Limrun

Teams building cloud-based coding agents (for example AI agents that write and fix mobile app code) need a way for those agents to actually build, run and validate mobile apps, which normally requires native Xcode or Android tooling that isn't available in a typical cloud sandbox. Without this, an agent can write code but cannot verify it builds, install it on a simulator, or show a human the result without someone manually downloading and running it. The page shows this concretely: an agent fixes a missing Combine import, triggers a remote xcodebuild, installs to a simulator, then drives the app through a game flow to confirm the fix. Reviewers would otherwise need to install the app themselves to check an agent's work, such as a pull request claiming a new scoring mechanism was validated.

Limrun provides native remote services, Xcode builds, iOS and Android simulators, and live preview, that a cloud agent can call on demand from its sandbox. Agents sync code and trigger remote builds via a CLI (shown as 'lim xcode build'), stream build logs in real time, and use an S3-backed cache for warm builds, with idle/build time pricing. For validation, agents can drive simulators remotely using a built-in accessibility tree, Appium, Maestro, or Detox, record 60fps video, stream logs from the running app, and batch actions for precise timing. To share results, the product generates a browser preview link, a demo video the agent can attach to a pull request, and WebUSB-based install to a physical device for real-device testing, all without the reviewer installing anything locally.