The short answer
If the video plays in a Chrome tab, Chrome can caption it and send those captions to Google to translate. YouTube makes translated subtitles from captions a video already has, so a stream with no caption track gives it nothing to work from. Windows 11 and macOS caption any audio on the machine, but Windows translates only on a Copilot+ PC, into English or Simplified Chinese, and Live Captions on a Mac does not translate.
Fraze is a Mac and Windows app that listens to whatever your computer is playing, in any app, and writes a translated line under each spoken line. Speech recognition and translation both run on your computer, so a desktop player, a stream in a native app and a webinar all work the same way, and nothing is uploaded.
Expect about half a second to a second of delay after a speaker pauses, and text only. Fraze shows subtitles, it does not dub the audio.
What each option can translate
| What it can caption | Translation | Where the text is translated | |
|---|---|---|---|
| Chrome Live Caption | Media playing in Chrome | Yes, with Live Translate | Captions go to Google |
| YouTube captions | Videos that have a caption track | Yes, from that track | YouTube does it |
| Windows 11 Live Captions | Any audio on the PC | Copilot+ PCs only | Microsoft says on-device |
| macOS Live Captions | Any audio on the Mac | None in Live Captions | Not available |
| Fraze | Any audio in any app | About 50 languages | On your computer |
Two gaps show up again and again. One is reach: a desktop player, a stream in a native app and anything that is not a browser tab sit outside Chrome. The other is the caption track: a platform can only translate words it already has, so a stream with nothing written down has nothing to translate.
Try the free routes first
Switch the language on a platform that has captions
YouTube's glossary defines a translation as a title, description or subtitle made by translating existing text, so this route needs a caption track to be there. Google's documentation also says automatic captions for live streams are available in English only, and the creator has to turn them on for each stream. A live broadcast in another language often has no track to switch.
Turn on Chrome's captions if the video is in a tab
Google's help page says Chrome can caption videos, live streams and other audio it plays, and send the captions to Google to translate. Live Translate is listed for Windows, Mac, Linux and Chromebook Plus, and you pick your language next to Translate to. For Chromebook Plus, Google puts the choice at over 100 languages. Live Caption has to support the spoken language first. The Chrome guide on this site has the steps. Audio outside the browser is not covered.
Use the captions built into your operating system
Microsoft's documentation says live captions are in Windows 11 version 22H2 and later, caption any audio on the PC, and need a Copilot+ PC on version 24H2 or later to translate, into English or Simplified Chinese. Apple's documentation says Live Captions on a Mac transcribes audio from all apps on Apple silicon, and lists no translation option. Apple's separate Live Translation feature works in Messages, FaceTime and the Phone app, not on a video or a stream.
What a system audio app does differently
- It works on a desktop player, a stream in a native app, a lecture, a webinar or a call, and on video with no subtitles at all.
- A small caption window can float over the video. The main window shows the original, the translation, or both, at the text size you want.
- Speech covers 25 European languages, and more on macOS 26 and later, including Japanese, Korean and Chinese. Translation covers about 50 languages to or from English, and other pairs go through English.
- Transcripts are saved on your computer in a history list, so you can read a lecture back later or export it.
- After the first download of the language files it keeps working with Wi-Fi off, and no audio or text is ever uploaded.
Fraze works from the sound your computer is already playing, so it does not care what is producing it. There is no caption track to wait for and no tab to be inside.
Hours of video, one fixed price
Speech recognition and translation run on your computer, not on a server, so there is no per-minute meter to feed. A semester of lectures or a full day of a conference stream costs the same as one short clip.
The first 30 minutes are free on every account, with no card. After that it is $29.99 once or $2.99 a month, both unlimited. A free Fraze account carries the plan, and it never sees your audio or your transcripts.
What to expect, honestly
- There is a delay. Lines land about half a second to a second after a speaker pauses, so fast back and forth reads slightly behind the picture.
- Machine translation of fast speech is imperfect. Names, numbers, jokes and people talking over each other are where it slips.
- It is subtitles only, with no dubbing and no synthetic voice, and there are no speaker labels or summaries.
- It is live only. You play the video and let it listen, you cannot hand it a file to process.
- It runs on macOS 14 or later on Apple silicon, and on Windows 11 on x64 with a 4-core AVX2 processor and 8 GB of memory. There is no browser version and no phone app.
If you need summaries, sharing links or a browser you can open anywhere, Verli (https://verli.app) is the cloud version from the same company. More languages and those extra features, but the audio goes to the cloud.
When the built-in option is the better pick
- The video already has a caption track. The platform can translate that text and you need nothing else.
- Everything you watch is inside Chrome and you do not mind the caption text going to Google. Google's page speaks of over 100 translation languages on Chromebook Plus, against about 50 in Fraze.
- You have a Copilot+ PC and want English captions for foreign audio. Microsoft's documentation says Windows translates over 40 languages into English, on the device.
- You want the video dubbed into your language, not subtitled. Fraze does not do that. Google's documentation says some YouTube videos carry automatically dubbed audio tracks that viewers can switch to.
Questions
Can you translate video in real time?
Yes, as running subtitles rather than dubbed sound. Chrome does it for media in one of its tabs, and a system audio app like Fraze does it for any video the computer plays. Both run a short step behind the speaker.
How do I get live stream translation when the stream has no subtitles?
Platform translation needs an existing caption track, and Google's documentation says automatic captions for live streams are English only. An app that listens to the sound needs no track. Fraze transcribes the speech on your computer and shows a translated line under it.
Does real time video translation work outside the browser?
Chrome's captions only cover media playing in Chrome. Windows 11 and macOS caption any audio, but Windows translation needs a Copilot+ PC and goes into English or Simplified Chinese, and Live Captions on a Mac does not translate. Fraze captions and translates any audio in any app.
How far behind the video are the subtitles?
About half a second to a second after a speaker pauses. Close enough to follow a lecture, a stream or a match, and noticeable in a fast interview.
Can Fraze dub the video into my language?
No. Fraze writes subtitles and shows them in a window you can float over the video. There is no voice output.
Can I translate a video file already on my computer?
Yes, by playing it. Fraze listens to whatever your player puts out, so a local file works like a stream. You cannot upload the file for it to process, because it only works live.