I know that there is a problem with creating subtitle files containing the original phrases (or even the translation of what has been said) and to release them.

There are platforms, eg. OpenSubtitles, where you can download all these subtitle files. Even though the providing of these subtitles is not legal, I would like to get only the words and their count in the movie.

If an developer sells an app, that uses a subtitle API, to retrieve the subtitle files for creating word-statistics about the movie, would this be a legal problem to any person involved?

  • Is the developer breaking the law, because his app downloads the whole subtitle file temporarily to the users device?
  • Is the user of the app breaking the law, because the app downloads the whole subtitle file temporarily to his device?
  • Are the resulting statistics (words and their count) in anyway copyrighted?
  • 2
    I wish that any downvoters would leave a comment indicting what they think is wrong with this question. In the absence of a comment, the poster cannot improve the post, others cannot edit it to fix the issue, and readers have no idea why someone objects to the post. Such a downvote seems pointless. Commented Jan 31, 2022 at 17:42

2 Answers 2


The subtitles would be protected by copyright, in general.

In the US, temporarily downloading the subtitle text to create statistics would, I think, constitute a fair use. It is transformative, it does not serve as a substitute for the original sub titles, and it does not harm the market for the film or for the sub titles. The existence of the API for the subtitles might or might not be evidence for fair use, depending on who supplied the data and under what terms.

In other countries that have an exception to copyright for analysis and criticism, such a download might be held to come under such an exception. One cannot be sure until there is case law on the point, and I know of none. The ruling might be different in different countries.

Statistics (words and their counts) about the subtitles for a particular film or video would be facts, and as such would not be protected by copyright at all. A text discussing those facts would be protected.

  • 2
    In Germany this might fall under the "Research" exemption.
    – Trish
    Commented Jan 31, 2022 at 19:21
  • Okay, thank you for the answer, I think that helps. @Trish, can you give me the German name for "Research exemption", so that I can maybe have a look? :) Commented Jan 31, 2022 at 21:18
  • @KevinGlier UrHG §60(c) allowing people to copy 75% of a work for own research and for data research especially UrHG §60(d), allowing to automatically process an entire work, under specific rules. In neither case, a commercial usage is allowed.
    – Trish
    Commented Jan 31, 2022 at 21:36
  • @KevinGlier The exemption in Germany you are looking for is generally searchable for with the combination "Urheberrecht Forschung und Lehre", and statistical text analysis has brought us things like diaware.de/ww/2003jan-tiz-faust.txt : 0.35% of all words in Faust I are "and", Mephisto talks about 30 times more than Faust, Gretchen only says something for every third time Faust says anything, who again says about twice as much as the next follow-up. Also, about every 5th word in the text has not been used in the text before - for about 6300 unique words in a 30000 word text!
    – Trish
    Commented Jan 31, 2022 at 21:48

Creating or providing subtitles is legal for disability access

Derivative works for the purpose of allowing access to people with disabilities, like subtitles or braille translations, are an exemption to copyright law.

So making and distributing these files for that purpose is not copyright breach. However, that isn’t your purpose.

You can make copies for research purposes

This is a fair dealing exemption where that doctrine applies and is almost certainly fair use in the Us. However, that only has a clear application to a researcher, not a commercial application that enables research.

If you don’t copy anything it’s not copyright violation

If the app doesn’t make a copy of the original (except for caching which is allowed) then there are no copyright issues.

  • So I think I shouldn't get problems in most countries, when I would create an API that temporarily downloads the subtitle to create the word-list and send just this list to the App, right? Would there be a problem if the subtitle author would add some word to the subtitle that would clearly prove that he created it? Like some hash in the credit text? Could that single word maybe be copyrighted? Commented Feb 1, 2022 at 8:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .