Claude Ralph Loop is INSANE!

Today we're going to be looking at Ralph Claude code which is basically like a powerful way to create a infinite loop and just let your agents run on autopilot and just autonomously go off cycle through and create stuff. For example, normally if you're in Cycl code, let me show you an example. So if you're in Cycl code here and you've got something that you're creating, right, it's not just going to run off in the background like it's going to require a lot of back and forth. You're going to have to go off and reply to it. it's going to be super annoying, etc. And it takes a long time as well, right? So, what you can actually do is if you open up Ralph Claude instead, then you can allow this to just autonomously run on an infinite loop until it improves stuff. The other cool thing about using this AI agent with Claude is that it can just it can track progress, but it can repeat stuff until the project is complete. So, it never gets stuck, right? It's just an agent running in the background for hours and hours until it finally gets stuff done. And as an example of this, you can see a tweet from Kazala here. It says, "This Ralph thing is insane. It took me months to develop the second runtime to generate this. Whereas Claude with Ralph just ran with the task and when they woke up in the morning after a few iterations, it solved all the issues, right? And this is the power of Ralph Claude. Like loads of people raving about it, but a lot of people don't know how to set this up yet. So I'm going to show you exactly how to do this. Right. So what we can do here is we can grab the GitHub li and then we can just set up a new folder. So let's add a new folder here. All right. So we'll call that Ralph. And then inside claw code, this is a by the way it's a free program, right? You can use it for free. We're going to open up the folder that we just created. So we'll open up Ralph here. I'm going to say okay, open this up. We trust that. And then inside claw code, which you can run inside the desktop, by the way, there was a new update inside Claude this week where if you go to claw desktop, then you go to the code section at the top. So you can switch between chat and code. You can run claw code inside claw desktop right here. See how for example I have to allow everything when claw code is running. with Ralph, it just runs off on autopilot. So, we'll come back to that in a second. Let's wait for it to load and open this up. The other option as well is you can just install it like this. So, if we go to terminal like so, we can just go inside the terminal, clone this, copy these commands, and then open this up as you can see right here. So, we can take this command, go into terminal, plug that in, and we're good to go. In fact, to be fair, it's way faster to just run it inside the terminal. Now, we're going to navigate to it as you can see right here, and then in the prompt MD file, you can describe your project requirements like you can see here. So, you can edit this. Then, we're going to hit save. And then once we've edited the prompt MD of our requirements, we can just start Ralph. Right. So let's go to the terminal here. If you get any errors, just paste them in to claude and then we'll say so we've we have to install TMUX as well. Once we've done that, we're going to run this. So we're going to say would you like to create a new Ralph project? I'm going to say yes. We'll just go with my app. This is a generic project name. We'll allow that. And then it says, do you want to So basically the prompt MD is the file that you want to edit like you can see in add to gravity right here to prompt Ralph into what it needs to do. Right? So you prompt it right here in terms of what it's going to build. And then from there you can edit the document, save it, and that's how Ralph knows what to build. Right. Now, if we go inside here and we'll say, "Okay, can you edit this is for a snake game that's 3D? " Something like that. Right. So, now it's opening up the prompt MD file. And as you can see right here, it's going to open that up. This is the default one with the templates, but then it's added this bit, right? Dr. Ralph, an autonomous AI development agent working on a 3D Snake game. As you can see right here, we're going to allow that. And it's beginning to write in like the technical specifications. So 3D Snake game overview, the text stack, core gameplay, the mechanics, etc. And then it's ready to start now. So set it all up. It's set up and configured the prompt. It's like, this will start Ralph autonomously building your 3D Snake game. Want me to start it? We say yes. And now it's just looping through

and we can monitor it. So if we go to the terminal here, we can open up Ralph monitor and see what it's working on. And this will refresh every 2 seconds with the details. So we can see the progress, the loops, the loop count, etc. Marine says when it runs in the background, is this running on theirs or our services? So it's running locally on your laptop, right? So you can see here it's running locally but obviously it's using claw code and your API with claw code which is online to loop it through and then it will just keep refreshing every two seconds like you can see over here in the terminal. Then if you want to stop it like you can see here we've actually stopped it just you can say can you restart it or can you stop it etc right inside your claw code session. So this claw code session is connected to the Ralph loop over here. And so if you ever want to stop, you can always press like I think it's command and C as well inside the chat here. But you can control it as well inside claude directly as you can see just by setting up that Ralph loop directly inside there from the GitHub. And then if you want to start it up again, you can just use a command like this as you can see. And you can just watch it live in your terminal using this command right here. But that's basically how you can set up claw code as you can see with Ralph. This is Ralph monitor the status dashboard etc. That will run on an infinite loop to edit the prompt and actually prompt Ralph to do the work. You just change the project MD file like you see right here. If you ever need to stop it, go back to your claw chat and then just tell it to stop the agent or just press command and C in the terminal. And then if you actually want to open up the GitHub directly inside claw code, then just paste the GitHub details and ask it to run it. That's the simplest and easiest way. If you have multiple projects like you see here, and you need the command for opening up like this, then you can just screenshot the page here, paste it into claude, and just say terminal command to run this. It will give you the ter terminal command like you see. And then if we want to run that as well, we can just press control and C inside the chat, paste in the command, boom, start running it like you can see. And then it will tell you how many minutes it's been running, what it's doing, what it's working on, etc. So, it's pretty cool. The benefits of Ralph are basically that it can automate coding and just run on a loop automatically in the background. You might be wondering, okay, what is this? Why would you use it? Raph Claude loop is a plugin for Claude that turns it into basically like a non-stop AI worker. So instead of generating code once and stopping it, Claude will just keep going, right, until the task is actually complete, fixing bugs, testing code automatically, and that sort of thing. And it's a huge shift from like the typical one and done approach that most AI tools use. Now, how it works, you basically write a prompt, add the completion signal, set a max retry limit, launch a loop, right? And then why would you use this? You could use it for example some of these examples. One of the most important ones I've seen a lot of people do is just build whilst they're sleeping, right? So they can have this coding out like you can see here and building stuff whilst they're not working at all which is pretty cool. I'm just going to terminate that. And then what makes this different versus just using something like claw code? Well, it's like you give it the end goal, not every step. Claude figures out the path. It checks his own work and self iterates. It learns from its previous tries. It works with live GitHub repos and real projects and it has built-in monitoring, error filtering, and rate limiting, right? And if you want to get all the setup instructions, plus an extra tutorial on exactly how this works, you can get that inside the AI success lab, completely free link in the comments description. This connects you with 45,100 people, as you can see. And also, if you want to get an amazing AI automation community that helps you grow and learn with AI and automate your business, check out the AI profit boardroom. Inside the community, you can post any questions, get help and support whenever you need it. You can see all the people winning with AI like you can see right here. And also inside the question section, if you have a question, people will help you a lot. So, you can see here, for example, Maddie had a question. 28 people got back to Maddie just to help her get her the support she needs. Also, inside the calendar, you get four weekly coaching calls. You can jump on a call with an AI automation expert, get help and support whenever you need it. And inside the classroom here, you can learn how to go from complete beginner to expert with AI automation in just 5 weeks. You'll learn how to build your first AI agent in under five minutes. You'll get my full processes for using AI avatar videos and automating those, plus how we automate, for example, Twitter and newsletters and shorts and everything else, right? And even emails. And then for example, if you want to learn how to get more clients and how to land more clients, you can check out the agency course. If you want to basically grow your YouTube channel like me, you can check out the six week road map right here. And if you want to rank number one like we do as

well, then you can check out the AI SEO automation section. So feel free to check that link in the comments description to the AI profitable. Cheers for watching. We got a few questions here. So let's see what we got. It says what is he working on? Ralph and claw code. Alson says coding with claw code regardless of claude opus consumes credit very fast. If I subscript to the plus van any better options. So it's the same thing, right? I made a Gemini gem for PowerShell running terminal commands. Pretty cool. AI brand building says you ever get issues with deploying sub agents properly. I keep trying to deploy subcoding agents for smaller features. Check out my video on automaker. So basically what that will do is if you have loads of features it will turn everything into a cambban board and then you can just press make right and that will just go off and make stuff for you. Let me show you an example just so I can be really clear on this. So if you go to my video on automaker here this one we can basically set up a camar board like you can see right here and whatever project you have if you have a backlog of features you can get these sub agents to create the tasks. So you just click on make here and it will go off and build separate tasks with multiple agents. That's one of the easiest ways to do it. You can set this up with claw code and automaker which is a free GitHub and then from there you've got a nice UI where you can host it and test it. Skully says Goldie Pop. That's what we're talking about. Thank you very much. What about advanced AI generalist? Not sure what you mean. You need to use obsidian mode. Yeah, I've heard a lot about obsidian. I've not tested it, but yeah. So, thanks for watching. Appreciate as always. And I'll see you on the next one. Cheers, peeps.