Google Meet Live Translation: 63 Languages, Auto-Detection & Plan Rules (2026)

All 63 Languages Supported by Google Meet Speech Translation

Before late October 2026, Google Meet's voice translation only handled a small set of language pairs, and every pair had to pass through English. The October 22, 2026 rollout for Rapid Release domains (and November 2, 2026 for Scheduled Release domains) removed the English requirement. A speaker in Tokyo, a client in São Paulo, and an engineer in Prague can now talk in the same meeting, and Gemini 3.5 translates directly between any combination of the 63 supported languages on web, Android, and iOS.

I grouped the complete official list from the Google Meet Speech Translation help page by region below so you can quickly check your team's languages:

There are a few practical details worth noting in this list. Google separates Cantonese (Yue Chinese) from standard Chinese (Mandarin) in the voice engine, which makes a huge difference for teams working across Hong Kong, Guangdong, and Mainland China or Taiwan. Portuguese is specifically listed as Portuguese (Brazilian) for voice synthesis, and Norwegian is supported via Norwegian Bokmål. If you need a language outside these 63—such as Kazakh, Uzbek, or Zulu—you cannot use voice translation yet, though several of those are available as text-only subtitles under Translated Captions.

---

How Automatic Language Detection ("Zero Manual Setup") Works

In the earlier version of Google Meet Speech Translation, someone had to configure a single meeting-wide language pair (for example, English to Spanish). If a third colleague joined and spoke French, or if someone switched languages mid-conversation, the translation pipeline failed.

With automatic spoken language detection, there is no meeting-wide language setting anymore. Every participant controls their own audio feed independently, and Speech Translation is off by default at the start of every call.

To turn on Speech Translation and let automatic detection handle the rest, follow these steps during a live call:

1. Hover over the Captions & translation icon at the bottom of the Google Meet window and click Translate speech.
2. If it is your first time enabling the feature, select the single target language you want to hear translations in, then click Next.
3. Click Translate me when prompted so other participants in the call have permission to hear your voice translated into their preferred languages.
4. Optional for multilingual users: Open Speech translation settings and add any additional languages you already understand fluently to your exclusion list so Meet does not translate them over the original speaker's voice.
5. Optional for text reinforcement: Click Turn on captions inside the same menu to display synchronized translated captions that match the exact language of the translated audio you are hearing.

Once you turn this on, you do not need to touch the menu again even if four different languages are spoken around the room. Meet detects the spoken language automatically and routes the translated audio only to the listeners who need it.

Preventing Automatic Detection From Translating Languages You Already Speak

One of the most annoying problems in earlier AI translation tools was being forced to listen to a robotic translation of a language you already speak fluently. If you speak both English and Spanish, and you join a meeting with English, Spanish, and Japanese colleagues, you only want the Japanese speech translated.

Google solved this with individual participant exclusions. Inside the translation control panel, you can list additional languages you already understand. When a colleague speaks Japanese, Meet translates their voice into English for you. When another colleague replies in Spanish, Meet recognizes that Spanish is on your known-language list and lets their original voice pass through untouched.

Dealing With the 2-to-3 Second Translation Buffer

Because Gemini 3.5 Live Translate waits for a clause or sentence structure to make grammatical sense before generating audio, there is a built-in delay of two to three seconds. When Speech Translation is active, a small translation badge appears in the corner of your screen and on the active speaker's video tile.

In practice, fast back-and-forth interruptions break the audio flow. If you start talking while the translation icon is still pulsing on the other person's video tile, they are still hearing the end of your previous sentence in their language. Waiting half a beat for that badge to clear before replying keeps the conversation from colliding. Also, if multiple people talk over each other at the same time, language detection can briefly stutter or fall out of sync with the on-screen captions.

---

Translated Captions vs. Speech Translation: Two Different Tools

Google puts both of its translation features under the same Captions & translation button in the bottom toolbar, which causes constant confusion. I often see users think Speech Translation is working when they have actually only turned on Translated Captions, or complain that their translated audio vanished from a recording when audio translation was never designed to be recorded.

Here is how the two features actually differ in day-to-day use:

One more distinction matters if you are migrating from another platform: Google Meet does not have human interpreter channels. On Zoom or Microsoft Teams, enterprise hosts can assign human interpreters to dedicated language audio tracks. Google Meet only offers AI-driven Translated Captions and AI-driven Speech Translation. If you are comparing how platform licensing works across tools, I broke down Zoom's per-user pricing rules in my guide to the Zoom Translated Captions Add-On: Which Plan You Need (2026).

---

Google Workspace and Google AI Plan Rules (90-Minute vs. 8-Hour Caps)

You cannot buy Google Meet Speech Translation as a standalone $5 monthly add-on. Instead, Google bundles it into specific Workspace editions and consumer Google One AI plans, and attaches a strict per-meeting timer to the license.

Which Plans Get 90 Minutes vs. 8 Hours Per Meeting

If your meeting runs past 90 minutes on a standard business plan, Speech Translation stops for that session, though standard text captions remain available. To get a full workday of continuous voice translation, the meeting host needs an Ultra or Expanded Access license:

• 90 minutes per meeting: Google Workspace Business Standard, Business Plus, Enterprise Standard, Enterprise Plus, Frontline Plus, AI Pro for Education, and the personal Google AI Pro subscription.
• 8 hours per meeting: Personal Google AI Ultra subscriptions, plus the enterprise AI Expanded Access and AI Ultra Access add-ons.
• Not eligible (0 minutes): Free personal @gmail.com accounts and Google Workspace Business Starter accounts do not include Translated Captions or Speech Translation.

The Host Rule: How Free Gmail Guests Can Use Live Translation

The most helpful rule in Google's licensing setup is how guest access works. You do not need to buy a paid Workspace seat for external clients, contractors, or overseas vendors just so they can hear you in their language.

• In scheduled meetings (Calendar invites or meeting links): Eligibility is determined entirely by the meeting host (organizer). As long as the person who created the meeting is on an eligible plan like Business Standard or Google AI Pro, every participant in that meeting—including external guests signed into free personal @gmail.com accounts or users from another company's Business Starter domain—gets the Translate speech button in their toolbar and can use two-way voice translation for up to 90 minutes (or 8 hours).
• In direct 1:1 calls: If you place a direct ring call to someone inside Google Meet rather than scheduling a room, only one of the two participants needs an eligible plan for both people to use Speech Translation.

If a free Gmail user creates their own meeting link and invites a paid Workspace Enterprise user into that room, Speech Translation will not appear because the host account lacks the license. Always make sure the paid account generates the calendar invite.

---

Participant Settings vs. Host and Admin Permissions

Even when the meeting host has an eligible Workspace Business Standard or Enterprise plan, participants still run into calls where the button is missing or a specific colleague's voice refuses to translate. That happens because Google splits translation control across four separate permission layers:

Why You Can't Hear a Specific Participant Translated

When three people in a call are being translated smoothly and the fourth person is still coming through in their untranslated native language, check these four culprits in order:

1. The speaker declined or skipped the "Translate me" consent prompt. Because Speech Translation synthesizes a voice resembling the speaker, Google requires per-user consent. If your colleague clicked past the prompt when turning on translation—or hasn't opened the translation menu at all—ask them to hover over the Speech translation badge on their own video tile and click Allow Translation (or go to Speech translation settings > Speech translation and switch from Don't translate me to Translate me).

2. They are speaking through Google Meet conference room hardware. This is a major blind spot for hybrid offices. If half your team is sitting in a physical conference room equipped with Google Meet hardware, the translation only works in one direction. The people inside the conference room can listen to remote participants translated through the room's speakers, but speech picked up by the conference room hardware microphone will not be translated for remote participants outside the room. For a person in a conference room to have their voice translated for remote colleagues, they have to mute the room hardware and speak into their own laptop or mobile phone on a separate Meet connection.

3. They are using Adaptive Audio. Google Meet's Adaptive Audio feature—which syncs microphones and speakers across multiple laptops sitting in the same physical room to prevent echo—is incompatible with Speech Translation. If participants in the same room have Adaptive Audio active, Speech Translation is disabled for their audio feeds.

4. Account or connection restrictions apply. Speech Translation is completely unavailable in Client-Side Encrypted (CSE) meetings, for participants dialing in by phone or third-party SIP/H.323 systems, and for any Google account belonging to a user under the age of 18. Mobile users on Android or iOS also need the updated late-October 2026 version of the Google Meet app for the 63-language selector to appear.

---

What Happens After the Call: Recordings, Transcripts, and Saving Text

Google's privacy rule for Speech Translation is strict: no audio is saved, and no AI models are trained on your voice. While that makes legal and compliance teams happy, it creates a serious workflow gap if your team relies on post-meeting records.

The moment you hang up a Google Meet call, the voice translation is gone forever. Specifically:

• Meeting recordings only keep the original voice: If you record a multilingual meeting to Google Drive, the MP4 video file only records the raw, untranslated audio spoken by each participant. The synthesized Gemini translation track is never mixed into the recording.
• Google Meet Transcripts and AI notes only capture one language: Many teams assume that turning on Google Meet Transcript: Plans, How to Start One, and Where It's Saved or Gemini's Take notes for me alongside Speech Translation will produce a translated Google Doc in Drive. According to Google's official documentation, neither Transcripts nor "Take notes for me" works across the 63 Speech Translation languages—they only capture speech spoken in the single primary language configured for the transcription tool.

If you need a written record of translated speech from a Google Meet session, you have two workarounds during the live call:

First, if you are recording the video, use Translated Captions and check the Record captions box before starting the recording; that embeds the translated subtitle text track into the recorded video file even though the audio track stays in the original language. Second, if you are an attendee without recording permissions and need to keep the translated text before the browser tab closes, you can capture the caption stream locally on your machine using the methods I tested in How to Save Live Captions on iPhone, Mac, Zoom and Teams.