Food360 AI

Photographs flatten food. Restaurants wanted diners to see portion, texture, and plating properly, which means shipping 3D models to a phone browser without a long wait or a hot device — and letting non-technical restaurant staff manage those assets themselves, per venue, per menu, without touching a 3D tool.

We split the product: a React + Vite front end tuned for progressive asset loading, and a Nest.js API owning menus, venues, tenancy, and asset pipelines. Models and imagery are stored and transformed through Cloudinary, Neon Postgres via TypeORM holds the relational menu model, and Redux keeps the builder's editing state predictable. MUI and Framer Motion give staff a builder that feels like a normal admin tool rather than 3D software, with Google Maps handling venue discovery.

Restaurant staff publish and reorganise 3D menus from a browser without design help, and diners load an interactive dish view on a phone rather than a static photo grid. The Nest.js API is a clean boundary, so new front ends can be added against the same contract.