Google Meet Adds Automatic Speech Translation Across 63 Languages

Google Meet’s speech translation is getting a substantial practical upgrade: participants can listen in their own preferred language without requiring the meeting host to choose one language pair for everyone. The feature supports 63 languages, detects speech automatically and is expanding across the web, Android and iOS. Google Meet hardware devices can play translated speech from remote participants, though room audio has a limitation.

The change addresses a weakness in Meet’s earlier voice translation. That version offered only a small selection of language pairs and applied a single pair to the whole meeting. In a call with people speaking several languages, that setup could leave some participants without a useful translation. Now, listeners can turn on translation for themselves, while Meet identifies the language being spoken.

Translation follows each listener

According to Google’s Google Workspace Updates announcement, the expanded feature is powered by Gemini 3.5 Live Translate. It can handle multiple languages during the same call and adapt when a speaker switches languages mid-sentence. Translation is intended to arrive in near real time, so participants don’t have to wait for someone to change a meeting-wide setting before they can follow along.

Each participant controls the feature using a translation button in Meet’s meeting toolbar. Turning it on affects that listener’s experience; it doesn’t impose one translation choice on everyone else. That makes the setup more useful for multilingual groups, where people on the same call may prefer different languages or may not need translation at all.

Listeners who understand more than one language can add those languages to their preferences. Meet can then leave speech in those languages untranslated, rather than placing a translated voice over words they already understand. When speech translation is active, live captions sync with the translated audio. That pairing gives listeners a text track that matches what they hear, which can help when audio is difficult to follow.

Google describes the feature as Gemini-powered speech translation, but the key change for users is how little setup a mixed-language meeting now requires. They can select their own preferences instead of coordinating a single choice for the whole call. The source doesn’t specify the precise delay between a speaker’s words and the translated audio, so “near real time” shouldn’t be read as a promise of perfectly simultaneous interpretation.

Plan limits and hardware restrictions

Access depends on the meeting organizer’s plan. If the organizer has an eligible Workspace or Google AI plan, everyone in the meeting can use speech translation, including guests from other domains. In a direct one-to-one call, either participant having an eligible plan is enough for both people to get access.

There are time limits. Meetings on a base Workspace tier or Google AI Pro plan can use speech translation for up to 90 minutes. AI Expanded Access, AI Ultra Access and Google AI Ultra raise that limit to eight hours per meeting. Teams planning a long session should check which plan covers the meeting before relying on translation for its full duration.

Meet hardware in a conference room has a specific audio limitation: people in the room can hear translated speech from remote participants, but speech picked up inside the room won’t be translated for people joining remotely. The feature also isn’t available for live streams, recorded meetings, client-side encrypted calls, dial-in calls, third-party interoperability calls or meetings using adaptive audio.

Privacy settings give users a choice about whether their speech can be translated for others. Google says it doesn’t store voice data or use people’s voices to train its models. Those details matter for teams deciding whether to enable translation across an organization, particularly because the administrative setting can be switched on for eligible users.

Availability and account setup

For Workspace administrators, speech translation controls are on by default when the broader Gemini for Meet admin setting is enabled. Admins can manage access by organizational unit or group in the Admin console under Apps > Google Workspace > Google Meet > Gemini settings. For individual users, speech translation starts off at the beginning of each meeting and must be enabled from the toolbar when needed. More details are available in Google’s admin controls for speech translation and Meet speech translation help page.

The rollout for Rapid Release domains began with full visibility starting October 22, 2026. Scheduled Release domains are set to begin a gradual rollout on November 2, 2026, with access taking up to 15 days to reach everyone.

Eligible accounts include Business Standard and Plus, Enterprise Standard and Plus, Frontline Plus, Google AI Pro for Education, AI Expanded and Ultra Access add-ons, and consumer Google AI Pro and Ultra accounts. The feature is coming to Meet on the web, Android and iOS; listening support is also arriving on Google Meet hardware devices. Mobile users should update the Google Meet app so the new toolbar controls appear once the feature reaches their account.

For distributed teams, the upgrade makes Meet translation easier to use across a call with several languages, while the time caps and conference-room audio restriction remain important planning details. Check your account’s plan and device setup before making it part of a long meeting.