NEW Nanobot DESTROYS OpenClaw?

So today I'm going to be testing out something called Nanobot. Now Nanobot is kind of like a smaller lightweight version of something like Open Claw. It's open source as you can see right here. You can get free access to it and you can set it up directly with your favorites chat. So you can use this for example on Telegram or WhatsApp etc. And the way that you use this is you just connect your token as you can see. So the easiest way to set this up is directly with Telegram. And then from here what you can actually do for example is you can discover new insights and trends. You can deploy different code as you can see inside this video. You can schedule, automate, and organize. And then you can learn as well. So you can actually set up a member uh a memory inside the chat as well. And then you can just use it directly inside your phone. So we're going to be testing out today. Let's get this set up. So I'm going to open up the

terminal and follow the commands. First, we need to install it. So we're just going to install from the source by cloning the git. There we go. Now once you've configured it, so we've actually already configured it. I'm going to stop that. And then we'll set up the local uh models, right? So, for example, you can set this up with Llama or whatever you want. And you can also connect it through the chat here, right? So, the way you can do this is you can link the device. If you have any questions as well as we go through, feel free to ask. I'm actually going to get uh claw code to add in the API key. And this is good if you need if you just need to replace an API key um and give it give the new one to Nanobot. This is what we're doing. So I'll just do that in the background and I'm going to grab the API key from open router. And then

if we go back to open router, this is how you get your key. So you just go to your profile once you've logged in and then go to keys here. Once you've done that, then we can go inside the terminal. If we open up the terminal here, you can see that we've said nanobot agent-m right? So you can communicate with it directly inside the terminal. It's very lightweight like that. However, what we really want to do is connect this to telegram. So I'm going to do that in the background. Now, Sam says, "Isn't it better to use on VPS? " I don't know if it runs local. It does run local. Um, but I wouldn't recommend setting up on VPS. Like, it I've just seen so many situations where running it on a VPS is not a good idea, whether that's any sort of AI agent. Usually, Open Core is the one that's had a lot of problems with VPS's, but either way, I just wouldn't recommend it. So, now what I'm

going to do is link it to Telegram. And there's two things you need for the Telegram integration. Number one is a bot token which you can set up via the bot father and number two is your user ID. So I'm just going to grab both of those in the background. Yo says what is tars ui tars. Yeah. We'll come on to that later. So if we go back to telegram now we've got nanobot connected as you can see right here. I'm actually going to create a guide so you can just follow the copy and paste instructions but essentially what you have to do is you go to the bot father you create a new bot for example like this one. Then you connect your user ID and your username and the token that you get from bot fiber directly inside the terminal. I did it with claw code which was just way faster and easier. And then once you've done that you can start connecting with it. And inside the terminal, as you can see right here, you want to make sure that you run the command nanobot gateway, right? Nanobot space gateway. And that's how you connect it. So if we say, okay, how are you? Let's see if this works. This is the first time I've really tried it inside telegram. I've used it inside the terminal before and it was pretty lightweight and fast. You can see how quickly it responds. If we actually have a look on open router, let's see how it's using the tokens. So if we go to keys here, we can see the usage so far. So it's pretty cheap to use. I think you can change the default model as well to make it even cheaper, but it's pretty fast as well when it replies. And then let's try some of the terminal commands for setting this up. So for example, discovery insights trends. We can say uh what happened in AI automation today? Let's try that. 5th of February 2026. Auto Clawude, not tested out yet. Let's see what else we got here. Where's the treadmill? It's the only way to tell if you the AI version or not. I got some comments saying, "Stop walking on a treadmill. " So, I thought I'd try it. See if the videos do better. I'm on a tight budget. Do you think Quen 2. 5 coder locally can do a good enough job for building websites and deploying them autonomously? So, if you're on a tight budget, just use anti-gravity for deploying websites, right? That can still do a pretty good job. Can you use a separate local model for the heartbeat? I'm not sure about that. I don't I've not seen anyone using multiple APIs with nanobot, but you can do that with uh something like open claw. So you can set that up with open core pretty easily. So it says, all

right, I'm going to say what API key do you need for that to search the web. So we need brave search set up. I wonder if we can get nanobot to install it directly for us. So you can get a free API and you get a certain limited number of uh searches per month and then we should be able to search the web once we've plugged this in. So, if we go to Brave Search here, and then we can grab a key from the API key section here. Now, I'm going to try and plug that API key into the chat to see if it can do it. Actually, looks like it can't. So, I'm going to try getting claw code to add it in instead. Then, we may have to restart the gateway. And claw code can basically add the web search API key. And then we just have to restart the nanobot gateway. The other cool thing, if you use uh something like clawbot, it will tell you which API keys you've used so far. and then what to delete afterwards, which is pretty cool. So, I'm just going to restart the gateway and boom, you can see that we've now got this information, right? So, it says yes, web search is working now. Here's what's happening in AI automation today, Feb 2026, today's automation news highlights. We got Elon Musk, year of the humans, um, all this new information that came right and it's like, do you want me to dig deeper into any of these stories? So, essentially what we've done now is set up nanobart. It's running with open router. We use claw code to set it all up and just run through the terminal commands. I think that's really good if you're not a coder like me. And then also we've plugged in Brave search so that it can search the web as well which is pretty awesome. What else can we do now? We can save um and deploy apps and code to our

computer. So if we go back inside the chat, we'll say okay, create a simple SEO calculator plus save it to my computer as a simple HTML. And Ray Sean says, "Hey, I use OpenClaw with OpenAI codeex 5. 1 and have a plus subscription. " Codeex shows token usage, but openclaw doesn't show extra credits. Is openclaw usage covered by plus or rather extra costs? So if you've used the OO off, if you use the OOTH of um chat GPT codeex, then you don't get charged for extra tokens if you're on like the plus or the pro plan, right? So for example, for me, I've set up open claw then used it with chat GPT codeex and then I don't get charged anything extra, right? Obviously there's a limit to that, but yeah, um so far so good. So this has created the calculator now and we should be able to run it from terminal commands. So if we copy that the thing that I do really like about nanobot is like it's just so much faster to respond than using something like um open claw and you can see we've got the SEO calculator right here. Right. So just to recap we have searched the web for the latest AI automation news using right. We've also set up and deployed an SEO calculator and saved the HTML file locally so that we can do that too. We

can also schedule and automate and organize tasks. So if we go back to Nanobot here and we'll say okay, schedule a daily task to search for the latest news and updates in AI automation within the last 24 hours. Do it at 8 a. m. Thai time each day. Happy to help. Now, if it's a more complex task, it will take a little bit longer to respond, but you can always check the terminal if it takes too long to respond if you're like, "Has it broken or not? " So far, it's not broken, though. And I like the fact that you can just run the UI with something like Telegram. Makes it easy. I would just be careful if you're connecting something like Telegram, though, cuz obviously um you need to be careful with this stuff. Like, you know, if you want to play it safe, just run it locally inside the terminal. And now you can see here it says, "Okay, done. Your daily AI automation news task is now scheduled. Here's a summary of what it's going to do. So, schedule 8 a. m. Thailand time. Search for the latest AR automation news. Reports are saved here. The logs are saved here. And then it's got some useful commands you can run inside the terminal. And it says, note, your Mac needs to be awake for the tasks to run. If it's asleep, MacOSS will run it when it next wakes up. And then we can do a test run as well. So, three good use cases there. And then also the final one

is memory, right? So, if I say inside here, okay, my cat's name is Arthur. Please remember this. G says, how good is Nanobot? Or it's good if you want a lightweight version of something, right? And you want something that's easy to set up and doesn't break so often. I would Yeah, I mean, Nanobots's pretty good so far. Pretty impressive. It does have limitations. It's not as good as Open Claw, but it's pretty easy to set up. So, it's kind of like pros and cons. And also, it's much more lightweight, right? So, if you're running this on an old computer, this might be an easier way and better way. I've also got video tutorials on YouTube about Mimuubot and Open Claw. So, depends whichever one you want to go with, right? Does it have persistent memory? Yes. So, you've got this option to save to your memory now. So, I said, "My cat's name is Arthur. Please remember this. " And said, "Got it. Arthur is now saved to the memory. I'll remember your cat's name across our future chats. Give Arthur a pet for me. " Right? So, it can remember stuff, too. So, that is basically it for setting up Nanobot. Pretty good. Pretty simple. I'm actually impressed by how good it was. First time I used it, I like all these tools. First time I use them, I'm like, "What is this? " And then the second time I'm like, "Actually, yeah, it's not too bad. Not bad at all. " Right? You can

Not bad. Not bad at all. " Right? You can see that we've got a full guide for the setup inside the AI Profit Boardroom. If you prefer to use Open Claw, no problem because we actually have a full 3-hour course on how to use Open Claw, Clawbot, Mobile, whatever you want to call it. We also show you three different ways to set this up for free with Clawbot as well. And then if you prefer something that doesn't require any sort of terminal usage at all, you can use something like mimubot. We've got a full and video tutorial inside the profit board as well. So link in the comments in description to that. That's my a profit boredom community on school. Shows you everything you need to learn about AR automation. We update it daily with new tutorials like you can see here. So everything that I learn I share with you. You can see all the updates that came just in uh February the 3rd. Also you can post inside the community, get help and support whenever you need it. We've got an amazing uh accountability group where you can ask questions and focus if you want to set your goals and make sure you're held accountable to them. You can post questions in here. You can introduce yourself. You can even meet cool people in your area. So, if you go to the map here, you'll see that in pretty much every location around the world, we have people uh locally, which is pretty cool. And then also, it's not just me winning with this, right? So, we've actually got 118 pages of testimonials. It's like an infinite scroll of people who are winning with AI automation based on what we taught them. Right? So you can see here how people are getting stuff. I mean for example Quinn said, "Wow, I thought I understood AI automation and joy join the air profitable boardroom. Every week there are new systems, live calls, and people sharing real results. If you're serious about learning with AI, this is where you start, right? And we've got tons of testimonials just like that from people who are winning, learning, and growing inside a supportive community for AI automation. So feel free to get that link in the comments description. We have four weekly coaching calls. And also inside the classroom here, we have the uh a fiveweek AI automation master class that takes you from beginner to expert with homework in just 5 weeks. Plus, you'll learn how to build your first AI agent in under five minutes. Randy says, "I don't know how to code and I have two agents running. Join the club, mate. I don't know how to code either, to be honest, you know. " Um, let's see what else we got here. Master Shake says, "I built Mimot for Linux yesterday, but I gave up reinventing the wheel. just went back to open claw. But Mimu alone is pretty good. Yeah, I would say Mimu is good. It's just like a simplified version. So like you don't have to use terminal or anything to set up. If you want to get all my best playbooks for AI automation, you can see we've got the X automation here, newsletters, shorts, Instagram AI avatar videos, even how we screen share with AI. We've got the full playbook and the step-by-step um guide here. Now, the cool thing about this is you can give it to your team. You can share the guides with them. If you don't want to do this yourself or if you want to set this up yourself, you got everything you need to win right here. Additionally, inside here as well, you got the full agency course. You've got a full section on AICO, YouTube, etc. There you go. Live says, "What do you think are the best use cases for agents? " So, when it comes to agents, the best use cases are just looking at what you already do and then figuring out what do you need to automate based on that. So for example for me the way that I do this and I actually have a full section and training on this right here right on the time audits so you can see how I do this personally but I look at where am I spending my time and then what do I need to automate based on that right so for example if I'm spending a lot of time on YouTube videos it's like okay well can I get my team to create AI avatar videos and that's how I started doing that right the same on X right a few years ago I was posting everything manually on X now my team does it for me and you can see that yesterday for example we reached 600,000 people in one uh 1 million people in one day on Sunday etc. and it's my team running these automations for me. So you want to look at okay where are you spending your time and then what can you automate based on that what's the one thing that you can automate this week that would save you know 10 plus hours what's the biggest constraint on your time and then how can you automate that and one of the best things you can do is run through that process every single week. So for me personally what I'll do is I'll actually look at where am I spending my time and then okay based on that data I know exactly what I need to automate next. So I tend to find that's the smartest way to do it. Master Shake says Mimu is actually separate from Mimu bot. It's their memory system. But yeah. So you got like the persistent memory API uh from their website as well. So thanks so much for watching. As I said, feel free to see hope to see you inside there. You can DM me personally as well and I hope to see you in that. Ch.