BREAKING
Sokuji: Real-Time Voice Translation
Cloud vs Local Inference
Cloud modeAccuracy
OpenAI, Gemini, Palabra.ai
Needs API key + internet
Local modePrivacy
WASM + WebGPU on device
No API key, works offline
On-Device Pipeline
1Speech recognition
2Translation
3Speech synthesis
Plugs Into Meet, Teams, Zoom
Open Source Sets It Apart
Rivals
KUDO, Soniox, LiveVoice
Cloud-focused, closed
SokujiOpen
Fully offline option
Published by Kizuna AI Lab
Balancing Cost and Privacy
AI NEWS BLITZ
An open-source app translates speech in real time, in the cloud or right on your device.