English to Chinese Live Translated Captions
Turn English talks, meetings, and courses into live Chinese captions, with the original right there.
English talks, international calls, online courses — your Chinese-speaking audience catches most of it, and the critical numbers hide in the rest. Avoce turns English speech into live Chinese captions: the original appears in about 0.8 seconds with the Chinese right behind it, side by side on one screen.
Nothing to install — it runs in the browser. Capture a live talk with the microphone, or caption an online meeting or video through tab audio. Register names and jargon in the glossary so the Chinese stays exact, let the audience scan a QR code to read captions on their own phones, and export the transcript as txt, Markdown, or srt when you’re done.
Where it shines
- English keynotes and seminars
- Cross-border video calls
- Online courses and tutorials
What to watch for in this direction
Big numbers change magnitude
A billion is 10⁹ and a trillion 10¹², while the Chinese 億 and 兆 are used loosely in speech. Financial figures, budgets, and population numbers are exactly what an audience writes down, and exactly where a digit goes missing. Keeping the original line visible (it is on by default) puts “2.5 billion” next to 「25 億」 so the numbers check themselves.
Names have more than one accepted rendering
Sydney is 雪梨 in Taiwan and 悉尼 in the mainland; the same speaker’s name can come out two ways inside one talk. Registering the speaker list, company names, and product names in the glossary locks one rendering for the whole session.
One sentence, three nested clauses
English can chain a sentence with which, that, and where; Chinese needs it broken into two or three. Captions settle once the structure is stable, so a habit of long sentences shows up as a slightly later translation. Splitting long sentences in two does more for perceived latency than any setting.
Three steps to start — nothing to install.
Put these in the glossary first
The glossary applies to both recognition and translation: terms you register up front stop the model re-guessing a rendering every time.
- Speaker and attendee names, with the Chinese rendering you want
- Company names, product names, and project code names
- Field-specific abbreviations, noting whether to translate them (ARR, SKU, KPI)
- Recurring units of measure (billion, basis point)
One point buys one minute of live translation — about $3.70 an hour on the $10.99 / 180-point pack. Signing up gets you a free 20-minute trial, no credit card needed.
Common questions
Do English accents affect accuracy?
Yes, though the microphone matters more. With the mic close to the speaker and low room noise, most common accents transcribe reliably; a phone left at the far end of the room loses words no matter how neutral the accent.
Can I keep only the English transcript?
Yes — transcript-only mode outputs the English alone and uses a quarter of the points (one point covers four minutes).
Does it work on an online meeting?
Yes. Tab audio captures the sound from a browser tab, which covers Google Meet, Zoom on the web, and YouTube — with no change to the meeting itself.