Two Macs, One Workflow

You have a MacBook Pro you carry out and shut down, and a Mac Mini that stays on at home. Here is how to make the Mini your home base, reach it from the laptop anywhere, and route Claude Code through it so the work keeps going even when the lid is closed.

1. It is two problems, not one

The mess feels like one thing, but it is two, and each has a different fix. Solve only the first and your files stay scattered. Solve both and the Mini becomes the place everything lives and runs.

Problem A · Access

Where the work runs. You want the heavy work to happen on the Mac Mini but to drive it from the MacBook Pro, even when you are out of the house.

Fix: a remote-access layer. Tailscale plus SSH and the Claude Desktop app.

Problem B · Storage

Where the files live. Some files are on the laptop, some on the Mini, and you are never sure which copy is current.

Fix: a single source of truth. Git for code, one synced store for everything else. Not remote desktop.

2. What Claude Code can do across machines now

The Desktop app can reach another machine in three ways. Each one has a different job, and the first is the best fit for daily use.

Desktop app, SSH to the Mini

Daily driver. Point a Claude Code Desktop session on the MacBook Pro straight at the Mac Mini. Files, tools, and the agent all live on the Mini. The laptop is just the window. SSH for Mac desktop sessions shipped in April 2026.

Use for: normal coding and agent work routed through the Mini.

Remote Control

Phone / away. Bridges a session already running on a machine to claude.ai/code and the iOS and Android apps. Outbound HTTPS only, no inbound ports. The catch: the terminal must stay open and it is one session per machine at a time.

Use for: checking in or steering from your phone while away.

Teleport & Dispatch

Handy extras. Start a task on the web or iOS app, then pull it into a terminal later with claude --teleport. Dispatch lets you fire a task from your phone and open the Desktop session it creates.

Use for: kicking off long jobs before you get to a keyboard.

3. The setup, in three layers

Same idea as a good project structure. One layer for the network, one for where things live, one for how you reach in.

The Mini is home. The laptop is a client.

Build it in this order and each layer stands on the one below it.

1. Network: Tailscale. Free mesh VPN on both Macs. Gives the Mini a stable name like macmini that works at home or from a cafe, with no port forwarding and nothing exposed to the public internet.
2. Home base: the Mini. Code lives in git, pushed to GitHub. Documents live in one synced store such as iCloud Drive. The laptop clones or syncs what it needs and never holds the only copy.
3. Access: pick the door. SSH or the Desktop app for terminal and agent work. Jump Desktop when you need the full screen. Remote Control or the web app from your phone.

  Anywhere                         Home (always on)
  ┌──────────────────┐             ┌──────────────────┐
  │  MacBook Pro     │   Tailscale  │   Mac Mini       │
  │  carry / shut    │ ───────────▶│   files + git    │
  │  down anytime    │   SSH /     │   Claude Code     │
  │  (thin client)   │   Desktop   │   runs here       │
  └──────────────────┘   app       └──────────────────┘
          ▲                                 ▲
          │  Remote Control / web / iOS      │
          └────────────── phone ───────────────┘

4. Which door for which task

All of these use Tailscale. Pick the lightest option that does the job, and save full screen sharing for work that needs the GUI.

5. The setup checklist

The full walk-through happens live when you are at both machines. This is the map so you know what is coming.

1. Install Tailscale on both Macs. Sign in with the same account on the Mini and the MacBook Pro. The Mini shows up under a name you can reach from anywhere. Turn on its always-on / login-item setting so it reconnects after a reboot.

2. Turn on Remote Login on the Mini. System Settings, General, Sharing, Remote Login. That enables SSH. Then from the laptop, confirm it works:

   ssh your-user@macmini   # the Tailscale name of the Mini

3. Make the Mini the canonical home for files. Put project and code work in git and push to GitHub. Pick one synced store for documents so “where is this file” has a single answer. Stop keeping the only copy on the laptop you carry out.

4. Install Claude Code on the Mini and connect from the laptop. Install the CLI on the Mini so the agent runs where the files are. Then open the Desktop app on the MacBook Pro and start an SSH session pointed at the Mini. Same CLAUDE.md, settings, and MCP servers as your local setup.

5. Add the phone path. Start a session on the Mini, attach Remote Control, and confirm you can pick it up from claude.ai/code or the iOS app. Now you can close the laptop and keep steering.

6. Trade-offs worth knowing

The honest edges, so nothing surprises you later.

Remote Control runs one session at a time

And the terminal must stay open. For several parallel agents on the Mini, drive them over SSH or Desktop sessions instead.

The file cleanup is the real work

It is the unglamorous step, but it is the one that fixes “all over the place.” Access is easy once there is a single source of truth.

Jump Desktop is the heaviest door

Full screen sharing uses the most bandwidth. Prefer SSH or the Desktop app when you only need a terminal or a diff.

Keep the Mini awake

Set it to never sleep on power, and enable wake on network. An always-on home base only works if it stays on.