Claude Code remote control: drive a local session from your phone

Last updated: November 6, 2026

TL;DR

Claude Code remote control lets you drive a Claude Code session that runs on your own computer from claude.ai/code or the Claude app on your phone. Run claude remote-control in your project folder, or type /rc in a session you already have open, then open the URL it prints or scan its QR code. Your code, files and MCP servers stay on your machine, so the machine has to stay on.

• ✅ Before you start. You need a Pro, Max, Ultra, Team or Enterprise plan and a claude.ai sign-in. API keys, Amazon Bedrock and LLM gateways don't work. On Team and Enterprise, an Owner turns it on first.
• 🚀 Start it. Server mode for a session you only steer from elsewhere, or /rc to hand off the session you're in.
• 📱 Connect. Open the URL, scan the QR code, or pick the session in the Claude app's Code tab.
• 🔀 Run several. One server holds up to 32 sessions, and pressing w gives each new one its own git worktree.
• 🔋 Keep it alive. Use tmux on a remote box. A stopped server's sessions come back for about four hours.
• 🧭 Pick something else when the repo isn't on your machine (a cloud session) or you want to send a task rather than steer one (Dispatch).

Before you start

You need an eligible plan, a claude.ai sign-in and a direct connection to Anthropic's API. Check these once; every item comes from Anthropic's Remote Control docs.

• Plan: Pro, Max, Ultra, Team or Enterprise. Remote control doesn't work with API keys.
• Team and Enterprise: an Owner enables the Remote Control toggle in Claude Code admin settings. It's off by default, which our Claude Team admin settings guide also flags. On Enterprise with custom roles, your roles need both the cloud sessions permission and the Claude Code remote control permission.
• Sign-in: run claude, then /login, and sign in through claude.ai. A long-lived token from claude setup-token can't open a remote session.
• API endpoint: the session has to talk to api.anthropic.com. Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, an LLM gateway or a custom ANTHROPIC_BASE_URL all block it.
• Environment: CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC or DISABLE_GROWTHBOOK turns remote control off. Unset them in your shell and in the env block of any settings.json.

Step 1: Start a remote control session

Start with claude remote-control for a session you'll drive from elsewhere, or /rc to hand off the session you're already in.

A session you'll drive from elsewhere (server mode). In your project folder, run:

claude remote-control

The first time, it explains what it does and asks Enable Remote Control? (y/n). In a folder you haven't trusted yet, it also asks Trust <directory>? [y/N]. Then it waits for connections, prints a URL and shows the connection status and tool activity. You don't type into this terminal; you work from the other device.

The session you're already in. Type this at the Claude Code prompt:

/rc

/rc is short for /remote-control. Add a name if you want one in the session list, such as /rc billing refactor. Your conversation history carries over, and a one-time dialog asks you to confirm the first time. The same command works in the VS Code extension and in a local session in the Claude Desktop app's Code tab. While it's connected, the terminal shows an /rc active indicator. Run /rc again to see the URL and QR code, or to disconnect while the local session keeps running.

You can also launch claude with remote control already on. That flag is in the CLI reference.

Step 2: Connect from your phone or browser

Connect by opening the URL Claude Code prints, scanning its QR code, or picking the session in the Claude app:

1. Open the URL Claude Code printed. It goes to claude.ai/code.
2. Scan the QR code to open the session in the Claude app. In server mode, press the spacebar to show or hide it.
3. Open the Claude app and tap Code, or open claude.ai/code, and pick the session from the list. A remote session shows a computer icon with a green dot while it's online.

No Claude app yet? Run /mobile in Claude Code for a QR code that opens the right app store.

Once connected, both sides stay in sync. You can type from the terminal, the browser and the phone interchangeably, attach a photo or file from the phone, answer permission prompts, and see a diff of your changes when the folder is a git repository. Typing @ on the phone autocompletes file paths from your local project.

Step 3: Get a push when Claude needs you

Turn on push notifications in /config, so your phone tells you when Claude is waiting instead of you checking it. Turn on one or both:

• Push when Claude decides, for things like a long task finishing.
• Push when actions required, for permission prompts and questions.

Sign in to the Claude app with the same account and organization you use for Claude Code, and allow notifications. If /config shows No mobile registered, open the app once so it refreshes its push token. On iOS, Focus modes can hold pushes back. On Android, exempt the Claude app from battery optimization.

Run several sessions from one server

One claude remote-control process serves up to 32 concurrent sessions by default, and you can start new ones from claude.ai/code or the app.

By default every session shares the current folder, so two sessions editing the same file can collide. Press w in the server's terminal to switch to worktree mode, where each new session gets its own git worktree. Claude Code remembers that choice for the project. Worktree mode needs a git repository or a WorktreeCreate hook.

Outside server mode, each Claude Code process carries one remote session at a time.

Keep the session alive

The phone is a window into a process on your machine. When the process stops, the session goes offline.

• Close the terminal or quit the app that runs the session, and it goes offline until you bring it back.
• Working over SSH? Start claude inside tmux or screen, so it keeps running after you disconnect.
• Laptop sleeps or Wi-Fi drops? Claude Code reconnects on its own when the machine is back online.
• Long outage on an awake machine: server mode gives up after roughly 10 minutes and exits. An interactive session keeps retrying and reconnects when the network returns.

Bring sessions back after stopping a server. If you stopped claude remote-control with Ctrl+C, run one of these in the same folder within about four hours:

For a session you started with /rc, resume the conversation with claude -c or claude -r. Claude Code reconnects it to the same remote session.

Turn it on for every session

To skip typing /rc, run /config and set Enable Remote Control for all sessions to true. The other values are false and default, which follows your organization's admin default if it has one.

To set it in a file instead, add "remoteControlAtStartup": true to your user ~/.claude/settings.json. A project's .claude/settings.json can turn auto-connect off with false, but Claude Code ignores a true there, so a checked-in file can't switch it on for everyone who opens the repo. Auto-connect uses your own claude.ai account, so the session appears only in your apps.

The VS Code extension has the same toggle in its command menu's Settings section. In the Desktop app it's Connect new sessions to Remote Control under Settings.

Common Claude Code remote control errors and fixes

Most errors are a sign-in or settings problem, and the message names it. Match the message Claude Code prints. claude doctor also reports remote control eligibility without starting a session.

"Remote Control disconnected" or "Can't reach" on the phone only means the device lost contact. Check the machine: if Claude Code is running with no failure shown, it reconnects by itself. If it isn't running, start it again and bring the session back as above.

Remote control, cloud sessions or Dispatch?

All three let you work away from your desk. The difference is where Claude runs and how you start the work:

For recurring work, scheduled tasks are the better fit.

What stays on your machine

Execution and file access stay local. The session makes outbound HTTPS requests only and never opens an inbound port. Traffic goes through the Anthropic API over TLS, with short-lived credentials scoped to one purpose each.

One thing does leave the machine. While remote control is connected, the transcript, including your messages, Claude's replies and tool activity, is stored on Anthropic's servers so devices stay in sync. It's kept under Anthropic's data usage policy; our post on Claude data privacy covers what that means on a Team plan.

Three controls are worth knowing:

• Turn it off on a machine with the disableRemoteControl setting, which IT can also push through managed settings.
• Organizations with zero data retention or the HIPAA configuration can't enable remote control at all.
• Trusted Devices, in beta, requires an enrolled device and a sign-in no more than 18 hours old before anyone can view or steer a remote session. On Team and Enterprise, an Owner turns it on under Organization settings, Capabilities, Remote sessions; on Pro, Max and Ultra you turn it on yourself. Members confirm with Face ID, Touch ID, Windows Hello or a passkey.

Frequently asked questions

What is Claude Code remote control?

Claude Code remote control connects claude.ai/code or the Claude mobile app to a Claude Code session running on your own computer. You can send messages, answer permission prompts and watch progress from another device, while the code, files and MCP servers stay on your machine.

How do I use Claude Code remote control from my phone?

Run claude remote-control in your project folder, or type /rc in an open session. Then scan the QR code with your phone, or open the Claude app, tap Code and pick the session marked with a green dot. Your computer has to stay on.

Is Claude Code remote control available on the Pro plan?

Yes. It works on Pro, Max, Ultra, Team and Enterprise plans with a claude.ai sign-in. It doesn't work with API keys. On Team and Enterprise, an Owner has to turn on the Remote Control toggle in Claude Code admin settings first.

Why is Claude Code remote control not available?

The usual causes are an API key or token taking precedence over your claude.ai login, an Owner who hasn't enabled it, an LLM gateway or Bedrock setup, or an environment variable that turns off feature flags. Run claude doctor to see which eligibility check failed.

Does my computer need to stay on for remote control?

Yes. The session runs on your machine, and the phone or browser is a window into it. If the computer sleeps or the network drops, Claude Code reconnects when it's back. If the claude process stops, the session goes offline until you resume it.

Can I run more than one remote control session?

Yes. A single claude remote-control server holds up to 32 concurrent sessions by default. Press w in its terminal so each new session gets its own git worktree instead of sharing one folder. An interactive session carries one remote session per process.

Is Claude Code remote control secure?

Your machine makes outbound HTTPS requests only and opens no inbound ports, and traffic goes through the Anthropic API over TLS. The transcript is stored on Anthropic's servers while connected. Team and Enterprise Owners can require Trusted Devices, and the disableRemoteControl setting turns the feature off on a machine.

What is the difference between Claude Code remote control and Dispatch?

Remote control lets you steer a Claude Code session that's already running on your machine. Dispatch lets you message a task from the Claude mobile app, and it can start a Claude Desktop session on your machine to handle it. Both run Claude on your computer.

Make the handoff worth it

Remote control is only as good as the session you hand off. If Claude needs your project's rules re-explained every time you pick it up on the phone, put those rules in the repo, where every session reads them. Our guide to compound engineering for Claude Code shows how to build that up one fix at a time.