OpenUI - The Open Standard for Generative UI
OpenUI is a full-stack, framework-agnostic Generative UI framework built around a compact, streaming-first language. A framework-agnostic core powers first-party runtimes for React, Vue, Svelte, and Angular, and React additionally ships built-in component libraries and ready-to-use chat interfaces. OpenUI Lang uses up to 67% fewer tokens than JSON.
Docs · Playground · Discord · Contributing
Important: OpenUI has no official cryptocurrency, token, or coin. Any asset using the OpenUI name is unaffiliated with this project and is not endorsed by its maintainers.
What is OpenUI
At the center of OpenUI is OpenUI Lang: a compact, streaming-first language for model-generated UI. Instead of treating model output as only text, OpenUI lets you define components, generate prompt instructions from that component library, and render structured UI as the model streams.
Core capabilities:
- OpenUI Lang - A compact language for structured UI generation designed for streaming output.
- Built-in component libraries - Charts, forms, tables, layouts, and more, ready to use or extend.
- Prompt generation from your component library - Generate model instructions directly from the components you allow.
- Streaming renderer - Parse and render model output progressively in React, Vue, Svelte, or Angular as tokens arrive.
- Chat and app surfaces - Use the same foundation for assistants, copilots, and broader interactive product flows.
Quick Start
npx @openuidev/cli@latest create --name genui-chat-app
This is the fastest way to start with OpenUI. The scaffolded app gives you an end-to-end starting point with streaming, built-in UI, and OpenUI Lang support.
What this gives you:
- OpenUI Lang support - Start with structured UI generation built into the app flow.
- Library-driven prompts - Generate instructions from your allowed component set.
- Streaming support - Update the UI progressively as output arrives.
- Working app foundation - Start from a ready-to-run example instead of wiring everything manually.
How it works
Your components define what the model can generate.
flowchart LR
A["Component Library"] --> B["System Prompt"]
B --> C["LLM"]
C --> D["OpenUI Lang Stream"]
D --> E["Renderer"]
E --> F["Live UI"]
- Define or reuse a component library.
- Generate a system prompt from that library.
- Send that prompt to your model.
- Stream OpenUI Lang output back to the client.
- Render the output progressively with Renderer.
Try it yourself in the Playground: generate UI live with the default component library.
Packages
| Package | Best for | Description |
|---|---|---|
@openuidev/lang-core |
Framework-agnostic parsing and prompt generation | Core parser, prompt-generation, runtime-evaluation, and type layer with no React, Angular, Vue, or Svelte dependency |
@openuidev/langchain |
LangChain and LangGraph agents | Agent transformer and server helpers that stream OpenUI through AG-UI |
@openuidev/react-lang |
React rendering runtimes | Define component libraries, generate prompts, and render streamed OpenUI Lang in React |
@openuidev/react-headless |
Bring-your-own React chat UI | Headless chat state, streaming adapters, and message format converters |
@openuidev/react-ui |
Fastest path to a full React chat experience | Prebuilt chat layouts, standalone UI primitives, and two built-in component libraries |
@openuidev/react-email |
Email generation and HTML export | React Email component definitions plus prompt options for model-generated emails |
@openuidev/assistant-ui |
assistant-ui apps | Tool UI renderers and instruction wiring for streaming OpenUI Lang inside assistant-ui |
@openuidev/vue-lang |
Vue integrations | Vue 3 bindings for defining model-renderable components and rendering streamed OpenUI Lang |
@openuidev/svelte-lang |
Svelte integrations | Svelte 5 bindings for defining model-renderable components and rendering streamed OpenUI Lang |
@openuidev/angular-lang |
Angular integrations | Angular bindings for defining component libraries and rendering streamed OpenUI Lang |
@openuidev/browser-bundle |
CDN, iframe, and no-build embeds | Prebuilt browser bundle that ships the renderer, UI library, React, and styles as script + stylesheet assets |
@openuidev/a2ui |
A2UI with OpenUI Lang component payloads | Framework-agnostic A2UI v1.0 protocol client with an optional React renderer |
@openuidev/server |
Server side utilities | Server utilities for OpenUI and OpenUI Gateway |
@openuidev/observability-cloud |
Hosted observability in the Thesys console | Cloud sink that batches OpenUI events and ships them to OpenUI Gateway |
@openuidev/cli |
Project scaffolding and prompt generation | CLI for creating new apps and generating system prompts or JSON schema from a library definition |
@openuidev/openclaw-os-plugin |
OpenClaw workspaces | OpenClaw OS plugin for serving OpenUI-powered OpenClaw workspaces |
Common starting points:
# React app with OpenUI rendering and prebuilt components npm install @openuidev/react-lang @openuidev/react-ui # Framework-agnostic backend or Edge prompt generation npm install @openuidev/lang-core # LangChain/LangGraph agent and server integration npm install @openuidev/langchain @langchain/langgraph # Vue, Svelte, or Angular runtime npm install @openuidev/vue-lang npm install @openuidev/svelte-lang npm install @openuidev/angular-lang # A2UI protocol with OpenUI Lang component payloads npm install @openuidev/a2ui zod
Why OpenUI Lang
OpenUI Lang is designed for model-generated UI that needs to be both structured and streamable.
- Streaming output - Emit UI incrementally as tokens arrive.
- Token efficiency - Up to 67% fewer tokens than equivalent JSON (see benchmarks).
- Controlled rendering - Restrict output to the components you define and register.
- Typed component contracts - Define component props and structure up front with Zod schemas.
Token efficiency benchmarks
Measured with tiktoken (GPT-5 encoder). OpenUI Lang vs two JSON-based streaming formats across seven UI scenarios:
| Scenario | Vercel JSON-Render | Thesys C1 JSON | OpenUI Lang | vs Vercel | vs C1 |
|---|---|---|---|---|---|
| simple-table | 340 | 357 | 148 | -56.5% | -58.5% |
| chart-with-data | 520 | 516 | 231 | -55.6% | -55.2% |
| contact-form | 893 | 849 | 294 | -67.1% | -65.4% |
| dashboard | 2247 | 2261 | 1226 | -45.4% | -45.8% |
| pricing-page | 2487 | 2379 | 1195 | -52.0% | -49.8% |
| settings-panel | 1244 | 1205 | 540 | -56.6% | -55.2% |
| e-commerce-product | 2449 | 2381 | 1166 | -52.4% | -51.0% |
| TOTAL | 10180 | 9948 | 4800 | -52.8% | -51.7% |
Full methodology and reproduction steps in benchmarks/.
Documentation
Detailed documentation is available at openui.com.
Repository structure
openui/
├── packages/
│ ├── react-lang/ # Core runtime (parser, renderer, prompt generation)
│ ├── react-headless/ # Headless chat state & streaming adapters
│ ├── react-ui/ # Prebuilt chat layouts & component libraries
│ ├── react-email/ # React Email component library for generated emails
│ ├── assistant-ui/ # assistant-ui Tool UI renderers for OpenUI Lang
│ ├── lang-core/ # Framework-agnostic parser, prompt, and runtime layer
│ ├── langchain/ # LangChain/LangGraph streaming integration
│ ├── vue-lang/ # Vue runtime bindings for OpenUI Lang
│ ├── svelte-lang/ # Svelte runtime bindings for OpenUI Lang
│ ├── angular-lang/ # Angular runtime bindings for OpenUI Lang
│ ├── browser-bundle/ # Script-tag bundle for CDN / iframe / no-build embeds
│ ├── a2ui/ # A2UI protocol with OpenUI Lang component payloads
│ ├── server/ # Server utilities for OpenUI & OpenUI Gateway
│ ├── observability/ # Framework-free observability event bus
│ ├── observability-cloud/ # Cloud sink shipping OpenUI events to Thesys
│ ├── devtools/ # Dev-only OpenUI Inspect widget
│ └── openui-cli/ # CLI for scaffolding & prompt generation
├── examples/ # Cookbooks and integration reference implementations
│ ├── agent-frameworks/
│ ├── app-frameworks/
│ ├── cookbooks/
│ ├── design-systems/
│ ├── harnesses/
│ └── miscellaneous/
├── docs/ # Documentation site (openui.com)
└── benchmarks/ # Token efficiency benchmarks
Good places to start:
- openui.com for the full docs
- Quickstart to scaffold a working app
examples/README.mdto find a focused reference implementation- Cookbooks for complete walkthroughs with runnable code
CONTRIBUTING.mdif you want to contribute
Community
- Discord - Ask questions, share what you're building
- GitHub Issues - Report bugs or request features
How OpenUI compares
| Feature | OpenUI | json-render (Vercel) | A2UI (Google) | CopilotKit OpenGenUI |
|---|---|---|---|---|
| Tokens | 1x | 3x | 3x | 4x |
| Latency (60 tok/s) | 4.9s | 14.2s | 14.2s | ~20s |
| Streaming | Yes | Yes | Yes | Partial |
| Consistent output | Yes | Yes | Yes | No |
| Components | Library + custom | Library + custom | Custom only | None |
| Multi-platform | Web, mobile, email | Web, mobile, PDF, email, video | Web, iOS, Android | Web |
| Built-in data fetching | Yes | No | No | No |
| Chat UI included | Yes | No | No | Yes |
For more details, refer to the official OpenUI Lang comparison documentation.
Adopters
A list of organizations and projects using OpenUI is maintained in ADOPTERS.md. If you're using OpenUI, please consider adding your organization; it helps the project gain momentum and helps other adopters find peers using OpenUI in similar contexts.
Contributing
Contributions are welcome. See CONTRIBUTING.md for contribution guidelines and ways to get involved.
Agent Skill
OpenUI ships an Agent Skill so AI coding assistants (Claude Code, Codex, Cursor, Copilot, etc.) can help you scaffold, build, and debug Generative UI apps using OpenUI Lang.
The skill is maintained in the thesysdev/skills repository.
Install
# With the skills CLI (works across all agents)
npx skills add thesysdev/skills --skill openuiThe skill covers component library design, OpenUI Lang syntax, system prompt generation, the Renderer, SDK packages, and debugging malformed LLM output.
Feedback
Found a bug or have an idea? Open an issue, or send anonymous feedback from your terminal:
npx @openuidev/cli feedback "Describe the problem or idea"Coding agents using the OpenUI skill can send feedback for you. They remove personal details and code from the message and ask for your approval before sending it.
License
This project is available under the terms described in LICENSE.