Meet's translated captions turn a speaker's words into on-screen text in the language you choose — a mature, steadily expanding feature that now sits beside a newer, voice-based speech translation for businesses.
2026-08-21 · by Moe Ameen
Live translated captions in Google Meet show a speaker's words as on-screen text in the language a viewer chooses — someone presenting in Spanish can be read in English in real time. The feature is not new: Google launched it in beta in 2021 and made it generally available in January 2022, initially translating English meetings into Spanish, French, Portuguese, and German. What has changed since is scope. Google has steadily widened the language set, and in June 2024 added 52 languages at once, pushing translated captions to dozens of spoken languages and thousands of translation pairs.
Recent updates kept the momentum. Google added Cantonese to translated captions in December 2025, gave mobile viewers of a Meet live stream the ability to pick their own caption language and change it mid-stream, and earlier in 2025 let participants scroll back through caption history rather than watching lines vanish. The feature stays gated by edition: translated captions require the meeting to be organized by someone on an eligible paid Google Workspace plan — Business Standard and Business Plus, Enterprise Standard and Enterprise Plus, plus select education tiers — while ordinary same-language closed captions remain free for everyone.
Translated captions are text. That distinguishes them from Google Meet's newer speech translation, a Gemini-powered feature that plays an AI-generated voice in the listener's language while the original audio stays faintly audible. Google previewed speech translation for consumers at I/O in May 2025 and began rolling it out to businesses on eligible Workspace plans in early 2026, starting with English paired with Spanish, French, German, Portuguese, and Italian, one language pair per meeting. Captions cover far more languages; speech translation covers far fewer but removes the need to read.
For all the reach, both features live inside the call. The captions render on screen for participants and disappear when the meeting ends; they are an accessibility layer for a live conversation, not an asset you can publish. Turning a multilingual Meet call into content an audience outside the room can watch is a separate job.
Meet's translated captions solve comprehension for the people in the call. They do nothing for the audience that was not in it — the captions are text on a live screen, not a post. If you run multilingual webinars, interviews, or panels, the reach you actually want lives after the meeting ends, in the clips and localized versions you publish. That is the gap [Kompozy](/) closes. Drop the recording in as a source and Kompozy cuts it into vertical [Clipped Shorts](/glossary/output-buckets), burns on word-synced captions for the sound-off feed, and — because the same source can be regenerated in another language's copy and voice — lets you ship a Spanish clip alongside the English one, instead of a caption that vanished when the call did.
From there it is a distribution problem, and that is the rest of the engine. One Meet recording becomes a batch — captioned shorts, a brand-exact [Carousel](/glossary/hyperframes) of the key points, a blog recap, and an email newsletter — each held to one [Persona Brief](/glossary/persona-brief) so the voice is consistent across languages and formats. Then [Autopilot](/glossary/autopilot) schedules and publishes the set across the eight social platforms plus blog and email, behind a per-post review gate. Meet made the meeting multilingual; Kompozy makes the content multilingual and puts it everywhere.
They are real-time captions that translate the speaker's words into a language you choose and display them as on-screen text. If someone presents in Spanish and you set captions to English, you read along in English while they speak. The captions are text only — they do not change the audio.
Yes. Translated captions require the meeting to be organized by someone on an eligible paid Google Workspace edition — such as Business Standard, Business Plus, Enterprise Standard, or Enterprise Plus, plus some education tiers. Standard same-language closed captions are available free to all users.
Dozens of spoken languages, with thousands of translation pairs after Google added 52 languages at once in June 2024 and continued expanding — Cantonese was added in December 2025. Exact coverage changes as Google rolls out more, so check Google's Meet Help page for the current list.
Translated captions are on-screen text in your chosen language and cover dozens of languages. Speech translation is a newer, Gemini-powered feature that plays an AI-generated voice in the listener's language and covers a smaller set (English with Spanish, French, German, Portuguese, and Italian at launch, one pair per meeting). Captions you read; speech translation you hear.