Most real-time translation tools still depend heavily on cloud infrastructure.

We wanted to build something different.

PolyTalk is an open-source, privacy-first platform for real-time speech-to-speech translation built for multilingual communication across meetings, live audio, browser tabs, and voice conversations.

Built around: — low-latency voice processing — self-hosted deployment — real-time multilingual communication — flexible open infrastructure teams can adapt to their own workflows

The goal is simple: make multilingual communication feel natural without giving up control over conversations and communication data.

The open-source core officially releases on May 28, 2026.