openui is a framework-agnostic Generative UI framework built around OpenUI Lang, a compact streaming-first language for model-generated UI that uses up to 67% fewer tokens than JSON. It lets developers define typed components and generate system prompts or JSON schema for their own LLM to produce progressively streamed, renderable UI.
Project overview
It addresses a concrete cost and rendering problem: model-generated UI as unstructured JSON is token-heavy and hard to render progressively while streaming, and OpenUI Lang's up-to-67% token reduction targets exactly that gap.
Project type
AI Agent · Prompt Engineering
Deployment
Refer to project documentation
License
MIT
Best for
Developers who can code and supply their own LLM, building agent interfaces with progressively streamed, structured UI.
Key capabilities
A compact, streaming-first language for model-generated UI, designed for streaming output with up to 67% fewer tokens than JSON.
Generate system prompts (or JSON schema) for the model directly from the components you define and register.
Limitations and risks
The project does not provide a model; prompts are sent to your own model, so you must supply and manage an LLM and its associated access.
Getting started
Run npx @openuidev/cli@latest create --name genui-chat-app, install with npm install @openuidev/react-lang @openuidev/react-ui, and use the scaffolded app with streaming and OpenUI Lang support.
Evidence and sources
README: OpenUI is a full-stack, framework-agnostic Generative UI framework built around a compact, streaming-first language.
README: Token efficiency** - Up to 67% fewer tokens than equivalent JSON
README: **Streaming renderer** - Parse and render model output progressively in React, Vue, Svelte, or Angular as tokens arrive.
README: 2. Generate a system prompt from that library. 3. Send that prompt to your model.