On Oct 29 Amp shipped macOS Satellites: threads keep running on a Linux orb, and when work needs macOS Amp boots a fresh cloud Mac with Xcode and simulators, uses it only to build, run and test, and returns screenshots, logs and screen recordings to the thread. mac.medium costs $2/hour and mac.large $4/hour; the beta is waitlisted. CEO Quinn Slack says agents need a Mac only about 35% of the time even for macOS-heavy work.
Key Takeaways
- ✓Amp says the cloud Mac is ready in about 30 seconds with Xcode, iOS simulators and Homebrew preinstalled
- ✓Two sizes: mac.medium (6 CPUs, 14 GB) at $2/hour and mac.large (12 CPUs, 28 GB) at $4/hour, drawn first from plan orb usage, then Amp credits
- ✓A Mac stays up 20 minutes after use so follow-ups reuse its build cache, then it is deleted and nothing is saved
- ✓CEO Quinn Slack posted on X that even macOS-heavy agent work needs a Mac only ~35% of the time (company figure)
- ✓Beta only: join the waitlist under Settings > Experimental Features
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Background and pain point
Most cloud coding agents run in Linux containers, so they stall on iOS and macOS apps: Xcode and the simulators only run on a Mac, and developers end up pulling that work back to their own machines. Amp already offered orbs, remote machines that run agents in parallel; this release fills the Apple-platform gap. Amp's post says it first tried "Morbs", which ran the whole thread on a Mac, but Docker and backends took minutes to start there and keeping a Mac disk per thread was too expensive.
Architecture and mechanism
The new design splits a thread across two machines. The thread, dev server and toolchain stay on a fast-starting Linux orb, and the Mac is a second machine attached to it, hence "Satellites". The agent decides when it needs a Mac, and Amp's server boots a fresh cloud Mac in about 30 seconds. Before each command on the Mac, the orb syncs changed files, while Git-ignored files such as .env stay on the orb; apps on the Mac reach the orb's backend through portals. A Mac tab in the thread streams the screen so users can watch or take control.
Benchmarks and comparisons
Amp has not published benchmarks such as build speed or success rates. The confirmed figures come from the official page: about 30 seconds to boot, 20 minutes of idle reuse, mac.medium (6 CPUs, 14 GB) at $2/hour and mac.large (12 CPUs, 28 GB) at $4/hour. CEO Quinn Slack posted on X that even macOS-heavy agent work needs a Mac only about 35% of the time; that is the team's own observation.
What developers can do now
The feature is in beta, so join the waitlist under Settings > Experimental Features. Once admitted, just ask Amp to work on your iOS or Mac app; you don't need to request a Mac. Every Mac starts as mac.medium, and you can ask Amp for mac.large. Archive the thread or ask Amp to stop the Mac to release it immediately. Mac time comes out of plan orb usage at a higher rate, then Amp credits, so estimate costs before long jobs.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.