Anthropic launches Remote Control to manage Claude Code from your phone

Anthropic brings mobile access to Claude Code with Remote Control

Anthropic has launched Remote Control, a new feature that transforms how developers interact with Claude Code by enabling them to manage coding tasks remotely from smartphones and tablets 1

1

. The mobile version of Claude Code addresses a key limitation that has restricted the AI coding assistant to desktop applications, terminal command-line interfaces, and integrated development environments since its launch 1

1

. Claude Code Product Manager Noah Zweben announced the feature as a synchronization layer that bridges local CLI environments with the Claude mobile app and web interface, allowing developers to initiate complex tasks in their terminal and maintain full control from a phone or tablet 1

1

.

How developers can control from mobile phone

Remote Control works by generating a one-time link or QR code from an active local Claude Code session 3

3

. To access the feature, users must update to Claude version 2.1.52 and execute the claude remote-control command or use the in-session /rc slash command 1

1

. Once active, the terminal displays a QR code that opens a responsive, synchronized session in the Claude mobile app 1

1

. When you open that link on another device, you gain live remote access and control to the same session, including its context, files, and command history 3

3

.

Security architecture keeps code local

Anthropic emphasizes that Remote Control does not move code or environments to the cloud 3

3

. The session continues locally on the user's machine rather than in the cloud, with Remote Control providing access from anywhere 2

2

. Your desktop machine initiates an outbound connection to Anthropic's API, meaning you aren't opening any inbound ports or exposing your computer to the open web 1

1

. Your files and MCP servers never leave your machine; only chat messages and tool results flow through the encrypted bridge 1

1

. Security is handled through temporary access tokens, and once the session ends, the remote link becomes invalid 3

3

.

Available for Claude Max subscribers first

Remote Control is currently available as a research preview for Claude Max subscribers, who pay between $100-$200 USD monthly 1

1

. Anthropic posted on X that Remote Control will also make its way to Claude Pro subscribers at $20 USD monthly in the future 1

1

. The feature remains notably absent from Team or Enterprise plans during this initial phase 1

1

. API keys are not supported either 2

2

.

Replacing community workarounds with official solution

Before this official release, developers relied on a patchwork of third-party tools like Tailscale for secure tunneling, Termius or Termux for mobile SSH access, and Tmux for session persistence 1

1

. Some developers even built complex custom WebSocket bridges just to get a responsive mobile UI for their local Claude sessions 1

1

. Remote Control replaces these unofficial solutions with a native streaming connection that requires no port forwarding or complex VPN configurations 1

1

. Claude Code was installed 29 million times in VS Code alone, and Anthropic hopes Remote Control will broaden its reach even further 2

2

.

Current limitations and what to watch

Being in research preview, Remote Control shows signs of being an early tool with three key limitations: users can only run one remote session at a time, the terminal must stay open, and if a machine cannot reach a network for more than roughly 10 minutes, the session will time out 2

2

. The messaging behind the release centers on preserving a developer's flow state, with Zweben framing the update as a lifestyle upgrade that encourages users to take a walk or see the sun without losing their flow 1

1

. This makes it useful for long builds, experiments, or deployments that don't need constant supervision 3

3

. Watch for the rollout to Claude Pro users and potential expansion to Team and Enterprise plans as Anthropic refines the feature based on early user feedback.