How to Use Claude Code From Your Phone With /remote-control

The desk-bound problem with AI coding

AI coding agents like Claude Code are at their best on long-running tasks: a multi-file refactor, a full test suite, a build-and-deploy pipeline. The catch is that these tasks don't finish on your schedule. You kick one off, then five minutes later you're in a meeting, walking to lunch, or already home — and the agent is sitting there waiting for you to approve a command or review a diff.

For most of Claude Code's life, the answer was "be at your keyboard." Remote Control changes that. With one command you can turn a session running on your machine into something you can monitor and steer from your phone, a tablet, or any browser.

This guide covers what the /remote-control command does, how to set it up, when to reach for it versus Claude Code on the web, and the security model that makes it safe to use.

Note: Remote Control is in research preview. The exact behavior and version requirements below are accurate as of June 2026 but may shift as the feature matures. Run claude --version to confirm you're on v2.1.51 or later.

What Remote Control actually is

The single most important thing to understand: Remote Control is a remote window into a session that keeps running on your own machine. It is not cloud execution.

When you enable it, your laptop or workstation keeps doing all the real work — reading and writing files, running your tests, talking to your local MCP servers. What travels over the network is just the conversation: the messages you send and the output and permission prompts that come back. Your code never leaves your machine.

That distinction is what separates it from the other ways you can reach Claude Code:

Remote Control isn't a separate "mode" you switch into — it's your normal local CLI (or VS Code) session, made reachable from your pocket.

Prerequisites

Before you start, make sure you have:

• Claude Code v2.1.51 or later (claude --version)
• A Pro, Max, Team, or Enterprise plan — Remote Control is included at no extra cost
• Account login, not an API key — sign in with /login. If you have ANTHROPIC_API_KEY set, unset it; API-key auth is not supported here
• The Claude mobile app (iOS or Android) if you want QR-code connection and push notifications
• Workspace trust accepted — run claude in your project at least once
• Team/Enterprise only: an admin must enable the Remote Control toggle in Claude Code admin settings

How to set it up

There are three ways to start, depending on whether you're already in a session.

1. Enable it on a session you're already running

If you're mid-task in the terminal and realize you need to step away, just type:

/remote-control

(or the shorter alias /rc). Your conversation history carries over, and Claude prints a link plus a QR code. In the VS Code extension, type /rc in the prompt box — note that VS Code doesn't render the QR code or accept a name argument.

2. Start a dedicated server for background tasks

For unattended or multi-session use, start server mode in its own terminal:

claude remote-control

This process stays running and displays a session URL and QR code (press the spacebar to toggle the QR display). Server mode can host many concurrent sessions and accepts flags like --name, --spawn, --capacity, and --sandbox.

3. Start a fresh interactive session, remote-enabled

claude --remote-control "My Project"

This opens a normal interactive session you can drive from both the terminal and remote devices simultaneously — anything you type in one place shows up everywhere.

Connecting from your phone

Once a session is live, you have three ways in:

1. Scan the QR code with the Claude mobile app (or any QR scanner, which opens the link in a browser).
2. Open the URL printed in your terminal in any browser.
3. Browse the session list under the Code tab in the Claude mobile app or at claude.ai/code.

From there your phone behaves like a thin client: you read output, send new instructions, and approve or deny permission prompts, all while the work happens on your machine.

When to use it — and when not to

Remote Control shines in a specific set of situations. Reach for it when:

• You've started a long task and need to leave your desk. Kick off the build at your desk, approve the deploy step from the couch.
• Claude is blocked on a permission decision. Instead of the agent idling until you're back, you approve the command from your phone and it keeps moving.
• You're monitoring a long test or migration run. Watch progress in real time from anywhere and intervene only if something goes wrong.
• You want a hands-off, notify-me workflow. Ask Claude to ping your phone when the suite passes, then walk away.
• You work across devices. Start in the terminal, continue on mobile, glance at it in a browser — the conversation stays in sync.

Reach for something else when:

• You have no local setup for the repo. If the code isn't on your machine, use Claude Code on the web, which clones and runs in the cloud.
• The task must run truly unattended for hours. The host process has to stay alive and reachable (see the ~10-minute timeout below). Cloud execution or a scheduled/cloud agent is a better fit.
• You need many parallel tasks at once. Server mode helps, but heavy fan-out is usually cleaner in the cloud.

Limitations and gotchas

A few things to keep in mind so you're not surprised:

• The local process must stay running. Close the terminal or quit VS Code and the session ends. There's no background daemon keeping it alive.
• ~10-minute network timeout. If your machine is awake but unreachable for more than about 10 minutes, the session times out and exits. Keep the device awake and on a stable connection.
• One connection per interactive process. Each claude instance supports a single remote connection; use claude remote-control server mode for concurrency.
• A few commands are terminal-only. Things like /plugin and /resume need the local picker, though most commands (/mcp, /compact, /context, /usage, and others) work fine from mobile or web on recent versions.
• Notifications can be throttled by the OS. iOS Focus modes and Android battery optimization may delay or suppress push alerts. If /config shows "No mobile registered," open the Claude app on your phone to refresh the push token.

The security model

Remote Control was designed so that opening your session to your phone doesn't open your machine to the world:

• Your files never leave your machine. Unlike cloud execution, all file operations and tool runs happen locally.
• No inbound ports. Your machine registers with Anthropic and polls outbound for work, so there's nothing new to expose or firewall.
• HTTPS with short-lived credentials. All traffic runs over HTTPS using single-purpose, short-lived tokens tied to your account login.
• You're still the approver. Permission prompts come to your device — you decide what runs, just as you would at the keyboard. The usual rule applies: don't enable auto-approve for anything destructive just because you're approving from your phone.

Bottom line

/remote-control removes the requirement to be physically at your keyboard to supervise Claude Code. It keeps the session — and your code — on your own machine while giving you a synced, secure window from your phone or browser. Use it to babysit long-running work, approve permissions on the move, and get pinged when a task finishes. When the work doesn't need a local machine at all, or has to run untended for hours, fall back to Claude Code on the web. Between the two, you rarely have a reason to sit and wait for an agent again.

Related Resources

• CLI vs IDE Extension vs Cloud: Which AI Coding Interface is Best? — Where Remote Control fits among the options
• Claude Code Pricing 2026: Free, Pro, Max & API Cost — Which plans include Remote Control
• Claude Code vs Cursor vs GitHub Copilot — How the tools compare
• Best Practices for AI Coding CLIs in Production — Safe approval and permission habits