Claude Code Beyond the Terminal

Start with a travel problem. A long migration is half done in your terminal. You're leaving for the airport, and you want it to keep going with your laptop shut in your bag.

Same engine, different seats

Every surface runs the same Claude Code engine and reads the same CLAUDE.md, settings and MCP configuration. The desktop app is the Code tab in the Claude app for macOS, Windows and (in beta) Linux: several sessions side by side with git isolation, visual diff review with comments, app previews and pull request monitoring. Run /desktop in a terminal session to continue it there (macOS and x64 Windows, with a claude.ai subscription). The VS Code extension adds inline diffs, @-mentions and plan review in your editor, and also installs in Cursor. The JetBrains plugin runs the CLI inside the IDE's terminal.

Cloud sessions

A cloud session runs on infrastructure Anthropic manages (or your organization's self-hosted environment), clones your repository from GitHub, and keeps running after you close your laptop. Start one at claude.ai/code, from the Code tab in the Claude mobile app, by choosing Cloud in the desktop app, or from your terminal with claude --cloud "Fix the flaky test in auth.spec.ts". Pull one back into your terminal with claude --teleport. Cloud sessions count toward your plan's usage, with no separate charge for the VM, and are available on Pro, Max, Ultra and Team, and on Enterprise with a premium or Chat + Claude Code seat.

The isolation is part of the appeal. Each session gets its own VM with limited network access by default. Your GitHub credentials never enter the VM: it holds a short-lived credential, and GitHub traffic goes through an Anthropic proxy that attaches the real one and refuses branch deletions. Your repository's branch protection rules still decide which branches a session can push to.

Remote Control: your machine, from anywhere

Remote Control is the opposite trade. The session runs on your machine, with your files, tools and MCP servers, and claude.ai/code or the Claude app becomes a window into it. Start it with claude remote-control (alias claude rc), which waits for connections and shows a QR code when you press the spacebar, or claude --remote-control for an interactive session you can also reach remotely, or /remote-control inside a session that's already running. It needs a claude.ai subscription (API keys aren't supported), an admin's go-ahead on Team and Enterprise, and a machine that stays on.

Your phone and Slack

Claude Code has no separate mobile app. In the Claude app for iOS and Android, the Code tab reaches your cloud sessions, your projects and any Remote Control session; on Pro, Max and Ultra you can also message Dispatch a task for the desktop app to run. Run /mobile for a QR code to the app.

In Slack, the answer depends on your plan. Claude Tag, in public beta for Team and Enterprise, puts @Claude in your channels under your organization's identity: anyone in a channel can tag it with a bug report and get a draft pull request back. It works in a temporary sandbox, reaches the services an admin has connected for that channel, and bills channel work to an organization usage balance with a spend limit. On Pro, Max and Ultra, the earlier Claude Code in Slack app answers @Claude mentions in workspaces that aren't connected to Claude Tag, running each task as a cloud session under your own account.

Work on a schedule

Create cloud scheduled tasks at claude.ai/code/routines or by running /schedule and describing what you want; Claude walks you through it. They need a claude.ai login. Two habits from the docs before you leave one running unattended: protect main with a GitHub branch rule the task can't bypass, and watch its first few runs before you trust it. For quick polling inside a session, /loop 5m check whether the deploy finished runs on a fixed interval, and leaving out the interval lets Claude pick one between a minute and an hour.