How it works

From a description to a native app: the AI team, the build, the cloud simulator, V2 Cloud and publishing.

OpenV2 brings an AI team, native builds, a cloud simulator, an optional backend and publishing into one workspace.

The AI team writes native code

You describe the app. The agent creates and edits your project files in Swift and SwiftUI (iOS) or Kotlin (Android), generates the images and the app icon, and runs a build check. In Team mode six specialists share the work; in Solo mode Leo, the engineer, does it all. See Solo and Team.

It builds and runs in the cloud

You do not install Xcode or Android Studio. Builds and the simulator run on cloud machines. The simulator shows an iPhone, iPad or foldable device in your browser. You turn it on when you want to test, and it turns itself off after 3 minutes without interaction. See Live preview.

You steer in chat, by voice or in the code

  • Chat for features and changes, with Power to make the agent think harder when the task is hard.
  • Voice to talk through ideas while the agent builds.
  • Code editor on paid plans to edit files yourself. See Code editor.
  • Version history to restore an earlier state. See Version history.

V2 Cloud is the backend

When your app needs accounts, shared data, file uploads, server logic or in-app AI, ask in chat for a V2 Cloud backend. Apps that keep data on the device do not need one. See V2 Cloud overview.

Test on a phone and publish

Companion installs a test build on a USB-connected phone. The Publish button starts the App Store or Google Play flow.

Usage, not credits

Plans include usage that resets every 5 hours and every week. Building, the simulator and generated assets draw from it; plain conversation does not. See Plans and usage.

Your project

Your generated code belongs to you. OpenV2 keeps your account and project data; your app's backend data lives in V2 Cloud, separately. Paid plans can export a ZIP.

Next steps