TL;DR: Anthropic just launched Claude Code Remote Control, a feature that lets you control your local Claude Code terminal sessions from your phone, tablet, or any browser. Run claude remote-control in your terminal, scan a QR code, and you’ve got full control of your coding session from anywhere in your house — or anywhere with an internet connection. Your code stays local. Nothing moves to the cloud. Available now for Max subscribers ($100–$200/month) with Pro plan access ($20/month) rolling out soon.

What Is Claude Code Remote Control?

Claude Code Remote Control is a new feature from Anthropic that creates a secure bridge between your local Claude Code terminal session and any remote device. Think of it as a live window into your running coding session that you can access from your phone, a tablet, or another computer’s browser.

The critical distinction here is that this is not cloud computing. When you use Remote Control, Claude continues running on your local machine. Your filesystem, your MCP servers, your tools, your environment variables, your project configuration — all of it stays exactly where it is. The remote device is simply a viewport and input mechanism for that local session.

This matters because many developers have complex local setups with custom tooling, private repos, and Model Context Protocol (MCP) integrations that don’t exist in the cloud. Remote Control preserves all of that context while letting you walk away from your desk.

Key Takeaways

Instant setup with minimal friction. Run claude remote-control or type /rc inside an existing session. A session URL and QR code appear. Scan the code with your phone and you’re connected.

Everything stays local. Your code, files, MCP servers, and project configuration never leave your machine. The remote device is just a control interface — Anthropic’s servers route messages between your devices over TLS, but your actual development environment stays put.

Conversations sync across all devices. You can send messages from your terminal, then from your phone, then from a browser on a different computer. The session doesn’t care where the input comes from. Everything stays in sync.

Auto-reconnect after interruptions. If your laptop goes to sleep or your network drops, the session automatically reconnects when your machine comes back online. You don’t lose your place.

One remote session at a time. Each Claude Code session supports a single remote connection. Your terminal must stay open — if you close it or kill the Claude process, the session ends.

Available for Max subscribers now. Remote Control requires a Pro or Max plan. API keys are not supported. Max users have access today, with Pro access rolling out soon. Team and Enterprise plans are not yet supported.

Not the same as Claude Code on the web. Claude Code on the web runs on Anthropic’s cloud infrastructure and doesn’t need a local machine at all. Remote Control runs on your machine and gives you remote access to that local session. Different tools for different situations.

How Claude Code Remote Control Works Under the Hood

When you start a Remote Control session, your local machine initiates an outbound HTTPS connection to Anthropic’s API. No inbound ports are opened on your computer. Your machine registers with the API and polls for work. When you connect from a remote device — phone, tablet, browser — Anthropic’s server routes messages between the web/mobile client and your local session over a streaming connection.

All traffic passes through Anthropic’s API over TLS using multiple short-lived credentials that are scoped to a single purpose and expire independently. Your files and MCP servers never leave your machine. Only chat messages and tool results flow through the encrypted bridge.

This architecture means your session URL is effectively a credential. Anyone with that URL can interact with your local Claude Code session, including approving file changes. Treat it like a password.

How to Set Up Claude Code Remote Control: Step by Step

Prerequisites

Before you start, make sure your environment meets these requirements:

Subscription: You need a Pro or Max plan on claude.ai. API keys won’t work here.

Authentication: Run claude in your terminal and use /login to sign in through claude.ai if you haven’t already.

Workspace trust: Run claude in your project directory at least once and accept the workspace trust dialog.

Claude Code version: Update to version 2.1.52 or later.

Option 1: Start a New Remote Control Session

Navigate to your project directory and run:

claude remote-control

The terminal displays a session URL and stays running, waiting for remote connections. Press spacebar to toggle a QR code display for quick phone access.

This command supports several flags:

--verbose shows detailed connection and session logs. --sandbox / --no-sandbox enables or disables filesystem and network isolation during the session. Sandboxing is off by default.

Option 2: Go Remote From an Existing Session

If you’re already deep in a Claude Code conversation and want to continue it from another device, type:

/remote-control

Or use the shorthand:

/rc

This carries over your entire conversation history — every message, file edit, and tool call — and generates the session URL and QR code. The --verbose, --sandbox, and --no-sandbox flags are not available with this in-session command.

Connecting From Your Phone or Another Device

You have three ways to connect:

Scan the QR code. This is the fastest path. Point your phone camera at the code and it opens directly in the Claude app if you have it installed.

Open the session URL. The URL is displayed in your terminal alongside the QR code. Copy it and open it in any browser.

Find it in the session list. Open claude.ai/code or the Claude mobile app and look for your session by name. Remote Control sessions display a computer icon with a green status dot when they’re online.

Pro tip: Use /rename to give your session a descriptive name before going remote. Something like “Auth refactor – Feb 2026” is much easier to find than the default “Remote Control session.”

If you don’t have the Claude mobile app yet, type /mobile inside Claude Code to display a download QR code for iOS or Android.

Enable Remote Control for All Sessions Automatically

By default, Remote Control only activates when you explicitly run the command. To make every session remotely accessible automatically, run /config inside Claude Code and set “Enable Remote Control for all sessions” to true.

Remote Control vs. Claude Code on the Web

Both Remote Control and Claude Code on the web use the claude.ai/code interface, but they serve fundamentally different purposes.

Remote Control executes on your machine. Your local MCP servers, tools, custom configurations, and entire filesystem remain available. Use this when you’re in the middle of local work and want to keep going from another device.

Claude Code on the web executes on Anthropic-managed cloud infrastructure. Use this when you want to kick off a task without any local setup, work on a repo you don’t have cloned, or run multiple tasks in parallel.

The choice is straightforward: if you have a complex local environment with MCP servers and custom tooling, use Remote Control. If you want zero-setup cloud execution, use Claude Code on the web.

Real-World Use Cases for Remote Control

Long-running refactors. Kick off a large refactoring task — say, migrating 40+ files from CSS modules to Tailwind. Instead of sitting there watching for 20 minutes, scan the QR code and monitor progress from your phone while you take a break.

Build monitoring. Start a complex build that’s been failing intermittently. Claude investigates, reads logs, and tries fixes. You head to a meeting and check results from your phone during a quiet moment.

Multi-session management. Run three separate Claude Code sessions — one fixing a production bug, one writing tests, one doing a dependency upgrade. Each gets its own Remote Control session. Switch between them on your phone like switching chat threads.

End-of-day code review. You’ve been at your desk for hours and don’t want to stare at a monitor anymore, but you have generated code to review. Connect from your phone on the couch, scroll through file changes, and leave follow-up instructions.

Limitations and Things to Watch Out For

Your terminal must stay open. Remote Control runs as a local process. If you close the terminal or stop the Claude process, the session ends. Period.

Your machine must be on and connected. If your home WiFi goes down, the session pauses. Claude isn’t doing work in the background while your machine is offline. It auto-reconnects when connectivity returns, but no work happens in the meantime.

One remote connection per session. You can’t have multiple remote devices controlling the same session simultaneously in separate connections — though you can switch between devices since the conversation syncs.

Mobile screen limitations. Reviewing diffs and detailed code on a phone screen has obvious constraints. Remote Control is great for monitoring, approvals, and simple instructions. Detailed code review is still better at your desk.

Session URL security. Anyone with your session URL has full control. This includes the ability to approve file changes on your local machine. Don’t share it carelessly.

Early bugs exist. As with any new feature, there are some edge cases. Some users have reported that remote sessions don’t always appear in the session list, making it hard to reconnect after navigating away from the app. Anthropic is actively working on fixes.

The Bigger Picture: Why This Matters

Remote Control might seem like a convenience feature, but it signals something important about where AI-assisted development is heading. Claude Code has hit a $2.5 billion annualized run rate as of February 2026 — more than doubling since the start of the year. It now powers an estimated 4% of all public GitHub commits worldwide and has reached 29 million daily installs in Visual Studio Code.

The move to mobile access reflects a shift in how developers interact with AI coding agents. These aren’t autocomplete tools that need you staring at a screen. They’re autonomous agents that work through multi-file tasks over minutes or hours. The natural next step is letting developers supervise and direct that work from wherever they happen to be.

Before Remote Control launched officially, developers were already hacking together mobile access using Tailscale for tunneling, Termius or Termux for mobile SSH, and tmux for session persistence. Some built custom WebSocket bridges. Anthropic essentially productized what power users were already doing — but with native integration, auto-reconnect, and proper security.

Competitors are approaching this differently. GitHub Copilot’s coding agent can be assigned from GitHub Mobile, but it runs entirely in GitHub’s cloud via Actions. Cursor shipped a third-party iOS companion app in January 2026 that relays prompts to a Mac running the Cursor IDE. Claude Code’s differentiator is clear: local execution with the full environment preserved, accessible from any device.

My Thoughts

This feature solves a real problem that most Claude Code users have hit at least once: you’re mid-task, something interrupts you, and you either abandon context or chain yourself to your desk. Remote Control eliminates that tradeoff entirely.

The fact that everything stays local is the right call architecturally. Developers with complex local setups — custom MCP servers, specific environment configurations, private tooling — would never trust a cloud handoff to preserve all of that context correctly. By keeping execution local and just proxying the interface, Anthropic avoids that trust problem entirely.

The security model is sensible too. Outbound-only connections, TLS encryption, short-lived credentials — there’s nothing unusual here, and that’s the point. The session URL being the single credential to protect keeps the mental model simple.

Where this gets really interesting is when combined with the auto-enable option. If every Claude Code session is automatically remotely accessible, developers can adopt a new workflow pattern: start tasks, walk away, check in periodically from wherever they are, and come back to completed work. That’s a meaningful change in how coding sessions are structured throughout a day.

For solo developers and indie hackers juggling multiple projects, the ability to monitor and manage several concurrent Claude Code sessions from a phone is genuinely powerful. It turns dead time — waiting rooms, commutes, coffee breaks — into lightweight supervision time.

The main concern is the early-stage bugs. Session reconnection issues and visibility in session lists need to be rock-solid for this to be a reliable part of anyone’s workflow. But those are solvable problems, and Anthropic’s pace of iteration on Claude Code has been consistently fast.

Bottom line: if you’re a Claude Code user on a Max plan, there’s no reason not to try this today. It takes 30 seconds to set up and it fundamentally changes how you can structure your development sessions.