Case study - build multiplayer game with ai

Vibe-coding a full-stack game with AI

I'm a digital-experience leader with just enough dev exposure to be dangerous. I wanted to push myself and try my hand at vibe coding — so I took an idea from concept to a fully functioning, shipped app on my own. The result: Bid Battle: Word Search War, a real-time competitive word-search game live on iOS and Android.

Android

See it in action.

Race a live opponent to find hidden words on the board before the clock runs out. Every match is a head-to-head battle. Find words faster than your rival to win.

All the videos, art, music and sounds were created with AI-driven design and media tools.

The initial idea & the challenge.

I've spent my career building and optimizing digital experiences, working alongside strong engineering teams. I understood how web applications are built, I'd just never shipped one entirely on my own. AI changed that.

By leveraging tools like Google Antigravity alongside AI-driven design and media tools, I took Bid Battle: Word Search War from a concept to a fully functioning, monetized app all by myself.

It wasn't painless. There were plenty of wrong turns, hallucinations, and hair-pulling debugging sessions. I don't think I could have done it quickly without years of collaborating with engineers and knowing how applications are actually built.

Quick Facts

• Product: Bid Battle: Word Search War
• Type: Real-time competitive word search game with tournament option
• Platforms: iOS and Android
• Build: Solo, AI-assisted vibe coding with Google Antigravity
• Scope: Full stack - design, app, backend, monetization, analytics

How it came together

A genuine full-stack build, start to finish: design and brand, then everything it takes to actually ship and run a product.

concept
build
ship

AI agents to QA the tournament feature.

Building the in-game multiplayer tournament mode was a big challenge. Anyone in app dev knows how critical testing is — so alongside beta testing with real people, I leaned heavily on AI.

I created lightweight AI agents to automate multiplayer tournament testing and accelerate QA. The agents simulate real player behavior in real time — joining matches, submitting bids and actions, and completing full gameplay loops concurrently.

How the agents work:

1. Built the agents as a local .js app using Antigravity.

2. Run my game on localhost in Chrome, create a tournament, and launch all agents at once via a PowerShell script.

3. Hit Start Tournament in the game — the bracket generates and a console window opens per agent, each logging its actions in real-time.

4.  Python script (also built with AI-assistance) parses every log into a single tracking spreadsheet.

5. Review the spreadsheet to confirm everything worked, or review any issues found.

why it matters

A middle ground between manual QA and enterprise automation

What's interesting is how transferable this is beyond gaming. The same lightweight agents adapt to:

• App onboarding flow testing
• eCommerce UX validation
• Conversion funnel testing
• Lightweight load testing

tech stack

Full-stack, end-to-end

Cloud and data, monetization, and analytics — all wired together through a GitHub-based development workflow.

Cloud & Data
Firebase (Firestore)

Real-time database powering matches, tourneys, game state, and player data.

Render AI Cloud

Hosting and backend services for the live, real-time game.

Monetization
RevenueCat

In-app purchases and subscription management.

AdMob

Ad monetization integrated into the game loop.

Analytics & Workflow
GA4 · BigQuery · Looker Studio

Event tracking, warehousing, and dashboards to understand gaming and monetization KPI's.

GitHub

Version control and a development workflow tying the whole build together.

The takeaway

The gap between idea and execution is smaller than it's ever been — if you're willing to learn, experiment, and leverage the tools available.

AI made it possible. Years of working alongside great engineers made it fast to ship. That combination is what turned a side-project idea into a shipped app.