Vercel Labs launches json-render/next to build entire apps from prompts

VercelVercel

ยท Updated

Vercel Labs released @json-render/next, a package that turns AI-generated JSON specs into complete Next.js applications with routes and layouts. By using a structured JSON schema instead of raw code generation, developers can build generative interfaces that are predictable, safe, and fully integrated with server-side features.

Vercel Labs expanded its json-render framework with @json-render/next, moving from individual components to full application architectures. This package allows a single AI-generated JSON specification to define an entire multi-page app, including layouts, routing, server-side rendering, and metadata. It treats the application structure as a data-driven schema.

Most generative UI tools rely on AI writing raw code, which risks unreliable or unsafe output. This framework uses a guardrailed approach where developers define a component catalog first. The AI only generates JSON instructions to assemble those components, ensuring the output matches your predefined components and remains type-safe.

Use this to build AI website builders, internal tool generators, or white-label SaaS platforms. The framework supports streaming via SpecStream for progressive rendering. The project is open-source under the Apache-2.0 license and includes pre-built components for shadcn/ui, React Three Fiber, and terminal interfaces via Ink.

Chris Tate
Chris Tate
@ctatedev
X

New: @๐š“๐šœ๐š˜๐š—-๐š›๐šŽ๐š—๐š๐šŽ๐š›/๐š—๐šŽ๐šก๐š Prompt โ†’ JSON โ†’ Full Next.js app Routes, layouts, SSR, metadata, data loaders, static generation. For AI website builders. Internal tool generators. CMS-driven apps. White-label SaaS. One JSON spec, entire multi-page app. https://t.co/2KeCIxkH1U

8retweets262likes
View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards โ†’

Share this update