Live translation for YouTube and YouTube Live
VoxTranslate for Chrome translates the audio of a YouTube video or live stream as it plays, with subtitles drawn over the player and an optional spoken translation. Live streams work the same way as recordings, because it translates the sound coming out of the tab rather than a caption file.
How it works
Open the video or stream and press the VoxTranslate toolbar button on that tab. VoxTranslate captures the tab audio, transcribes and translates it, and draws subtitles in an overlay that follows the player into fullscreen. On tiers that produce audio it speaks the translation while the original ducks underneath on its own volume control, which is usually what you want for a talk and rarely what you want for music. Because it works from the audio rather than from YouTube's caption track, it does not depend on the creator having uploaded subtitles, and it does not wait for a live stream to finish and be processed.
Live streams are the interesting case
A recorded video will eventually get auto-captions, and often auto-translated ones. A live stream will not — not while it is running, which is exactly when you need it. A keynote, a results call, a government briefing, a conference track in a language you do not read: these are live, in a browser tab, and unsubtitled at the only moment they matter.
VoxTranslate translates what is playing now. There is nothing to wait for, because it listens to the audio rather than to a caption file that does not exist yet. See how to translate a YouTube live stream.
When YouTube’s own captions are the better answer
Say this plainly: if the video already has decent auto-translated captions in your language, use them. They are free, they are instant, and paying per minute for something YouTube gives away is a bad trade.
The extension is for the cases where that fails — a live stream, a language pair YouTube handles poorly, a video with no captions at all, or a talk you would rather listen to in your own language than read while trying to follow slides.
Reading versus listening
For a forty-minute conference talk with slides, subtitles compete with the screen for your attention. Spoken translation does not. The original stays audible underneath at whatever level you set, so you keep the speaker’s timing, laughter and emphasis while understanding the words.
That trade goes the other way for anything where the original audio is the point. Music is the obvious one.
What it does not do
This is not free — it bills your VoxTranslate credit per minute of speech, so it suits a conference talk or a lecture rather than an afternoon of casual viewing. YouTube's own auto-translated captions cost nothing and are the right choice when they exist and are good enough. The spoken language is detected once at the start, so a video that switches language partway is not re-detected. It is Chrome only, and it translates speech — music, sound effects and overlapping crowd noise are not what it is for.