OpenClaw + Hermes is INSANE

**Eric Siu** (0:00)
I have hired a lot of marketers over the years. Some were great, most were average, all of them had limits. They forgot things, they needed managing, they stopped improving the moment they got comfortable. My OpenClaw and Hermes agents don't do any of that. They run every morning before I wake up, they get better every day, and they cost less than one month of a junior hire. Here's why I would take them over the vast majority of marketers I've ever worked with. So first, we gotta talk about what OpenClaw and Hermes actually do. Hermes is the newer one, and Hermes is faster, Hermes self improves. Hermes is a great brain for what you have going on with OpenClaw. OpenClaw is an autonomous agent, and a lot of people talk about it, and OpenClaw is great for being the execution. Okay, you want it to be your arms, you want Hermes to be your brain. And the cool thing is you could have each of them hold each other in check. So sometimes my OpenClaw will forget things, but Hermes will kick it back into gear and say, hey, yo, like you actually forgot this. No, you actually can't do that right now. I know actually hold it accountable to make sure it's getting done. The other cool thing is sometimes OpenClaw will go down because we use it pretty heavily at my company. Let's say OpenClaw goes down, Hermes will catch it, and it'll just do a gateway restart and it'll reset it. I used to have to do that myself. You don't want to spend time doing that. You should let Hermes do that type of work for you to make sure that the system is continuing to run and that's why you want it to be the brain. And both of them together are a lot stronger than them separately, right? A lot of people like to say, oh, it's Hermes versus OpenClaw, it's OpenClaw versus Hermes. No, you want to have both because there's a chain of accountability that wasn't there previously. And now, if you have this on your agent fleet, things are going to run a lot more smoothly. Now, I do want to emphasize this again. Hermes learns from every session. So imagine this, after 20 to 30 tasks in any one domain, it's going to be measurably better than it was before. OpenClaw doesn't do that. And I've tried building a lot of that into my OpenClaw. I have a lot of skills and things like that to recursively improve, to do all these things. It doesn't do that reliably, right? Whereas Hermes is just naturally wired up to do that. So let's talk about what they actually replace inside of my company as an example. When Hermes and OpenClaw work together, I have actually had them doing an end to end SEO campaign. So Hermes might say, okay, here is an SEO, AEO analysis on the single grain website. Let's take a look at that. And then what will happen is OpenClaw will do the work. OpenClaw is working on the best model, okay? The best large language model in terms of the version. So in this case, it's Opus and it's using Opus and it's coming up with all the ideas. It's doing all the analysis. And then not only that, it takes the API that we have with our SEO software, ClickFlow. It takes that and it generates the content. And so what you want to be doing here, by the way, is Hermes, you can have it running on a lower model if you want, but then you can have from a builder standpoint, because if you're coding up a bunch of stuff, then you want to make sure that your OpenClaw is running on the best possible model, on a frontier model at the very best. And then you can have your Hermes running on a bit lower, because you're just saying, hey, ask it to do this, make sure it does this, and then OpenClaw will handle the heavy lifting. So I thought it was reversed before, where you should have the brain on it, or maybe you should have both of them on the strongest model, but I'm looking to optimize on token costs right now, and it's doing a pretty good job. So look, you can have it do SEO work end to end, so it'll do the analysis, it'll create the content for you, it can even publish the content for you, it can do that. Or let's say you're doing outbound, for example. So OpenClaw can handle the cold email infrastructure, it'll buy the domains for you, it'll figure out the naming conventions for you, it'll create the sequences for you, it'll do all of these things for you, it'll even send the emails for you, even optimize the emails for you, but you have Hermes coming to make sure that all that stuff is going, because what happens sometimes is you might have a cron job wired to fire on Sunday at 8 a.m. Sometimes it'll forget. The problem with OpenClaw is a lot of these cron jobs break. Hermes has self-improvement where it'll continue to fix these over time and it'll continue to monitor. So again, each of these agents has their job. I look at Hermes as the brain that makes sure that the trains are running on time, and OpenClaw is this really strong builder that sometimes doesn't know what to do and sometimes gets lost. And so again, one plus one equals 10 in this situation. By the way, if you're looking to implement this in your business right now, literally this is what we do. With our world intelligence, our single brain unified intelligence, it literally takes all the software that you're using, puts it onto a single brain so your agents can execute on it and your team can execute on things as well. Just go to singlebrain.com with a B, and we'll see you over there. So let me tell you why the vast majority of human marketers can't compete with this. Number one, when you have these two working together, they're not going to forget things. They're going to help each other get better. They're going to improve every single day because of Hermes. I mean, OpenClaw is getting better every day. There's not to say that OpenClaw won't be able to do that on its own. Not only that, so it checks and balances huge. But when you have a central intelligence, and in this case, we use Obsidian for that. So Obsidian, we have all of our daily logs, our shared decisions, all these things happening within our agents, our ecosystem, these agents can pull directly from that. And that allows them to move a lot faster, that allows them to give you the right information. That's something that we didn't have before. So if we're going to have this brain and you're going to have the execution side of things, you're going to need to have something from a memory standpoint. And I've tried all the plug-ins in the world, but once you have Obsidian added in, it makes it easier to manage everything. So look, it obviously can't be all sunshine and rainbows, right? So there are some things, there are some limitations here. So like I mentioned earlier, OpenClaw tends to go down quite a bit. You have to restart the gateway, especially when it becomes overwhelmed. I will also say that this does become costly over time. Even if the token costs are coming down over time, you're going to want to use this stuff more and more, because it's helping you get a lot of work done. I would say that the open models are pretty limited. So even us, when we try to optimize with our DGX sparks, for example, and we try to put them on open models, they're getting better, but they're not quite the best, right? And so just know that when it comes to really strategic work, you're probably going to be wanting to run on the frontier models, which are going to cost you money. Securities is an issue too. Securities, permissioning, that's an issue too. So that's exactly why you need to have something that will help from a security standpoint. So for example, if one of your agents gets, let's say there's a prompt injection attack, or that agent gets compromised somehow. Well, you want to make sure that your other fleet of agents doesn't get compromised. Well, how do you do that? You might want to use something like a NemoClaw, okay? So there's a lot of things to consider here. Okay, are we going to use NemoClaw over here? What are we going to use to plug in all our apps? How do we make sure that there's good memory over there? And how do we make sure that we're pulling the right data and the data is accurate? That's a huge thing too. Most people don't think about data integrity. They don't think about keeping their CRM clean. They don't think about any of this stuff, okay? And so you're going to have to keep that clean. The other thing too is, if you don't have good skill.mb files, meaning that you don't have good processes within your company and you don't pass them over to like your Obsidian to have that central memory continue to get better over time in terms of how you make decisions, well, then this is not going to be that helpful because garbage in garbage out, right? So if you have bad data, you have poor process documentation or you have like light documentation, it's not going to end up working out for you. So you get what you put in with this ultimately. If you want to grow your AEO and your SEO faster, you have to check out ClickFlow. ClickFlow will track your AEO rankings. Not only that, it will produce high quality production grade content. We've had a bunch of people come over from other content creation tools and other SEO tools, and they say they're replacing the publicly traded companies that are out there. So check it out. It's clickflow.com. Once again, it's clickflow.com, and there's a free 14 day trial that you can check out. So the question is, how do you set this up yourself? So I would just pick one to start with initially. I would personally pick OpenClaw because that's, I started with initially, I would set up one chief of staff agent. And then from there, maybe I might even add in Hermes, and then have Hermes help me with some of the things that OpenClaw has been falling short on, so they can hold each other accountable. And then from there, I might decide to add to my agent fleet and just grow it bigger and bigger. And then from there, you can start to decide whether you want to buy local infrastructure, you want to be running some of your tokens on local models. You got to think about permissioning for your team as well. If your team is going to be accessing this fleet of agents, you have to be thinking about security as well. If you're going to be doing this for your clients. And I'll also say this, if you don't want to set up OpenClaw and Hermes, you want something that's maybe lighter. There are plenty of solutions out there that you can get started with. Claude has, for example, they have launched their routines. So you can have these routines repeat over time. These are just scheduled tasks that can have repeat. You can use Claude Cowork too, that's fine. Or you can use Perplexity Computer. And then Claude has an extension as well. So you can get started there with a Perplexity Computer, a Manus, for example, using some of these general agents. And then if you want to be able to do more, then you might consider moving over to an OpenClaw or a Hermes, and then you want to combine them together. Again, you can decide if you want to run them on a virtual private server. I just like running it on my little devices over here. And that's what's worked for us, right? And so it's worked so well for us, in fact, that our team can't live without it. And that's why we do this for our clients as well now. And so I think the world is converging into doing things this way in terms of you being able to 100X yourself as a marketer, right? This was not possible even a few months ago. So if you start taking advantage of this right now, you're going to be ahead of everyone else. And by the way, if you want to see how urgent and important this is and what industries are being affected by this right now, you want to check this video somewhere over here.