How it works

Understand how OpenV2 turns conversations into running mobile apps.

OpenV2 combines an AI code generator, an in-browser preview runtime, and optional cloud services into one workflow.

The big picture

When you send a message in chat, OpenV2:

Understands your request using a large language model
Plans file changes (screens, components, theme, backend schema)
Writes real React Native / Expo source code
Applies changes to your project workspace
Bundles the app with browser-metro and shows it in the preview iframe

No local Node.js install is required for preview — everything runs in your browser.

Core components

Chat builder

The chat is your primary interface. You describe features, and the AI emits structured actions: create files, edit code, update PocketBase schemas, and more.

See Chat builder for prompting tips.

Preview runtime

OpenV2 uses browser-metro with a virtual file system (VFS) to bundle Expo apps entirely in the browser. The preview renders inside a simulated device frame (iPhone, Pixel, etc.).

Workbench

The workbench includes:

Preview — live app in a phone frame
Code editor — full access to every project file
V2 Cloud dashboard — manage backend when enabled
Build tools — native Android/iOS builds (Pro+)

V2 Cloud (optional)

When your app needs accounts, shared data, or file uploads, enable V2 Cloud. OpenV2 provisions a PocketBase instance per user with:

Authentication
Collections (database)
File storage buckets
Server hooks

Learn more in V2 Cloud overview.

AI models

OpenV2 routes requests to state-of-the-art models via OpenRouter. Pro and Ultra subscribers get access to premium models with stronger reasoning and longer context.

Each chat turn consumes credits based on complexity. See Plans & credits.

Data & ownership

Your code belongs to you — export anytime as a standard Expo project
Platform data (account, projects, billing) is stored securely on Supabase
App backend data (V2 Cloud) lives in your dedicated PocketBase instance