The iOS Simulator Inside Cursor, Claude Code, and Codex
Article Summary
Antoine van der Lee, creator of RocketSim, is rethinking the iOS Simulator's role as AI agents take over coding. His latest move? Embedding the Simulator directly into Cursor, Claude Code, and Codex.
As AI agents handle more coding tasks, developers are shifting from writing code to orchestrating and reviewing agent output. Van der Lee, who's been building iOS apps since 2009, argues the Simulator remains critical for visual validation even as agents write the code. He's now integrating the Simulator into AI IDEs to keep developers in flow.
Key Takeaways
- Developer time shifted from coding to orchestrating multiple AI agents across projects
- WWDC 2026's Device Hub enables streaming single Simulator to multiple environments
- RocketSim 17 streams Simulator to localhost browser inside AI IDEs
- Agents navigate apps and validate changes using RocketSim's CLI and Agent Skill
- Visual feedback loop lets developers annotate UI and communicate directly to agents
The iOS Simulator is evolving from a testing tool into an integrated feedback mechanism that keeps developers inside their AI IDE while agents build and validate apps.
About This Article
AI agents write code and build features well, but they tend to over-engineer solutions. This means developers have to review everything carefully before it ships, which slows down the whole workflow.
Van der Lee built RocketTrace as a simpler alternative to Xcode Instruments. It lets agents check how well an app performs without all the complexity, while developers manage multiple agents across different projects.
RocketSim and RocketTrace will work together as a connected set of tools. Developers can focus on coordinating the agents and making high-level decisions instead of spending time debugging performance issues manually.