A multilingual meeting no longer needs one translation setting to suit everyone. Google Meet’s expanded speech translation supports 63 languages and lets each participant select a listening language. Automatic detection follows the language being spoken, including changes during the conversation.
The translation choice belongs to the listener
Participants turn translation on from the meeting toolbar. They can also specify languages they already understand, leaving those untranslated. Captions follow the translated audio. That creates a practical distinction: people share the same conversation without having to share a single translation preference.
Three people may own three different settings
The host needs an eligible Workspace or Google AI plan. An organisation’s administrator manages Gemini for Meet availability. Each participant then enables translation for their own use. If the control is missing, checking those responsibilities in that order gives the team a clearer troubleshooting path.
Guests can use translation in a meeting hosted by an eligible user. For a direct one-to-one call, eligibility on either side is sufficient. Depending on the plan, translation is available for up to 90 minutes or eight hours per meeting.
Check the meeting format as well as the account
Support covers the web, Android and iOS. Meet room hardware can play translated speech from remote participants, but speech originating in the room is not translated for everyone else. Recording, livestreaming and certain specialised call formats are outside the supported scenarios. Scheduled Release rollout began on November 2 and can take up to 15 days.
Finley’s assessment
There’s a simpler handoff here. A short invitation note can tell people who checks access and where to choose their listening language. Making those responsibilities explicit should save the first few minutes of a call for the discussion itself. The useful automation is the language detection; the useful human decision is how each person wants to listen.
