Google Meet gets real-time speech translation support: How it works, availability and more

At its annual Google I/O 2025 developer conference, Google announced that Meet is gaining new live translation capabilities powered by Gemini AI. The feature will enable users to conduct voice conversations across language barriers by translating spoken words into another language in real time.

How it works
Unlike traditional caption-based translation, the new system will translate the speaker's voice into the listener’s preferred language while retaining the original tone, emotion, and speaking style. This is made possible by low-latency processing and natural voice synthesis powered by Gemini AI. During a live demonstration at the event, Google CEO Sundar Pichai showcased a conversation between English and Spanish speakers. The translated voice preserved the nuances and vocal inflexions of each speaker, making the interaction feel more natural.
In practice, a user speaking English will be heard in Spanish by their conversation partner, and vice versa, without either needing to read subtitles. The goal is to create a seamless multilingual video call experience where both participants feel like they’re hearing the original speaker.

Availability
According to Google, the real-time speech translation feature is rolling out in beta to Google AI Pro and Ultra subscribers, with initial support for English and Spanish. The company plans to expand support to additional languages in the coming weeks. Business customers using Google Workspace will be able to test the feature later this year.