Claude Code Remote Control: Run Your AI From Your Phone

You’re mid-flight. Your laptop is in the overhead bin. Your Claude Code session back home is halfway through a content pipeline that needs a decision before it finishes. Two months ago, that would mean waiting until landing, pulling the session state out of your head, and starting over.

Not anymore. As of February 25, 2026, Anthropic shipped Remote Control. Professionals running 20-60 minute AI workflows can now supervise them from anywhere—without losing the session, the context, or the work in progress.

The short version: start a Claude Code session on your machine, walk away, and pick it up from your phone anywhere in the world. Nothing migrates to the cloud. Nothing stops running. You just get a new window into the same session.

What Claude Code Remote Control Actually Does

Claude Code Remote Control connects your claude.ai/code browser tab or the Claude mobile app directly to a Claude Code session running on your local machine. Your computer keeps doing all the work. You just gain the ability to monitor, direct, and interact with that session from any device with a browser or the Claude app.

The distinction matters. This isn’t “Claude Code in the cloud.” Your files stay on your machine. Your environment stays on your machine. Your tools, credentials, and local context—all still on your machine. Remote Control is a window, not a migration.

Think of it like this: your AI worker is at their desk. Remote Control is a way to call them from your phone and say “keep going on that, but pivot to X first.” They’re still at their desk. You’re just not in the same room.

Quick Setup: What You Need to Start

Requirement	Details
Plan	Max (Pro rollout coming soon; Team/Enterprise not supported)
Claude Code version	2.1.52 or newer
Devices	Your machine running the session + any browser or Claude mobile app
Security	Outbound HTTPS only—no inbound ports opened on your machine

Start a remote-accessible session two ways:

1. Launch with claude remote-control from your terminal
2. Inside an existing session, run /remote-control (or the shorthand /rc)

Your terminal displays a session URL. Open it on any other device to connect. Press spacebar to generate a QR code—useful if you’re handing off to your phone quickly.

Enable it automatically: Run /config inside Claude Code, then set “Enable Remote Control for all sessions” to true. Every session you start will be remotely accessible from that point forward.

Don’t have the Claude mobile app? Run /mobile inside Claude Code to display a download QR code for iOS or Android.

Why This Matters More Than It Looks

Most coverage treats Remote Control as a convenience feature. For professionals who’ve built real workflows on top of Claude Code, it’s an operating model shift. The difference matters.

Here’s why the timing is interesting. The value of Claude Code scales with complexity—long-running agents, multi-step content pipelines, background workers that take 20-40 minutes to complete. Until now, all of that meant staying tethered to your desk. Step away and you lose the thread. Take a call, get on a flight, or just close the laptop and you break the session.

Remote Control removes that constraint. A content pipeline running on your Mac at home can now be supervised from your phone during your commute. A research agent spinning up on your office workstation can be redirected from a client meeting. That’s not a small thing for anyone running AI systems as actual business infrastructure.

One practical detail that doesn’t get enough attention: session reconnects automatically after sleep or network drops. You don’t have to babysit the connection. Close the app, reopen it, and you’re back in the same session. That matters when you’re bouncing between WiFi networks or your phone goes to sleep.

Who This Is For (And Who Should Wait)

Remote Control is a Max plan feature right now. That’s the $100/month individual plan, not Pro ($20/month) or any team/enterprise tier. Pro is listed as “coming soon”—no committed timeline as of this writing.

You’ll get immediate value if you:

• Run Claude Code for business workflows, not just one-off coding tasks
• Regularly need to check on or redirect long-running sessions
• Work across multiple locations or devices
• Have set up anything resembling an AI-powered operating system for your work

You can wait if you:

• Use Claude Code primarily for quick in-session coding help
• Stay at your desk for the full session lifecycle
• Are on Pro and won’t upgrade to Max

Pro users: This feature is worth watching. If Anthropic rolls it out to Pro at existing pricing, it substantially increases what the $20/month plan delivers for anyone running workflows.

Practical Workflow Patterns: Where Remote Control Actually Earns Its Keep

The feature is genuinely useful in a handful of specific scenarios. Here’s where I see the most leverage.

The Airport Pivot

You queued up a Claude Code session to run your weekly content pipeline before heading to the airport. Halfway through, the agent hits an ambiguous decision—two content angles both valid, needs direction. In the old model: job pauses, waits, resumes when you get back.

With Remote Control: you get a notification (if you’ve set Claude to alert on decision points), open the Claude app from your gate, provide direction, and the pipeline continues while you board. Job finishes before you land.

The Client Meeting Override

You’re running a background research agent on your office workstation. An hour in, new information from your client meeting makes part of the research direction obsolete. Rather than letting the agent burn time on work you know you’ll discard, open claude.ai/code on your phone, interrupt the session, redirect.

This is the scenario where Remote Control pays for itself in AI token savings alone.

The Multi-Device Workflow

You work on a laptop during the day and a desktop at home. Previously: context transfer between machines was clunky, session state didn’t migrate, starting fresh meant re-entering context.

With Remote Control: one session runs on whichever machine started it. You access it from any device. Continuity becomes the default.

The SMB Owner Scenario

If you’re running a small business and have built any level of AI workflow infrastructure—even something as simple as a regular Claude Code session for customer outreach drafting, competitor analysis, or content generation—Remote Control lets you supervise and direct that work without being physically present.

The workflow I’ve seen work well: start heavy-duty tasks on a capable desktop, step away, check in from your phone during natural breaks. You get the processing power of your workstation plus the flexibility of mobile oversight.

The Security Setup You Should Understand

This is the question I hear most: “Is it safe to expose my local session remotely?”

Here’s how it actually works. Your machine makes outbound HTTPS requests only. Remote Control does not open any inbound ports. There’s no port forwarding, no VPN requirement, no firewall rules to configure.

The connection model is similar to how many remote desktop tools work—your machine calls out to Anthropic’s relay infrastructure, and your phone connects through that relay. Your session never becomes a publicly accessible endpoint.

What does move through the relay: the conversation interface. Commands you send, responses you receive. Your actual files and local environment stay entirely on your machine. The relay sees the conversation, not your filesystem.

For most professional and SMB use cases, this is a reasonable security posture. If you’re in a regulated industry or have compliance requirements around AI tool usage, confirm with your security team before enabling. Anthropic publishes full technical documentation on Remote Control including the security architecture. The AI compliance considerations that apply to your business don’t change just because you’re accessing Claude Code from a different device.

How Remote Control Differs From Claude Code on the Web

Anthropic also offers Claude Code accessible via claude.ai/code in the browser—which can look similar from the outside. The difference is fundamental.

Claude Code on the web: Runs in Anthropic’s cloud environment. Clean slate, no local context, no access to your files or tools.

Claude Code Remote Control: Runs on YOUR machine. Full access to your local environment, files, installed tools, credentials, APIs, and anything else your terminal session can reach.

If you’ve invested time setting up Claude Code with custom skills, environment configurations, or integrations—all of that is only accessible via Remote Control. The web version can’t touch it.

This is why the feature matters for anyone who treats Claude Code as infrastructure rather than a one-off assistant. Your agent teams run on your hardware. Remote Control keeps you connected to them.

How to Set It Up in 10 Minutes

What you’ll need before starting:

• Claude Code installed and working locally
• Max plan subscription active
• Latest Claude Code version (run claude --version to check; update with npm install -g @anthropic-ai/claude-code if needed)
• Claude mobile app installed if you want phone access (iOS or Android)

Step 1: Update Claude Code

npm install -g @anthropic-ai/claude-code

Confirm you’re on 2.1.52 or newer before proceeding.

Step 2: Start a Remote-Accessible Session

claude remote-control

Or if you’re already in a session:

/remote-control

You’ll see a URL in your terminal. That’s your access point from other devices.

Step 3: Connect From Your Phone

Open the URL in any browser, or open the Claude mobile app and connect to the session. Press spacebar in the terminal to get a QR code if you want the fastest path to your phone.

Step 4 (Optional): Enable Auto-Start

If you want every Claude Code session to be remotely accessible:

/config

Set “Enable Remote Control for all sessions” to true. Done.

Checkpoint: After connecting from a second device, send a simple command and confirm it executes on your local machine. You should see the response in both places simultaneously—terminal and phone.

Common Failure Points

Session isn’t accessible after setup: Verify you’re on version 2.1.52+. Earlier versions don’t support the feature even if you have a Max plan.

Mobile app doesn’t show the session: The Claude mobile app needs to be signed into the same Anthropic account as your terminal session.

Connection drops and doesn’t reconnect: Auto-reconnect handles most network interruptions, but if your machine goes to sleep (not just the screen—full sleep mode), the session may need to be restarted. Configure your machine to stay awake during long sessions.

Pro plan gets the “not available” message: Max only for now. The Pro rollout is listed as coming but no date is committed.

The Practical Reality for Professionals

I’ll be direct about what this feature does and doesn’t change.

If you’re running Claude Code primarily for session-based coding assistance, Remote Control adds moderate convenience. Interesting, not essential.

If you’ve built workflows that run Claude Code for 20-60 minute tasks—content pipelines, research agents, data analysis runs—Remote Control adds something meaningfully different. It removes the “stay at your desk or lose the thread” constraint that limits how you structure AI-assisted work.

The professionals who’ll get the most value: anyone managing AI workflows the way they’d manage an employee doing a complex task. You want to be reachable if a decision needs to be made. You don’t want to physically supervise every step. Remote Control makes that oversight model viable.

One concrete efficiency gain: Long-running sessions that previously required babysitting can now be started before leaving the office. You supervise from wherever you are. The session runs on your hardware—full power, full context, full tool access—without requiring your physical presence.

For SMB owners who’ve started treating Claude Code as part of their operations stack, this is a meaningful capability upgrade that arrived quietly alongside a lot of louder AI announcements this week.

Your Next Step

If you’re on Max: update Claude Code and run claude remote-control today. Spend 10 minutes connecting your phone and testing a real workflow. The setup is trivial. The value shows up in the first session you run from a different device.

If you’re on Pro and this sounds useful: it’s worth tracking. The Pro rollout is coming. Building the habit of structured, supervisable workflows now—so you’re ready when Remote Control becomes available at your plan tier—is time well spent.

The pattern that makes Remote Control valuable is the same pattern that makes Claude Code valuable at all: treating AI sessions as ongoing work, not one-shot queries. Start building that way and the remote access capability pays off immediately when it arrives.

---

Related Reading:

• Claude Code as Your Business Operating System
• Agent Teams That Ship: Claude Code’s Multi-Agent Approach
• AI Implementation Guide: Bridge Learning to Real Results
• Save 20 Hours Weekly With n8n’s No-Code AI Automation