DeepSeek V4.1 Flash + Hermes Agent (FREE!)

DeepSeek V4.1 Flash and Hermes Agent. I want to show you number one how to connect it to Hermes, how to use DeepSeek V4.1 Flash inside Hermes Agent. So you can see we've got Hermes Agent ready to go over here and we can plug it into all of our favorite workflows with DeepSeek. But the other cool thing about this is that you can actually get a free API and you can use that as the brain with DeepSeek V4.1 Flash and I'll show you how to do that in a second. So basically with Token Harbor for a limited time only, you can see here that you get DeepSeek V4.1 Flash free on the API. Why would you want the API for this? Because you can plug it in as a free brain into Hermes Agent and use it agentically or you can just use a normal version of DeepSeek which is probably better than this but either way you can use it for free. So let me guide you through exactly how to use it step by step. So what you're going to do first of all is you're going to go to Token Harbor. So TokenHarbor.ai, this is where you can get the free API key and you can see here I've created one in my settings. I just created a free account and then signed in and then got the free API key. And once you've got that API key as you can see right here, you can plug it into whatever you want, right? So for example, you can use it inside DeepSeek Harness, you could use it inside Hermes Agent, whatever you prefer. Okay. Now for me, I've already plugged it into DeepSeek Harness. It's pretty decent. So you can see we've got DeepSeek Harness over here inside our Agent OS. Link in the comment description if you want to get the Agent OS. And you can see here that we're using Token Harbor with the free API. Now let's plug it into Hermes Agent as well. So the way that I like to do this is I like to create a separate profile for each brain that I use. When I say brain, I just mean, for example, API. So we've got a GPT-6 Astra. We have the Research Agent. We have LFM 2.5, 2.6b, which is a free local model. We've got GPT-5.6. And what we're going to do now is plug in Token Harbor, the V4.1 Flash into Hermes Agent. How do we do that? The easiest way that I've found, and you can get it ready-made if you want inside the Agent OS system in the AirProp volume. The easiest way I've found to set this up is basically we're going to go inside ChatGPT, GPT-6 Astra. And we're going to say, okay, can you create a new profile for Hermes Agent with DeepSeek V4.1 Flash on Token Harbor so that I can test it out inside the Agent OS as a separate profile? Just make sure it actually works as well. And then you can see here I've added the API key and I've said, can you set this up? And also one thing that's quite important to do here is we've got the documentation from Token Harbor on exactly how the AI can understand to use this. This is really important because it trains the AI on how to use that free API and set it up inside Hermes Agent. So add in the details right there. And if you're wondering, okay, where'd you get that documentation from? So inside Token Harbor here, if you go to connect and then integration in the menu bar, you can see all the details and you can plug that into any AI agent that you prefer. And then it basically teaches your AI agents how to use it. You can also see there's like optional details here on how to connect it. You can use it inside like any time, Goose or OpenClaw, whatever you want. But for me personally today, we're going to use it inside Hermes. Boomshakalaka. So it's now adding a new Hermes profile for Token Harbor. And by the way, if you're wondering, okay, is DeepSeek V4.1 Flash actually good? Let me show you some examples of what it created for me earlier today. So here's an example of a game we created. And I will say the outputs are super nice, right? Like it creates some pretty cool stuff, as you can see right here. For example, it created this website as well. It's like a 3D generated website with a parallax scroll and everything that looks quite nice. And then for example, we've got this full tool that we created, like productivity tool, where we can use it as a Pomodoro timer. It's got the details of, you know, our progress, our streak so far, and our to-do list over here, and we can mark it as done. We can also import tasks and export them as well, or we can clear them all as well. So it creates like some really, really nice stuff, is basically my point here. And it's all inside one page. So if we now go back to GPT-6 Astra, you can see that it has now created the profile for DeepSeek V4.1 Token Harbor. So now what we're going to do, so we're going to go back to the agent OS. We'll go over to the Hermes section, you see on the left-hand side here. We're going to refresh a page here. And then we are going to scroll down to DeepSeek 4.1 Token Harbor, as you can see right here. I'm just going to make sure it works. Sometimes it doesn't work when you set this up, so I just want to make sure this actually works. And look at that. Wow. It is really fast to reply. This is a free API. And bear in mind, like, this is just something to bear in mind. Like, sometimes these APIs get rate limited, or for example, they're not, they're not like free forever. Just use it once you can. I think it's for like a couple of weeks or something like that. But it's another free option for a free brain. And it's super fast inside Hermes, as you can see right here. Let's just do another little test. So what I'm going to do is I'm going to go over to this documentation on Token Harbor. And what I'll do from here is we'll go inside the agent OS, run the forward slash learn command, which basically teaches Hermes agent a skill from a guide. So if you point at a guide, you can look at that guide and then create a new skill for it. So I'm going to say forward slash learn, and then I'm going to plug in the details of how to connect Token Harbor. And let's just see if it can actually run agentically. One thing to note here, just make sure you have selected DeepSeek Token Harbor, as you can see right here. So we'll plug in forward slash learn as a command, and then you can see it's running. Now I will say when I pointed it, when I asked, does it work? It literally replied in like two seconds. So it's super, super fast, which is absolutely amazing. Absolutely loving it right now, to be fair. And by the way, I will delete this API key. I'm just showing you as a little demo. Obviously, I'm going to delete the API key after this has been done, but it's a pretty cool feature. The other cool thing that you could do with this is like, for example, inside our agent OS, you can build like custom workflows. So for example, we have like Hermes Oracle here, which refreshes every 24 hours, pulls in the latest news. We have Hermes Apollo, which is a voice activated agent. With Hermes agent, we have Hermes Muse, which basically analyzes our competitors and gives us new content ideas. And then we have Hermes Astros, which does something similar, right? So Hermes Muse actually looks at my content, and then gives me more ideas based on what just worked. And then Hermes Astros actually pulls in from our competitors and looks at what our competitors are talking about, and then gives us ideas based on that, which is pretty cool too. So if we go back to Hermes here, you can see it's working his magic. On token Harbor. And yeah, really cool. And also it's pretty good inside deep seek harness as well. The other thing that I would say is if you've looked at the benchmarks, really high performing model, like it's right up there with V4 Pro. I don't know if they're going to release like a new version of V4 Pro as well soon, but you can see like just a month on since it's released. So deep seek V4 flash was released on the 31st of July. V4.1 flash was released and it's quite a big step up as you can see right here. And the paper is actually really interesting about it as well. So it's just a lot more efficient as an architecture, which means you get lower API costs as well, which is probably why like token Harbor can give us the API for free. So that is basically it. That is the whole agent OS, how to plug in Hermes agent with deep seek V4.1. We test it out actually is pretty good when we've tested out as well. It works really well. And also you can use it in deep seek harness. And the cool thing about token Harbor as a free API is like you can plug it into any agent that you want, any agentic harness. It could be open core, for example, whatever you prefer. So that's basically, if you want to get this full agent OS system, that's available inside the AI Profit Boredom link in the comment subscription or go to the AI Profit Boardroom.com. You can see we're getting loads of great reviews, which is absolutely amazing. So I appreciate everyone who's done that. And also inside the community, you can ask questions, get help and support. I personally create video tutorials inside there personally for the questions that you ask as well, just to help you as much as I can. So for example, you can see all these great questions here and lots of people. It's just a really active communities where everyone helps each other out as much as they can. Inside the classroom, you can get access to all our new trainings, best lessons, et cetera. And then inside the calendar, you can jump a weekly coaching course, ask questions, get help and support, and everything else. Inside the map, you can meet people in your local area and if it's available inside the AI Profit Boardroom link in the comment subscription or go to the AI Profit Boardroom.com. We also have full training, a full one hour course on how to use deep seek harness and also a full course on Hermes agents. All my best trainings on deep seek and harness are inside there as well.