Google Meet is expanding its live speech translation from a single language pair to multilingual conversations covering 63 languages. Powered by Gemini, the feature automatically detects the language being spoken—even when a speaker switches languages mid-conversation—and renders translated speech in the language selected by each listener.

The change removes the previous requirement for a meeting host to choose one language combination for an entire call. Translation is enabled individually from the call controls, and participants can choose not to receive translations for languages they already understand. Captions are synchronized with the translated audio, providing a text reference when names, technical terms or fast speech are not interpreted accurately.

Speech Translation is supported in Google Meet on the following platforms:

  • Web
  • Android
  • iOS

People using compatible Google Meet room hardware can also hear translations of remote participants. However, speech captured in the meeting room is not translated for other attendees.

Plans, time limits and exclusions

Access depends on the meeting organizer’s Google plan, although external guests can use the feature when the organizer is eligible. Google says one participant in a direct one-to-one call needs an eligible plan.

  • Up to 90 minutes per meeting: Google AI Pro; Workspace Business Standard and Plus; Enterprise Standard and Plus; Frontline Plus; and AI Pro for Education.
  • Up to eight hours per meeting: Google AI Ultra, AI Expanded Access and AI Ultra Access.
  • Direct one-to-one calls: one participant needs an eligible plan.

The feature is unavailable in live streams, recordings, client-side encrypted calls, phone calls, third-party interoperability sessions and meetings using Adaptive Audio. Google says it does not store participants’ voice data or train its models on that voice data.

As with other live AI translation systems, Google warns that output can misidentify grammatical gender, produce unexpected accents and arrive with several seconds of latency. Meet began beta testing speech translation between English and Spanish in 2025; the broader rollout started October 22 for Rapid Release domains, with Scheduled Release domains due to begin receiving it from November 2 after the Meet mobile app is updated.

SOURCEworkspaceupdates.googleblog.com
Previous articleUS Banks Reportedly Restrict AI Agents After Blocking Some Transactions
Next articleSamsung Galaxy Buds On Could Reach Europe in December, Retail Listings Suggest