Speech translation in video calls isn’t brand new, but Google is giving its meeting platform a major capability boost. The tech giant is expanding Google Meet’s live speech translation feature to support 63 languages with no fixed language pairs required. Powered by the Gemini 3.5 Live Translate model, the updated system automatically detects whichever language a participant speaks across web, Android, and iOS devices.
Instead of restricting users to preset language pairs, Meet now handles any language combination on the fly during live calls. Participants turn on translation from the bottom toolbar, select their preferred listening language, and let the software handle spoken audio alongside live captions. The app can even deliver translated audio using a synthetic voice designed to match the original speaker’s vocal tone.
Privacy controls require explicit attendee consent
Before translating any voice during a call, Google requires each participant to give explicit consent before their audio stream processes. Anyone who declines consent simply stays untranslated in their original voice. Users can also list the languages they already speak fluently so the system leaves those portions of the meeting completely untranslated.
Alongside the 63-language expansion, Gemini takes over post-call organization by generating automated summary notes and action items. The software builds quick bulleted lists of decisions, creates full transcripts with embedded slide screenshots, and routes follow-up assignments straight into Google Tasks and Workspace Studio. These integrated workflow tools help remote teams turn casual conversations into actionable projects without taking manual notes.
Subscription tiers determine maximum translation length
Because the translation engine waits for a speaker to complete a sentence, audio translation arrives with a brief delay of a few seconds. Taking turns during group discussions prevents overlapping voices from tripping up the underlying AI model during fast conversations. However, the expanded language tool requires an eligible host account to unlock. Guests joining from outside domains will still hear the translated audio feed.
Translation limits are tied directly to the organizer’s subscription level, with standard Workspace and Google AI Pro subscriptions capped at 90 minutes per session. Premium AI Ultra subscriptions can have translation access for up to 8 hours for longer international workshops and corporate seminars. The rollout began hitting scheduled-release domains in early November, and global availability is expected to be completed for all eligible accounts by mid-month.
The post Google Meet Expands Live Speech Translation to 63 Languages with Gemini appeared first on Android Headlines.
​Â
