Claude Code Remote Control lets you keep coding from anywhere

Claude’s new(ish) Remote Control mode for Claude Code is designed for a specific kind of AI-assisted coding workflow: keep a session running locally on a development machine, but continue the same conversation from a browser or phone when stepping away. It’s currently available as a research preview on Pro and Max plans, and it’s not available on Team or Enterprise plans.

Unlike “Claude Code on the web,” which runs in Anthropic-managed cloud infrastructure, Remote Control keeps execution on the local machine. The web and mobile UI effectively become a thin window into the existing local session—handy for checking progress, nudging a task along, or resuming work without moving the environment itself.

What Remote Control is (and what it isn’t)

Remote Control connects claude.ai/code and the Claude app on iOS and Android to a Claude Code session already running on a machine.

Anthropic highlights a few practical benefits that map neatly to real development setups:

• Full local environment remains available, including the filesystem, MCP servers, tools, and project configuration.
• Session sync across surfaces, so messages can be sent from terminal, browser, or phone interchangeably.
• Automatic reconnection after interruptions like sleep or network drops.

Requirements to get started

Remote Control has a few prerequisites:

• Subscription: requires a Pro or Max plan; API keys are not supported.
• Authentication: run claude and use /login if not already signed in via claude.ai.
• Workspace trust: run claude in the project directory at least once to accept the workspace trust dialog.

Starting a Remote Control session

Remote Control can be launched either as a fresh session or by extending an existing one.

Start from the terminal

From a project directory, run:

claude remote-control

The process stays active in the terminal and shows a session URL. It also supports:

• --verbose for detailed logs
• --sandbox / --no-sandbox to enable or disable sandboxing (off by default)

A QR code can be displayed for phone access (spacebar toggles it).

Start from an active Claude Code session

Inside an existing session, use:

/remote-control (or /rc)

This carries over conversation history and displays a session URL and QR code. (Notably, the --verbose/sandbox flags aren’t available via this in-session command.) Anthropic also notes that using /rename beforehand can make the session easier to identify across devices.

Connecting from another device

Once running, a session can be reached in several ways:

• Open the session URL in a browser (lands in claude.ai/code)
• Scan the QR code to open in the Claude app
• Find the session by name in the session list (Remote Control sessions show a computer icon with a green status dot when online)

Connection and security model

Remote Control is designed to avoid hole-punching the local network: the local Claude Code session makes outbound HTTPS requests only and never opens inbound ports. When Remote Control starts, it registers with the Anthropic API and polls for work; once connected, the server routes messages between web/mobile and the local session over a streaming connection. Transport is over TLS, and the connection uses multiple short-lived credentials, each scoped and expiring independently.

Limitations worth noting

Remote Control comes with a few operational constraints:

• One remote session at a time per Claude Code session
• The terminal must stay open; closing it or stopping the claude process ends the session
• If the machine can’t reach the network for roughly 10 minutes, the session times out and the process exits

Original source