Anthropic's Claude Code Desktop can run an iOS app in a simulator pane beside the coding conversation, letting the developer watch and interact with the same simulated device the AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → is checking. The documentation describes installing and launching the app, tapping through it and reading its screen to inspect changes.
The pane controls the simulator directly and keeps a separate device for each session. That makes a proposed interface change visible during the coding workflow, rather than leaving the developer with code alone. This public beta requires a local macOS session with Xcode and its simulator tooling; the device being controlled is a simulated iPhone or iPad.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.