Google Translate conversation mode vs real-time translation in calls

Google Translate conversation mode will not translate your video call. It translates what its own microphone hears in the room and plays the result from the phone speaker, so the person on the other side of Zoom, Meet or Teams never receives it. For a call you need a translator that sits inside the meeting.

How the two voice modes work

Conversation mode is designed for two people sharing one phone. You choose two languages, tap the microphone for yours, speak, and the app shows and reads out the translation. With auto-detect on, it switches direction as speakers take turns. It works well in a taxi, a pharmacy or a hotel lobby.

Transcribe mode is for a long stretch of speech, such as a lecture. It shows a running translation on screen for you to read. At launch it covered a short list of languages (English, French, German, Hindi, Portuguese, Russian, Spanish and Thai). Treat that as the launch list and check the app for the current one.

Both modes share one assumption: the audio source is the air around the phone. Offline does not change that. Language packs allow offline text translation, but conversation mode has needed a connection until now, and the newer offline live translation is limited to a few languages on certain Pixel phones.

Why a phone next to your laptop fails

In a video call, everyone's audio follows a single path: your microphone goes into the meeting, and the meeting comes out of your speakers. Google Translate sits outside that path.

• You can understand them. Hold the phone near your speakers and the app translates what the other person says. That half works, with some delay and errors from room noise.
• They cannot understand you. Your reply reaches them through Zoom in your original language. The phone's translation never enters the meeting.
• Playing the translation aloud makes it worse. Your laptop microphone hears your voice and the phone speaker at once. The other person gets two languages on top of each other, plus echo.
• Headphones remove the echo but not the one-way problem. And then the phone can no longer hear the other person.

The root cause is that the app has no seat in the call. It is not a participant, so it cannot receive the meeting audio or send audio back.

What Meet, Teams and Zoom offer natively

Platforms have been closing the gap, so check your plan before adding another tool.

• Google Meet: translated captions on supported Workspace editions, and spoken translation between English and a few other languages. Only one language pair can be active per meeting.
• Microsoft Teams: translated live captions and transcripts, plus an interpreter agent that translates each participant's spoken audio. Availability depends on the host's license.
• Zoom: translated captions, available on certain plans or as an add-on. Captions are text only.

These work well when everyone is on the same platform with the right plan. They fall short when you join a client's Zoom as a guest, need languages outside the supported pairs, or need more than two languages at once.

What a call translator adds

A call translator joins the meeting as a participant and listens to the meeting's own audio. Akume joins Google Meet, Microsoft Teams or Zoom from an invite link. Live captions appear in its tile inside the call. Anyone who opens the listener link in a browser hears the translation spoken in their language and can read it too, with no account and nothing to install. One meeting handles 2 to 10 languages at once, and Akume detects who speaks which.

Limits worth knowing: spoken translation arrives through the listener link rather than mixed into the meeting audio, and Zoom Webinar is not supported.

Comparison

Yandex Translator and Papago

Yandex Translator's voice mode and Papago's dialogue mode follow the same pattern as Google Translate: one phone, two people taking turns. If you work with Russian-speaking clients, Yandex Translator is the one you will meet most; if you work in Taiwan or with Korean partners, it is Papago. In a call the problem is identical. They are useful on the street or at a counter, but none of them joins a meeting.

Which to choose

• Same room, two people: Google Translate conversation mode is fine.
• Call on a platform whose plan includes translation: use the built-in feature first. To compare tools, see the best real-time translator for calls.
• Client calls on someone else's platform, or three or more languages: use an in-call translator. The article on live translation in Zoom meetings covers the Zoom case in detail.
• A recording you want to read later: transcribe mode or a general AI tool will do, but it is not live. See ChatGPT as a conversation translator for what it can and cannot do, and voice translators for phones and apps for handheld and app-based options.

Scene: the video tour nobody understood