How to get a YouTube transcript in Japanese, Korean, Hindi or any language

Paul By Paul, who built Transcript Any Video
Published

Open the video on a computer, expand the description and click Show transcript: if the video has captions in Japanese, Korean, Chinese, Hindi, Arabic, Spanish or French, the transcript comes in that language. YouTube makes automatic captions in all seven, but only in the language of the video's original audio, and an automatic transcript can get words or characters wrong, so check it against the audio before you study or quote it.

This guide is about getting the text in the language spoken in the video. Turning it into English is a separate step, covered in how to change transcript language in YouTube.

Which language will the transcript be in?

The language of the original audio. YouTube Help says that when a video has several audio tracks, automatic captions are in the default one. Its page on multi-language videos says the same: automatic captions are available in the language of the original audio track.

Two things can hide that language from you:

Play a few seconds with the original audio before you look for the transcript. That's the language you will get.

How do you get the transcript in the video's language?

  1. Check the tracks. Click the gear icon on the player, then Subtitles/CC. A language followed by “(auto-generated)” is YouTube's speech recognition. A language without that label was added by the creator.
  2. Pick the creator's track if there is one. Automatic captions are machine-made, and YouTube Help itself asks creators to add professional captions first.
  3. Open Show transcript. Expand the description under the video and click Show transcript.
  4. Copy it. Select the text in the panel and copy it into your notes.

On a phone or an iPad, the button sits somewhere else: how to see transcript on YouTube covers the app and the tablet.

Why does a Japanese transcript show the wrong kanji?

Because Japanese has many words that sound alike and are written differently, and automatic captions pick the characters from the sound. In January 2026, a learner on the Language Reactor forum posted a screenshot of a Japanese video. The speaker says 魚や貝が, “fish and shellfish”. The captions show 魚屋絵画, read the same way, sakanayakaiga, but meaning “fish shop” and “painting”.

The learner had seen the same error in other videos. The forum's team answered that Language Reactor doesn't make subtitles: it shows the ones YouTube already has, so the mistake was in YouTube's captions.

That holds for every tool that reads YouTube's captions, mine included when a video has them. YouTube Help warns that automatic captions “might misrepresent the spoken content due to mispronunciations, accents, dialects, or background noise.” For a sentence you plan to learn or quote, listen to it and compare.

Does a Korean transcript work the same way?

Yes. Korean is on YouTube's list of languages for automatic captions, so a Korean video gets them unless one of the exceptions in YouTube Help applies:

The checks are the same as for Japanese: prefer a track without “(auto-generated)”, and compare a line with the audio before you rely on it.

What about a Chinese transcript?

YouTube Help lists Chinese among the languages for automatic captions. Some older guides still say otherwise. One from shoots.video says the CC button on a Mandarin video shows “Subtitles/closed captions unavailable”. That no longer matches YouTube's help page.

The script is the other catch. If you transcribe a Chinese video yourself with Whisper, OpenAI's free speech recognition model, it can return Simplified characters when you wanted Traditional. On Whisper's GitHub, users steer it with a first prompt written in the script they want:

whisper video.mp4 --language Chinese --initial_prompt "以下是普通話的句子。"

One user still got Simplified with some accents, and fixed it by adding 請以繁體輸出 (“output in Traditional”) to the prompt. Converting afterwards isn't a clean fix, another points out: several Traditional characters can share one Simplified form.

Why is there no transcript on some Hindi videos?

Hindi is on YouTube's list. But among the reasons a video gets no automatic captions, YouTube Help gives “multiple languages at the same time”. A Hindi video that keeps switching into English can fall into that case.

No CC button means there's no track to pick, in Hindi or anything else. The text then has to come from the audio, as explained below.

How good is an Arabic transcript?

YouTube lists Arabic as a single language for automatic captions. Its help page names no dialects, and it counts dialects and accents among the reasons automatic captions can misrepresent speech. Treat an automatic Arabic transcript of a dialect-heavy video as a draft, and prefer a creator track when the Subtitles/CC list has one.

And a Spanish or French transcript?

Same steps, same checks. Both are on YouTube's list. If you need the Spanish or French text in English, Auto-translate and copying the transcript into a translator are covered in the language guide linked at the top.

What if the video has no captions in its language?

No CC button and no Show transcript means YouTube made no captions. The free route is to transcribe the audio yourself with Whisper. It can identify the language itself, and you can set it with an option like --language Japanese. Its README is honest about the limit: “Whisper's performance varies widely depending on the language.” Downloading the audio and running Whisper are in how to transcribe YouTube audio.

No captions, and no time to set up Whisper? That's the case I built Transcript Any Video for. You open the video on YouTube, and the extension transcribes its audio, with the spoken language detected on its own. You get the text in the language spoken; it doesn't translate it. Videos without captions can run up to 2 hours, at 1 credit per 30 minutes started. Ten credits cost $1.99, paid once, and they never expire. No subscription, and refunds are possible. Copy it all, or download a .txt or .md file, each line starting with its time, like [0:31]. Sign-in with just your email. If the video has captions, the extension reads YouTube's track, the creator's first: you'd get the same text as Show transcript, wrong kanji included.

Add to Chrome

FAQ

Do Japanese or Korean live streams get automatic captions?

No. YouTube Help says automatic captions on live streams are available in English only. They can also only be turned on stream by stream, not for a whole channel.

Is Cantonese captioned as Chinese on YouTube?

No. YouTube Help lists Cantonese/Hong Kong as its own language for automatic captions, separate from Chinese. Both are on the list, so a Cantonese video can get automatic captions of its own.

Why does a Hindi video show an English title?

YouTube can translate titles and descriptions automatically. Its help page says videos whose original audio is in one of your preferred languages are not translated. Add Hindi under Settings, then Languages, then Preferred languages, and those videos keep their original title.

Sources