How to Setup Claude Code's NEW Remote Control Feature (Step-by-Step)

Hey everyone. So Anthropic released a new feature yesterday called Remote Control. This basically allows you to kick off a task on your terminal of cloud code, and then you're able to pick it up on your phone when you're doing other different things like walking, for example, or just doing any other errands or sitting in a bathroom, taking your long bathroom break. You could be building something while you're sitting down, right? So this is not new in a sense. You could be able to do this even with OpenCLaw, for example, if you're using OpenCLaw. And there's also being able to use tmux and Tailscale to connect this. I did one with just connecting to WhatsApp and triggering from WhatsApp so long as my terminal is open. So same concept here, but they're integrating this directly into Claude Code. I think we all kind of knew this was going to come. And this is something that will continue happening. They built out the tool, which is Claude Code, put it out there. Everybody's playing around with it. Everybody's coming up with different ways to use Claude Code in different ways. They're observing all that and getting that feedback. And then they eventually just build that thing and add it natively into Claude Code. And so this is kind of the world we're living in right now. So let's get back to how this thing works, right? So they've added some, a few documentation here. I'll show you how you can set it up real quickly on your machine. All right. So the first thing you want to do is navigate to your project that you're working on or that you want to work on. So go to that folder. For me, this is the folder, Post Roll, and this is my project. The project is really, you're able to pass in a YouTube video and it generates a blog, generates a guide, a companion. It's a project that is inspired by Karpathy and he posted this like 2 years ago, I think to the date almost. And the idea was to pass in a YouTube video, generates a guide companion that walks you through the entire project as well. So the project is not important here. This is just for demo purposes, but I just wanted to show you Navigate to your project first that you want to work on. And then from there, make sure first you update— you upgrade Claude. Make sure you have the latest Claude version. So the current one should be 2.156 that rolled out yesterday. And it will take just a few minutes to check this or a few seconds to check this. Once it's done, the next thing is running that command that will trigger it to create that connection to happen to Anthropic. So the command is Claude rc and it will pick up from that. Or just if you want to, if you're fancy enough, just write the full thing, remote control. And if you run that, you will see that it will pop up a link that you can then open on Claude UI on your desktop and continue the process there. But that's not fancy because you're not trying to use it on your local system. You can use it on the web, on your phone for sure. But where you want to use it is really on your on your phone, on your app, Claude app on your phone. So to do that, you can just do spacebar and you have the QR code. You can scan the QR code with your phone and it will open up your Claude app and then you can continue from there. But for me, I don't want to do that directly there. And also I want to show you all how it looks like when I'm interacting with it on a phone and how it's going to show up on the terminal because the terminal should be open when I'm continuing building the project, even when I'm far away from my computer so that it could work. So the terminal needs to be open. So we'll be sending messages, piping it in here, and then it will continue running on the same terminal process that you have. So we've seen how to do that. So the next step that I want to fire up so you can see is just start Claude. So Claude. And from here, this is just a regular screen. But now you type /remote-control and you should see it pop up right there at the end. So run remote control. That should start the process and it says it's already started it. You can navigate to this on the web and you'll start working with it. Or for, for me, I want to use it on my phone. So go to your phone, open your Claude app on your phone, and from there you navigate to the code section. So on Claude you have chats, projects, artifacts. Go to code, and from there just refresh the page. You'll see the interactive session right there, and that's the one that you're working on. So this is in progress. Click on that one and it should show you a blank screen right there. So you can start chatting with it directly there. So for example, let me ask it, tell me about this project, and I'm going to type it in here. Tell me about this project, and I'm going to send it in. And there you go, I sent it in. It's already typing. You can see on, on my phone and you can see also on the terminal that it's actually piped in that entire prompt and it's running on the phone. And so it's going to run through some processes. So while it's going through, it's showing some errors there. It's probably going to ask for permissions and the permission will pipe back to my phone. It seems like it didn't need to. Maybe I added proceed with the YOLO mode, but probably not. So it already ran there. It's generated the summary here on the UI, but it has it on my phone. Now it has— there's a short lag, but essentially this is what went ahead and brought it for me too. So similar to what you're seeing on the main terminal is what you're seeing on the phone. And so I could instruct it to add more features onto the app and it will do it directly and show me on the phone as well. If there's any permission that it needs also, it will pop up on my phone as well. And for example, we could tell it modify the README to include the projects we have already generated. So sent that in, should take a minute there to show up. If you come down here, there you go, it's going to kick off. And let's see if it's going to ask for permissions. It didn't. I think it already inherited from the previous one, or it's either that or I have it on YOLO mode. But if I didn't have that, it would pop up here and I would have to accept it for it to continue running within the session. So that's very important. But when this finishes, at the same time when that's running— oh yeah, there you go, now it popped up. So you can see on the, on the main screen it's showing that I need to accept it, and also on my screen here it's showing me that I need to Allow once or allow for this session. So I'm going to allow once and you'll see there's just a tiny lag, but then not so much of a big lag and then it just continued from there. So it's pretty cool. You can do it on your phone and continuously there. So that is done. Now there's another way you can actually set this so that it's available on all the sessions you're running. So if you forget to start the remote control and you're running, maybe you got to pick your phone, you got to run. You can actually go to the config and here you do /config. And from there, just search remote and you will see there enable remote control for all sessions. And you can come here, you could say true, and that's it for the most part. And you're done. And that's configured now. And once that's done, really, you can go back and continue your main processes. So each new session you start, it will spin up a remote control session. And so if you have to walk away and run, you can go back and pick it up from your phone and use it. So that's the main concept here that I wanted to share. The other thing here that I wanted to talk about a bit is this is not new, like I mentioned earlier, and Claude is going to keep eating up on these extra features that we keep adding. So we get more creative, we're just contributing ideas to them and they're going to build it, which is great. I'm not complaining. But also I think there's people that are going to ask probably, When does work stop? When do you stop working? Because it seems like this is continuous work. Yes, it's just going to be continuous work. I think we're at this age in time also that things are happening so fast that you will want to be updated. Even if you decide you're logging off work, there's still that element of like you trying to still continue doing work. And so it's an enabling factor, but also if you want to actually turn yourself, turn everything off, it's very tempting. You have to find a way to do it yourself, but it's interesting. All these capabilities are coming out. My thinking here is Claude will continue, or Anthropic will continue adding more of these features that we see with ClaudeBot or OpenCLaw, which I'm sure OpenAI is also going to be doing the same thing. So everybody's racing towards the same thing here, which is more autonomy, more token consumption in a sense as well. There's so many ways to look at this, but I think it's an exciting world where we live in. So many options out there at the moment that you can actually build. So if you build anything cool with this, which is, I think you will, let me know. I'm going to be experimenting with this a lot more and seeing what other things we can build on top of it and see what features we could add that are not there already. I'm sure already some other people are thinking about it. For example, you could already, since you've connected it to within Claude Code, you've connected your Claude Code, maybe you have MCPs for Slack, for all these other different things that you have access to, you can directly still use it on there. So if you have specific connectors that you've added to your Claude code locally, it will be able to use those still within your sessions. So you'll be able to pick that up and use it to do different things within your system. So that opens up a lot of opportunities for automations and actually being in tune with what things that you're trying to build real quick. So ideas will be just a matter of think about it, start building real quickly from wherever you are. So see you in the next one.