thesys-generative-ui

Streaming LLM responses convert to real-time React UI components using TheSys Generative UI API, which parses LLM output and renders interactive forms, charts, and tables as they arrive. This Skill provides production-ready integration patterns for Vite+React, Next.js, and Cloudflare Workers.

Yes. This Skill supports tool calling via Zod schemas, enabling AI providers like OpenAI and Claude to invoke typed functions within streaming UI components. Integration templates handle schema validation and tool responses across all supported frameworks.

Generative UI works with Vite+React, Next.js App Router, and Cloudflare Workers. The Skill includes 15+ framework-specific templates and multi-frame composition patterns to reduce deployment errors across these stacks.

Backend integration templates for Express, Next.js, and Cloudflare Workers handle API calls to OpenAI, Claude, or Workers AI. Responses stream directly to React components, which render UI updates in real time as tokens arrive.

Streaming UI requires handling partial renders, race conditions between frames, and error recovery mid-stream. This Skill includes robust error handling and reproduction-ready templates to address these constraints in production environments.

ThemeProvider theming is supported but optional. This Skill includes templates with and without theming, allowing you to apply custom styles to streamed components or use unstyled output depending on your design system.