Open AI Lab unveils a text-to-layout model tuned for product design workflows

AI · 5 min read

The new model, described as a layout-first multimodal network, takes plain-language product briefs and returns annotated wireframes with component hierarchies, responsive breakpoints, and accessibility annotations. It also produces a variant set for desktop, tablet, and mobile that preserves content priorities across breakpoints.

Built with explicit constraints in the training loop, the model reportedly understands standard UX patterns (forms, dashboards, modals) and maps them to platform-appropriate components. Designers can ask for changes using natural language and get updated layouts and component specs in seconds.

Integrations with popular design tools are part of the rollout: the vendor shipped plugins that export generated layouts as editable vectors and component trees. Companies say the model is intended to accelerate early-stage ideation, not replace detailed design decisions or research-driven UX work.