I Tried Claude Code Remote Control So You Don't Have To

Anthropic recently released remote control for cloud code and the main idea behind it is it will allow you to execute your terminal sessions from the mobile app while you're on the go. All you have to do is click a link, scan a QR code, and you're in. Now, in my last video, I walked you through step by step how I created my own clawed code operating system that I can manage from Telegram, and I can create, send videos, images, manage my email, and everything in my workspace. And some of you seem to really like it. And for others, you wanted a more turnkey way to get that same experience. So in this video, I'm going to walk through three things. First, I'm going to walk you through how remote control works and how you can set it up. And then I'll show you how it can actually be a compliment if you are building your own cloud code operating system, not a competitor, and how they can work together. And third, this is the fun part. I'm going to show you a small prototype of how you could use this remote control to get a similar experience to what I created just by using the terminal session and combining it with this feature because with a bit of creativity, anything is possible. Let's get into it. Now, to get started in using this feature, all you have to do is just make sure you're on the absolute latest version of Claude Code. Now, you can do this in two ways. You just say claude update, and once you send that over, you should get back your current version and then have it execute it for you. But for whatever reason, if you're having problems, my trick of the trade is to use warp, not affiliated, go into the agent mode that you can just click on right here. And then you can just say update my claude code to the latest version. And for whatever [music] reason, if you have conflicting files on your computer that are stopping the autoupdating or updating period, this will help you troubleshoot that. And there we go. We just double checked everything's good to go. If for whatever reason you had conflicts, this would be able to go into agent mode and execute and fix it for you. So once life is good and you're on the recent update, you should be able to do slash and then remote and then you'll see right here you have control and then you can click on this and then you can click on this particular link on your mobile app and be able to continue the conversation there. So if you click on this link right here, this will open it in a cloud code session. In this case, it's in my desktop browser, but you could run this on the mobile app itself as well. So, I could just have a conversation. The one thing it's missing is just YOLO mode, but I can just say, "Hi, this should be able to submit the exact same request to my terminal here." So, I have a mirror of my existing terminal, which is why you're seeing the exact same request land in the exact same place. And that's pretty much it in terms of the setup. Once you have those two things, you're good to go. And you can also ask it for a QR code. I've tried that as well. Once in a while, it will be a bit buggy where I'll use my phone, I will scan it, it will open my claw app, but not directly take me to the active session. So, if you use that method, then you want to just refresh your browser and refresh your chat and make sure you can see the word active. And just in case, you can take a look at the status right here and you can confirm it's active. And it shuts down as soon as you shut down your instance of cloud code. So that's exactly how you remove that mirror because it's just basically cloning what is in the terminal at that point in time. So outside of being able to let you take cloud code on the go, the main use cases I see as of today is being able to just double check and watch what's happening on your computer in your terminal for an agent run, especially if it's something that's longer running, let's say beyond 15, 20 minutes. It's unbelievably helpful. Another way it helps is if you put Claude Code into accept edits mode versus bypass permissions mode, then you don't want to spend four or five hours outside only to think that Claude Code actually did something to come back and learn that it asks for your permission or asks for your opinion for a plan before actually executing. So, I could summarize this video by saying remote control helps you remove your cloud code performance anxiety. On the security side, it's meant to be a secure line between both interfaces with only outbound HTTPS requests, meaning that no credentials are exported from your terminal to execute it and use it elsewhere. And like I said, you can run a command, scan a code, and you're in either on mobile desktop, could be another computer, and you can just port it over there. Now, I've seen some comparisons to OpenClaw, and honestly, I don't see the resemblance. I just think that as more and more people are getting acquainted with what Open Claw and Similar are doing underneath the hood, they're just comparing apples to oranges. The idea of the Open Claw revolution is that you have a personal assistant with some brain with an orchestration layer with a series of access to different tools. Experentially, remote control doesn't give you a personal assistant touch and feel. It's very procedural. It is that terminal for that session in that project at that point in time. So, in a way, it's meant to replace your need for a remote desktop app where you can just pour it in and just go on your desktop and type and see what that terminal session is doing. And one big picture to keep in mind is that Anthropic isn't in the race for consumer, which is why they abandoned the whole Cloudbot revolution the moment it started. They are always thinking about the enterprise. So when it comes to enterprise use cases, remote control is made for developers and people like you and I who want the flexibility of working on the computer, walking away, and as long as both are turned on, executing everything that we want. We can actually use it as a pseudo personal assistant if you really want to. We can maggyver it, which I'll show you shortly, but it's not a consumer chat app, and it doesn't look like it was ever designed to be or plans to be. So they're using it for developer productivity. Even if you want to use it for personal assistant usage. Now if I compare this to my last video where I showed you my version of OpenCloud that I operate using my pure cloud code on my cloud code instances that gravitate across my whole operating system. There's a huge difference here as well. As of today, remote control is for a particular project in a particular terminal. meaning if I ask it to go outside that project or that repo, I would have to port in and remote in to another specific one. So if you want access to the skills, the resources, the context, the files of one particular folder and you want to work purely in that context, this is very helpful. Now if you happen to create one mega folder and you happen to be one of those users that creates a workspace full of folders each of which can execute different tasks then assuming you have a longer running session let's say you use a 1 million sonnet you can technically work on multiple projects from multiple repos at the same time it's really just based on how you've designed your workspace in my case I can just tell my cloud code to navigate to a different directory go to that project and I have the ability ility to traverse across all those different folders. If you're on the quest like I am to find the revolutionary ROI friendly personal assistant, you would use the best of both worlds. You would use the remote control if you're zeroed in on a particular session with all the files contextually aware. And then if you want to be able to freestyle, go and navigate between different skills, different functionalities, combine them, and not have to worry as much about specific context, then creating your own version of remoting into your cloud code instance could be the perfect companion to this. So one could have access to your active sessions and one could have access to memory that you designed. All of these additional features I doubt will come out natively for the personal assistant use case. funnily enough, as I record this, the team actually released a brand new automemory feature that lets it learn across different sessions. So, we could get to a place where we could even tack this off for only being on the remote control side and port it over here. But if you want that personal assistant experience, that extra layer, then for now, you still have to find a way to build that yourself. So, the TLDDR is that the core remote control feature is the lightest setup, easiest to use, and it's great for checking on tasks already running. But naturally, if you want more firepower and the ability to create that pseudo assistant, then you'd have to inject your own layers yourself. So, let's say you wanted to create your own version of a light personal assistant using this feature. This is what you could do. You could open a brand new folder right here and you could say something along the lines of, "Hey, could you create a memory file in this folder that will always remember how I like to communicate with you? I want you to create a warm and fuzzy experience in our back and forths. I don't want you to be very corporate. I want to talk to you about anything from my business to what's bothering me. And I want you to create a soul file. You can call it a soul.md file. And then you can spin up a claude.mmd file that can tell claude how to use this soul file and how to use this memory file. So now we will just ask it to execute this just to start us off. And after 2 minutes it comes back with all the files even though weird spelling mistake right here. So I got that sorted immediately. And now we can ask for something else. Okay. Can you import a couple skills from the official anthropic skills, but specifically for this project? I just care for the PowerPoint skill and the CSV skill. And yeah, that's pretty much it. So, we can just pull that in. And now all we're doing is just creating our environment here where we have access to all the files, folders, functionalities that would give us that pseudo experience of the personal assistant. And after we have all the skills that we need, we can check our cloud MD here and pull this over to the left hand side. You can see that we have the soul file. It is routing to how to use it. What this means in practice, how to use the memory file. So we're basically copying the structure of let's say open claw or something similar. We have the memory file where my values are stored. I value honesty over comfort. Apparently, we have the soul file of who I am, how to talk to me in this conversation, etc. Right? So, if we were to start a brand new session, just so that everything can materialize. And now I do slash remote control right here. And then we click on this bad boy and we pull over this specific experience. So now you could do something like port this on your phone and say, "Hey, I had a hard day. I saw Claude released another 15 updates. What do I do? And then you send that over. And then it should be able to take the context of that specific repository that you're in, that specific project, take that into account, and ideally, there we go. Check the soul file to make sure it's nice to me. And then it should come back with very soothing therapeutic discourse. And there we go. We have my therapist up and running. But you could use this for a lot more business proactive purposes. You can connect different APIs, MCPs, skills, and have this on the go to give you that pseudo experience, even if it's not the full version. And that's pretty much it. So, you can see that Anthropic is really investing in building all the infrastructure that you need as a developer and a non-developer to build on something, execute on the go, and most importantly, being able to put all the pieces together to recreate whatever experience you want. If you found this video helpful, please leave me a like and a comment on the video. It really helps the video and the channel. And to the rest of you, I'll see you in the next

Join My Community: https://www.skool.com/earlyaidopters/about --- Anthropic just shipped Remote Control for Claude Code -- the ability to control your terminal sessions from your phone or any browser. In this video, I walk you through the full setup, show you exactly how it works, and give you my honest take on what it is and what it isn't. Spoiler: it's not an OpenClaw killer. They're complementary. Then near the end, I show you a quick prototype of how you can MacGyver Remote Control into a pseudo personal assistant using soul files, memory, and skills. One feature, two use cases. --- Timestamps 00:00 - Anthropic just released Remote Control 01:04 - How to update Claude Code to the latest version 01:57 - Using /remote control and connecting your phone 03:10 - Monitoring long-running agent tasks remotely 04:12 - Why this is NOT a competitor to OpenClaw 05:31 - Enterprise vs consumer: what Anthropic is actually building 05:50 - Remote Control is scoped to one project at a time 06:45 - Best of both worlds: Remote Control + your own system 07:48 - Building a pseudo personal assistant prototype 08:05 - Creating a soul file, memory file, and Claude MD 09:00 - Importing skills into your assistant terminal 09:57 - Live demo: Remote Control with the personal assistant 10:56 - Final thoughts --- Book a Consultation: https://calendly.com/d/crfp-qz3-m4z #ClaudeCode #RemoteControl #ClaudeCodeRemoteControl #OpenClaw #AIAssistant #AnthropicAI #ClaudeCodeTutorial #AIWorkflow #CodingFromPhone #ClaudeCodeMobile #PersonalAI #AIAutomation #ClaudeCodeSetup #DeveloperTools