thesysdev/openui

What it solves

OpenUI provides a way for LLMs to generate interactive user interfaces instead of just plain text. It solves the problem of high token costs and latency associated with using JSON for structured UI generation, while allowing developers to maintain strict control over which components are rendered on the client.

How it works

The framework centers around OpenUI Lang, a compact, streaming-first language. Developers define a component library (using Zod schemas for props), and OpenUI generates a system prompt based on that library. The LLM then streams output in OpenUI Lang, which a client-side renderer parses and renders progressively in real-time as tokens arrive.

Who it’s for

It is designed for developers building AI assistants, copilots, and interactive product flows who want to integrate generative UI into React, Vue, or Svelte applications.

Highlights

• Token Efficiency: Uses up to 67% fewer tokens than JSON-based formats, reducing costs and latency.
• Streaming-First: Renders UI components incrementally as the model generates them.
• Renderer-Agnostic: Official support for React, Vue, and Svelte, with a framework-agnostic core parser.
• Library-Driven: Automatically generates model instructions from the defined component library to ensure consistent output.
• Ready-to-Use: Includes built-in component libraries (charts, forms, tables) and prebuilt chat interfaces.