Bill's Computer

A Pokemon trainer's desktop OS in the browser
Format Web App
Discipline Art Direction, Code, Design
By LUKEABC Date 2026

Overview

A fully interactive desktop operating system in the browser, themed as a Pokemon trainer's workstation. "OakOS" simulates a retro 90s PC where you manage your Pokemon Trading Card Game collection, chat with AI characters, explore the Kanto map, and more. The whole thing is one HTML file, no build step.

OakOS desktop
View project →
The Idea

It started because I collect Pokemon cards and every app or site for logging your collection is obsessed with prices — everything's about portfolio value, market drops, what your cards are "worth." I just wanted somewhere to organise my collection without constantly being reminded of money. So I built my own thing.

The premise became: what if Professor Oak gave you a computer? OakOS is a love letter to 90s desktop interfaces and the Pokemon universe. It takes the nostalgia of Windows 95 chrome, Game Boy boot screens, and AIM-style messaging and wraps it around a functional card collection tool. Everything is designed to feel like a real OS from 2001 that a Pokemon trainer would actually use.

OakOS with multiple windows open
Apps
App Detail
Dex Full Pokemon TCG card browser with search, filters, and 3D card rendering
Binders Collection manager — create custom binders, drag-and-drop sorting
Messenger AI chat with in-character personas (Oak, Ash, Gary, Team Rocket) via Groq/Llama
Map Pannable, zoomable Kanto region map
Music Album player with equalizer visualizer
trainer_notes.txt Persistent text editor
Settings Wallpaper, data management
PokeClock Real-time clock
Dex app showing Charizard card
Messenger app — chatting with Prof. Oak
Technical

The entire app is a single HTML file (~9,000 lines). React 18 is loaded via CDN with Babel transpiling JSX in-browser — no bundler, no build step. Card rendering uses perspective transforms with mouse-tracked tilt, shine, and holographic overlays. The window manager handles drag, resize, minimize, maximize, focus tracking, and z-index stacking. A CRT-style boot sequence plays on load with phosphor green text, scanlines, and glass reflections.

Layer Technology
Framework React 18 (CDN, in-browser Babel)
Card data TCGdex API + PokeAPI
AI chat AWS Lambda proxy → Groq / Llama 3.1
Storage localStorage, no backend
Hosting Vercel
Credits
Role Name
Art Direction, Code, Design Lukeabc

Navigation