Remote Control is Anthropic's own way to steer a Claude Code session on your machine from claude.ai/code or the Claude app; use it when Claude Code on a Pro, Max, Team or Enterprise plan is all you run. Reemoat is separate, open-source, self-hosted software that runs Claude Code, Codex, OpenCode, Kimi Code, Grok Build and Cursor on machines you own, supervised from one app.
Where Remote Control is the better choice
- It notifies you
"When Remote Control is active, Claude can send push notifications to your phone." Reemoat sends nothing: no push, email or bot.
- Any browser will do
A session opens "in any browser" at claude.ai/code. Reemoat has no browser client.
- It can sandbox
"To sandbox the sessions the server starts, turn on sandboxing in a settings file." Reemoat cannot: the agent runs as your OS user.
Where Reemoat fits better
- Agents from several vendors
All six in one app. Remote Control covers Claude Code.
- Claude Code on an API key
Anthropic says "API keys are not supported". Reemoat runs Claude Code on its own sign-in or a saved API key.
- Where the transcript lives
With Remote Control it "is stored on Anthropic servers". Reemoat's daemon keeps it on your machine, and the relay carries ciphertext.
Row by row
| Question | Remote Control | Reemoat |
|---|---|---|
| What it is | A Claude Code feature: "a window into that local session" | A daemon, a control plane and relay, and an app |
| Agents | Claude Code | Claude Code, Codex, OpenCode, Kimi Code, Grok Build and Cursor |
| Paying for the agent | "Pro, Max, Team, and Enterprise plans. API keys are not supported." | Each agent's own sign-in, a saved API key, or a provider such as OpenRouter |
| Cloud providers and gateways | Not with "Amazon Bedrock, Google Cloud's Agent Platform, or Microsoft Foundry", ANTHROPIC_BASE_URL at "a host other than api.anthropic.com", or sign-in "through an enterprise Claude apps gateway" | Not tested; no claim |
| Organisations | Team and Enterprise: off until an Owner enables it. Zero Data Retention or HIPAA organisations "can't enable Remote Control" | None, and no access log; a machine's owner shares it by grant |
| Where the agent runs | Your machine; "the claude process has to keep running" | Your Linux or macOS machine |
| Headless Linux server | Anthropic says to "start it inside tmux or screen", and after an extended network outage in server mode "Claude Code gives up after roughly 10 minutes and the claude remote-control process exits"; an interactive session keeps retrying | The daemon is a systemd user service that keeps redialling the relay. It needs lingering switched on to outlive your logout: the installer does that where it can, and otherwise prints the sudo loginctl enable-linger command |
| What you hold | A browser at claude.ai/code, or the Claude apps | A native desktop or mobile app; no browser client |
| Install on the host | Nothing beyond claude; the setup is "Run claude remote-control or /remote-control" | The daemon, from a shell installer; no Docker |
| Several sessions | Server mode: --spawn same-dir by default, --spawn worktree for a worktree each; --capacity: "Default is 32" | A git worktree each; none outside a git repository |
| Several machines | Messages pass "between your own sessions on different machines" | Every machine in one app, with messages between sessions |
| Who can read the session | Anthropic stores the transcript | The daemon keeps it; the relay "moves bytes it holds no key for" |
| What both share | Claude Code's data to the model "includes all user prompts and model outputs" (Data usage) | The same; Reemoat runs the vendor's own CLI |
| Extra party to trust | None beyond Anthropic: "All traffic travels through the Anthropic API over TLS" | The control plane's operator, who can act as you, unless you run it: the control plane and the relay are in the repository |
| Notifications | Mobile push | None |
| Sandbox | Optional, for server-mode sessions | None; the agent runs as your OS user |
| Device control | Trusted Devices, off by default: an enrolled device and a sign-in "no more than 18 hours old", renewed by a "biometric step-up" | Tokens bound to a device key; a device can be retired; no biometric check |
| Maturity | First-party; only Trusted Devices is marked beta | Pre-1.0, one maintainer |
| Licence | Claude Code: "© Anthropic PBC. All rights reserved." (LICENSE.md) | AGPL-3.0-only |
Checked on 12 October 2026 against Anthropic's Remote Control documentation and Reemoat v0.13.0. Written by Reemoat's author; where Remote Control is the better choice, this page says so.
Pick
| Pick | If |
|---|---|
| Remote Control | Claude Code is your only agent, your plan includes it, and you want push, a browser or a sandbox |
| Reemoat | You run several vendors' agents, pay for Claude by API key, or want the transcript on your machine |
Both on one machine is untested here. Remote Control steers sessions
claude registers with Anthropic; Reemoat steers the ones its daemon starts,
and takes Claude Code's own remote-control tool away from those.
If Remote Control tells you "Remote Control requires a claude.ai subscription"
Anthropic's fix: "Run claude auth login and choose the claude.ai option", and
remove an API key variable the message names
(their entry).
If an API key is how you pay for Claude, a separate tool is the remaining route.
If Remote Control tells you "Remote Control is only available when using Claude via api.anthropic.com"
Anthropic's fix: "If you have an eligible claude.ai login, unset the named variable" and restart (their entry). If a cloud provider or gateway is how you pay for Claude, a separate tool is the remaining route; Reemoat is untested there and makes no claim.
If Remote Control tells you "Remote Control is disabled by your organization's policy"
Anthropic's entry
lists four causes. A stale Team or Enterprise sign-in on a Pro or Max plan is
cleared with claude auth logout, then claude auth login. If your
organisation turned it off, that is policy and no tool changes it.
When this is the wrong tool
- You need to be told when a run finishes or blocks. Reemoat notifies nobody.
- You need the agent isolated from your machine. There is no sandbox.
- You want to work from a browser. There is no browser client.
- Claude Code on an eligible plan is all you run. Remote Control has fewer parts.
- You will not run a control plane, nor trust whoever does.