Inspiration
People who care deeply about culture often maintain separate accounts for anime, films, games, music, books, and television. That fragments the story of what they love. TasteStack turns those scattered lists into one living, shareable profile — inspired by the clarity of MyAnimeList, but designed for a whole life of interests.
What it does
TasteStack lets someone create a public profile and build a cross-media stack across anime, manga, movies, TV, games, albums, and books. From Discover, users can filter or search the curated catalog, add a title with a meaningful status (planned, in progress, completed, on hold, or dropped), and mark favourites. Their profile then turns that activity into media-by-media stats, favourite picks, and shelves of recent titles. Profile settings persist a display name, bio, colour, and public/private preference.
The goal is simple: one beautiful link that represents a person's taste — not seven different accounts.
How we built it
TasteStack is a responsive Next.js 14 and TypeScript application, styled with Tailwind CSS. It uses Prisma and SQLite for local persistence, including profiles, media items, favourites, progress, activities, and a follow-ready social model. NextAuth credentials authentication protects editing while allowing public profile sharing.
The local discovery catalog gives judges a dependable demo without requiring third-party API keys. The data architecture is already source-aware for future AniList, TMDB, RAWG, Last.fm, and Open Library enrichment.
How Codex and GPT-5.6 helped
Codex with GPT-5.6 accelerated the project from a starter app into an end-to-end product flow. It helped shape the information architecture, implement the responsive landing page, discovery controls, authenticated add-to-stack API, public profile shelves, activity feed, and persisted settings. It also surfaced and resolved TypeScript and routing issues through production-build verification. Human direction defined the central product insight: bring MyAnimeList-style tracking to every hobby, not only anime.
Challenges we ran into
The main design challenge was making seven different media worlds feel coherent without reducing them to a generic checklist. We solved that with a shared item model and a consistent set of actions — status, progress, favourites, and profile visibility — while retaining labels that make sense across media.
What's next
We plan to add live catalog search and metadata from AniList, TMDB, RAWG, Last.fm, and Open Library; richer reviews and ratings; social following; and cross-media recommendations built from a user's actual taste profile.
Run it locally
Clone the repository, copy .env.example to .env.local, set DATABASE_URL, NEXTAUTH_URL, and AUTH_SECRET, then run npm install, npm run db:push, and npm run dev. The README includes complete setup guidance and explains how Codex and GPT-5.6 were used.
Built With
- codex
- gpt-5.6
- next.js
- nextauth
- prisma
- react
- sqlite
- tailwind-css
- typescript
