS, you can use Claude CLI, or you could use GLM 5.2 CLI. And the CLI options are a lot more stable and reliable than relying on something that's open source like OpenClaw. So that's the way that I would look at it. You don't have to rush into doing anything, but I think it's always best to rely on your intuition and it's down to each person.

So for example, someone who has a lot of experience of agents, they're not going to worry about using OpenClaw at all. Someone who's new to this and a beginner would probably have a lot of doubts about using OpenClaw and they should listen to that and just stay away from it, you know, until they are comfortable with it. We've got another question here from, I think it's pronounced Jerd, but I'm not 100% certain. So sorry for that.

I'm building a board game app for friends. They say developing a game, any tips on how to make good games with AI? So one thing that we recently used is this open source GitHub skill. It's called 3JS game skills.

This seems pretty good. When we installed this into Claude to build games for us, the quality of our games became a lot better. And I'll show you an example. So for example, if we go in to, let's have a look here.

I think probably Hermes mixture of agents would be a good example. So if we go to Hermes mixture of agents, we built out all this different stuff using the 3JS skill. A lot of it looks really cool. I mean, like that's really cool as you can see right here.

And then also we have this one too, and this one too, right? And so like, you can see the quality of this stuff looks pretty cool. Really interesting to play, definitely helped level up the stuff that we're creating if it's like a 3D generation. So I'd recommend that to you too.

Really good skill. And it's free and open source to use. If you want to see more of our creations, what we've built, we've actually got every model and what we built with it on GoldieBench. So you can see, for example, if we scroll down here, you see all the examples, and then if you want to test them out and have a look yourself, you can check them out right here.

So Daniel was talking about, you know, GPT 5.6 is in preview. It's not released yet. This is why I'm focusing more on the system than the model. So with the Agent OS, we can achieve better levels of intelligence than what's already out there using mixture of agents or Sakana Fugu or Fusion.

Because they're all panels of agents working together, which means you get better outputs overall. So that's what I'd recommend for anyone watching this is just focus on the systems, not the models, because the system's yours, but the models can get taken down or removed or blocked at any time. So it's always better to focus on the system instead. And that's basically it.

We've answered every single question from the AI Profit Boredom community. I do this every single day, as I mentioned, if you want to get the agent operating system that you saw today, you can get it inside the classroom, just go to the new daily updates and you'll find the Agent OS section over here, we update it daily. You get a video tutorial and a guide on how to install it, and then you get the zip file to install it too. And any time there's something useful that comes out, we create a video tutorial and a step-by-step guide on exactly how to get the most out of it.

And also you can post questions inside the community. Also, the cool thing about this is you don't just get help from me. You get it from everyone online. So for example, there's 3,800 members inside here and there's 113 people online right now, which means 24 seven, you can pretty much get the answers that you need to win with AI.

Now inside the calendar, you can jump on with coaching calls, share your screen, ask questions in real time. On the map, you can actually meet people in your local area who are building with AI agents like you. And then inside the classroom, you get access to all of our best training. So that's all inside the AI Profit Boredom, link in the comments description, or just go to the AIprofitboredom.com.

Thanks for watching.