Remote Control never opens an inbound port on your machine, and your code never leaves it. But the session transcript is stored on Anthropic servers while the connection is live. Both halves of that sentence matter.
Network posture
Your local Claude Code session makes outbound HTTPS requests only and never opens inbound ports. When you start Remote Control, it registers with the Anthropic API and polls for work. When you connect from another device, the server routes messages between the web or mobile client and your local session over a streaming connection.
Practically: nothing on your machine is exposed to the open internet, and you do not need to touch your firewall or router. You do need outbound access to the Anthropic API on port 443.
Transport security
All traffic travels through the Anthropic API over TLS, the same transport security as any Claude Code session. The connection uses multiple short-lived credentials, each scoped to a single purpose and expiring independently.
What stays local and what does not
| Where it lives | |
|---|---|
| Your source files | Your machine only |
| Code execution | Your machine only |
| MCP servers and local tools | Your machine only |
| Session transcript: your messages, Claude’s responses, tool activity | Stored on Anthropic servers while connected |
This is the detail most summaries get wrong. Execution and filesystem access stay on your machine, but the transcript is stored server-side. That is what keeps the conversation in sync across your devices and lets the session reconnect after a network drop. Stored transcripts are retained under Anthropic’s Data usage policy.
If you have seen Remote Control described as “end-to-end encrypted” or “Anthropic never sees anything”, that is not what the documentation says. The accurate framing is: your code stays on your machine; your conversation passes through and is stored on Anthropic servers.
Turning it off
Use the disableRemoteControl setting to switch Remote Control off entirely on a machine. Organizations with compliance requirements such as Zero Data Retention cannot enable Remote Control at all. The admin toggle appears greyed out.
Trusted Devices
Trusted Devices is an organization-wide beta setting on Team and Enterprise plans, off by default. When on, members must verify their device before they can view or steer Remote Control sessions from claude.ai, the mobile apps, or Desktop. It ties access to a known device and a recent authentication rather than just a signed-in account.
Two conditions apply:
- An enrolled device. Each browser, phone, or desktop app enrolls its own credential. Enrollment is only offered shortly after a full sign-in, so a device joins the trusted list as part of a real authentication rather than silently.
- A recent sign-in. No more than 18 hours old. Instead of signing in daily, members confirm presence with Face ID, Touch ID, Windows Hello, or a passkey.
Biometric checks run on the device through the operating system or browser, the same mechanism as passkey sign-in. Anthropic never receives fingerprints, face data, or any other biometric information. Only the device public key and basic metadata such as display name, platform, and enrollment time.
The setting applies only to Remote Control. Regular Claude chat, Claude Code in the terminal, and API usage are unaffected. Members review and revoke their own devices at claude.ai/settings/account under Trusted devices; unused credentials expire on their own.
Sessions already running when an admin enables the toggle are not retroactively protected and continue without the device requirement until they end.
Sandboxing
Server mode accepts --sandbox and --no-sandbox to enable or disable filesystem and network isolation. Sandboxing is off by default. These flags are not available with the /remote-control slash command.