I went into Microsoft’s documentation expecting the usual answer: that translated captions need Teams Premium for everybody, and a 50-person organisation is therefore looking at $6,000 a year. That’s not what it says.
It says the organiser needs the license, and everyone in their meetings gets translated captions without one. Which for most organisations turns a $6,000 line item into a $600 one. That’s the single most useful thing in this article, so I’ve put it up front.
The rest is what Teams does out of the box, how to switch it on, and where it stops. Everything links to Microsoft’s own docs, and prices are US list, paid yearly, correct as of writing.
What you get without paying anything extra
Live captions. Real-time captions with speaker attribution, for every Teams user, on every plan. Microsoft’s admin guide, Manage transcription and captions for Teams meetings, shows the default meeting policy as “Not enabled, but users can enable”. Captions are off when the meeting starts, and any participant can switch them on for themselves. No add-on anywhere in that sentence.
Spoken-language coverage is broad: around 55 locales, including six flavours of English, both Gulf Arabic variants, German (Germany) and German (Switzerland), Dutch for both Belgium and the Netherlands, and small ones like Maltese, Icelandic and Welsh.
Live transcription. A transcript is not the same thing as captions, and this trips people up constantly. Transcription is a per-organiser and per-user policy, and it’s on by default in new meeting policies. Transcripts get saved next to the recording in OneDrive and SharePoint, and Stream plays the recording with the transcript beside it.
Captions and transcripts are not interchangeable, and the difference bites people at audit time, and there’s a section on it further down.
Human interpretation. Teams also does language interpretation, with up to 16 language pairs, with interpreters on a separate audio channel. This is not machine translation and Microsoft doesn’t pretend otherwise: “Microsoft does not provide interpreters. Organizations will need to hire their own interpreters.” Worth knowing that recordings capture only the original audio, not the interpretation.
How to turn on live captions and translation in Teams
As an admin: Teams admin center → Meetings → Meeting policies, pick your policy (or Global for everyone), set Live captions to Not enabled, but users can enable. Transcription is its own toggle in the same policy. In PowerShell:
Set-CsTeamsMeetingPolicy -Identity Global -LiveCaptionsEnabledType DisabledUserOverride
Set-CsTeamsMeetingPolicy -Identity Global -AllowTranscription $trueWebinars and town halls live somewhere else entirely, under Meetings → Events policies, as Transcription for webinar and Transcription for town hall (-TranscriptionForWebinar, -TranscriptionForTownhall). Both default to on. I’d check them anyway, because “it works in meetings” tells you nothing about your town halls.
As a participant: More actions → Language and speech → Show live captions. On mobile, More options → Turn on live captions.
To translate: with captions running, open Caption settings → Language settings, check the Meeting spoken language is right, then turn on Translate to and pick your language. If the spoken language is wrong you get confident nonsense, and this is far and away the most common reason translated captions look broken.
Does everyone need a Teams Premium license? (Only the organiser)
No. This is the part that’s worth reading twice.
Live translated captions need Teams Premium or Microsoft 365 Copilot. Same for live translated transcripts. But the support page is specific about who: if the meeting organiser has Teams Premium or Microsoft 365 Copilot, all participants can use translated captions without a license of their own. The entitlement follows the meeting, not the attendee.
So you license the handful of people who schedule external events, not the whole company.
How much does Teams Premium cost?
Current US list prices, paid yearly:
| License | Price per user per month |
|---|---|
| Microsoft Teams Essentials | $4.00 |
| Microsoft 365 Business Basic | $7.00 |
| Microsoft 365 Business Standard with Copilot | $23.50 |
| Microsoft Teams Enterprise | $8.55 |
| Microsoft 365 E3 | $39.00 |
| Microsoft 365 E5 | $60.00 |
| Teams Premium (add-on) | $10.00 |
| Microsoft 365 Copilot (add-on) | $30.00 |
So, 50 people on Business Basic at $7. That’s $350/month for the base licenses.
- Teams Premium for everyone: 50 × $10 = $500/month, $6,000/year. The translation costs more than the entire Microsoft 365 subscription underneath it, which is the point at which most people stop reading and go looking for alternatives.
- Teams Premium for the five people who actually organise external events: 5 × $10 = $50/month, $600/year, and every attendee at those events still gets translated captions.
Factor of ten, purely from reading the licensing rule carefully.
The catch (there’s always one): this only works for meetings your people schedule. If a partner organisation hosts and their organiser has no Premium license, your attendees get same-language captions and that’s the end of it. You have no way to fix this from your side.
Teams Premium vs Teams Enterprise after 1 April 2026
Microsoft changed the packaging on 1 April 2026: some features that had been Premium-only moved into Teams Enterprise ($8.55/user/month). Premium kept what Microsoft calls the “intelligence capabilities”, and that’s where live translated captions and translated transcripts live.
Two practical consequences. If you’re buying now, Teams Enterprise does not get you translation, so don’t let anyone talk you into it as a cheaper substitute. And if you bought Premium before that date, Microsoft says existing licenses keep access to everything they previously included until they expire, so what you have and what a new license buys may not be the same thing. Worth checking against your own tenant rather than the marketing page.
Are Teams live captions saved?
No. Microsoft’s wording is “Teams doesn’t save captions.”
Captions and transcripts are separate features and people conflate them constantly:
- Captions are the live text on screen. Per-user, each participant turns them on for themselves, nothing is kept.
- Transcripts are the saved record, stored beside the recording in OneDrive and SharePoint, and only produced if someone started transcription.
If you need text after the meeting, transcription has to be running before it ends. There’s no way to recover captions afterwards.
Live translated captions in Teams town halls
Town halls and webinars are governed by their own policies, and the language handling is more restricted than in meetings.
Organisers of events optimised for large audiences can preselect up to six languages for attendees, rising to ten with Teams Premium. That’s the ceiling. There’s no option to let attendees pick freely from the full list.
Multilingual speech recognition, which lets different people speak different languages, is not supported in town halls at all. If your event has speakers switching between languages, a Teams town hall won’t caption it correctly no matter what you pay.
Where the limits actually are
Around 30 target languages, not 100. Microsoft’s support page lists what captions can be translated into: roughly 30, covering Arabic, both Chinese scripts, the major European languages, Hindi, Japanese, Korean, Thai, Turkish, Ukrainian, Vietnamese. The much larger numbers you’ll find online come from Microsoft Translator’s document translation catalogue, which is a different product with a different list.
Mixed-language meetings are a Premium feature with a much shorter list. By default a Teams meeting has exactly one spoken language. Multilingual speech recognition lets each participant nominate what they’re speaking, so a genuinely multilingual conversation still captions correctly. It needs Teams Premium or Copilot, and it does nine languages: English, Spanish, French, German, Italian, Portuguese, Chinese, Japanese, Korean. It’s not supported in town halls. And the translated captions and transcripts it produces are only available during the meeting. Nothing is kept.
Town halls cap the language menu. For events optimised for large audiences, organisers preselect up to six languages, or ten with Teams Premium. Ten is the ceiling, which is fine for most events and not fine for a genuinely international one.
Captions still aren’t saved. Saying it twice because it’s the assumption that catches organisations at audit time. Captions vanish; only transcripts persist, and only if someone started transcription.
Microsoft reserves the right to throttle. The docs say Microsoft may “restrict or place limitations on the use” of transcription and translation to prevent fraud and maintain performance. No published quota, which also means no contractual floor.
Accuracy has the usual ceiling. Microsoft’s own advice is to speak clearly into the mic, the docs acknowledge that simultaneous speakers reduce accuracy, and profanity comes out as asterisks. Machine translation of machine transcription multiplies two error rates. One person on a good headset: genuinely useful. Four-way panel with cross-talk: not.
Would I rely on it?
For internal meetings where a colleague wants captions in their own language, yes. And if you scope Premium to your organisers it’s cheap enough that I’d struggle to argue against it.
For events, four questions:
- Who’s organising? The entitlement follows the organiser. Meetings hosted by partners, customers or an association won’t have it.
- Is anyone speaking a second language? Then you need multilingual speech recognition. Nine languages, Premium, and not in town halls.
- Do you need the text afterwards? Transcription has to be running before the meeting starts. Captions leave nothing.
- Is this compliance or convenience? A toggle in a meeting client isn’t an accessibility programme.
That last one is where platform captions and dedicated tooling genuinely part company. In-meeting captions are built for people already sitting in the call on their own laptop. Making an event accessible usually means something else entirely: a caption feed on a room display, a transcript people follow on their own phone in their own language, and a record you can point at afterwards. Our use cases section covers what that looks like in practice, and the accessibility statement sets out the standards we hold ourselves to.