Back to blog
·8 minAnnouncement

Introducing UIX Protocol — The Universal UI Schema

We built UIX Protocol because the design-to-code pipeline is fundamentally broken.

The problem

Every team building cross-platform UI faces the same set of compounding issues:

  • Design-code drift. What ships never matches the mockup. Every handoff introduces subtle (and not-so-subtle) inconsistencies.
  • Platform fragmentation. A button on web, iOS, Android, and Flutter means four separate implementations. Four times the surface area for bugs.
  • Manual QA. Visual regressions are caught by screenshots and human eyeballs — if they're caught at all.
  • AI can't participate. Without a structured, machine-readable representation of UI, AI agents have no way to read, write, or validate your interface.

What UIX Protocol does

UIX Protocol is a universal schema for user interfaces. You describe your UI once as a UIXDocument — a structured JSON document containing nodes, tokens, layout rules, and styling — and UIX renders it to seven platforms:

  • React (.tsx)
  • Flutter (.dart)
  • SwiftUI (.swift)
  • Vue (.vue)
  • Angular (.ts)
  • React Native (.tsx)
  • HTML (.html)

But code generation is just the beginning. UIX Protocol also provides:

  • Semantic QA — regression detection based on schema properties, not pixel diffing
  • Design token sync — tokens flow between Canvas, code, and your CI pipeline
  • AI integration — 58 MCP tools that let AI agents read, write, and transform any UIXDocument
  • Real-time collaboration — a WebSocket sync server for live multi-user editing
  • A CLI toolchain — uix init, uix push, uix diff, uix watch — Git for UI

Performance by design

UIX Protocol is built for speed at every layer:

  • Schema parsing completes in under 1ms — validated by Zod with zero allocations on hot paths
  • Code generation for 7 platforms runs in ~50ms for a typical document
  • Diff engine computes structural differences in under 5ms
  • Token resolution resolves nested references in under 0.5ms
  • The runtime pipeline interprets a document across 9 deterministic steps

Every operation is designed to be fast enough for real-time collaboration and CI pipeline integration.

What's next

We're currently in Phase 10 — production hardening. The core protocol is stable, the tooling is comprehensive, and the architecture is proven. We're focused on:

  • CI/CD pipeline automation
  • Docker deployment for the server and kernel
  • Public documentation site
  • Beta program for early adopters

If this resonates with how your team builds UI, we'd love to hear from you. Star the repo on GitHub or reach out.