Judify

What if your taste was put on trial?

Anyone who's shared a Spotify Wrapped knows how catchy music stats can be. But the average wrap is, at the end of the day, a chart. Judify flips that: instead of neutral numbers you get an opinionated, sharply written verdict on your listening behaviour — a piece of theatre instead of a spreadsheet.

Taste is subjective; the verdict isn't.

The flow

You log in via Spotify, pick a judge mode, and the app pulls in top artists, top tracks, recent plays, playlists and library. That raw API data is normalised first into a compact taste profile — a deliberate intermediate step to save tokens and improve AI output. After that OpenAI does the rest: a verdict in the chosen tone and language, ready to share.

Why it's fun to build

Judify sits on a sweet spot where three things meet: a third-party API with scopes and rate limits (Spotify), an AI layer that rewards prompt engineering and language craft (OpenAI), and a social mechanic where the output itself becomes the marketing channel (shareable results with OG images). That forces you to think across all three axes simultaneously: data, language and distribution.

Under the hood

Stack: Next.js 16 with App Router, NextAuth v5 for the Spotify OAuth flow, Firestore for profiles and history, and @vercel/og for dynamic preview images. Server Components do the heavy lifting; client components only show up for judge mode selection and the duo compare flow. Framer Motion handles transitions and shadcn/ui (Base UI) provides the base components.

The Duo Compare trick

The duo compare mode adds a social layer: send a link, have a friend log in, and the app builds a joint verdict that magnifies the differences in taste. It's a feature that earns its keep marketing-wise — every comparison brings in a new user.