PicYep: an AI photo generator built with shadcn/ui

[Liujiacheng-yuexin]

Duplicate of #12623 — please ignore this post.

Hi everyone! I wanted to share PicYep, an AI photo generator that turns a sentence or one reference photo into realistic photos—headshots, portraits, selfies, social photos and pet portraits.

Why shadcn/ui

• shadcn/ui gives the app a clean, accessible foundation without locking the UI into a component package.
• I can keep the components in the codebase and adapt them to the generator workflow: ready-made looks, prompt and reference-photo inputs, model selection, credit-cost display and the photo library.
• The same primitives keep the account and billing screens consistent with the generation experience.

The rest of the stack

• TanStack Start, React and TypeScript
• Tailwind CSS
• Cloudflare Workers, with D1 for the database and R2 for image storage
• Drizzle ORM and better-auth
• Paraglide JS for English and Chinese

A couple of product details I tried to make explicit in the UI: the credit cost is shown before generating, and credits are returned automatically when a request ends without a photo. New accounts can try it with free credits and no card: https://picyep.com/

I’d love feedback from others using shadcn/ui for product interfaces, especially how you organise and extend the components as a workflow gets more involved. Thanks for the great tools!