Claude Code Remote Control: A 3-Minute Overview and When to Actually Use It¶

Audience:

Developers using Claude Code (Max/Pro plan) in their daily workflow. Engineers evaluating remote operation options or wanting a quick overview before reading the official docs.

Key Points¶

30-Second Setup
claude rc → scan QR → continue on your phone. It just works
Clear Decision Framework
A decision flow for choosing between Remote Control and SSH+tmux+Tailscale
Gotchas Covered Upfront
Practical notes the official docs don't address

Just Do This: Quickest Setup¶

Skip the theory. If you're on a Max plan, these 3 steps are all you need.

Step 1 — Navigate to your project directory in the terminal and run the command.

claude rc

Step 2 — A QR code and URL appear in the terminal. Scan the QR code with the Claude app on your smartphone.

Step 3 — Your Claude Code session appears on your phone. Continue working right there.

That's it. Files, MCP servers, project settings — everything stays on your local machine. No code moves to the cloud.

Already in a session?

Type /rc in the Claude Code input field to switch your current conversation — including full history — to a remote session. Run /rename feature-A-refactor beforehand so you can identify it easily in the session list on your phone.

Setup is done. Next, let's understand what's happening behind the scenes in 30 seconds.

What's Actually Happening¶

┌─────────────────────────┐
│  Local Machine (Home/Office)  │
│  ┌───────────────────┐  │
│  │ Claude Code Running │  │
│  │ File System         │  │
│  │ MCP Servers         │  │
│  │ Project Settings    │  │
│  └───────┬───────────┘  │
│          │Session Sharing │
└──────────┼──────────────┘
           │(Encrypted)
    ┌──────┴──────┐
    ▼             ▼
 Browser      Mobile App
(claude.ai/code) (Claude App)

The key insight: this is "session sharing," not "remote desktop." It's not streaming your terminal screen. Instead, Claude Code's conversation and tool execution state are synchronized at the application level. That's why the mobile experience feels native rather than a clunky terminal-over-touch interface.

With this architecture in mind, the decision criteria for when to use Remote Control become clear.

When Should You Use Remote Control?¶

The official documentation focuses on feature descriptions and is thin on helping you decide whether this is right for your situation.¹ Here's a decision flow.

What's your situation?
│
├─ Want to monitor/operate a running Claude Code task away from your desk
│  └─→ ✅ Remote Control is ideal
│
├─ Need full access to your home dev environment from outside (beyond Claude Code)
│  └─→ SSH + Tailscale + tmux is more appropriate (higher versatility)
│
├─ Can't or don't want to keep your local machine always on
│  └─→ Running Claude Code on a VPS is a better fit
│
└─ Want shared sessions for a team
   └─→ Remote Control is currently for individual use. Team/Enterprise not supported

So what does "ideal" look like in practice?

Scenarios Where It Shines¶

Scenario 1: Monitoring a large refactoring job

You hand off a test suite run or codebase-wide rewrite to Claude Code and head to lunch. When an approval prompt arrives, you handle it on your phone. Previously, you'd SSH into a tmux session, but terminal UI on a touchscreen was mostly "look, don't touch." Remote Control provides a mobile-native UI where approving and giving instructions is actually practical.

Scenario 2: Background work during meetings

A Claude Code task you kicked off in the morning completes during a meeting. You can't open your laptop, but you check the results on the Claude app and send the next instruction. When you return to your desk, the conversation is synced to your terminal.

Scenario 3: Approving code reviews on the commute

Claude Code finishes preparing a pull request. From the train, you review the diff summary and approve it. Detailed code review happens later at your terminal — a natural division of labor.

That said, Remote Control isn't a silver bullet.

Where It Doesn't Work¶

Environments where the local machine shuts down — Remote Control connects to a local process. If the machine powers off, the session ends. Sleep/wake auto-reconnection is supported, but power-off is not
API key authentication — Remote Control requires Pro/Max plan account authentication. It doesn't work with API-based usage
Enterprise environments with strict security policies — Team/Enterprise plans aren't supported yet, making organizational security approval difficult

With these constraints in mind, here's how Remote Control stacks up against the alternatives.

Comparison at a Glance¶

Criteria	Remote Control	SSH + tmux + Tailscale	VPS	Third-party (Happy, etc.)
Setup time	30 seconds	30 min–1 hour	1–2 hours	10–30 min
Mobile UX	◎ Native UI	△ Raw terminal	△ Same	○ Tool-dependent
MCP/settings carry-over	◎ Complete	○ Env-dependent	△ Rebuild needed	△ Partial
Local machine required	✕ Required	✕ Required	◎ Not needed	✕ Required
Team/Enterprise support	✕ Not yet	◎ No restrictions	◎ No restrictions	◎ No restrictions
Versatility (beyond Claude Code)	✕ Claude Code only	◎ Anything	◎ Same	△ Tool-dependent

In one sentence: Remote Control is the optimal choice when you want the best mobile experience with minimal setup, scoped to Claude Code. For a general-purpose remote dev environment, SSH+Tailscale remains the solid pick.²

Things That'll Save You from Getting Stuck¶

For those about to read the official docs, here are the easy-to-miss details.

Know the difference between claude rc and /rc. claude rc is a terminal command that starts a new session and accepts flags like --sandbox. /rc is a slash command inside an existing session — it carries over conversation history but doesn't accept flag options. Use the former for new tasks, the latter for taking ongoing work mobile.

Name your session before going remote. Run /rename before /rc. Remote Control sessions default to your last message content or "Remote Control session," which makes them hard to identify when juggling multiple sessions.

Sleep recovery is automatic; power-off is not. Closing your laptop lid triggers sleep, and reconnection happens automatically on wake. But OS shutdown or crash kills the session. If you're stepping away for a while, check your power management settings.

Remote Control via claude.ai/code and Claude Code on the web are different things. The naming is confusing, but Remote Control connects to a process running on your local machine. Claude Code on the web uses cloud infrastructure. The file system location is fundamentally different.

Summary¶

Remote Control's core value is application-level session sharing for Claude Code. Compared to SSH-based terminal sharing, the gap is clear in mobile UX quality and seamless MCP/project settings carry-over.

That said, the requirement for a running local machine and the lack of Team/Enterprise support are constraints you can't ignore when making adoption decisions. If your workflow matches "I want to operate Claude Code tasks away from my desk," it's worth trying. If not, traditional methods work fine.

Looking ahead, if Team/Enterprise support and cloud-hosted session integration materialize, Remote Control could become the de facto way to operate Claude Code from mobile. Getting your workflow ready now, while it's still optional, gives you a head start.

Run claude rc once — you'll know in 30 seconds whether it's for you.