The problem
The client needed reliable multi-party voice calling with live transcription, on a platform that could scale without heavy per-call infrastructure overhead — and without building telephony signaling from scratch.
What we built
Built on Twilio for SIP/VOIP signaling and media handling, with a captioning pipeline layered on top of live audio streams. The backend tracks multi-party call state — who joined, who's speaking, call events — and streams captions to the frontend as event-driven updates instead of polling, keeping latency low as participant count grows.
The result
Shipped as a production calling platform with live captions, running as one of three core platforms owned end-to-end by a 5-engineer team.
Delivered by NovaBuildz engineers before the studio was founded — the same team now works under one roof.
Start a similar project