At the Google I/O conference in 2025, Google announced that it would introduce real-time voice translation to Google Meet. The company stated that this feature utilizes large language audio models from Google DeepMind, enabling natural and smooth conversations between users speaking different languages.

The voice translation feature in Meet can instantly translate spoken words into the listener's preferred language while preserving the original voice, tone, and expression.

This tech giant mentioned that the new feature has various application scenarios. For instance, it can help English-speaking grandchildren communicate with Spanish-speaking grandparents or be used by companies operating across regions so global colleagues can connect and interact in real time.

According to Google, the latency of voice translation is extremely low, allowing multiple people to chat simultaneously—a capability that was previously unattainable.

While the other person is speaking, you can still faintly hear their original voice, with the translated speech overlaid on top.

Voice translation in Google Meet will roll out in beta to consumer AI subscribers starting Tuesday. Initially available for English and Spanish, more languages will be added in the coming weeks, including Italian, German, and Portuguese.

Google stated that it is developing the voice translation feature in Meet for businesses and will conduct early tests with Workspace customers later this year.