Skip to content

AI/ML Full-Stack Monorepo

Product Decoder

Web app that identifies products from a photo using Google Cloud Vision and aggregates product context from Google Custom Search, GNews, and YouTube (Next.js 15, Bun).

When
Jun – Aug 2025
Length
3 months
Status
Completed

2 screens open one

Where it sits

EarlierLinkShortApr 2025LaterGDG on Campus FSC WebsiteJun 2025

The project

Led product-decoder, a full-stack monorepo web application that identifies products from a photo using Google Cloud Vision.

Built with a Next.js 15 App Router frontend, Clerk authentication, and a Bun/Elysia.js backend integrating Google Cloud Vision, Google Custom Search, GNews, YouTube, Prisma, MongoDB, Effect-TS, and ArkType.

Screens

  • Fig. 2Product Decoder landing page with Ollie, an illustrated four-step workflow and a product metadata section.

Built with