Google announced at Google I/O 2025 that it’s bringing real-time speech translation to Google Meet. The feature leverages a large language audio model from Google DeepMind to allow for a natural, free-flowing conversation with someone in a different language, Google says.
Speech translation in Meet translates spoken words into the listener’s preferred language in real time. Voice, tone, and expression are all preserved in the translation.
The tech giant says the new feature has a variety of use cases. For instance, it can be used for English-speaking grandchildren talking to their Spanish-speaking grandparents. Or, it could be used by companies that operate across different regions to allow global colleagues to connect and chat in real-time.
The latency for speech translation is very low, according to Google, allowing for multiple people to chat together, which the company says hasn’t been possible until now.
When the person on the other side speaks, you will still faintly hear their original voice, with the translated speech overlaid on top.
Speech translation in Google Meet will begin rolling out to consumer AI subscribers in beta starting Tuesday. The feature will first be available in English and Spanish, with more languages coming in the next few weeks, including Italian, German, and Portuguese.
Google says it’s building out the speech translation in Meet for businesses, with early testing coming to Workspace customers this year.