Voice
Follow a conversation with live voice translation
Speak, review recognized speech and translation, then play the result aloud when supported.
Looking up every sentence interrupts a conversation. The voice workflow is designed to capture speech, show the translation, and support turn-by-turn communication from one session screen.
Set the conversation languages
Choose the source and target languages for the conversation. Automatic source detection is available in supported configurations; check the selected language and recognition engine before starting.
Speak and review each turn
The session displays recognized speech and its translation as the conversation proceeds. Microphone access and a suitable audio environment are needed, and recognition quality varies with noise, accents, and speaking style.
Voice features depend on the setup
Speech recognition, translation, and spoken playback may use different engines. On-device operation depends on compatible models being available, and some live voice capabilities are marked beta in the app.