Live translated captions for the whole audience: what Zoom, Teams and Google Meet require, and how each attendee reads in their own language (2026)

An all-hands, a webinar, a partner briefing, a town hall. One person talks; the audience is in ten countries. Everyone asks for the same thing — let them read along in their own language — and every platform answers it, so the feature looks settled.

It is not settled. The three platforms answer a different question each, and the difference only shows up when the event starts:

• Who decides which languages the audience may read in — the organizer, in advance, from a capped list; or each attendee, at any moment, for themselves.
• What the host has to own for the attendee to see anything: a licence, an edition, an add-on.
• What is left of it afterwards. Captions are live text. On most platforms they stop existing when the event does.

This guide lays out what each vendor's own documentation states, checked September 2026, then how the same job is done in InterMIND. Every external fact links to the vendor's page at the end; nothing here is our test of someone else's product.

Microsoft Teams: the organizer preselects the languages, and Teams Premium is on the organizer

Microsoft calls it live translated captions. From the Teams admin documentation: "Live translated captions allow your users to see captions translated into the language they're most comfortable with. By default, live captions are displayed in the language spoken during a meeting or event."

For an audience event, the important sentence is the next one: "For events optimized for large audiences, such as town halls, organizers can preselect up to six languages for attendees to use during the event. With a Teams Premium license, this limit increases to 10 languages."

And the licence sits on the organizer's side, not the attendee's: "The meeting or event organizer must have a Teams Premium license for attendees to use live translated captions." An admin also has to leave the door open — live translated captions require Live captions set to "Not enabled, but users can enable" in the meeting policy; setting Live captions to Off turns translated captions off too.

Microsoft is also explicit about what survives: "Live captions provide real-time subtitles during the meeting, but Teams doesn't save captions." The saved artefact is the transcript, which is a separate policy, stored in OneDrive and SharePoint.

Zoom: the host's account settings fix the language pairs, the attendee picks inside them

Zoom's documentation describes translated captions as a per-attendee choice inside a set the host has already fixed: "Available caption languages are determined by the host in web settings before the live session, but participants can freely enable captions and select the language they want to use for translation without the need of the host."

Two facts follow from that sentence, and both are stated on Zoom's pages. The attendee is autonomous inside the set — "Any participant in the meeting or webinar can self-enable captions, as they are not dependent on the host to enable", and "Other participants can choose a language for their captions independently of your chosen language". But the set itself is an account setting: "Language pairs (the input speaking language and the output caption language) available to participants are determined by the host's settings", configured under In Meeting (Advanced) → Translated captions → Edit translation languages, and lockable by an admin for the whole account.

The entitlement is on the host: "Translated captioning requires the host to be a member of a Zoom Workplace Business Plus, Zoom Workplace Enterprise Essentials, Zoom Workplace Enterprise Plus, or Zoom Workplace Enterprise Premier account, or assigned the Zoom Translated Captions add-on." Zoom's page adds two limits worth knowing before you promise anything: "Some accounts in regulated industries like Healthcare and Higher Education may not have access to translated captions", and dialects do not translate into each other — "such as translating between French (France) and French (Canada)".

Zoom lists 35 languages that translate between most of each other, plus Catalan and Welsh as translation targets only, and Bengali and Greek with restricted pair lists.

Google Meet: each attendee picks, if the organization is on the right edition

In Google Meet the choice is the attendee's from the start: in Settings → Captions the participant selects the Language of the meeting, turns on Translated captions, and then selects "the language you prefer to translate into". Google also keeps the scrollback: "If translated captions are turned on, you can review the translated captions of the previous parts of a conversation."

The gate is the edition. Google's page: "Translated captions are currently only available for the following Workspace editions" — Business Standard, Business Plus, Enterprise Starter (listed as available until 30 June 2025), Enterprise Standard, Enterprise Plus, and Google AI Pro for Education. Google's list of languages captions can be translated to or from runs to roughly seventy entries, with the note that "This feature is gradually rolling out and may not be available yet."

The axis: who owns the language list, and what is left the next morning

Put the three side by side and the same two lines separate them.

Who owns the list. In Teams, the organizer owns it and it is capped at six languages, ten with Teams Premium — an attendee whose language was not on the list has no option to add it during the event. In Zoom, an account admin owns the pair list; inside it, the attendee is free. In Google Meet, the attendee owns the choice, and the organization owns the edition that makes the menu appear at all.

What is left afterwards. Teams states it plainly — captions are not saved; the transcript is a different feature under a different policy. Zoom's translated captions are a live view, with recording and transcripts as separate settings. Google Meet keeps a scrollback inside the session. In every case the thing your audience read is a live overlay, and the record that survives is in one language: the one the meeting was held in.

That second line is where a translated-captions feature and a multilingual communication space stop being the same product. Captions solve the hour. They do not solve the week after it, when someone who read along in Portuguese opens the recap and finds English.

InterMIND: the caption language is a setting on the attendee's side, and the room is translated by voice too

In InterMIND there is no caption-language list for an organizer to prepare. Each participant has a translation language of their own, and the room renders to it:

• The attendee sets it, including guests who never sign in. A signed-in user sets it in Settings; someone opening a meeting link without an account sets it on the join screen, where it defaults to their interface language — which itself defaults to the browser's. The picker holds the full list of 23 meeting languages (30 for documents, a separate pipeline).
• Subtitles are on by default. subtitlesEnabled and translationEnabled both default to true in the device settings, so an attendee who changes nothing still sees the speech in their own language; the subtitle field can be dragged, closed and re-enabled per device, and right-to-left languages render right-to-left.
• Reading is not the only channel. The same translated stream is voiced on each participant's device, so the audience hears the talk in their language as well as reads it. Nobody in the room has to be told which languages were "made available" — the language is the listener's setting, not the event's configuration.
• The size of the audience is the host's plan limit, not a separate webinar product: 50 participants on Basic, 100 on Pro, 300 on Business, 1500 on Enterprise. There is no per-language surcharge and no interpreter booth to staff.
• After the event, the record is multilingual too. Every signed-in participant gets the recap in their own language, and a channel's whole history is read by each member in theirs. One honest limit, worth stating because it is the opposite of a marketing line: an attendee who joined anonymously as a guest gets full translation in the meeting and nothing after it — there is no address to deliver a recap to and the anonymous identity is swept. An account is free; it is what makes the after-the-event half reachable.

The comparison, in one table

Running a multilingual all-hands without surprises

1. Count the languages you actually need before you pick the platform. Six is a real cap on Teams town halls without Premium; if your audience spans more, the list is the constraint, not the feature.
2. Check who holds the licence. On Teams and Zoom the entitlement is on the organizer's or host's account, so a colleague scheduling the event from an unlicensed account silently removes the feature for everyone.
3. Check the pairs, not just the languages. Zoom's unit is a pair — source language plus caption language — and dialects of one another are excluded.
4. Tell the audience where the switch is. On all three platforms it is the attendee who turns captions on; an audience that does not know the menu exists reads nothing.
5. Decide what happens the next morning. If the recap, the notes and the follow-up chat go out in one language, the multilingual part of the event lasted exactly as long as the event.
6. Rehearse with one speaker in each source language. Automatic captions are generated from speech, and every vendor says accuracy varies; Zoom states plainly that "translated captions may not be accurate."

See it instead of reading about it

• Try the live demo — no signup: speak, and hear the room translated per listener.
• See the benchmark — per-pair translation quality measured monthly on real traffic, with the methodology public.
• Read the language breakdown — which list applies to voice, chat, documents and the interface, and why they differ.

FAQ

What are live translated captions? Captions generated from the speaker's speech and translated in real time into another language, shown as an overlay during the meeting or event. They are distinct from a transcript, which is a saved record, and from interpretation, which is audio.

Do all attendees see the same caption language? No, on all four products discussed here the language is chosen per attendee. What differs is the menu: Microsoft Teams offers, for large-audience events, only the languages the organizer preselected — up to six, or ten with Teams Premium; Zoom offers the language pairs configured in the host's account settings; Google Meet and InterMIND let the attendee pick from the platform's full list.

Does the audience need a paid account to read translated captions? The entitlement sits on the host's side. Microsoft requires a Teams Premium licence on the meeting or event organizer. Zoom requires the host to be on Workplace Business Plus, Enterprise Essentials, Enterprise Plus or Enterprise Premier, or to hold the Translated Captions add-on. Google Meet requires the organization to be on one of the listed Workspace editions. In InterMIND joining needs no account at all.

Are translated captions saved after the meeting? Microsoft states that "Teams doesn't save captions" — the saved artefact is the transcript, a separate feature with its own policy and storage. Zoom's translated captions are a live view; recording and transcription are separate settings. Google Meet lets you scroll back through translated captions within the session. In InterMIND the post-event record itself is multilingual: every signed-in participant gets the recap in their own language.

How many languages do translated captions support? Zoom lists 35 languages that can be translated between most of each other, plus Catalan and Welsh as translation targets only and restricted pair lists for Bengali and Greek. Google lists roughly seventy languages to or from, noting the feature is rolling out gradually. Microsoft publishes its language list per feature. InterMIND covers 23 languages for meetings and 30 for document translation.

Can the audience hear the talk in their language, not just read it? Translated captions are text on all three incumbent platforms. Audio is a different feature there: Microsoft's language interpretation "allows professional interpreters to convert what the speaker says into another language in real-time", enabled and staffed before the meeting starts; Zoom's lets a host "designate up to 20 participants as language interpreters", each providing their own audio channel, on a Pro, Business, Education or Enterprise account and only in scheduled meetings with an automatically generated ID. In InterMIND the translated stream is voiced on each participant's device by default, and the subtitles are the same stream written down.

What is the difference between translated captions and interpretation? Captions are machine-generated text produced from speech and translated on the fly. Interpretation is audio: on Teams and Zoom it is people — interpreters the organizer invites or designates, one channel per language, with the attendee selecting a channel. That is why the caption feature is often switched on while the interpretation feature is not: one is a licence, the other is a roster.

Sources: Microsoft Learn — Manage transcription and captions for Teams meetings, Zoom — Viewing captions in another language, Zoom — Enabling and configuring translated captions, Google Meet — Use translated captions, Microsoft — Use language interpretation in Teams meetings, Zoom — Using Language Interpretation in your meeting or webinar. Vendors change licensing, plans and language lists over time; check their pages for the current state. All facts checked September 2026. InterMIND facts: per-surface language breakdown, live benchmark.