AI for Automation
Back to AI News
2026-03-22AI toolsgenerative UIopen sourceweb developmentAI automation

OpenUI just made AI 3x faster at building live apps

OpenUI is an open-source framework that lets AI respond with interactive charts, forms, and dashboards instead of plain text — now 3x faster after a major rewrite.


What if ChatGPT or Claude could answer your question with an interactive chart, a working form, or a live dashboard — instead of just a wall of text? That's exactly what OpenUI does, and it just got 3x faster.

OpenUI is an open-source framework (a ready-made toolkit developers can plug into their apps) that lets AI models generate real, interactive user interfaces as they respond. It already has 2,300+ GitHub stars and hit 278 points on Hacker News this week after the team revealed a major performance breakthrough.

OpenUI demo showing AI generating interactive UI components in real-time

AI That Responds With Apps, Not Text

Today's AI chatbots respond with text. You ask for sales data, you get paragraphs. You ask for a comparison, you get bullet points. OpenUI changes that — the AI responds with actual interactive components:

What AI can generate with OpenUI:

📊 Live charts and data visualizations

📝 Working forms with validation

📋 Interactive tables you can sort and filter

🏗️ Full dashboard layouts with multiple widgets

🎨 Cards, lists, and custom layouts — all interactive

The key insight: instead of asking AI to spit out code that a developer then wires up, OpenUI lets AI generate UI components directly — and they render live as the AI streams its response. You see the chart building itself in real time.

The 3x Speed Breakthrough

The Thesys engineering team behind OpenUI originally built their parser (the engine that converts AI output into visual components) in Rust compiled to WebAssembly. Rust is known for being blazing fast — so they expected great performance.

They were wrong. The bottleneck wasn't the code itself — it was the boundary overhead. Every time the parser ran, data had to be copied back and forth between JavaScript and WebAssembly memory, serialized to JSON, and deserialized again. All that copying ate up the speed advantage.

So they rewrote the entire parser in TypeScript (the language browsers already speak natively). The results:

Simple table: 9.3µs vs 20.5µs → 2.2x faster

Contact form: 13.4µs vs 61.4µs → 4.6x faster

Dashboard: 19.4µs vs 57.9µs → 3.0x faster

Streaming dashboard: 255µs vs 840µs → 3.3x faster

For real-time streaming — where the UI updates with every token the AI generates — the improvement is even more dramatic.

67% Fewer Tokens Than JSON

OpenUI also saves money. Most AI apps that generate structured output use JSON (a data format computers read easily). But JSON is verbose — lots of curly braces, quotes, and key names that all cost tokens (the units AI providers charge you for).

OpenUI's custom language is up to 67% more token-efficient than JSON alternatives. For companies running AI at scale, that's a significant cost reduction on every API call.

Works With Every Major AI Provider

OpenUI isn't locked to one AI company. It works with:

OpenAI (GPT-4o, GPT-5)

Anthropic (Claude)

Google (Gemini)

Mistral, and any OpenAI-compatible API

It also integrates with popular AI development frameworks like Vercel AI SDK, LangChain, and CrewAI.

Try It Right Now

You can try OpenUI in two ways:

1. Live playground — no install needed:
Visit openui.com/playground to generate UI live in your browser.

2. Build a full app — one command:

npx @openuidev/cli@latest create --name my-genui-app
cd my-genui-app
npm run dev

This scaffolds a complete chat app with streaming UI generation, built-in components, and AI integration ready to go.

Who Should Care

If you're a developer building AI-powered apps, OpenUI lets you skip weeks of custom UI rendering code. Define your components once, and the AI handles the rest.

If you're a product designer, this is worth watching — it points to a future where AI assistants don't just describe solutions, they show them with real interactive elements.

If you're a business user, imagine asking your company's AI assistant "show me Q1 sales by region" and getting an actual interactive chart you can hover over and drill into — not a paragraph describing the numbers.

OpenUI - The Open Standard for Generative UI

The Bigger Picture

OpenUI is part of a growing trend: AI is evolving from text-only responses to multimodal output. Google recently launched Stitch for turning text into app designs, and other projects are exploring Markdown-based live app generation.

But OpenUI's open standard approach — MIT licensed, framework-agnostic, and 67% cheaper on tokens — positions it as the infrastructure layer that could become the default for generative UI across the industry.

The project is actively developed by the Thesys team, with 434 commits and growing community support on Discord.

Related ContentGet Started with Easy Claude Code | Free Learning Guides | More AI News

Stay updated on AI news

Simple explanations of the latest AI developments