thesysdev/openui

The Open Standard for Generative UI

OpenUI – Generative UI Framework

OpenUI is a full‑stack, framework‑agnostic library that lets you build user interfaces that are generated by large language models (LLMs) and streamed to the client token‑by‑token. It introduces OpenUI Lang, a compact, streaming‑first language that describes UI components more efficiently than JSON (up to 67 % fewer tokens). The core parses the model’s output, turns it into live UI, and can be used from React, Vue, Svelte, Angular, or even a plain browser bundle.

How it works

  1. Component library – Define a set of UI components (charts, forms, tables, etc.) with typed props (using Zod).
  2. System prompt generation – OpenUI creates a prompt that tells the LLM which components it may emit.
  3. LLM call – Send the prompt to your model.
  4. OpenUI Lang stream – The model streams a compact UI description.
  5. Renderer – The appropriate runtime (React, Vue, …) parses the stream and incrementally renders the UI as tokens arrive.

The flow is visualised in the README’s Mermaid diagram.

Main packages (monorepo)

Package What it provides Typical use
@openuidev/lang-core Parser, prompt generator, runtime evaluation, type layer – no framework dependency Backend services, edge functions, custom runtimes
@openuidev/react-lang React bindings for parsing and rendering OpenUI Lang Any React app that wants streaming UI
@openuidev/react-ui Pre‑built chat UI, component library, ready‑to‑use layouts Fastest way to get a full chat interface
@openuidev/react-headless Headless chat state & streaming adapters Build your own UI on top of the streaming logic
@openuidev/vue-lang, @openuidev/svelte-lang, @openuidev/angular-lang Framework‑specific bindings Vue / Svelte / Angular projects
@openuidev/langchain Helpers for LangChain / LangGraph agents to stream OpenUI Lang Build LLM agents that produce UI
@openuidev/browser-bundle Single‑file bundle (script + stylesheet) for CDN/iframe embeds No‑build, embed‑in‑any‑page scenarios
@openuidev/a2ui A2UI protocol client that carries OpenUI Lang payloads Interoperate with A2UI‑compatible back‑ends
@openuidev/server Server utilities and gateway helpers Host the prompt generation or event observability
@openuidev/observability-cloud Cloud sink that ships OpenUI events to the Thesys console Production‑grade telemetry
@openuidev/cli CLI for scaffolding apps, generating prompts, converting schemas npx @openuidev/cli@latest create …

Quick start (React example)

npx @openuidev/cli@latest create --name genui-chat-app

The command scaffolds a React app that includes:

  • OpenUI Lang support
  • Library‑driven prompt generation
  • Streaming UI updates
  • A ready‑to‑run chat surface

You can also install only the core parser for a backend service:

npm install @openuidev/lang-core

Why use OpenUI Lang?

  • Streaming‑first – UI appears progressively as the model generates output, reducing perceived latency.
  • Token efficiency – Benchmarks show ~50 % fewer tokens than JSON‑based streaming formats across common UI patterns.
  • Controlled output – The model can only emit components you have defined, preventing arbitrary markup.
  • Typed contracts – Component props are described with Zod schemas, giving static guarantees and easier debugging.

Documentation & community

License

OpenUI is released under the MIT License.

TL;DR

OpenUI gives you a compact streaming language and a set of framework‑agnostic runtimes so LLMs can generate UI that appears instantly, with far fewer tokens than JSON. Install the core parser for backend work, or pull in the React/Vue/Svelte/Angular bindings for a full front‑end experience, and use the CLI to scaffold a working app in seconds.

Related

  • Project
    wandb/openui

    An open-source tool that lets you describe UI components in plain language and see them rendered live, with the ability to convert them to React, Svelte, or Web Components.

  • Project
    vercel-labs/json-render

    json-render is a generative UI framework that lets AI create interfaces from prompts while staying safe and predictable. It works by constraining AI output to a catalog of defined components and actions, which are then rendered across many platforms.

  • Project
    thesysdev/appless

    An experimental generative UI operating system that replaces traditional apps with live-streamed, native mobile interfaces generated on the fly based on user requests.

  • Project
    AGenUI/AGenUI

    AGenUI is a production-grade SDK that implements Google's A2UI v0.9 protocol to render LLM-generated interactive UI in real time on iOS, Android, and HarmonyOS. A shared C++ core handles parsing, layout, and diffing while native renderers draw the components, giving 22 built-in UI components, custom component registration, Function Call support, a self-contained JSON catalog schema for constraining LLM output, an agent skill for turning natural language into UI protocol messages, and a local Studio workbench for prototyping.

  • Project
    thesysdev/open-intelligent-ui

    Open Intelligent UI is an open‑source demo that recreates ChatGPT’s “Intelligent UI” – LLM‑generated, interactive answers (maps, photos, itineraries, dashboards). Built on the OpenUI generative‑UI standard, it provides a TypeScript component library, 59 built‑in functions, stateful inputs, and can run against OpenUI Cloud, any OpenAI‑compatible API, or a local Ollama model.