China’s GLM 5.2 + Hermes AI Agent is INSANE!

GLM 5.2 plus Hermes, how do they perform? So we actually created some amazing stuff already by plugging the new GLM 5.2 update into Hermes Agent. It's super powerful. Here's an example of a video we actually created using this new update.

And basically what we've done here is we've got a team of Hermes Agents to create an avatar of me, to create the voiceover, to create the script, to edit it. Even for example, like the effects on this, also the editing right here, and even the portfolio of stuff it's showing, this was all done using Hermes plus GLM 5.2, which is pretty crazy in itself. So it was researched, scripted, edited, and judged by a team of GLM 5.2 Hermes-powered Agents. And this is really powerful stuff.

It's actually very impressive. Now, the difference here, of course, is like if you're using something like Claude, Claude is great, but the problem with Claude is like you can't plug it into Hermes unless you're using the API, which is quite expensive. Whereas what you could do with GLM 5.2 is you can actually get the coding plan and then plug that into Hermes, and you can also create teams of Agents. Now, if you want to indicate like, is GLM 5.2 any good, et cetera, let me show you some examples of what we've built with it so far.

Some of it with Hermes Agents, some of it not, but you can see an example of what we've actually done directly with Hermes here inside this video with the team of Agents working together. So how does this work? Well, I mean, if you look at some of the server stuff that we've built out here, this is another example of a 3D game we built. Here's another one.

We created a full operating system with individual apps inside here as well, using the system. This was kind of like a fun game that we created. It's pretty fun. Synthwave style, background, these particles, et cetera.

And the other interesting thing about this was I actually compared them side by side. So if you look, we've got Opus 4.8 versus KimiK217. GLM 5.2 creates the best stuff if you have a look here. This is the most fun game to play out of all of them.

And the same, for example, if we compare them on this liquid bowl example, this creates something way more interesting and fun to use. It was also better at creating websites as well. So for a lot of agentic tasks, GLM 5.2 is way better, right? It looks way more interesting.

So if you're wondering how to use Hermes with GLM 5.2, the way to do this is you go to your model settings over here and you can add a new model, right? So you can configure it right here. The other option that you have is that you would go to your terminal and inside your terminal, you would just make sure that you have, you type in Hermes model and then you can switch from there. So that's another option to do it.

And once you've done that, you can then change the model. So if we type in Kimi here, you can see that we can select KimiK217. Obviously we want to use GLM 5.2, which we've plugged into our models and we've created separate profiles using that. So if we go inside our profile section in the Hermes dashboard, we've got GLM 5.2 plugged in right there with ZAI.

And what we've actually done is created multiple different profiles with Hermes agent so they can work together as a team. Now you might be wondering, okay, why would you do that? Well, for example, what you can do is you can have like a team of agents that work on one work that you'll focus on improving. So for example, if I'm trying to create content with GLM 5.2 and Hermes, I can have a GLM SEO content writer, I can have a writer, I can have a researcher, I could have a judge that actually looks at the work and tells my team of agents to iterate until the work is done.

So these are all powerful ways that you can actually use GLM 5.2 with Hermes agent, which is pretty amazing itself. Then what we also have is I have a separate chat for each of my agents here. I think you can also do the same inside Hermes desktop too, but I have it inside my agent operating system. And so what we can do from here is we can have GLM 5.2 as a separate chat and then we can speak to it directly.

Now, if you're wondering, okay, how do you organize your teams of agents? There's two ways I like to do it. So one way is that we will actually have a setup that looks something like this, where we can basically plug in a task, it will triage it, and then it will get done. So if you look at these previous outputs here, we can see everything that we've created with our team of GLM agents, right?

They're all working together to create and build something. So they actually create a full blog post and a website, which I'll show you in a second. And then also the other option you have is you can run them with Paperclip. So you could create a team of agents like you see in this org chat here, all running on Hermes, and they can be plugged in with GLM 5.2 as the model for each of those agents.

And that's pretty amazing itself because then you can build stuff, you can see everything that we've built with our Hermes agents and teams of them. So there's many different ways to coordinate it. You can have Hermes inside the chat, you could have Hermes Jarvis with GLM 5.2 plugged in, this is a voice activated version of Hermes. You can plug it into Paperclip and have a team of agents working right there.

And then you could also have a Kanban board where your agents just tackle one task together and to get stuff done. That's another amazing way to do this. The great thing I like about that as well is like we can watch it back later and we can come back to that later. So for example, if we have a look at this, we've got all the files right there ready to go and we can just come back and watch them and use them whenever we need to.

So we also created this like open world game with GLM 5.2. This one as well was created with GLM 5.2 so you can create like all sorts of interesting visual stuff as well. Here's another example. And the thing I like about using Hermes with this is like, because you've got separate profiles, you can have that judge who tells them to just keep going and improving until the work is done properly.

Whereas if you're just going back and forth inside the CLI, you don't really have a team of agents that are gonna work and improve together, right? It's just gonna be like, okay, here's a job done. Now, if you wanna see an example of that SEO project that I just showed you a second ago, you can see it right here. So this is a full website built out with multiple agents and it actually talks about the process that it created.

So for example, you can see here that it's created a blog post on how we have a crew of agents with Hermes working together. So it's kind of like a self-aware agent that can create a website and a guide about what it does and how it works so that anyone can understand it. So you can see here that we had a researcher, a writer, a builder, and a judge to score it. And then that's how it worked step-by-step.

And the design of this website is super nice. Now, was this created in one go? No, but literally within like 30 minutes, we'd created a full guide for this as well. It's pretty amazing.

Like the website looks super nice. The blog posts are nice. The other thing that I noticed when I built out a website with GLM 5.2 is that it's really, really fast. So if we switch between pages here, look how fast it switches, like it's crazy.

And you might say, well, why would you wanna build out so many websites with Hermes agents and that sort of thing? Well, this is actually the system that we're using for this website and it's growing really nicely with SEO using the similar sort of process. So you can create awesome websites. You can grow them.

As you can see here, this website went from like two clicks a day all the way up to, it's getting now 73 clicks per day. And the trajectory of that website is really good for SEO. So, you know, this is great for SEO or any sort of automation that you wanna build out. By the way, if you want a free one-to-one SEO strategy session, go to Goldie Agency and we can look at your website.

There's also a free course link inside the comments and description. So that's basically it. So, you know, we had a team of agents working together. They built out an amazing website.

They built out some awesome projects like you saw today and very, very visual, nice designs. I tested it versus Opus 4.8 as well. And it seems to work really nicely. And you can set this up easily as well.

And then you can have GLM 5.2 sitting next to all your other AI agents. We can switch between the model, the chat. We can save stuff inside our workspace and see it later. So what I like to do is have a workspace separated by the API I've actually used.

And the cool thing about that is we can preview and look back at, okay, which API do I like the most? And I think that's a great thing about having an agent operating system, like you can see. Now, you might say this is technical or whatever, but you can see here, loads of people inside the AI Profitable Boardroom are building out similar things and they're getting awesome results. So I know if I can and they can, then you can too.

And, you know, if you look at this in terms of the old way versus the new way, if you're just using like Claude, well, you're kind of stuck inside Claude. You know, you've got one coding tool, one model really, or one or two models. It's hard to change. If, for example, like Fable 5 gets taken down, which it did recently, then the whole conversation you're having just gets lost.

Also, without a system like this, where you don't have agents together, they're just in different tabs with no shared memory, there's nothing saved in one workspace. And also, if you're using Claude with agents particularly, then you have to pay, right, which can get expensive. Whereas, for example, if you have a crew of like GLM 5.2 or Kimmy K 2.7, agents working together, or every model is an agent you pick, right? You can add a new profile in minutes, as I've shown you today inside the dashboard.

You can have one chat, one workspace, one Kanban for each of your agents. Every build is saved inside the workspace. And also because GLM 5.2 is on a flat plan, you don't have to use, you don't have to worry about API tokens. And then if a new model comes out, like for example, Kimmy K 2.7 comes out, or for example, GLM 5.2 just came out recently, no problem, right?

You can just plug it into your system, which saves a lot of time. Now you also might think, okay, well, I've already got an agent, why would I add another one? Because with this system, you just let your crew pick the best job. So what I like about this as well, is like if we're using the Kanban board, when we assign a task to Hermes, it will automatically pick the best profiles based on their skills and their profiles.

So if we have like a video writer, well, it's gonna pick the writer if it's a task related to videos. And that's how powerful it is. That's how smart Hermes is. Now, if you want to indicate how do you set this up with Hermes, again, you can go to terminal and you just point it at, with the model.

You can put it on a board with the rest of your AI agents. And also GLM 5.2, this is interesting, it's so new that it's not even out inside the chat, but you can switch the model on the API, which is what we've done here with the coding plan. Now, some people say, well, I should pick one AI and stick with it. I think honestly, like the models change all the time, right?

The models are always changing. So the great thing about this is whatever comes out next, you can just swap the models in. I always, other people say, well, you know, running multiple agents, it sounds complicated, but it's just one dashboard and you pick an agent as you can see here. So whichever agent you want to use, you can do that.

Other people say, well, I need, I'll set it up later, but later is really how you stop making progress. You want to start setting this up today, really and getting the most out of this stuff. So that's how to set up a crew of Hermes agents with GLM 4.5. You've seen all the cool stuff we've built with it.

Pretty amazing. Let me just recap and show you here. So we created that AI avatar video. We created like a full game.

Here's another one. The individual apps inside the system. And it's a lot of fun to build and play with, as you can see. So thanks so much for watching.

If you want to get my full setup for the agent operating system with Hermes, you can get that inside the AI Profit Boarding. Link in the comments description or go to the AIProfitBoarding.com. So you can get the agent operating system here. You can see we've just updated it today.

You can get the zip file to install it. We add new daily tutorials, as you can see. Inside the community, you can ask questions and I personally answer them as well. And then also you can jump on weekly coaching calls and get help and support in real time.

Inside the map, you can connect with people in your local area who use an AI agents like you. And this is all inside the AI Profit Boarding. Thanks for watching.