Claude Code now runs an iOS simulator right next to your chat
Anthropic just added an interactive iOS simulator directly into Claude Code Desktop on macOS in public beta. Instead of juggling windows, the simulator pops up right next to your conversation so Claude can build the app, tap through screens, and verify its own fixes live. It's available on Pro, Max, and Team plans with Xcode installed.
Anyone who has built an iOS app knows the constant window-juggling tax: tweak a line of Swift, command-tab into Xcode, wait for the simulator to boot, and manually tap through three screens just to see if your button aligned. Rinse and repeat a hundred times a day.
Anthropic wants to kill that context switching. On July 21, the company launched an interactive iOS Simulator pane right inside Claude Code Desktop for macOS in public beta.
Claude does the clicking for you
The workflow is refreshingly seamless. Open your iOS project in the Code tab, ask Claude to run or verify the app, and an interactive simulator pops up directly beside your conversation. You do not need to hunt through settings to open it.
Once it's up, Claude does not just dump code on you. It builds the app, installs it, taps through buttons, and reads the screen using screenshots to verify its own UI changes. When it is actively driving the device, a badge warns you that Claude is using it so you do not fight over the cursor.
Between tasks, you can take over completely. You can tap, swipe, and use standard macOS simulator shortcuts like Cmd+Shift+H for Home or Cmd+L to lock. You can even grab screenshots and screen recordings directly to your desktop, or dial down the frame rate and resolution in the display menu if your Mac is breaking a sweat.
Each session gets its own isolated devices, handling up to 4 simulators at once, and Claude automatically shuts them down 10 minutes after detachment to save your battery.
Native integration over blunt screen control
What makes this technical implementation clever is how it avoids Anthropic's general computer-use feature.
According to MacRumors, the simulator pane drives Apple's tooling directly under the hood. Because it talks straight to the simulator rather than pretending to be a human sitting at your desktop, it completely skips the intrusive macOS Accessibility and Screen Recording permissions usually required for AI screen control.
To run it, you will need macOS with Claude Desktop v1.24012.0 or later, plus Xcode 26.x or Xcode 27 with iOS platform files installed. If you keep multiple Xcode versions around, Claude respects whatever path xcode-select points to.
It is available across Pro, Max, Ultra, Team, and Enterprise plans, excluding HIPAA-configured enterprise environments. Anthropic does note that device screenshots are sent back and retained under standard policies, so you should definitely avoid logging into real personal accounts inside the simulator.
Why it matters
We think this is one of the smartest product moves Anthropic has made for mobile developers.
Most coding assistants today act like backseat drivers: they give you directions, but you still have to step on the pedals and verify every turn. By closing the loop between code generation and visual verification inside a single pane, Claude shifts from a simple text assistant into an active QA teammate.
Choosing a direct harness into Apple's developer tools instead of brute-force pixel matching is also a huge win. It gives developers a faster, more reliable experience without asking them to surrender sensitive root-level macOS screen recording privileges.
That said, running Claude, local dev servers, and Apple's heavy simulator side by side will easily test the thermal limits and memory of base-model Macs. And because those screen captures ping Anthropic's servers, teams handling proprietary apps will need to think twice about what data gets rendered on screen.
Sources (2)
- code.claude.com(opens in a new tab)Originalcode.claude.com
- macrumors.com(opens in a new tab)Reportmacrumors.com