YouTube creates automatic captions in roughly 70 languages, and the transcript you open under a video is built from those captions. Anything outside that list, or any video where the creator turned captions off, has no transcript by default. Subtitle translation sits on top of the captions and depends on them. This guide explains where each limit sits and what you can do when the language you need is missing.
TL;DR
- YouTube’s automatic captions cover about 70 languages, from Afrikaans to Zulu. Live streams are English only.
- Automatic captions are produced in the video’s spoken (default) language. Translation into other languages is a separate step that needs a caption track to start from.
- Accuracy varies by language, accent, audio quality and vocabulary, so check names, numbers and jargon before you quote anything.
- When a transcript is missing or in the wrong language, you can switch the caption language, download the captions, or use a tool that transcribes the audio itself.
- Claras transcribes YouTube videos, including ones without captions, in over 70 languages.
Table of Contents
- What “language support” actually means on YouTube
- Languages YouTube supports for automatic captions
- How translation fits on top of captions
- How accurate are transcripts across languages?
- What to do when the transcript isn’t in your language
- Who needs which kind of language support
- Our perspective
- Getting a transcript in your language with Claras
- Practical workflows by use case
- FAQ
- Sources
What “language support” actually means on YouTube
People use “transcript,” “captions” and “subtitles” as if they were one thing. YouTube does not. Its glossary defines a transcript as unformatted, untimed text transcribed verbatim from the video. A caption covers both same-language transcriptions and translated subtitles shown on the video. Subtitles are text tracks in a language different from the one spoken.
That gives you three separate layers of language support:
| Layer | What it is | Where the language comes from |
|---|---|---|
| Spoken-language captions | Speech turned into text in the language of the audio | Automatic speech recognition, or captions the creator uploaded |
| Translated subtitles | A caption track converted into another language | The existing captions, plus translation |
| Translated metadata and audio | Titles, descriptions, dubbed audio tracks | The creator, or YouTube’s automatic dubbing |
When a reader asks whether YouTube “supports” their language, the useful question is which layer they mean. Reading a Spanish lecture in Spanish needs layer one. Reading it in English needs layers one and two. If you want a primer on how these terms work together with AI tools, this YouTube AI tools course covers the basics for creators.
Languages YouTube supports for automatic captions
According to YouTube’s automatic captioning help page, automatic captions are available for long-form videos and Shorts in these languages:
Afrikaans, Albanian, Amharic, Arabic, Armenian, Azerbaijani, Basque, Belarusian, Bengali, Bulgarian, Burmese, Cantonese/Hong Kong, Central Khmer, Chinese, Czech, Danish, Dutch, English, Estonian, Farsi, Filipino, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Javanese, Kannada, Korean, Lao, Latvian, Lithuanian, Macedonian, Malay, Malayalam, Marathi, Mongolian, Nepali, Norwegian, Polish, Portuguese, Punjabi, Romanian, Russian, Sinhalese, Slovak, Spanish, Sundanese, Swahili, Swedish, Tamil, Telugu, Thai, Turkish, Ukrainian, Uzbek, Vietnamese and Zulu.
Three points matter more than the list itself:
- Live streams are narrower. The same page says automatic captions for live streams are available in English only, and are still rolling out to English channels streaming at normal latency.
- Captions follow the spoken language. YouTube’s subtitle help says automatic captions are created in the video’s default language only. If a creator sets the wrong default language, the captions can come out wrong or not appear.
- Lists change. YouTube updates these pages over time, so treat the page above as the current reference, not a fixed list.
How translation fits on top of captions
Once a video has a caption track, viewers can usually ask for it in another language through the player’s captions settings. YouTube also tries to match a video’s language to each viewer. Its page on translated videos says it uses signals such as your language, location and recently watched videos for that. To see a different language version, it suggests opening an incognito window, going to the watch page, clicking your profile picture, choosing Language and picking the one you want.
Creators have a second route. Multi-language audio lets them upload extra audio tracks, and viewers can switch between them in the player settings. Automatic captions in that case are available in the language of the original audio track. YouTube also offers automatic dubbing. For videos in English, dubs are available in languages including Arabic, Bengali, Dutch, French, German, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Tamil, Telugu and Ukrainian.
For a deeper look at turning a transcript into another language, see our guide to YouTube transcript translation.
How accurate are transcripts across languages?
Language support means a language is on the list. It does not mean every video comes out equally clean. YouTube itself says automatic captions come from machine learning, so quality may vary, and that they can misrepresent speech because of mispronunciations, accents, dialects or background noise.
Research points the same way:
- A 2025 study from the University of Turin found that speech recognition struggles with homophones, named entities and different accents, which affects automatic subtitle translation.
- A 2024 study of academic video presentations found that most machine-translated subtitle errors came from poor caption segmentation and missing context. Viewers understood more when the captions had been edited first.
- For languages with little training data, the gap is bigger. A 2025 review of African low-resource languages found only about 11,206 hours of speech data across 111 languages, with some languages having fewer than 10 hours of usable audio.
YouTube’s dubbing page adds that proper nouns, idioms and jargon are hard to translate, and that quality is not the same across all languages.
What this means in practice
Treat a transcript as a strong first draft. Before you quote it, cite it or publish it:
- Check names of people, brands and places.
- Check numbers, dates and units.
- Listen again to any passage with technical terms or heavy accents.
- If you translate, read the translation against the original for idioms.
What to do when the transcript isn’t in your language
This is the situation most readers searching this topic are really in. Work through these steps in order.
- Open the transcript and check the language selector. On desktop, open the description area and choose the transcript option. If the video has more than one caption track, a language menu appears at the bottom of the transcript panel. Our guide to getting the transcript of any YouTube video walks through each method.
- Try the captions menu. Click the CC button, open settings, and look under subtitles for other languages. If the creator or YouTube offers translated tracks, you can pick one there.
- Check the video’s default language. If the spoken language isn’t on YouTube’s list, or the creator disabled captions, no automatic track exists to translate from.
- Download the captions you can get. A caption file in the original language can be translated elsewhere. Our YouTube subtitle downloader saves SRT and TXT files, and this guide to exporting transcripts covers copying and saving as text.
- Transcribe the audio directly. If there are no captions at all, you need a tool that listens to the audio instead of reading a caption track. We cover the options in YouTube transcript without captions.
Keeping timestamps when you switch languages
Students and researchers often need to return to the exact moment a speaker said something. Timestamps make that possible, and a transcript with timestamps stays useful even after translation because you can jump back to the original audio. See how to get and use timestamped transcripts.
Who needs which kind of language support
Different readers care about different parts of the stack.
| If you are… | What you usually need | Watch out for |
|---|---|---|
| A student following a lecture in another language | Captions plus translation, with timestamps | Technical terms and names mistranscribed |
| A researcher reviewing interviews | A verbatim transcript in the spoken language, saved as text | Accents, dialects and overlapping speakers |
| A creator reaching new audiences | Accurate source captions, then subtitles or dubs | YouTube advises focusing on one or two languages and dubbing as much as possible in them |
| A professional scanning many videos | Fast summaries and search across transcripts | Summaries built on weak captions |
| An AI builder feeding transcripts into a pipeline | Clean text exports | Encoding of non-English text |
On that last point, YouTube’s transcript file tips recommend saving non-English transcript files with UTF-8 encoding so characters display correctly.
Our perspective
Language support is usually described as a list, but readers experience it as a question: can I get usable text from this video, right now? From what we see, the answer fails in three places. The video has no captions. The captions are in a language the reader doesn’t read. Or the text is there but too long to use.
That is why we built Claras around the transcript itself rather than around the caption track. It is designed to transcribe videos even when they have no captions, to work across more than 70 languages, and to sit directly on the YouTube page, so you don’t have to leave the video to get the text.
We also think a transcript is only the start. Once you have the words, you want chapters to scan, a summary to decide if the video is worth your time, and the option to ask the video a question and get an answer drawn from the full transcript. For learners and busy professionals, that turns a language barrier into a reading task you can finish in minutes.
Our advice for anyone working across languages: get the transcript in the spoken language first, check the names and numbers, then translate or summarize from there.
the author
Getting a transcript in your language with Claras
If YouTube’s built-in options don’t cover the video you need, Claras is built for that gap. It provides instant transcripts, AI summaries, comment summaries and clickable chapters on the YouTube page, and it is designed to work on videos without captions. You can chat with a video to get answers from its full transcript, export results to common formats and knowledge tools, and process playlists in bulk.
There is a free plan with limits, and paid plans on the pricing page: Pro is $4 per month or $48 per year, and Unlimited is $10 per month or $120 per year. You can also try the free tools on the Claras tools page, such as the video summarizer and chapter generator, or install the Chrome extension from the Claras homepage.
Practical workflows by use case
For studying
Open the lecture, get the transcript, and skim the chapters first. Search the text for the terms on your syllabus. Our guide on searching a YouTube video transcript shows how to find any spoken moment without scrubbing through the video.
For research and analysis
When you review many videos, consistency matters more than speed. Pull transcripts in the spoken language, store them as text, and note the source video and timestamp for each quote. For wider approaches, see methods for YouTube video analysis. Comments can add a second layer, covered in our comment sentiment analysis guide.
For creators publishing in more than one language
Start with accurate source captions. YouTube’s own advice is to review and edit automatic captions, and the studies above suggest edited captions are understood better. Then add translations or dubs for the languages where you want to grow, rather than spreading thinly across many.
FAQ
Does every YouTube video have a transcript?
No. A transcript depends on captions, and a video has none if the creator disabled them or the spoken language is not on YouTube’s automatic caption list. Newly uploaded videos can also take time to get captions. Tools that transcribe the audio directly can cover videos with no caption track.
Can I read a YouTube transcript in a different language?
Often, yes, if the video has a caption track to translate from. Check the CC settings for subtitles in other languages, or the language menu in the transcript panel. If neither offers your language, download the captions and translate them, or use a transcript tool with translation.
How accurate are automatic YouTube captions in other languages?
YouTube says quality may vary and that accents, dialects, mispronunciations and background noise can cause errors. Research shows names, homophones and low-resource languages are the hardest. Review key passages before you rely on them.
Do live streams have transcripts in languages other than English?
According to YouTube’s help page, automatic captions for live streams are available in English only. For live content in other languages, creators need to provide captions themselves, or you can transcribe the recording after the stream ends.
Sources
- Translation & transcription glossary, YouTube Help
- Use automatic captioning, YouTube Help
- Add subtitles & captions, YouTube Help
- Check translated videos, YouTube Help
- Add multi-language features to your videos, YouTube Help
- Use automatic dubbing, YouTube Help
- Tips for creating a transcript file, YouTube Help
- Automatic generated intralingual and interlingual subtitling, University of Turin
- Automatic closed captions and subtitles in academic video, 2024
- Automatic Speech Recognition for African low-resource languages, arXiv
