Google’s PaLM-Next adds live layout-aware image generation for UIs

Design · 4 min read

PaLM-Next’s latest update introduces spatially-aware synthesis: designers can provide a mockup or grid constraints and receive images or illustrations that fit specified bounding boxes, color palettes, and aspect ratios. The system promises to preserve focal points and maintain photographic perspective across different breakpoints, addressing a common friction where generated assets didn’t adapt cleanly to responsive layouts.

Google released an API and a Figma plugin that wire generated assets directly into layers, automatically resizing and exporting multiple density versions for production. There’s also an option to generate table-based or SVG shapes instead of raster images, which designers can style programmatically to match brand tokens.

Product managers and design leads said the tool reduces handoffs between visual designers and front-end teams by producing production-ready assets with consistent cropping and padding. Google added controls to enforce license-safe generation, and a “brand lock” feature that holds a chosen palette and visual style constant across sessions.