Hermes + Antigravity 2.0: Build ANYTHING!

Key Takeaways

Imagine having Hermes agent and anti-gravity working together in one place, building, automating, and creating for you 24/7. I just ran a full masterclass breaking down exactly that, and in this video I'm going to show you the exact setup, how to combine Hermes and anti-gravity, so your agents know who you are, understand you, and get smarter every single day. I'm also answering the biggest questions people have right now about context limits, model choices, migrating from other tools, and what actually works in the real world. If you've ever felt stuck or overwhelmed, or like you're not getting the results you should be getting with AI, this is a video that fixes that. Let's get into it. What I'm going to do is just answer some questions inside our community, and let's see what we got. Ah, so this is interesting. So, for example, Alan has just posted inside the AI Profit Bot in about anti-gravity, and he says be careful with anti-gravity 2 cuz it's a major change from the old version, and Google removed the built-in Visual Studio Code style editor and terminal. So, it's now a standalone agent only app, designed purely for chatting with agents managing projects. So, you can actually use this guide to go back step-by-step. That's why I've set up the Agent Operating System really, because if you look at this, it's more agentic, the apps, right? Anti-gravity just it is like a standard chat, as you can see, but if you want to turn it into something fully agentic and work with all your AI agents, then you can use the OS instead. What else you got here? Kenny says, "Can anyone suggest like an AI agent chat workflow with NA10?" We actually got training on this over here. So, we actually have an example tutorial on how to do that right here, to build your first sort of chat agent essentially. Let's see what questions we got here. Sumit says, "Anti-gravity 2 with context 7 consider more powerful." Honestly, I think the best thing you can do is just give it more context about you, who you are, etc., and then it'll be way more agentic, right? That's what we've done with the memory system here, and all the skills that we have on the left-hand side. For example, like the SEO tool, the goals section, the journal, the Kanban board, the notebook, the studio. I think that makes it way more powerful instead. And then this is Ola. Nice to see you thanks for joining. Jason says thanks for the updates. Happy to help. Anti-gravity is needed now for But let's see. Yeah, so basically anti-gravity IDE removed from anti-gravity. That's what's happened here. So if you go into anti-gravity, it doesn't have an IDE, right? So for anyone watching, IDE is like a kind of way of seeing the code and everything else. And you can see it's much more like the code a setup of ChatGPT's super app rather than the original version of anti-gravity which was more like Visual Studio Code if you've come across that before. So Las says she just got Open Claw with Hostinger. Which content to use to start automating the content they have, right? It really depends what type of content you're trying to automate. But for example, if you're trying to automate SEO content, then I would use a WordPress with Open Claw training. So basically what you can do with Open Claw is you can use it So you can use it to, for example, have your WordPress login details and then deploy content to your website, right? So you can give it a keyword. If you want Open Claw to create video content for you instead, then what I'd recommend is that you set up Hyperframes with it or Remotion. Now with Hyperframes, the training that we have is with Hermes, but you can do exactly the same process with Open Claw. It's literally the same process. So Neil was saying he doesn't have many resources. What should he be using if you don't have much? So like a MacBook, this is a good setup actually. So a MacBook, Owlpha, and Hermes is really good. So if you're using Hermes, you can use it for free with a MacBook plus with Owlpha, the API, and they are completely free, which is pretty cool. Let's see what other questions we've got inside the comments here. So Jason says I came from Visual Studio Code with its anti-gravity, so I'm going to need to the need ID for now until I get adjusted adjusted." Yeah, I think for a lot of people used to the old setup, it's going to be a bit of a transition on the UI. Gaz says, "I didn't like anti-gravity 2.0." I think a lot of people don't like it actually from what I'm seeing in inside the comments and inside the list questions. I think they don't like the setup, but again, that's why we built the agent operating system, right? Cuz you can make it much more agentic and customize it exactly how you want. Jason says, "Is N10 relevant anymore?" That is a really good question. For me personally, I don't personally use N10 anymore. And the reason for that is because with N8N, it's a bit more technical and harder and more time-consuming to set up than if you just go straight to Hermes Agent. So, for example, if I'm using N10, it's going to take me ages to set up an automation. Whereas if I want to set up an automation with Hermes, I just type exactly what I want. For example, let's say I'm trying to do computer use. That literally took me about 3 hours last year to set up with N10. When I set up with Hermes, it took me about 20 minutes, right? And now I can use computer use anytime I want inside the agent operating system. So, that's the way that I look at it. Let's see what other questions we've got here. By the way, if you ever want me to answer your questions in video tutorial like today, then you can just join the A Pro for all of them and just post inside the community and then we can help you. I do this every day for members. So, this is a really questions. So, Connor says, for example, excited to learn I'm new to AI, got a basic setup, would love to hear some tips from you all on what to do and what to avoid early on in the AI automation journey. In terms of what to do, I would just focus on one automation per week and just simplify as much you can. The number one mistake that I see most beginners make with AI automation is that they struggle because they try to do too much and feel overwhelmed. If you can simplify as much as you can and just stay focused on one thing at a time, you'll get much better results, it will be much less stressful, and also you'll enjoy the journey more. So, that's my biggest tips to you. Great question. And we actually have a full section on how to do that right here. Jason says, "Thank you, Abisoye." Beyond says, "How you handling context window limits in long automations? Running into a problem." So, when he has to process a long document, the context fills up and the quality degrades. So, it depends what model you're using. So, for example, if you are using Claude desktop for this, then it tends to handle context pretty well, especially longer documents. I give it quite long documents to deal with when it comes to SEO content. The other thing that I would say here is that if you can use compacting, that helps a lot as well. So, for example, if I want to compact anything in inside Claude, I can just type {forward slash} compact, and then it will reduce the context window of the whole conversation, which allows me to do more in the future. And finally, you can always split it up into separate sections. So, you could have one section that deals with part of the document, then that passes on to the next part of the automation, and so that sort of chunks it down as well. This is a really good question, right? So, for example, Dan says, "Fair play to Julian here. His content has pushed me more to think less about AI tools and more about building actual systems, which is what we're doing with the agent operating system." So, I'm glad that's helped. But he says, "If the system doesn't understand these things, then the problem is not Hermes itself, it's actually teaching the system how the business works, right?" So, this is actually why we set up the infinite context engine, so that I can train AI agents and plug it into any AI agent all of my memory and systems and everything about me. So, it has a lot more context. That way you get better outputs when you're actually speaking to the AI. So, for example, So, for example, if you're using Obsidian, Obsidian has a map and a full knowledge base about you, the more you use it, see the training above. And so, what that allows you to do is then plug that into any AI agents you use in the future, so you can get the best out of them. So, for example, you can see here like it understands everything about me, all the tools that I use, what I'm using, how I work, who I work with, everyone in my life. It's just even like the groups and the people that are part of my life are just organized inside this, right? And so that makes everything 10 times easier. So, what I'd recommend doing is setting up context engine inside Obsidian, and then you can use that to train your AI agents in all the problems above, so that it understands you context better. So, we've got a question from Philip who says, "What are the best open-source models on Ollama?" All right. "And what are the best ones for running with Hermes?" So, he has a DGX box, which is like the perfect setup for this sort of thing. But this is interesting as well. So, when he runs multiple agents concurrently on Qwen 3.5, his agents respond slower. So, I don't tend to find like amazing outputs from local models, but you're on a better setup than I am. So, you're always going to get better results. I would look at quantized models. So, there's a few available on LM Studio as well that you could set up. And these are like smaller versions of the more powerful models. There's actually one on LM Studio that's designed specifically for Hermes, which might be worth a try. Try and find that one now. Let's see here. So, if we're going to LM Studio, I think if you type in Hermes, yeah, you'll see a bunch of different models designed for Hermes as you can see right here. And so some of those are worth checking out, right? You can see all of these different models. Some are actually from Nous Research directly, but some are actually from the community, which are worth checking out, too. So, that's the way that I would look at it. There's loads of different models here. And also, you can see from LM Studio which ones are the best to set up because they actually tell you based on your setup which ones are too large. One model that's really lightweight and easy to use with Claude or Hermes or whatever you want to use is actually using Gemma 4. It's pretty lightweight. It's not super intelligent, but if you just want an AI to do sort of heavy tasks, but ones that don't require too much thinking, that's a good option, too. Oh, wow. Look at this. So, Tom actually built the agent operating system that we set up as well, which is pretty amazing. So, that is awesome. Congrats to you, Tom. This is an interesting question. So, Julian was asking, "What's the primary difference between Notebook LM and Obsidian? And should you use one or the other?" Both with Claude and Hermes, right? The way that I would look at this is that you want to use Obsidian for your memory, not really use Notebook LM. Notebook LM is great for like generating media or generating videos or podcasts or infographics, but it's not great for really storing personal context for you. And also, it's not open source. Whereas Obsidian, with Obsidian, all of your files are stored locally, they all belong to you, it's open source, and also it's easy to connect to your AI agents. Whereas Notebook is harder to connect to your AI agents. So, if I had to pick one, I'd go with Obsidian. This is asking, you know, how to migrate from Open Claude to Hermes, right? He said, "I'm in the middle of building something. Open Claude doesn't seem to be very efficient. I was considering migrating off to Hermes." So, what you can actually do is a migration option over here. So, you can run, once you've installed it, it's not going to be 100% smooth, but it is a way that Hermes have designed making switching over from Open Claude to Hermes much easier than it should be. So, that's a good option. And then also, just try them and see which one you prefer. Personally for me, everything has been smoother since I switched over to Hermes. All right. So, I think we've covered every single question, which is awesome. If you want your question answered, and you want the agent operating system and everything else I've talked about in the guides today, you can get that inside the AI Profit Room. Let's see what questions we've got. How to use Claude code inside Antigravity. So, the way that I do it is this is the simplest way that I know, it's like just have it inside your operating system like this. You have Antigravity CLI, you have Gemini, you have Cloud CLI plugged in, right? What you can actually do is if you go over to anti-gravity, let's just double check here if they have a I don't know if they have a terminal option anymore. I think you'd have to roll back to the previous version cuz they don't seem to have a terminal. I don't know if you can. Or you might have to set up Cloud MCP to integrate them. And that's why we have the agent operating system so you can just customize the format a lot more, which you can see right here. Is this AI? No, mate. No, it's not AI. This is me live. Danny, good to see you here, sir. Danny is an absolute legend, Danny Barrera. What are you using in your agency right now? So we actually have all of our processes set up inside the playbook here that I personally use. The ones I use day-to-day are number one, I use the agent operating system and that just helps me deploy SEO content and build out websites, right? Super useful, really powerful. You can see all of our websites are set up over here and we have an agent operating system which allows us to deploy and create content for SEO and then we can build out websites very quickly and start ranking them on Google. That That works really well. The next one that I use is Twitter, right? So actually if you have a look at Twitter, for example, let's have a look here. This is my account and you can see here that we're getting this amount of reach per day and it's working really well using this process. So we use a lot of that stuff inside my business. Next up, we have AI avatar videos, of course. So we have a full process for this as you can see a full set up and step-by-step guide and tutorial. But basically we use AI avatar videos, that's great for marketing. And we also have AI avatar influencers, too, which we've got with this full process. We've broken it down right there for you. So yeah, all those are like the biggest ones that we use. There's a lot more to it. I could go on all day, but those are like my my top examples of how we use this. Should I switch from ChatGPT Plus to Gemini? I would just test it out, see what you think first. Test it out, see if you like it. If you do, great. If you don't, no problem. You don't have to switch. And just see what you feel. Which one do I personally use? I I tend to use ChatGPT a lot more than I use Gemini, honestly. But I use them all, right? I'll use Hermes for some things, I'll use Claude for some things, I'll use Anti-Gravity for other stuff. So, thanks so much for watching. If you want to get more training like this, you can get it all inside the AI Profit Bot Army. If you want me to answer your video tutorials like we did today, you can just post inside the community here and I answer questions for you like you've seen inside this video. And what I do is I basically give you the tutorial with the video and then I also break down the four steps into a nice quick paragraph that you can read and implement if you want to save time. Also, we do four weekly coaching calls. We have the classroom. We have the Agent Operating System, which you can see over here. And basically, we have everything you need to really win with AI and learn and grow as much you can. You might be saying, "That's great for you, Julian. What about everyone else? Is Is anyone else getting results with this?" And that is a great question if you're asking that. So, we actually have over 159 pages of testimonials, as you can see right here. So, lots of people winning, learning, growing, people who've never used AI before are just getting awesome results, which is amazing. And if you want to join the community, if you want to connect with me personally, if you want to get all of this stuff and just be a part of a a positive, happy community who are all learning and growing and helping each other, then feel free to join us, right? Link in the comments and description or go to the AI Profit Bot Army.com. Thanks for watching. I will see you on the next one. Cheers. Bye-bye.