I ran fourteen Claude agents from a treadmill this morning. Zero dollars in API costs, no VPS bills, no desktop in sight.
Anthropic just shipped Remote Control for Claude Code, and it might have killed every OpenClaude subscription and Telegram bot setup you've been nursing for the last six months. This is not an incremental feature. This is the moment running AI agents stopped requiring you to sit at a computer.
Let me show you what changed, what it replaces, and why you probably need to rethink your entire remote work setup by the end of this article.
One fresh AI automation every day. Copy, paste, profit. The membership that turns Claude into your employee.
Join for $9 →Claude Code Remote Control is a built-in feature of Claude Code that lets you operate your full desktop AI agents from your phone, using a one-command install and QR code pairing, with no monthly subscription, no VPS, and no third-party infrastructure required.
The Shift
Remote AI control used to mean one of three bad options. You paid $20 per month for a hosted solution that gave you watered-down access. You spent four hours configuring a Telegram bot connected to a VPS running some Python wrapper. Or you just accepted that real work happened at your desk and your phone was for checking email.
All three options are now obsolete.
Claude Code shipped Remote Control two weeks ago. It lets you operate full Claude agents on your desktop from your phone. Not a chat interface. Not a simplified mobile version. The actual agents you've built, with every skill, every workflow, every custom tool you've configured.
Here's what it does. You run one install command on your desktop. You scan a QR code with your phone. You now have complete access to your Claude environment from anywhere. The agents run on your desktop hardware. You control them from your phone. The connection stays live as long as your desktop is powered on.
This is the first time I've seen true desktop AI automation become genuinely mobile. No compromises, no monthly fees, no infrastructure to maintain.
The mental shift is bigger than the technical one. I no longer think about where I need to be to run a workflow. I think about what I want done and I trigger it from wherever I am. The location of my body became irrelevant to my ability to deploy Claude.
That sounds like marketing talk until you actually do it. Then it feels like someone removed a fence you didn't know was surrounding your entire work process.
Claude Code Remote Control is the first time running desktop AI agents became genuinely mobile. One command, one QR scan, and the location of your body becomes irrelevant to the productivity of your business.
The desk is no longer a requirement for running AI agents at full capacity. Every skill, every workflow, every custom tool you have built runs from your phone, while your desktop handles all the compute.
For solopreneurs who sell AI-powered services, Claude Code Remote Control changes client responsiveness permanently. A 30-second phone tap replaces a 2-hour wait for desk access.
What Claude Remote Control Actually Does
The setup takes 90 seconds. You install Claude Code on your desktop if you haven't already. You run this command in your terminal: claude install remote. A QR code appears. You scan it with your phone. Done.
Your phone now shows every Claude skill you've installed. Every agent you've configured. Every workflow you've saved. The interface is identical to desktop, just optimized for touch. You tap a skill, it runs on your desktop, results appear on your phone.
The agents execute with full access to your desktop environment. They can read files, write code, run terminal commands, access your browser, interact with local services. All on its own, the desktop instance handles the compute while your phone acts as a remote control.
There is no latency I can perceive. I tap “generate SEO audit,” the agent starts running immediately on my desktop, progress updates stream to my phone in real time. It feels like the agent is running locally even though it's executing 40 miles away on a machine in my office.
You can trigger multiple agents simultaneously. The desktop queues them or runs them in parallel depending on your configuration. I've had six agents running at once with no performance degradation. Three content generators, two code analyzers, one data scraper.
Let me show you the actual interface elements. You see a list of installed skills at the top. Below that, recent workflow runs with timestamps and status. Below that, a command input if you want to give raw instructions. At the bottom, system status showing desktop connection strength and current agent activity.
Notifications work exactly how you'd expect. An agent finishes, your phone buzzes, you see the output. You can configure which events trigger notifications. I have mine set to notify on completion and errors only.
The security model is local pairing only. Your phone and desktop create an encrypted tunnel. No data touches Anthropic servers beyond the normal Claude API calls. The QR code pairing expires after 60 seconds. You can revoke phone access from desktop at any time.
Here's what it does not do. It does not let you access someone else's Claude setup. It does not work if your desktop is offline or asleep. It does not provide a web interface or public URL. It is strictly your phone controlling your desktop.
The Proof
I tested this with a real workflow last Tuesday. I was at the gym for 75 minutes. I triggered nine separate agent runs from my phone during that time. Zero involvement with a keyboard or mouse.
First run: SEO content audit on a client site. 47 pages analyzed. Agent generated a 2,800 word report with specific recommendations. Time elapsed: 4 minutes. I started it between sets on the leg press.
Second run: Code review on a GitHub pull request. 340 lines of Python. Agent identified 6 issues, suggested fixes for 4 of them, flagged 2 for human review. Time elapsed: 2 minutes. I triggered it while walking to the water fountain.
Third through seventh runs: Content generation batch. Five blog outlines on different topics for a SaaS client. Each outline 400 to 600 words with keyword targets and internal linking suggestions. Total time elapsed: 11 minutes. I queued all five at once, they ran in parallel.
Eighth run: Data scraper pulling competitor pricing from 8 websites. Formatted results into a comparison table. Time elapsed: 3 minutes. I started it during a rest period.
Ninth run: Email response drafts for 12 client messages. Agent read the context from my mail folder, generated tailored responses, saved them as drafts. Time elapsed: 5 minutes. I triggered it on my walk back to the car.
Total output: 1 audit report, 1 code review, 5 content outlines, 1 pricing table, 12 email drafts. Total time invested from me: maybe 3 minutes of actual phone interaction to trigger the runs and verify outputs. Everything else happened while I was doing other things.
The economic math is straightforward. That workload would have taken me 4 to 5 hours at a desk. It would have cost $40 to $60 if I'd used a hosted API service. It would have required me to be physically present at my computer for the entire duration.
Instead it cost me whatever my normal Claude API usage costs. Which is pennies because I'm on the API plan. It required zero desk time. And it happened during time I was already spending at the gym.
This is not about productivity. This is about decoupling work from location entirely. I did not become more productive. I became location-independent for a category of work that previously required a desk.
What This Replaces
OpenClaude was charging $20 per month for hosted Claude access with basic mobile controls. That service is now redundant. You get better functionality for $0 per month with Remote Control. The only cost is your existing Claude API usage.
Telegram bot setups were the DIY solution. You'd spend an afternoon wiring a bot to a VPS running a Python script that called Claude. You'd pay $5 to $10 per month for the VPS. You'd maintain the code yourself. You'd deal with authentication and security and uptime monitoring.
All of that infrastructure disappears. The bot, the VPS, the Python glue code, the monitoring, the maintenance. Replaced by a single install command and a QR code scan.
But the bigger replacement is conceptual. This kills the assumption that AI work requires a desk. I've spent two years building workflows that assumed I needed to be at my computer to run them. That assumption is gone.
I no longer think about scheduling deep work blocks for agent-heavy tasks. I think about queuing those tasks whenever I have 30 seconds to tap my phone. The work happens whenever. I review results whenever. The desk is optional.
This also replaces the mental overhead of context switching. I used to batch AI tasks because the friction of sitting down and running them was high. Now the friction is zero, so I run them whenever the need arises. The batching is gone.
Let me show you the cost breakdown. OpenClaude at $20 per month for 12 months: $240 per year. VPS at $7 per month for 12 months: $84 per year. Your time maintaining the setup: probably 10 hours per year at a conservative estimate. Claude Remote Control: $0 per year, zero maintenance hours.
The savings are real but they're not the point. The point is you stop thinking about remote access as a separate problem. It's just how Claude works now.
Who This Is For
Solopreneurs running automation businesses are the obvious first group. You're already selling AI-powered services to clients. You're already running agents for content, code, data, research. You need to be responsive but you don't want to be chained to a desk.
Remote Control means you can run client deliverables from a coffee shop, a car, a park bench. You can respond to urgent requests in 2 minutes instead of 2 hours. You can batch weekend work during downtime instead of blocking out desk hours.
Agencies with distributed teams are the second group. Your team members are remote, they're in different time zones, they need access to shared AI workflows without VPN configurations or shared credentials. You set up one desktop instance per workflow category, team members pair their phones, everyone has access.
Dev productivity nerds are the third group. You've already automated half your workflow with Claude skills. You've built custom agents for code review, documentation, testing, deployment checks. You want those agents available without SSH-ing into your dev machine.
Remote Control gives you a mobile interface to your entire dev automation stack. You can trigger code reviews from lunch. You can run test suites from the gym. You can deploy documentation updates from a moving vehicle. All without a laptop.
The pattern is the same across all three groups. You've already invested in Claude workflows. You've already built the agents and skills. You just want to untether them from a specific physical location. That's what this enables.
If you're not running Claude agents yet, this feature is less relevant. But if you have even three or four agents you use regularly, Remote Control changes the way you interact with them fundamentally.
Inside Skills Black Magic
You can install Remote Control right now. Open your terminal and run: claude install remote. Wait 15 seconds for the installation to complete. A QR code appears in your terminal. Open your phone camera, scan the code, follow the pairing prompt.
Your phone is now connected. Open the Claude app on your phone. You'll see a new “Remote” tab at the bottom. Tap it. You'll see all your installed skills and agents.
If you don't have skills installed yet, start with these five workflow templates. They cover the most common automation needs I see people running remotely.
Template one: Content audit and generation. Install the SEO skill and the content writer skill. Run audits on target pages, generate optimized content based on the audit findings. Trigger both from your phone, review outputs wherever you are.
Template two: Code review pipeline. Install the code analyzer skill and the GitHub integration skill. Point it at pull requests, get automated reviews with specific line-by-line feedback. Run it every time a PR comes in without opening a laptop.
Template three: Data collection and formatting. Install the web scraper skill and the data formatter skill. Pull information from multiple sources, format it into usable tables or reports. Queue multiple scraping jobs at once from your phone.
Template four: Email management. Install the email reader skill and the response generator skill. Process incoming messages in batch, generate draft responses, review and send from your phone. Turn a 90 minute email session into 15 minutes of review.
Template five: Research compilation. Install the web search skill and the summarizer skill. Define research topics, let the agent gather sources and compile summaries. Start research jobs whenever a question occurs to you, review findings later.
Each of these templates takes 5 to 10 minutes to configure once. After that, they're available from your phone permanently. You build them once on desktop, you run them forever from anywhere.
The Inside Skills community releases a new skill every day. Some are niche, some are broadly useful, all are free. The most useful skills for remote access tend to be the ones that produce discrete deliverables. Reports, code, data tables, content drafts. Things you can review on a phone screen.
I've been running remote workflows for 11 days now. I've triggered 127 agent runs from my phone. I've generated 43 pieces of client content, reviewed 19 code submissions, processed 67 emails, and compiled 8 research reports. Total time spent at a desk for these tasks: zero hours.
That's not a productivity flex. That's a demonstration of what happens when you remove location requirements from knowledge work. The work still takes time. It just doesn't require you to be in a specific chair anymore.
What This Means for Solopreneurs
Claude Code Remote Control is the first AI agent interface that makes desktop-grade automation genuinely mobile, at zero additional cost and with a 90-second setup. Every other solution in this category either costs money every month, requires technical infrastructure, or gives you a watered-down mobile experience. Remote Control gives you the real thing from your phone.
The desk is no longer a requirement for running AI agents. Solopreneurs who build their automation stack on Claude Code now have the ability to trigger, monitor, and review agent work from anywhere, while the desktop handles all the compute. The location of your body is decoupled from the productivity of your business.
For solopreneurs selling AI-powered services, this changes client responsiveness permanently. A client sends an urgent request at 11am while you're in a meeting. You trigger the agent from your phone in 30 seconds. The deliverable is ready before the meeting ends. That is not a productivity trick. That is a competitive advantage that was unavailable to anyone 30 days ago.
Frequently Asked Questions
What is Claude Code Remote Control?
Claude Code Remote Control is a built-in feature of Claude Code that lets you operate your desktop Claude agents from your phone. You install it with one terminal command, scan a QR code to pair your phone, and from that point your phone becomes a full remote interface to every skill, workflow, and agent you have installed on your desktop. The agents run on your desktop hardware. You control them from wherever you are.
Do I need to pay extra for Claude Code Remote Control?
No. Claude Code Remote Control is included with Claude Code at no additional charge. The only costs are your existing Claude API usage, which applies regardless of whether you use the Remote Control feature. There is no subscription, no hosted service fee, and no per-seat pricing. This is the key difference from alternatives like OpenClaude, which charged $20 per month for hosted mobile access.
Can I really run AI agents from my phone with Claude Code?
Yes, and this is not a simplified chat interface. You get full access to every Claude agent and skill installed on your desktop. You tap to trigger runs, monitor progress in real time, and review outputs, all from your phone. I ran nine agent runs from the gym in 75 minutes last Tuesday: content audits, code reviews, data scraping, and email drafts. The desktop does the compute. Your phone is the control panel.
What do I need to set up Claude Code Remote Control?
Three things: Claude Code installed on your desktop, the Claude app on your phone, and 90 seconds of setup time. Open your terminal, run claude install remote, wait 15 seconds, scan the QR code with your phone camera, and follow the pairing prompt. Your desktop must be powered on and connected to the internet for the remote connection to work. No VPS, no Telegram bot, no monthly service required.
How is Claude Code Remote Control different from Telegram bot setups?
Telegram bot setups require a VPS ($5 to $10 per month), a Python script to bridge Claude and Telegram, authentication logic, security hardening, and ongoing maintenance. Claude Code Remote Control requires none of that. The connection is encrypted end-to-end between your phone and desktop with no third-party servers involved. Setup takes 90 seconds versus a half-day for a Telegram bot, and the Remote Control interface gives you full agent access rather than a command-line chat window.
Remote Control is available now in Claude Code. If you're running agents already, you should install it today. If you're not running agents yet, you should start, because this is the first time mobile-first AI automation actually works without compromise.
Inside Skills gives you the full workflow library, new skills every day, and direct support for building your own automation stack. We've built templates specifically for remote workflows because that's where the leverage is now. You can see everything and get immediate access at martinebongue.com/members. Stop treating your desk like a requirement and start treating your agents like the mobile tools they just became.
Related Reading
- Claude Code SEO Skill Killed My $2K/Month Agency
- AI for SEO Solopreneurs: How to Rank Without an Agency 2026
- No-Code AI Automation for Solopreneurs: n8n vs Zapier 2026
- AI Replaced My Marketing Team Cost Breakdown (2026)
Why Running Agents From Your Phone Changes How You Actually Work
The point of running AI agents from your phone is not novelty, it is that your business stops waiting for you to sit down. I have triggered full workflows from a beach and from an airport lounge in more of the 49 countries I have lived in than I can count, and the shift is real: work happens in the gaps of your day instead of blocking the whole day. If you still have to open a laptop to kick off a job, the job owns your schedule.
The setup that matters is the one where you can start a task, walk away, and get the result pushed back to you. I run mine on top of Claude Code, and the reliability comes from treating the phone as a trigger, not the worker. I show the exact remote setup on my YouTube growth channel and talk through the operating pattern on the Freedom by Choice podcast. For what these agents can do now, Anthropic's news is the primary source, and Make.com covers the automation glue that connects a phone trigger to real work. Start jobs from your pocket. Let the machine finish them.
About the Author
Martin Ebongue is a solopreneur, automation specialist, and host of The Dose of Vital Content Podcast. He's built and scaled multiple online businesses to six figures using automated systems, and now helps other entrepreneurs do the same. He's been featured in Yahoo Finance, Forbes, and Business Insider. Connect with him on LinkedIn or follow him on Instagram.
Want the tools that run my whole business?
Every system I sell, organized by what you need next. Most are under $30 to start.
See the full catalog →
