Claude Code Remote Control: Run Your Local Session From Your Phone
October 26, 2026
Claude Code Remote Control lets you start a task in your terminal, walk away, and keep steering it from your phone. The session never leaves your computer. Your phone or a browser tab is just a window into it.
That's the whole idea, and it's why people search for ways to control Claude Code from a phone. Here's the short version:
- Start it with
claude remote-control,claude --remote-control(or--rc), or/remote-controlinside a running session. - Connect by opening the session URL, scanning the QR code with the Claude app, or picking the session from the list at claude.ai/code.
- Plans: Pro, Max, Team, and Enterprise. API keys don't work.
- Where the code runs: on your machine, the whole time. Your computer has to stay on.
The rest of this guide covers the three ways to start it, how phone access works, what you can and can't do remotely, and how Remote Control differs from Claude Code on the web.
What Claude Code Remote Control actually does
Anthropic's Remote Control documentation describes it as a connection between claude.ai/code (or the Claude app for iOS and Android) and a Claude Code session running on your machine.
Because Claude keeps running locally, everything you set up still works:
- Your filesystem and project config. Typing
@from your phone autocompletes file paths from your local project. - Your MCP servers and tools. Nothing has to be reinstalled in a cloud environment.
- Both surfaces at once. The conversation, plus the progress of subagents, stays in sync across your terminal, browser, and phone. You can type from any of them.
- Attachments. Send a photo or file from your phone. Photos go to Claude directly. Other files are downloaded to your machine and passed in as
@file references. - Interruptions. If your laptop sleeps or the network drops, Claude Code reconnects when the machine comes back online.
The catch is the flip side of the benefit. The web and mobile interfaces are only a view. Close the terminal, quit the app running the session, or shut the lid for good, and the session goes offline.
Remote Control is not Claude Code on the web
This is the mix-up behind a lot of confused setup attempts. Remote Control and cloud sessions use the same claude.ai/code interface, but they run in different places.
| Remote Control | Cloud session (Claude Code on the web) | |
|---|---|---|
| Where Claude runs | Your own machine | Cloud infrastructure, Anthropic-managed by default |
| Your computer must stay on | Yes | No |
| Local MCP servers, tools, config | Available | Not unless configured in the cloud environment |
| Needs a GitHub repo | No | Yes, the repo is cloned into the cloud |
| Best for | Continuing work already in progress locally | Starting tasks with no local setup, or running several in parallel |
If you want to hand off a task and close the laptop entirely, that's a cloud session. Our guide to Claude Code on the web covers how those work. If you want to keep your local environment and just move the conversation to your phone, that's Remote Control.
Requirements before you start
Remote Control has a few hard requirements, and most "it doesn't work" reports trace back to one of them.
| Requirement | What the docs say |
|---|---|
| Plan | Pro, Max, Team, or Enterprise. API keys are not supported |
| Team and Enterprise | Off by default. An Owner has to turn on the Remote Control toggle in Claude Code admin settings |
| Sign-in | You need a claude.ai login. Run claude and use /login if you haven't |
| API endpoint | Not available on Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, or when ANTHROPIC_BASE_URL points at a gateway or proxy |
| Compliance setups | Organizations with Zero Data Retention or the HIPAA configuration can't enable it |
The most common failure: an ANTHROPIC_API_KEY in your shell or settings overrides your claude.ai login. Remove it and run claude auth login. When you're not sure which check is failing, claude doctor reports them one by one. Still getting Claude Code itself running? Start with our install guide.
How to start Remote Control: three ways
The CLI gives you three modes. The Desktop app and the VS Code extension use the slash command.
| How you start it | Command | What you get |
|---|---|---|
| Server mode | claude remote-control | A process that waits for remote connections and can serve several sessions |
| Interactive session | claude --remote-control or claude --rc | A normal terminal session you can also drive from your phone |
| From an existing session | /remote-control or /rc | Your current conversation, history included, made available remotely |
The first time you use any of them, Claude Code asks you to confirm before it connects. In a folder you haven't trusted yet, claude remote-control also asks whether to trust the directory first.
Server mode and its flags
Run claude remote-control in your project directory and it stays running, showing a session URL. Press the spacebar to show a QR code for your phone.
The flags you're most likely to use:
| Flag | What it does |
|---|---|
--name "My Project" | Sets the session title you'll see in the list at claude.ai/code |
--spawn worktree | Gives each new session its own git worktree, so parallel sessions don't edit the same files |
--spawn session | Serves exactly one session and rejects extra connections |
--capacity <N> | Caps concurrent sessions (the default is 32) |
--permission-mode <mode> | Sets the starting permission mode, such as acceptEdits |
-c, --continue | Brings back the session the last server in this folder started with |
The default spawn mode, same-dir, puts every session in one working directory, so parallel sessions can trip over each other's edits. worktree mode fixes that. Press w while the server runs to switch.
If you stop the server with Ctrl+C, running claude remote-control again in the same folder brings the sessions back. The docs say that works for about four hours.
The --permission-mode flag matters more than it looks. When you're approving actions from a phone screen, the mode decides how often you get asked. Our Claude Code permissions guide explains what each mode allows.
Interactive mode and the slash command
claude --remote-control "My Project" opens a regular interactive session that's also reachable from your phone. Unlike server mode, you can keep typing locally.
Already deep into a session? Type /remote-control. It carries your conversation history over. Run it again later to open a status panel with the URL, the QR code, and a disconnect option. While connected, the terminal shows an /rc active indicator.
To make every interactive session connect automatically, run /config and set Enable Remote Control for all sessions to true.
How to control Claude Code from your phone
Once a session is active, you have three ways in:
- Open the session URL in any browser. It takes you straight to the session on claude.ai/code.
- Scan the QR code with your phone. It opens the session in the Claude app.
- Open the Claude app and tap Code, then find the session by name. Online Remote Control sessions show a computer icon with a green dot.
Don't have the app yet? Run /mobile inside Claude Code. It shows a QR code that sends your phone to the right app store.
Push notifications
This is the part that makes phone control practical. With Remote Control active, Claude can push a notification when a long task finishes or when it needs a decision from you. You can also ask for one in your prompt, like "notify me when the tests finish."
To set it up, install the Claude app, sign in with the same account and organization you use in the terminal, allow notifications, then run /config in the terminal. Turn on Push when Claude decides, Push when actions required, or both.
Claude Code skips pushes while you're typing in the connected terminal, so you won't get buzzed while you're sitting right there.
What works from your phone, and what doesn't
Messages, permission prompts, and model or effort changes all work from a connected device. Some commands don't, because they only run in the terminal interface.
| Works from mobile and web | Local terminal only |
|---|---|
/compact, /clear, /context, /usage | /plugin |
/model sonnet, /effort high (pass the value as an argument) | /resume |
/rename, /color, /fast | /claude-api (Claude can still load that skill on its own) |
/mcp, returning a text summary in the mobile app | A line starting with ! goes to Claude as a message, not to your shell |
A connected device also shows a diff pane when the project is a git repository, so you can see what changed before you approve the next step.
Remote Control security: what leaves your machine
If you're wary of opening your laptop to your phone, the connection and security section of the docs is short and specific.
- No inbound ports. Your local session makes outbound HTTPS requests only. It registers with the Anthropic API and polls for work.
- TLS throughout. All traffic goes through the Anthropic API over TLS, the same transport as any Claude Code session.
- Short-lived credentials. The connection uses several credentials, each scoped to one purpose and expiring independently.
- The transcript is stored. While connected, your messages, Claude's responses, and tool activity are stored on Anthropic servers to keep devices in sync. Execution and file access stay local. Retention follows Anthropic's data usage policy.
- An off switch. The
disableRemoteControlsetting turns the feature off entirely, and an IT team can set it through managed settings.
Anthropic's security overview adds one point: no cloud VMs or sandboxing are involved in Remote Control. Your local setup is the boundary. If you want Bash commands fenced in, turn on sandboxing in a settings file.
Trusted Devices
Trusted Devices is labeled beta in the docs. When it's on, viewing or steering a Remote Control session requires an enrolled device and a sign-in no older than 18 hours. Instead of signing in daily, you confirm with Face ID, Touch ID, Windows Hello, or a passkey.
It's off by default. On Team and Enterprise, an Owner turns it on. On Pro and Max, you turn on Require trusted devices yourself in your settings. Anthropic says it never receives biometric data, only the device's public key and basic metadata.
Remote Control vs Dispatch, Channels, and cloud sessions
Claude Code now has several ways to work when you're away from the terminal. They differ in what starts the work and where it runs.
| Feature | How work starts | Where Claude runs | Best for |
|---|---|---|---|
| Remote Control | You drive a running session from claude.ai/code or the app | Your machine (CLI, Desktop, or VS Code) | Steering work already in progress |
| Dispatch | You message a task from the Claude mobile app | Your machine, via the Desktop app | Handing off work with minimal setup |
| Channels | Events pushed from Telegram, Discord, or your own server | Your machine (CLI) | Reacting to CI failures or chat messages |
| Cloud session | You start a task on the web | Anthropic's cloud by default | Work that shouldn't depend on your laptop |
One limitation to plan around: outside server mode, each Claude Code process supports one remote session at a time. If you run Claude Code on a remote box over SSH, start it inside tmux or screen so the session survives when you disconnect.
Where Taku fits
Remote Control is great once Claude Code is installed, signed in, and configured the way you like. For a lot of people, that setup is exactly where things stall. They see a working agent setup shared online and never get it running locally.
Taku is built for that gap. It mirrors a working AI setup into a desktop workspace and runs it, so you can use a power user's configuration without reproducing their environment first. Taku is in Beta, and the Mac app is available now.
FAQ
What is Claude Code Remote Control?
A feature that connects claude.ai/code or the Claude mobile app to a Claude Code session running on your own computer. You can send messages, approve actions, and watch progress from another device while the code keeps running locally.
How do I control Claude Code from my phone?
Start a session with claude remote-control, claude --remote-control, or /remote-control. Then scan the QR code with the Claude app, open the session URL in your phone's browser, or tap Code in the Claude app and pick the session by name.
Which plans include Claude Code Remote Control?
Pro, Max, Team, and Enterprise. It doesn't work with API-key authentication. On Team and Enterprise, an Owner must enable the Remote Control toggle in admin settings first.
Does my computer need to stay on?
Yes. Remote Control sessions run on your machine, so the computer has to stay on and the claude process has to keep running. If your laptop sleeps briefly, Claude Code reconnects when it wakes. For work that doesn't depend on your machine, use a cloud session instead.
Is Claude Code Remote Control secure?
Your machine makes outbound HTTPS requests only and opens no inbound ports. Traffic travels over TLS through the Anthropic API, using short-lived, narrowly scoped credentials. The session transcript is stored on Anthropic servers while connected. For stricter control, organizations can require Trusted Devices (in beta) or disable Remote Control entirely.
What's the difference between Remote Control and Claude Code on the web?
Where the session runs. Remote Control drives a session on your computer, with your local files, tools, and MCP servers. Claude Code on the web runs the session in the cloud from a GitHub repo, so your computer can be off.