Live captions vs. spoken translation: which to use
Live captions show speech as text in the same language; spoken translation says it aloud in another. See who each helps and which to pick per meeting.
Published · 6 min read
In short
- Live captions turn speech into text in the same language. Translated captions put that text in another language. Spoken translation says it aloud in the listener's language.
- Only text works for deaf and hard-of-hearing participants, so audio-only translation is never enough on its own. Listeners who cannot read along comfortably, or who are driving or presenting, need audio.
- Phone and browser captions (Android, iOS, Chrome) are built for one device and mostly one language. For a mixed-language meeting, use a platform's translated captions or a translator that gives text and voice together.
In this article
Live captions show what a speaker says as text, in the same language, while they are talking. Spoken translation says it aloud in another language. In between sit translated captions: the same text, shown in a different language. Pick text when someone cannot hear, audio when someone cannot read along, and both when the room is mixed.
When deciding between live captions vs translation—or more precisely, live captions vs translation and captions together—the choice depends on who needs what. The three options are often sold as one feature, which is why a "captions" setting in a video app rarely answers the question you actually have: can everyone in this meeting follow it? Below is how they differ, what built-in phone and browser captions can and cannot do, and what to choose for each type of meeting.
What each option does
Live captions (same language)
Speech is turned into text in the language being spoken. Nothing is translated. Automatic captions come from the device, the meeting app or a bot. CART (Communication Access Real-Time Translation) is the human version: a trained captioner types a verbatim transcript, and it is the standard when accessibility is a formal requirement.
Who benefits: deaf and hard-of-hearing participants, people in noisy rooms, and anyone following a speaker with a strong accent or fast delivery in their second language.
Translated captions
The same text, shown in another language. The listener reads. This is what Google Meet, Microsoft Teams and Zoom offer on their paid tiers, and what Chrome and some Pixel phones offer on-device.
Who benefits: people who read a language more comfortably than they hear it, and deaf or hard-of-hearing participants whose meeting is in a language they do not use. For them, translated text is not a convenience. It is the only way in.
Spoken translation
A voice says the translation in the listener's language, usually in earbuds or on a second audio channel while the original speaker keeps talking. Human simultaneous interpreters do this at conferences. AI translators do it on any call. See how a voice translator works for the mechanics.
Who benefits: people who are driving, presenting, taking notes or simply tired of reading, and listeners with low literacy or dyslexia. It does not help someone who cannot hear it.
Device-level captions: what they can and cannot do
Phone and browser captions are free, private and on-device. They are also built to caption audio playing on that one device, not to run a multilingual meeting.
- Android Live Caption. Captions media and calls on the device and, per Google, processes audio and captions locally. Language coverage depends on the phone, with the widest on Pixel 6 and later. Translation (Live Translate) is limited to Pixel 6 and later. Typing replies in calls is same-language only.
- iPhone Live Captions. Part of Apple's accessibility settings, it captions audio from apps and FaceTime. Apple's own translation features for calls depend on device, iOS version and language, so confirm on Apple's page before promising it to a client.
- Chrome Live Caption. Captions audio played in the browser, which includes a web meeting. Google lists a set of caption languages, and translating captions currently works on Windows, Mac and Chromebook Plus, not on every device.
What all three share: the captions are visible only to you, they are not part of the meeting, and nobody else gets a translation because you turned them on. They are a good accessibility baseline for one person and a poor tool for a team.
Platform captions and translation in Meet, Teams and Zoom
- Google Meet offers translated captions on selected Workspace editions. Each participant has to switch them on in caption settings.
- Microsoft Teams offers live translated captions in meetings when the organizer has Teams Premium or Microsoft Copilot. Microsoft lists 50+ spoken languages and a smaller set of translation languages.
- Zoom offers translated captions as an add-on or on higher plans. For spoken translation it supports a separate human interpreter audio channel.
All of this is text. None of it produces a translated voice you can listen to on your own, which is where a dedicated translator comes in. Our guides on real-time voice translation and English-Spanish live translation cover the voice side.
Comparison
| Live captions | Translated captions | Spoken translation | |
|---|---|---|---|
| Output | Text, speaker's language | Text, listener's language | Voice, listener's language |
| Helps deaf and hard-of-hearing | Yes | Yes | No, on its own |
| Reading load | High | High | None |
| Eyes free | No | No | Yes |
| Two-way conversation | Not needed | Each side needs their own language shown | Each side hears their own language |
| Built-in on Meet, Teams, Zoom | Yes | Paid tiers or add-ons | No (a human interpreter channel, or an external translator) |
| Good for | Accessibility baseline | Mixed-language readers | Hands-free, long calls |
In a two-way conversation the question changes from "what output?" to "who needs which output?". A deaf designer and a hearing client who speak different languages each need a different thing, and one shared captions window serves neither well. That is the scene in the example below.
How a translated call goes
workshop with a deaf designer and a Portuguese-speaking client
Scripted example, not a recording
Speaker 1 (English)Before we sign off the layout, can the client tell us which screens worry them most?
Antes de aprovarmos o design, o cliente pode nos dizer quais telas mais o preocupam?
| Said | Translation | |
|---|---|---|
| Speaker 1 (English) | Before we sign off the layout, can the client tell us which screens worry them most? | Antes de aprovarmos o design, o cliente pode nos dizer quais telas mais o preocupam? |
| Speaker 2 (Portuguese) | A tela de pagamento. Os botões parecem pequenos demais no celular. | The payment screen. The buttons look too small on a phone. |
| Speaker 1 (English) | Understood. We will enlarge them and send a new version today. | Entendido. Vamos aumentá-los e enviar uma nova versão hoje. |
What to pick for each meeting type
- Internal meeting, one language, a deaf or hard-of-hearing colleague. Same-language live captions. Use CART if accuracy is a formal requirement. Automatic captions are fine for most daily work.
- Client call, two languages, nobody with a hearing need. Spoken translation, with captions on as a safety net for names and numbers. An example: Italian-English calls rely heavily on the voice because people stay on camera and do not want to read.
- Same call, one participant with a hearing need. Captions, in whichever language that person reads, plus voice for everyone else. Audio-only translation fails them.
- Webinar or all-hands with a large audience. Captions for the whole room in the speaker's language, plus translated captions per language. Voice for people who cannot read along.
- Workshops, interviews and research. Text, because a transcript is useful afterwards and exact wording matters. Add voice only if the guest asks for it.
The rule behind all five: choose the output by what each participant can do, not by what the platform calls the feature.
Compare them in the live demo
Reading about the difference takes longer than hearing it. In the free live demo you can speak, watch the text appear, and hear the translation, then decide which format your own meetings need.
FAQ
Are live captions the same as translation? No. Live captions write down what is said in the same language. Translation changes the language, as text (translated captions) or as speech (spoken translation). A meeting can use one, two or all three at once.
Which one do deaf and hard-of-hearing participants need? Text. Same-language captions if the meeting is in their language, translated captions if it is not. Spoken translation alone excludes them, so offer text whenever you offer audio translation.
Can Android Live Caption or iPhone Live Captions translate a meeting? They are made to caption audio playing on your own device. Translation is limited: Google offers it on some Pixel phones, and Apple's translation features depend on device, app and language. Check the vendor's current list before relying on either for a client call.
Can I use captions and spoken translation in the same meeting? Yes, often the safest setup for mixed audiences. Meet, Teams and Zoom all support both simultaneously. For example, Akume shows captions on its tile while listeners hear translation through a separate link.
Is automatic captioning the same as CART? No. CART is a trained human captioner producing a verbatim transcript, and it is the usual choice when accuracy is a formal accessibility requirement. Automatic captions are faster to switch on but can miss names and jargon.
Questions and answers
Are live captions the same as translation?
No. Live captions write down what is said in the same language. Translation changes the language, as text (translated captions) or as speech (spoken translation). A meeting can use one, two or all three at once.
Which one do deaf and hard-of-hearing participants need?
Text. Same-language captions if the meeting is in their language, translated captions if it is not. Spoken translation alone excludes them, so offer text whenever you offer audio translation.
Can Android Live Caption or iPhone Live Captions translate a meeting?
They are made to caption audio playing on your own device. Translation is limited: Google offers it on some Pixel phones, and Apple's translation features depend on device, app and language. Check the vendor's current list before relying on either for a client call.
Can I use captions and spoken translation in the same meeting?
Yes, and for mixed audiences it is the safest setup. Akume, for example, shows captions on its tile in the call, and anyone who opens the listener link hears the translation spoken in their language.
Is automatic captioning the same as CART?
No. CART is a trained human captioner producing a verbatim transcript, and it is the usual choice when accuracy is a formal accessibility requirement. Automatic captions are faster to switch on but can miss names and jargon.
Sources
Prepared by the Akume editorial team with the help of AI tools. Platform facts come from official documentation, dated in the text.