Claude Code Remote Control connects a session running on your computer to claude.ai/code or the Claude app on iOS and Android, so you can read along, type, approve prompts and attach files from your phone while the code keeps running locally. Start it with claude remote-control, claude --remote-control, or /remote-control inside a running session. As of October 2026 it needs a claude.ai Pro, Max, Team or Enterprise login, and your computer has to stay on.
How Remote Control works
Nothing moves to the cloud except the conversation. Claude Code keeps running in your terminal, desktop app or VS Code, with your files, MCP servers, tools and project settings. The phone or browser is a window into that session:
- Your machine makes outbound HTTPS requests only and never opens an inbound port. It registers with the Anthropic API and polls for work; messages between the device and your session travel over TLS.
- While Remote Control is connected, the transcript is stored on Anthropic's servers so every device stays in sync and the session can reconnect after a network drop.
- The conversation, subagents and workflows stay in sync everywhere, so you can type from the terminal, a browser and your phone interchangeably.
- From the device you can see a diff of your changes, pick the model and effort level, and attach photos or files. Claude Code downloads attached files to your machine.
Set it up in two minutes
You need Claude Code signed in with your claude.ai account (/login), not an API key. On Team and Enterprise plans an Owner first has to enable the Remote Control toggle in the Claude Code admin settings.
There are three ways to start, depending on where you are. Server mode waits for connections and can host several sessions:
claude remote-control --name "checkout refactor"
The flag starts a normal interactive session that you can also drive remotely:
claude --remote-control "checkout refactor"
Already in a session? Type /remote-control (or /rc). It carries over the conversation so far. The same command works in the desktop app and the VS Code extension.
Then connect from your phone in one of three ways:
- Scan the QR code. In server mode, press the spacebar to show it.
- Open the session URL that Claude Code prints, in any browser.
- Open the Claude app, tap Code, and pick the session from the list. Remote Control sessions show a computer icon with a green dot while they are online.
No Claude app yet? Run /mobile in Claude Code for a QR code that opens the right app store.
Useful server options
claude remote-control takes a few flags worth knowing:
| Flag | What it does |
|---|---|
--spawn worktree | Each new session you start from the phone gets its own git worktree, so parallel sessions don't edit the same files |
--capacity 8 | Caps concurrent sessions (the default is 32) |
--permission-mode acceptEdits | Sets the starting permission mode for the server's sessions |
--continue | Brings back the session the last server in this folder started, within about four hours of stopping it |
To connect every interactive session automatically, run /config and turn on Enable Remote Control for all sessions, or set it in ~/.claude/settings.json:
{
"remoteControlAtStartup": true
}Turn on push notifications
Remote Control can push to your phone. Install the Claude app, sign in with the same account and organization, and allow notifications. Then run /config in Claude Code and enable Push when actions required (permission prompts and questions), Push when Claude decides (usually when a long task finishes), or both. Pushes are skipped while you are typing in the connected terminal. If /config says No mobile registered, open the Claude app once so it refreshes its push token.
Limits worth knowing
As of October 2026, the documented limits are:
- Your computer must stay on and the
claudeprocess must keep running. On a remote server over SSH, start Claude insidetmuxorscreenso it survives the disconnect. - One remote session per interactive process. Use server mode to host several from one process.
- Not every provider. It does not work with Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, a custom
ANTHROPIC_BASE_URL, or organizations with Zero Data Retention or the HIPAA configuration. - Permission modes from the app are Manual, Accept edits, Plan and Auto. You can't select Bypass permissions from the phone.
- Some commands are terminal-only, such as
/pluginand/resume. Text commands like/compact,/contextand/usage, and/model sonnetwith an argument, work from the phone. - Long outages end server mode. If your machine is awake but offline for roughly ten minutes,
claude remote-controlexits; an interactive session keeps retrying until the network returns.
Other ways to follow Claude Code from your phone
Remote Control is the right tool when the work needs your local files and tools. It isn't the only option:
| Option | Where Claude runs | Best for |
|---|---|---|
| Remote Control | Your computer | Steering a local session you already started |
| Cloud sessions at claude.ai/code | Anthropic's cloud | Tasks that should keep running with your laptop closed; claude --cloud sends one from the terminal |
| Dispatch | Your computer, through the desktop app | Messaging a task from the phone and letting the desktop app run it (Pro or Max) |
| Channels | Your computer | Pushing Telegram, Discord or iMessage messages into a running session (research preview) |
| SSH and tmux | Any machine | Full terminal access from a phone terminal app, with no Claude features in between |
Channels are worth a look if you live in Telegram. After installing the official plugin with /plugin install telegram@claude-plugins-official and pairing your bot, you start Claude Code with the channel turned on, and messages to your bot arrive in that session:
claude --channels plugin:telegram@claude-plugins-official
Channels only deliver while that session is open, so they suit a long-running session in tmux or a background process.
Get a ping when any session needs you
Remote Control's pushes cover the Claude Code sessions you connected. If you run several at once, start some without Remote Control, or mix in Codex or Cursor, you want one place that tells you which one is waiting.
That is the job Briefkin does. A one-command installer adds a hook to each agent on your computer, and every session becomes a card on one board: working, waiting for permission or input, or done. You get a Telegram ping only when a session is waiting on you, never for routine progress. When a local Claude Code session ends its turn on a question, you can answer on the board or by replying to the Telegram message, and the session wakes up with your answer as if you had typed it. Permission prompts still have to be answered in the session itself, which is where Remote Control and Briefkin fit together. Setup takes a minute and is described in the docs.
For the wider picture, see how to get notified when Claude Code finishes and how to watch many sessions in one place. If you plan to leave a session running with fewer prompts, read the guide to skipping permissions first.
Questions
Is there a Claude Code mobile app?
Not a separate one. Claude Code sessions live in the Code tab of the regular Claude app for iOS and Android, which can open cloud sessions, projects and Remote Control sessions running on your computer.
Does Remote Control work with an API key, Bedrock or Vertex?
No. As of October 2026 it needs a claude.ai Pro, Max, Team or Enterprise login, and it is unavailable on Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, or when ANTHROPIC_BASE_URL points somewhere other than api.anthropic.com.
Does my computer have to stay on?
Yes. The session runs on your machine, so the claude process has to keep running. If the laptop sleeps, Claude Code reconnects when it wakes; for work that must continue with the lid closed, use a cloud session instead.
Can I approve permission prompts from my phone?
Yes. Permission prompts and Claude's questions are forwarded to the connected device and stay open until you answer. You can switch between Manual, Accept edits, Plan and Auto from the app, but not to Bypass permissions.
Is Claude Code Remote Control secure?
Your machine only makes outbound HTTPS requests and opens no inbound ports; traffic goes through the Anthropic API over TLS with short-lived, single-purpose credentials. While connected, the transcript is stored on Anthropic's servers so devices stay in sync. Admins can turn the feature off with disableRemoteControl.