Microsoft Translator's group conversation vs a phone on the table: which one for a real dinner?
Microsoft Translator has had a multi-person conversation feature for years: one person starts a room, the others scan a code and join from their own phones, and everyone reads the conversation in their own language. It is free and genuinely clever. It also asks every single person at the table to join. That is the whole difference.
Microsoft Translator's conversation feature works when every participant joins a shared session on their own device and speaks into it; it then shows each message in every participant's language. GotThat takes the opposite approach: only the person who needs subtitles runs anything, the phone lies on the table, and it listens to the whole group, labelling speakers A, B and C. One is a shared chat room; the other is a listening device.
How the Microsoft approach works
- One person opens the Translator app (or the web page), starts a conversation and gets a code or QR.
- Everyone else opens the app or the web page, enters the code, picks their language and name.
- Each person taps the microphone when they speak, or types. Their words appear for everyone, translated into each reader's language.
- Because each phone knows who is holding it, speaker names are perfect. Because each phone has its own microphone, noise is handled well.
For a planned meeting, a classroom exercise or a multilingual family that is happy to set it up once, this is excellent, and it supports far more languages than GotThat.
Why it rarely survives a dinner
- Everyone has to join. Four coworkers at an izakaya will not scan a QR code and hold their phones for two hours so that one person can follow along. The social cost falls on you, the person asking.
- Everyone has to tap. Speech is captured per device, per turn. Natural overlapping conversation does not map onto “tap, speak, release”.
- The group adapts to the tool. People slow down, speak into phones, wait. The conversation you wanted to join stops being the conversation.
Side by side
| Feature | Microsoft Translator conversation | GotThat |
|---|---|---|
| Who installs or joins | Everyone, on their own device | Only you |
| How speech is captured | Each person into their own phone, per turn | One phone listens to the table |
| Speaker labels | Exact (one device per person) | Automatic, occasionally wrong in noise |
| Natural overlapping talk | Not designed for it | Handled; labels less reliable |
| Languages | Dozens | Subtitles in 5; spoken language auto-detected |
| Catch-up / topic / reply helpers | No | Yes |
| Price | Free | 10 free minutes a day, no sign-up |
| Best for | Planned multilingual meetings where everyone cooperates | Dinners, seminars, club rooms, hostels: situations where nobody will cooperate |
Honest advice
If you can get everyone to join, Microsoft's approach gives you better labels and more languages, and you should use it. If you are the one foreigner at a table of locals who did not sign up for a translation exercise, you need the listening device. That is the case GotThat was built for.
Nothing for the others to scan. Put the phone on the table and press start.
Questions people ask
Is Microsoft Translator's group conversation still available?
At the time of writing (October 2026) the conversation feature is listed on Microsoft's Translator feature page for the app and the web. Check Microsoft's page for the current state.
Can GotThat also give each person their own language?
No. GotThat shows subtitles on one screen, in one language, for the person who needs them. It is a listening tool, not a shared room.
Which is more accurate?
Per-device capture gives Microsoft an edge on speaker names and in noisy rooms. GotThat trades some of that for requiring nothing from anyone else.