Google released Gemini 3.1, an AI model showing strong performance in UI and web design tasks. The model was tested for generating mobile application interfaces, specifically a home screen for a food ordering app similar to Uber Eats or DoorDash. Using markdown-formatted prompts proved effective because markdown provides natural content hierarchy through headings, logical grouping of UI sections, and improved readability for both humans and AI models. The experiment involved creating a detailed prompt for an iOS app called Foodiez with structured sections including status bar, header, and other UI components.
"Google recently released a new AI model, Gemini 3.1, that demonstrates great results in UI and web design tasks. I've already tested this model for web design tasks and in this article, I want to experiment with Gemini 3.1 and generate UI for a mobile application."
"Why markdown? Markdown naturally has a content hierarchy (with headings like ## Header), offers a logical grouping of UI sections, and provides better readability for both humans and AI models."
"I will generate a home screen of a mobile app for food ordering (similar to Uber Eats or DoorDash), and use a prompt in markdown format with sections including Status Bar, Header Section, and other structured UI components."
Read at Medium
Unable to calculate read time
Collection
[
|
...
]