A multiplayer step sequencer with polyrhythmic patterns, built for real-time collaboration. Create/Collaborate. Remix. Share.
Keyboardia — Multiplayer Music in the Browser
A real-time collaborative step sequencer with polyrhythmic patterns, built on Cloudflare Durable Objects and WebSockets.
About Keyboardia
Keyboardia is a browser-based step sequencer where up to 10 players can jam together on the same session. Each player adds tracks, sets patterns, and hears the result in real time. Every change syncs across all connected browsers through WebSockets.
The sequencer supports polyrhythmic patterns: each track can have a different step count (from 3 to 128), so you can layer a 7-step melody over a 16-step drum pattern. The app ships with 64 sound generators, including 32 Web Audio synths, 11 Tone.js FM/AM synths, and 21 sampled instruments (grand piano, 808 kit, acoustic drums, vibraphone, and more). An effects chain with reverb, delay, chorus, and distortion syncs across all players.
Sessions can be shared via link or QR code. You can remix other players’ sessions or publish your own as immutable snapshots.
Frameworks & Technologies Used
- Frontend: React 19 with Vite (sequencer UI and controls)
- Audio: Web Audio API and Tone.js (synthesis, sampling, effects)
- Backend: Cloudflare Workers with Durable Objects (session state and sync)
- Real-time: WebSockets with Hibernation API (cost-efficient persistent connections)
Why Cloudflare for This Project?
- Durable Objects for Session State: Each jam session is a single Durable Object instance. Mutations save to DO storage immediately, and KV gets updated on disconnect. No external database needed for real-time collaboration.
- WebSocket Hibernation: The Hibernation API keeps WebSocket connections alive without burning CPU while players are idle. The Durable Object wakes up only when a message arrives, keeping costs low for long sessions.
- KV for Session Persistence: Published sessions are stored in KV for fast global reads. Anyone with the link can load a session from the nearest Cloudflare data center.