Speak languages you don't know — in your own voice
Two-way interpreting in your cloned voice: you speak Chinese, they hear you speaking English. A glossary locks proper nouns, barge-in keeps it conversational, and one URL adds live bilingual subtitles to OBS.
- Cloned-voice output · it's still you they hear
- Glossary lock · zero term errors
- OBS bilingual subtitles · SRT export
Private deployment · data stays in-house · USDT settlement · verifiable output
One voice, two languages
Hear the Chinese source, then the English interpretation — same cloned voice.
Hard capabilities, verifiable live
Cloned-voice interpreting · term lock
Two-way interpreting that keeps your voice; glossary-locked terminology with a TM cache that speeds up over time; barge-in supported.
Glossary + TM cache · barge-in
OBS live bilingual subtitles
One URL in an OBS Browser Source adds live bilingual subtitles; export SRT afterwards for publishing.
SSE live push · one-click SRT
Tri-engine voice foundation
Built on VoiceX: ten languages in one voice, natural prosody, sub-second first packet that keeps the conversation flowing.
10 languages · sub-second first packet
Meetings, streams and calls
Video meetings, cross-border streams and phone bridges on one pipeline, with a one-tap call package.
One pipeline for all three
Three steps to start
Clone your voice
One 10-40s sample powers both directions of interpretation.
Pick languages + import glossary
Ten languages; hand the engine your term list and proper nouns never break.
Meet, stream or call
Real-time interpretation with barge-in; live subtitles are one OBS URL away.
FAQ
How much latency?
Sub-second first packet keeps a natural pace, and barge-in support means the engine yields instantly when someone cuts in — like a human interpreter.
Which languages?
Ten languages including ZH/EN/JA/KO, both directions, all in your one cloned voice.
Does it work with calls and meeting apps?
Yes — phone bridging with a call package, meeting apps via virtual audio, and live streams get bilingual subtitles on top.
Bring your glossary and try it
30-minute live demo: two-way interpreting with your voice and your term list — data never leaves the room.