Karam Abu MokhEmail me

Case study Live

GourmetAI

A web app that turns an ordinary phone photo of a dish or product into a picture lit and styled like a studio shoot.

The original phone photo of a steak
Before
The same steak after GourmetAI, editorial look
After
The live GourmetAI site on a phone, scrolling to its before and after slider
My role
Freelance engineer, built alone
When
Since March 2026
Built with
  • React
  • TypeScript
  • Vite
  • Tailwind
  • Node.js
  • Express
  • PostgreSQL
  • Cloudflare R2
  • Gemini
  • OpenAI
  • Expo
  • Vercel
See it
gourmetai.it.com

Who it is for

Restaurants and small shops need good pictures for their menu, their shop page and Instagram, but they cannot pay a photographer for every one. GourmetAI gives them that picture from the phone photo they already have.

What I built

Everything, alone: the web app, the API, the mobile app, the AI prompts and the category logic, sign-in, credits and monthly plans, the admin console and the deployment.

How it works

  1. The owner uploads or takes one phone photo.
  2. A light AI vision model recognises the product type, and the owner confirms it.
  3. The owner picks a look, for example Studio Premium, Bright & Fresh or Dark & Dramatic. Nobody writes a prompt: the app builds it from the choices.
  4. The AI re-renders the photo and the owner downloads it. The original photo is not stored.
  5. A second tool, Branding, takes a brand's logo, photos and a short brief and renders a matching set of posters and product shots.

13product categories, from food to jewelry, perfume, flowers and real estate

9interface languages, with right-to-left Arabic and Hebrew

3codebases: web, API and mobile

163 of 164commits are mine

Live today

The web app and its API are live at gourmetai.it.com, with sign-in by Google, Apple or a phone code. The mobile app for iPhone and Android is built on the same API.