Fanpire

The FIFA World Cup 2026 is here, and fan energy is everywhere. The idea was born from brainstorming sessions with two close friends — what if fans could claim territory for their team? Vote in your district, and the map changes color based on who's winning. The rivalry IS the product.

While I designed and built the entire product solo with AI, those two friends remained crucial throughout — acting as the first peer reviewers, pressure-testing ideas at every stage, and shaping the direction of each new feature.

The core mechanic is simple: one tap to vote, and a live map shows the result. But building the infrastructure to support real-time voting, fraud prevention, gamification, and a 3D interactive map — that's where it gets interesting.

Every architecture decision was made to balance simplicity with scalability. The goal was: ship fast, handle real traffic, and keep costs near zero.

The map is the hero of the product. The idea was to take an existing open-source map renderer — MapLibre GL — and build an entirely custom data visualization layer on top of it. Instead of building maps from scratch or paying for expensive tile servers, I used free CARTO dark basemap tiles as the canvas and layered custom GeoJSON boundaries, 3D extrusions, and interactive overlays on top.

This approach — external map + custom data layer — means the app gets a professional, beautiful basemap for free while owning the entire interactive experience above it. The GeoJSON boundary files for 759+ Indian districts are served as static assets, parsed on the client, and rendered as MapLibre fill-extrusion layers with dynamic coloring based on real-time vote data.

Beyond voting, the platform has a deep engagement layer designed to keep fans coming back.

The visual identity was crafted to feel premium, immersive, and unmistakably football. A dark-first design with rich accent colors creates the atmosphere of a stadium at night.

The app is built for real traffic with production-grade infrastructure.

The Supabase Postgres database has 30+ tables designed with Drizzle ORM — covering core voting, competitions, matches, gamification (XP, ranks, challenges, predictions), arena contests, push notifications, analytics, and admin infrastructure. Key patterns include idempotent ledgers for XP and Fang Points, trigram search indexes for places, and trigger-maintained aggregate tables.

This project demonstrates that a designer with strong product thinking can ship production-grade software by using AI as a development partner. Every line of code was written through AI-assisted development — but the product decisions, architecture choices, UX flows, and design system were all driven by human judgment.

The AI handled implementation details; I handled the "what" and "why." The result is a complex full-stack application with 30+ database tables, real-time maps, gamification systems, push notifications, and multi-platform deployment — shipped by one person.

Fanpire wasn't designed in isolation — it evolved through continuous feedback from real users and peers throughout development. Every major feature addition was a direct response to what people actually asked for after using the product.

Fanpire launched just one day before the FIFA World Cup kicked off — far later than planned. Key features like match predictions, fan zones, and the new user flow were still being built during the Round of 16. The late timing meant the product never got the early-tournament window to build an audience organically. Despite that, it crossed 400+ votes and saw real momentum — users were returning, streaks were being maintained, and the map was coming alive with team colors.