If your team runs weekly calls with a Japanese client, you have probably seen this happen. The client side discusses a point in Japanese, someone summarizes it in English, and the detail that mattered (a date, a number, a condition on the estimate) gets lost on the way. Live translated captions fix part of that by putting every sentence on your screen in your language while the speaker is still talking. This guide covers what the built-in caption features in Microsoft Teams, Zoom and Google Meet need before they work, what to check on privacy and screen sharing, and how MeetTune, the live meeting translation app MOHA builds, handles the same call.
A note on who is writing this: MOHA Software is a Vietnamese software company in Hanoi with a Japanese entity, MOHA Japan, in Yokohama, and many of our published case studies are projects for Japanese clients. MeetTune is our own product, so read the product section with that in mind. The comparison with the built-in features is based on Microsoft, Zoom and Google's own help pages, linked below.
What live translated captions do in a client call
Plain captions turn speech into text in the language that was spoken. Translated captions go one step further and show that text in another language. In a call with a Japanese client, that means you read English (or Vietnamese, or any language you choose) while the client speaks Japanese.
That helps in three places where meetings across a language gap usually go wrong:
- You catch numbers and dates the moment they are said, instead of in a summary later.
- You can follow a side discussion on the client's side without waiting for someone to interpret it.
- You have a written record to check against when the minutes go out.
Captions do not replace a human interpreter for a contract negotiation. For the weekly sprint review, a requirements call or a status update, they are often what decides whether your team understood the request on the first pass.
Built-in options in Teams, Zoom and Google Meet
All three platforms have translated captions. The catch is that each one ties the feature to a paid plan, and in most cases to the account of whoever runs the meeting.
| Platform | What the vendor's help page says is required | Who turns it on |
|---|---|---|
| Microsoft Teams | Live translated captions in meetings are part of Teams Premium and Microsoft Copilot | If the meeting organizer has one of those licenses, every participant can use translated captions |
| Zoom | A Zoom Workplace Business Plus or Enterprise account, or the Translated Captions add-on; automated captions must be enabled | An account owner or admin enables it; the host sets the caption languages, and participants can switch on captions themselves |
| Google Meet | A supported Google Workspace edition, such as Business Standard, Business Plus, Enterprise Standard or Enterprise Plus | Turned on from the Captions menu in the meeting, choosing "Translate from" and "Translate to" |
Sources: Microsoft's guide to live captions in Teams meetings, Zoom's article on enabling and configuring translated captions, and Google's page on translated captions in Google Meet. Japanese appears in the supported languages on all three.
Why the organizer's plan matters more than yours
With Teams, the license that counts is the organizer's. With Zoom, the host decides which caption languages are available. So if your Japanese client sends the invite from their own tenant, whether you get translated captions depends on their licenses and their settings, not on what your company pays for.
For a vendor team, that is the normal case. The client often owns the meeting. You can ask them to change their setup, but most teams would rather not open a call with a request about licenses.
The built-in features also stay inside one platform. If one client uses Teams and another uses Zoom, your team learns two caption setups, and an in-person visit to the client's office gets neither.
Five questions to ask before you choose a caption tool
Third-party tools fill the gap, and they work in very different ways. These are the questions that matter most for client meetings.
Does a bot join the call?
Some meeting assistants join as an extra participant. The client sees an unfamiliar name in the attendee list and may have to admit it. In a meeting with a Japanese client, an unannounced extra attendee is a bad start. A tool that listens to the audio on your own computer avoids that, because nothing new appears in the meeting.
Does it show up when you share your screen?
You will share your screen in most client calls: a demo, a ticket board, a spec. If the caption window appears in the shared view, the client sees your translations and notes. Check whether the tool keeps its window out of screen shares and recordings.
Where does speech recognition run?
Speech recognition turns audio into text. Some tools send the raw audio to a server; others do this step on your computer and send only text onward. If your client has rules about where meeting audio may go, this is the question their security team will ask.
Can it handle a call that switches language?
Calls with Japanese clients rarely stay in one language. The client side speaks Japanese, your PM answers in English, and an engineer adds a point in Vietnamese. A tool that can only translate one fixed language pair will miss parts of that conversation.
What do you have when the call ends?
Captions disappear when the meeting does. What your team needs afterwards is minutes: what was decided, who owns each action item, and the transcript to check a quote against. Ask where those records are stored and who can read them.
How MeetTune handles a call with a Japanese client
MeetTune is a desktop app for Mac, Windows and Ubuntu that gives you live translated captions for Teams, Zoom, Google Meet, FaceTime and any other app that plays sound on your computer. Here is how it answers the five questions above, based on the MeetTune product page.

Nothing joins the meeting
MeetTune works from the sound on your own computer. It listens on two channels: the computer's audio for the other side and your microphone for you, so their words and yours stay apart in the transcript. The client does not install or accept anything. Because it does not depend on the meeting platform, it works the same way whether the client's invite is a Teams link or a Zoom link.
It stays off your shared screen
On Mac and Windows, the MeetTune window and its floating captions stay out of screen shares and screen recordings. You can share a demo while reading the translation of the client's questions on the same screen.
Speech recognition runs on your computer
Speech becomes text on your own Mac or PC. Only the text goes to the AI for translation, key points, replies and minutes. That AI is Claude, from Anthropic. You can run it on your own Claude Pro, Max or Team plan (MeetTune ships with Claude Code, so there is no API key to manage), or through MeetTune's AI server with a monthly allowance set on your activation code. Transcripts, minutes and recordings are saved on your computer, with an optional backup to your own Google Drive.
What you see during the call
MeetTune translates each sentence while the speaker is still talking. The translation is set large and the original small, so you can still check the exact Japanese words. It tags questions for you, decisions, dates and numbers as they come up.
Say the client asks, in Japanese, whether your team can send a new estimate by next Friday. MeetTune shows the question in your language, tags the date, and drafts an answer in Japanese with the meaning in yours. You can also type or say your answer in your own language and MeetTune gives you the sentence to say.
A side panel lets you ask Claude about the call so far: who said what, what a term means. You can pin a line as important or note a question to ask the other side, then turn that question into their language with one click.
Languages
MeetTune covers ten meeting languages: English, Japanese, Korean, Chinese, German, French, Spanish, Italian, Portuguese and Vietnamese. For calls that switch language, turn on automatic detection when you start the meeting; MeetTune recognizes each sentence in its own language and translates it into yours. When everyone speaks your language, it skips translation and still takes the key points and writes the minutes.
Minutes when the call ends
When the meeting ends, MeetTune writes the summary, the decisions and the action items with owners, in your language. You can export the minutes to Word or copy them as Markdown and send them to the client. Every meeting stays in a searchable library on your computer, and you can ask Claude about decisions, numbers and who promised what weeks later. Click a line of the transcript to hear that moment of the recording again.

Client visits and meetings in the room
Not every meeting is online. For a client visit, a workshop or an interview, pick "In person" when you start. MeetTune then listens through your microphone only, so put the laptop where it can hear everyone at the table. Captions, key points, suggested replies and minutes work the same way as in an online call.
Between meetings: chat with the client
Much of the work with a Japanese client happens in chat and email, not in calls. Outside meetings, MeetTune sits in the menu bar. Select text in any app and press a shortcut:
- Option + Shift + T rewrites your message in the other person's language.
- Option + Shift + R shows a message you received in your language.
- Option + Shift + G corrects your grammar without changing what you meant.
On Windows the shortcuts use Alt + Shift. They work in apps such as Slack, Microsoft Teams, LINE, Gmail and Outlook. You can add your project's terms once, so translations use the same words your team and the client already use.
Live translated captions with no bot in the call
MeetTune works with Teams, Zoom, Google Meet and meetings in the room, on Mac, Windows and Ubuntu.
Limits to know before you start
A fair guide covers what a tool does not do. From the MeetTune page:
- It needs an internet connection. Speech recognition runs locally, but translation and minutes come from Claude.
- On Windows, speech recognition uses Whisper, so captions arrive a little later than on a Mac. It is fastest on macOS 26, about a second behind the speaker.
- The Mac app needs Apple silicon and macOS 13 or later; the Windows app needs Windows 10 or 11, 64-bit; the Ubuntu package also installs on Debian.
- Recording laws and your company's rules still apply. Tell people when you record a meeting.
- In browser versions of Meet and Teams, real speaker names need the MeetTune Chrome extension.
If your team works with Japanese clients and wants a wider view of the working relationship, our guide to IT outsourcing for Japanese businesses covers communication style and quality expectations. For how English proficiency differs between offshore locations, see our Vietnam vs India vs Philippines comparison.
Getting MeetTune
MeetTune runs with an activation code, issued for a person or a team. To get one, use the request form on the MeetTune page and tell us who will use it. Install the app, open it and enter your code; updates download on their own. If you have questions about using MeetTune with your team, contact us.


