Hermes Agent: Automate ANYTHING! 🤯

Hermes agent is already one of the most powerful AI tools you can use right now, but most people are only using about 10% of what it can actually do. In this video, I'm going to show you how to make your Hermes agent 10 times more powerful. We're talking about turning it from a basic chat tool into a fully autonomous AI system that runs tasks, builds things for you and gets work done without you having to manage it. I'm going to show you the exact settings I use personally, the ones that make Hermes work harder, think smarter and actually run your business in the background whilst you focus on other things.

There's one setting most people have never touched that completely changes how your agent behaves and I'll show you exactly how to use it. And I'll show you how to run a full team of Hermes agents working in parallel at the same time for free. By the end of this video, your Hermes agent is going to be doing 10 times more work for you than it is right now. Let's get into it.

So let's kick off this Q and A on Hermes agents from the AI profit boardroom community. So the first question that we had was about Mac mini and performance issues. Now Mike Powell actually replied on this and said that one of the best ways he uses Hermes agent and he has a Mac mini as well is he doesn't set it up locally, right? So he sets it up on Hostinger and then OpenCourt runs really well like that.

So I think that's a great way to run agents. If you're having problems running Hermes locally, then instead what you can actually do is run it on something like Hostinger and with Hostinger it's cloud based, right? So it's much easier to host and it's much easier to start using. Let's have a look at the next one.

We had a lot of questions about Hermes agent today. So Karen says, quick context, I am a non dev building a personal AI system to support life in general, meals, fridge, et cetera, and business too. I'm a health coach and I have my own specific framework I want infused in the support. The vision is an agent team that holds my method and runs in the background.

Here's the infrastructure is where I'm stuck, right? It feels less autonomous than expected. This is interesting. And then Karen says, you know, where I'm stuck, it feels less autonomous than expected, more like just a basic chat with file access.

The orchestrator doesn't respond well and still noticing the patterns, not being proactive. Basically, that's the biggest issue with Hermes agent right now. So she says, it feels like I need to keep asking and chasing. I used a chat agents and this is what I've been doing so far provided me support, but obviously now I want to upgrade it with agents.

So I'm wondering if part of the problem is that migrated into a fully built vault from a previous platform. So for example, like Claude co-work and fed the orchestrator, the existing details. So here's the question. Have others found this and does GPT 5.5 just need heavier prompting to be proactive or does Hermes work better when you let the agent build its own context from scratch through use, considering archiving everything and restarting.

Also, if I do read archive and restart, what's the best way to do this? Also open to other stuff like this. Okay, cool. All right.

This is interesting. So here's the way that I would look at it. If you want your agents to be autonomous, right? Let's say for example, you've got Hermes agent.

Like there's a few ways you can make it much more autonomous when you're working with this. So let me pull this up right here. The first method is the sole MD file, right? So this changes how your agent response and you can easily, you can actually navigate to your sole MD file and view it.

So if we say to the Hermes agent here, okay, where is my sole MD file? Basically the sole MD is how your agent responds plus thinks. So it's like a custom prompt for the agent and its thinking methods. So the way that I would organize this is that you want it to be super proactive, right?

That's what it sounds like. You want it to be autonomous. You want it to be proactive. So what you can say here is edit this.

So it's more proactive, more autonomous. Plus also ask your Hermes agent, how can I make you more autonomous? Right? So that it responds better and it replies quicker and that sort of thing.

Leonard says, what about being able to use Hermes agent on a PC directly? Yeah, you can do that directly, right? Like you can just set this up from the GitHub. You haven't seen that before.

If you want to set up Hermes agent directly, which is what I do, right? I have it running locally as you can see right here inside the agent running locally on my Mac studio. And so you can run it on windows. So you can run it on Mac locally.

It's pretty easy. You just follow the instructions right here. So if you're running it on a PC, you just copy this command, paste it into your terminal or whatever. And then you can go from there.

That's pretty much it. Actually going to switch models over here. So you can see, for example, this is my sole MD file for Hermes agent and you see how it basically is be genuinely helpful, have opinions, be resourceful. And you could say, for example, be autonomous or be more proactive.

So that's the way that I would look at it is improve your sole MD file and just edit it. You can actually give Hermes agent instructions inside the terminal and say be more proactive or edit your sole MD file so that you're more autonomous, that sort of thing. Right? So just coming back to this, your sole MD file changes how your agent thinks and responds, right?

Also you can ask your Hermes agent, how can I make you more autonomous? That's another option. The other way that you can do this is you can use something like forward slash goal, right? So if you have, this is not so much for a new daily workflow, but if you've got like a one-off task, you can use forward slash goal here.

And this equals like a autonomous persistent loop. Now we actually have a training on this inside the AI Profit Boarding, which you can see right here on Hermes goals. And you can use this to resolve tasks way faster because it has 20 loops to get it done. So that's another one.

What else would I say here? Yeah, I think those are it. And then you got the sole MD file, you got your, how you can ask it and then just make changes based on that. But GPT 5.5 can be more proactive, but again, edit the sole MD file, ask it directly, do the forward slash goal.

I wouldn't restart. I know you mentioned that too. So don't restart or delete your setup. I don't think that's required.

And I think that would cost you a lot of time. Just improve it. And that's way faster. And that way you still got all the context.

All right. So those are some really good ways you can make Hermes much more autonomous and much more powerful. Let's have a look here. What else we got inside the questions now?

It says, Hey Jules, good to see you here. Rue says hello. Good to see you too. Do you need a Mac mini to have agents or can I just use my MacBook?

Yeah. You can just use your MacBook. Well, the first time I used OpenClaw, I just used my MacBook. It was super chill, right?

So you don't need a powerful setup to run Hermes. It's pretty lightweight, pretty fast and easy to use. So you don't need like super powerful way for this. We actually have a lot of training on that inside the classroom here.

So you can see this section right here. That's perfect for this sort of stuff. Can you show how to change the model and point it to a local LLM? Yeah, sure.

So if you want to set up Hermes agent with a local LLM, what you do is you just type forward slash model here, right? So forward slash model, and then you pick your model. For example, you can select whichever one you want. Now, one thing that I would say here is when you're doing this, you need to go to Hermes setup inside your terminal.

And then from here, you would select LLM studio, or you would use Ollama, right? So you have to add LLM studio or Ollama with your Hermes agent first, before you can do that. So if we wanted to set up LLM studio, we just configure it like, and then we're good to go. Right.

And that's pretty much how it works. You don't need to add API key. You can use the base URL usually if you don't have any custom setup, and then you can add like the provider name and everything else. Right.

That's it. And make sure that you have LLM studio running in the background. And then inside the chat, just do forward slash model anytime you want. We actually have a full setup on LLM studio here.

So full training on LLM studio with Hermes agent here, if you want to run local LLMs with it. And then also we have another one on Ollama with Hermes over here, right? So you can run Ollama and LLM studio for free with Hermes agent. It's pretty chill.

Great question. Can Hermes run multiple agents in parallel? Yes. So what you do is you go to something like Hermes workspace.

Let me show you an example. All right. So we've got Hermes workspace here. This is a free extension for Hermes agent that you can get it on GitHub.

And what you would do is you've got two options, but my favorite is to go to the swarm section, right? So you can set up agent swarms and then you can see here, for example, we have 17 agents from Hermes that are ready to go, ready to start building and creating. You can see them all in the little office right there. So you can have multiple agents working in parallel, just like this.

The other option that you have is you can actually switch between profiles and you can use the conductor set up to with Hermes agent workspace. But my personal favorite is a swarm feature. I think that's the most powerful for running Hermes agents in parallel. I can't get Hermes to work on windows.

So if you're not sure how to run it, follow the instructions on the GitHub. If you've already followed the instructions on Hermes agent, GitHub to send up on windows. What I would do in that situation is I would get Claude code locally, right? So I'd open up Claude code locally like this.

And I just say it's Claude code and I paste in this information here. And I'd be like, based on this, help me set up Hermes agent, right? And just keep iterating until it's done. That's the way that I would usually approach stuff like that.

I'm trying to decide if I should use Hermes agent over open code. So open code is great for coding tasks, right? It's like an idea. It's great for like building apps or tools or websites of that sort of thing.

But if you want an actual agent, then you wouldn't use open code, right? Open code is just for coding tasks. But if you want like a personal assistant sort of thing with AI, then you would use Hermes instead directly. Thanks Ben.

We're going to access your platform. So just go to the AI Profit Boarding link in the comments description or go to the AI Profit Boarding dot com to get access to the platform. Do you want to have a go? I'm going to try a Linux on a spare desktop.

I want it. Yeah. It could help with that. I believe you've discussed multicore in the past.

Why did you switch to Aion UI? I just test everything out, right? So for me, I think it's important to try new things, to test things out. So I test Aion, I test multicore.

I like to switch between them and just see what's working the best. So yeah. And also they have different, different benefits really. But yeah, I like to switch between them and just test.

And I wouldn't say one's better than the other. So we've answered Karen's question here about how to make Hermes more autonomous. Let's have a look at the next one. Also answered how to fix Hermes agents if you're running on a Mac mini or if you're having performance issues.

Let's have a look here. Anyone running OpenCore and Hermes? I feel like it's a Mac versus PC type of thing. Thinking about trying Hermes, actually I take it back.

So for me personally, here's why I use Hermes instead of OpenCore. Easier to use, great open source community. And I see a lot more, when I say open source, like they both have amazing open source communities. But the thing is, with Hermes, I see a lot more interesting projects coming out versus OpenCore.

So that's why you'll see, for example, I test out multicore, I test out Aion UI, I test out Hermes Web UI, Hermes Workspace, etc. The reason I do that is because there's loads of cool projects with Hermes. And I don't see that much coming out for OpenCore specifically. Then also what I'd say here is like smoother plus works every time.

Whereas, for example, OpenCore actually breaks a lot, especially with every update. And also I just find like Hermes actually does what I want. OpenCore seems to struggle. And also if I look at the way that the world is going, and I think this is interesting to pay attention to, if you look at the way the world is going, like when we look at OpenCore and we look at Hermes Agent, what you can see here is that Hermes is growing massively in terms of trajectory and OpenCore is just dropping.

And so what does that tell me? It tells me the world thinks the same thing. Now, both of them are great, but I personally prefer Hermes because of this. And honestly, since I've switched from OpenCore to Hermes, I've used AI agents way more and I've got way more powerful use cases out of them.

So that's why I recommend this stuff. Also, what I like seeing here is everyone just helping each other out inside this community. It's awesome. Look at the quality of responses and the way everyone just helps each other and raises each other up.

And that's awesome. It is a great community of people. Let's see what else we've got here. This is interesting.

So Joe says, Hermes Workspace is awesome, right? He deleted OpenCore, installed Hermes Agent, installed Hermes Workspace. Huge win for the weekend. It was a lot of work and many hours of frustration, but he made it.

He said, thank you, Julian, for posting that Hermes Workspace is amazing. Now, how to implement it. Finally feel like I'm accomplishing something. So relieved.

That's amazing. So fair play to Joe. Number one, shout out to Joe for just being an awesome action taker. But yeah, it's amazing.

I'm actually, in terms of Hermes Agent and Workspace, let's have a look at this and how to set up. So if you've never used Hermes Workspace, basically what this is like a really powerful way to manage your agent. So instead of using the terminal, you can just use this nice user interface. And it's a lot smoother.

It's a lot easier to use, et cetera. Right. And so when you're running this, there's some really cool ways you can do this. And I'll tell you about some actual use cases in terms of how I use Hermes Workspace.

Right. And actually I'll include some links to this stuff as well. So number one, you can actually connect it to your phone, right? So you can connect Hermes Workspace to your phone, which is pretty nice.

So that's super useful. Also, it's a much, it's much easier and more fun to use Hermes Workspace than it is to use directly inside the terminal, which I think this is enough to put 99% of people off. Look at the terminal and all the code that comes out here. Also, they look like a really cool Hermes Agent game.

I'll show you that in a second and what it looks like, but it's really cool. And then also what I would say here is Agent Swarms are crazy powerful, right? For example, we actually built out a full website, a full 50 page website with Agent Swarms. So there's so much you can do with it.

And also it's not really about the use cases, but more about making things easier, plus more fun to use. That's what I would say here. So if we go to Hermes Workspace here, right, you've got the chat, you've got the chat, you have the files. Also, it's really good for just managing your files and that sort of thing.

Right? So for example, I can edit my memory. We were talking about sole MD, so you can search for anything here and find it, which is pretty useful. You can also manage any skills that you have.

So it's really good for just managing your skills. You can chat with it directly. You can use the Kanban board, which is pretty useful. If you've got a project that you're working on.

And then also we have, I think this is the most useful feature, which is the Swarms feature, super powerful. That's how we build the website. And then also you can easily manage like your memory skills, MCP, et cetera. How do you keep up to date?

By just share, I think just by documenting my process along the way, right? Every day, documenting, recording what I'm working on, how I'm using it, et cetera. And also I would say X is really good. X as a platform is amazing for staying up to date, super useful and really easy to keep up to date if you're using X.

And then also the AeroProfit board and community, there's people always posting and sharing new cool stuff. So it's easy to stay up to date inside that too. Can I configure Hermes agent so that different Telegram groups use different dedicated agents? Yeah, for sure.

So what you would do here is you're going to approach it like this, right? So essentially you would step number one, set up different profiles inside Hermes agent, right? So you set up different profiles inside Hermes agent. So you can say, for example, okay, Hermes profile create Julian, right?

And then you set it up and you set up a separate profile for marketing support, et cetera. Right. Step number two. And then step number three is you connect each one separately to different Telegram agents profiles, and then step number four, add them to the groups.

So that's how you do it. So you would set up different profiles inside Hermes agent, set up a separate profile for marketing and support, et cetera, connect each one separately to different Telegram group profiles, and then add them to groups. That's basically how it works. Step-by-step how to set up a self-hosted agent.

So the way that I would do this, if you want a basic one, go with Hermes and just follow the quick start instructions here, then ask it. Also, you could set up a team Claude account to share, then you can make different projects private. And the third option is that you can set up something like Maltworkers. Right.

So Maltworker is pretty good as well, but it's quite time consuming to set up. So just to recap there, right. Daybiz is like how to set up a self-hosted agent for a company. The self-hosted is go-to resource.

People can ask questions about events, marketing performance, et cetera. Each department has a specific channel in which to communicate with the agent, but there's also restrictions, et cetera. Right. And so for me personally, I think actually I'd go with number one.

So if you want a basic one, you could go with Hermes. Then you could add to discord with separate discord channels for each department, and you could edit the instructions of the agent by speaking to it directly where you host it and ask it to be configured based on your use case. Right. Also, you could set up a team Claude account to share, right.

And then the other option is setting up Maltworkers. So there's three different options that you could take right there. Awesome. All right.

We've answered all the questions inside the community with a custom video and step-by-step arrangements. We've covered how to set up a self-hosted agent for a business and coordinate across different channels with different access, depending on what department they're in, we've covered how to fix performance issues with Hermes agent. We have also covered open core versus Hermes and why I switch between them. And then we've also covered Hermes workspace and the best use cases for that.

Right now, if you want your question answered like this, I do this every day for the AI Profit Boredom community. So feel free to join link in the comments description or go to the AI Profit Boredom. You can ask questions inside there. You also get access to 3000 people who are also using AI agents just like me.

Inside the calendar, you get four weekly coaching calls. You can get help and support whenever you want to. You can connect with people inside your local city who are also using AI automation and AI agents like Hermes and open core. And then also inside the classroom, you get new daily tutorials and all of my best trainings.

Right. So we have a class on how to go from beginner to expert with AI automation. And we also have new daily advanced tutorials right here. Right.

So if you want to learn this stuff, you can get it all inside the AI Profit Boredom. We update this all the time. We're just sharing new cool stuff. Plus you can connect with me personally inside there.