40 / AI-Native Platforms
Cross-Platform AI-Native UI Language & Compiler
We built a single UI language that compiles one source into native iPhone, Android and web apps at once, so a product built for AI to write needs one codebase instead of 3.
- Built by
- 360 Labs Research
- Stack & approach
- Programming LanguageCompilerSwiftUIJetpack ComposeRust
The problem
Building the same product across iOS, Android and web means writing 3 separate UI codebases in 3 paradigms. Cross-platform frameworks compromise on native feel and bloat token counts when used as LLM context, while design tokens and security constraints live outside the code and drift.
Solution & architecture
One compact source language that compiles to all 3 native targets. Aura's grammar is designed for both human and AI authoring: typically 70% fewer tokens than equivalent TypeScript + React Native. Design tokens and security types are first-class language constructs, enforced by the compiler before any platform code is generated.
Research highlights
Built for AI authoring
70% fewer tokens than React Native
Security as a type
Enforced by the compiler, not convention
True native output
SwiftUI and Compose, not a webview
