OpenPolySphere

Realtime Call Translator

Translates both sides of a video or voice call live — you speak your language, the other person hears theirs, and vice versa.

OpenSSF Best Practices passing macOS 14+ Linux Windows 10/11 29 languages MIT

How it works: Mic audio → speech-to-text → translation → TTS → routed back into the call. Local mode uses Whisper + Opus-MT by default; cloud backends optional.

Quick start

git clone https://github.com/org-event/OpenPolySphere.git
cd OpenPolySphere
./scripts/bootstrap
cargo run --release -p translator -- setup   # first time: models
cargo run --release -p translator            # open http://127.0.0.1:5050

Use Google Chrome for the call app and the web UI.

Downloads

Pre-built binaries for macOS, Windows, and Linux:

Latest release on GitHub

macOS: macos-arm64 or macos-x64 .dmg → drag to Applications.

Windows: windows-x64-setup.exe installer (or portable .zip). Run OpenPolySphere Setup once to download models.

Linux: .deb (Ubuntu/Debian) or .rpm (Fedora), or portable .zip. Command: openpolysphere setup then openpolysphere.

Links