Claude Code Remote Control: Run Sessions From Any Device

Claude Code remote control is a research-preview feature that connects claude.ai/code or the Claude mobile app to a Claude Code session running on your own machine, so you can start a task at your desk and keep steering it from your phone, tablet, or any browser. Claude keeps executing locally the entire time; the web and mobile interfaces are just a window into that local session, so your filesystem, MCP servers, and project configuration all stay available. This guide covers setup, the three ways to start a session, how it differs from teleport and Claude Code on the web, the security model, and the limitations, verified against Anthropic's official documentation as of July 2026.

What Claude Code Remote Control Actually Is

Remote Control is session hand-off, not remote desktop: the Claude Code process, your files, and every tool execution stay on your machine, while claude.ai/code and the Claude mobile app become synchronized remote interfaces to that one session. Anthropic shipped it in research preview in February 2026, and Simon Willison's write-up from launch week captures why it landed: long agentic runs no longer chain you to the terminal that started them.

Because the session is local, everything your terminal session can do, the remote view can do: your MCP servers keep working, typing @ on your phone autocompletes file paths from your local project, and the conversation stays in sync across all connected devices, so you can send messages from terminal, browser, and phone interchangeably. If your laptop sleeps or the network drops, the session reconnects automatically when the machine comes back online. Searches for "Claude remote control" mean this same feature; there is no separate product.

Requirements

Remote Control needs four things, per the official Remote Control documentation:

• Claude Code v2.1.51 or later: check with claude --version.
• A claude.ai subscription: Claude Pro, Max, Team, or Enterprise. API-key authentication is not supported, and on Team and Enterprise plans an Owner must first enable the Remote Control toggle in Claude Code admin settings.
• Direct Anthropic API access: it does not work on Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, or when ANTHROPIC_BASE_URL points at a gateway or proxy.
• Workspace trust: run claude in the project directory at least once to accept the trust dialog.

How to Start a Remote Control Session

There are three CLI invocation modes plus a VS Code command, each suited to a different workflow:

# Server mode: waits for remote connections, supports many sessions
claude remote-control

# Interactive session with remote access enabled (alias: --rc)
claude --remote-control "My Project"

# Already mid-session? Hand it off without losing history
/remote-control

Server mode is the workhorse: the process sits in your terminal serving up to 32 concurrent sessions (tune with --capacity), and --spawn worktree gives each on-demand session its own git worktree so parallel sessions cannot clobber each other's edits. Press spacebar to show a QR code for your phone. The interactive flavors are simpler: claude --remote-control starts a normal terminal session that is also reachable remotely, and /remote-control (or /rc) converts the session you are already in, carrying over the conversation history. In VS Code, type /remote-control in the prompt box (requires v2.1.79+). To make every interactive session remote-capable automatically, run /config and enable "Enable Remote Control for all sessions." The full command list lives in our Claude Code commands cheat sheet.

Connecting From Your Phone or Browser

Once a session is active, connecting takes seconds: open the printed session URL in any browser, scan the QR code with your phone to jump straight into the Claude app, or open claude.ai/code (or tap Code in the mobile app) and pick the session from the list. Remote Control sessions show a computer icon with a green status dot when online. Sessions are named from the --name flag, a /rename, your conversation, or an auto-generated name like myhost-graceful-unicorn; renaming from the phone updates the local title too. If you do not have the app yet, the /mobile command displays a download QR code for iOS or Android.

How It Works Under the Hood

The security model is outbound-only: your local Claude Code session makes outbound HTTPS requests and never opens inbound ports on your machine. On start, it registers with the Anthropic API and polls for work; when you connect from another device, Anthropic's servers route messages between the web or mobile client and your local session over a streaming connection. All traffic travels over TLS through the Anthropic API, the same transport as any Claude Code session, and the connection uses multiple short-lived credentials, each scoped to a single purpose and expiring independently.

Your code never leaves your machine; what flows through the bridge is the conversation, tool activity, and results. Permission prompts appear on the remote surface for approval, and the available remote modes are Ask permissions, Auto accept edits, and Plan mode (bypass permissions is deliberately unavailable remotely). For Team and Enterprise organizations, the beta Trusted Devices setting adds a device-level gate: members must enroll each browser or phone and re-verify with Face ID, Touch ID, Windows Hello, or a passkey once their sign-in ages past 18 hours. If you want machine-level remote access instead of session-level, a VPN approach like the one in our Tailscale remote access guide solves a different, complementary problem.

Remote Control vs Teleport vs Claude Code on the Web

Anthropic now ships several ways to work away from your terminal, and the differences come down to where Claude runs and what triggers the work:

The rule of thumb: Remote Control pushes an existing local session out to your other devices, teleport pulls a cloud session down to your terminal, and Claude Code on the web skips your machine entirely. Dispatch and channels trigger work rather than mirror it; we cover them in the Claude Dispatch guide and the Claude Code channels guide.

Mobile Push Notifications

With Remote Control active, Claude can send push notifications to your phone: typically when a long-running task finishes or when it needs a decision to continue, and you can request one explicitly ("notify me when the tests finish"). Setup requires Claude Code v2.1.110+, the Claude mobile app signed into the same account, and enabling "Push when Claude decides" and/or "Push when actions required" in /config. Claude skips pushes while you are typing in the connected terminal, so you are not notified about things happening right in front of you. This combination (long task, plan approved, phone in pocket) is exactly why pairing Remote Control with plan mode works so well: approve the plan at your desk, walk away, and answer the occasional permission prompt from your phone.

Limitations

Remote Control's constraint is inherent to its design: the local process must keep running, so closing the terminal, quitting VS Code, or shutting the laptop ends the session. If your machine is awake but offline for more than roughly 10 minutes, the session times out and the process exits. Other limits as of July 2026: each interactive process supports one remote session at a time (server mode is the exception), starting an ultraplan session disconnects Remote Control because both occupy the claude.ai/code interface, and commands that open interactive terminal pickers (like /plugin and /resume) remain local-only, though /mcp, /config, /compact, /usage, and other text-output commands work remotely. It is also unavailable on Bedrock, Vertex-based Agent Platform, Foundry, and behind LLM gateways. If you need an always-on agent that survives your laptop closing, that is a job for a VPS-hosted agent rather than Remote Control; our remote vs self-hosting comparison walks through that decision.

Related Guides

• Claude Code Plan Mode: How It Works and When to Use It
• Claude Code Channels: Pipe Chat Apps Into Your Sessions
• Claude Dispatch: Delegate Tasks From Your Phone
• OpenClaw Tailscale Remote Access Guide