Who said what??

Inspiration

We wanted to build something that feels human — something chaotic, social, and funny enough to break the awkward silence of every online hangout. Most browser games rely on text or visuals, but voices? Voices are personal. So we thought — what if we distorted them, scrambled them, and turned them into a guessing game? That’s how Who Said What? was born — a party game powered by confusion, laughter, and a touch of audio madness.

What it does Players join a lobby, get assigned random phrases (from memes to tongue-twisters), and record themselves saying them. The app instantly distorts each recording using randomized effects like pitch shifting, static, or reversal. Everyone then listens to the mystery clips and votes on who they think said each one. Points are awarded for correct guesses — or for fooling everyone else. It’s Broken Telephone meets Among Us, powered by chaos.

How we built it Frontend: Built with Kaplay.js for the interactive game loop and audio visualizations. Backend: Colyseus manages real-time multiplayer rooms, voice data, and scoring synchronization. Audio Engine: Web Audio API handles live recording and distortion (pitch shift, reverb, bit-crush). Deployment: Hosted on Vercel + Node.js for easy browser-based play.

We focused on smooth, low-latency communication, ensuring all clips and votes sync perfectly — even during rapid rounds.

Challenges we ran into: We built Who Said What? completely from scratch — learning every tool as we went. None of us had touched these technologies before the hackathon, so every part of the stack was a crash course in something new.

🕹️ Frontend: Built with Kaplay.js, which powered the game flow, UI animations, and integrated audio playback directly into the browser.

⚡ Backend: Managed with Colyseus, handling all multiplayer rooms, state synchronization, and real-time communication between players.

🎧 Audio Engine: Implemented using the Web Audio API, where we experimented with live mic input, pitch shifting, reverb, reversal, and bitcrushing effects to make each clip sound uniquely chaotic.

🔥 Hosting: Deployed through Vercel + Node.js, allowing players to instantly join via browser — no setup, no installs.

Accomplishments that we're proud of: Built a working, real-time multiplayer voice game entirely in the browser. Achieved near-instant audio distortion and playback using native APIs. Created phrase packs that made every playthrough unique and genuinely funny. Designed a game that brings people together — literally through laughter.

What we learned How to integrate Colyseus for scalable multiplayer state management. Deep understanding of the Web Audio API for live effects. The importance of UX pacing in multiplayer games — especially when voice and humor are central. That even the simplest idea — “guess who said this?” — can turn into hours of fun when built right.

What’s next for Who Said What? We’re planning to add: AI Voice Filters – mimic celebrities or characters for next-level chaos. Speech-to-Text + Re-Speech Mode – create a “Broken Telephone” chain where each voice gets re-synthesized. Mobile App Port – built with Capacitor for cross-platform play.

💬 Community Phrase Packs – custom themes for parties, schools, or memes.

Our goal? Make Who Said What? the go-to web game for anyone who wants to laugh with friends — and question everything they just heard.