文A
How your meeting audio is processed
Oneshot AI Translator joins this meeting to provide live translated captions, so every participant can follow the conversation in their own language. This page explains exactly what happens to the audio.
What is processed
- Speech from meeting participants is streamed, per speaker, to the Oneshot translation pipeline while the translator is active.
- Speech is transcribed (speech-to-text) and machine-translated into the languages participants have selected.
- Silence and background noise are filtered out before leaving the meeting — only detected speech is sent for processing.
What is NOT done
- No video is captured or processed — audio only.
- Captions are generated in real time; the translator does not publish or share meeting content outside the meeting's caption surfaces.
- Audio is processed transiently for translation; it is not used to train models.
How you know it's active
When translation starts, an announcement is posted to the meeting's caption track and to every live-caption view: "Oneshot AI Translator is active: audio in this meeting is transcribed and machine-translated to provide live captions." The translator also appears in the participant list as Oneshot AI Translator when the visible bot joins.
Your choices
- If you do not consent to this processing, tell the meeting host — the host can stop the translator (it stops automatically when the meeting's media stream is closed).
- You choose the language you read captions in; that choice is stored with the name/email you provide in the lobby, only to route your captions.
Questions
Contact your organization's administrator, or Oneshot Translation at support@oneshottranslation.com.