I will not be investing significant effort in the very near future to review and address issues on this repository. However I do want my software to be useable!
If you have an issue that must be resolved for your work, please open a pull request to fix it, and send me a direct email to make sure that I see it. I ignore most messages from GitHub these days.
I'm also happy to help out if you have a question about how to use the library.
My email can be found at the top of this commit.
Keep in mind that I have a full-time job and a personal life as well as other hobbies that have taken priority over open source, so I might not respond immediately. But don't hesitate to follow up after a few days if you think I've missed your email.
Extract and save WebVTT (.vtt) closed caption files from YouTube videos.
YouTube videos don't use a standard closed caption format so this script parses that format and converts it into the WebVTT format. The exported caption files can be used to display native captions in any browser supporting the HTML video element.
-
Open a YouTube video with closed captions in a web browser
-
Open the JavaScript console (in Chrome this is Ctrl+Shift+J/Cmd+Option+J/).
-
Paste the contents of
save-vtt-files.jsand hit Enter. -
Run the command to export and save a .vtt file for each caption track:
a. To export with default settings, just run:
saveVttFiles();
b. By default we're making only one caption display at a time, but YouTube saves the captions in overlapping (two-at-a-time) fashion, which makes sense for the way YouTube shows captions. If you'd like to preserve the overlapping durations, you can run this instead:
AVOID_CONCURRENT_CAPTIONS = false; saveVttFiles();
c. If you'd like your captions to be auto-translated into a different language by YouTube, you can specify the language code as an option:
saveVttFiles({ translationLanguageCode: 'zh-Hans' });
-
For each caption track, a file will be saved called
[Video Title]-[Language Code].vtt.
Alternatively, you can use a CLI which allows you to trigger downloads in a more automated fashion.
npm install -g youtube-vttYou also must have the Google Chrome browser (not Chromium) installed on your system or the commands below will fail.
Download captions for a video (vtt files go into a downloads directory under the current working path):
youtube-vtt https://www.youtube.com/watch?v=XXXXXXXXXXXTranslate downloaded captions to Simplified Chinese:
youtube-vtt https://www.youtube.com/watch?v=XXXXXXXXXXX --translation zh-HansAllow concurrent timespans for captions (disabled by default):
youtube-vtt https://www.youtube.com/watch?v=XXXXXXXXXXX --concurrentWait only one second for downloads to complete (default wait time is 5 seconds):
youtube-vtt https://www.youtube.com/watch?v=XXXXXXXXXXX --wait 1000Wait 15 seconds for downloads to complete (if you have a slow connection):
youtube-vtt https://www.youtube.com/watch?v=XXXXXXXXXXX --wait 15000Run in debug mode. The browser will open a window rather than running in the background, and it will wait for you to close it manually, allowing you to interact with the browser and inspect the page.
youtube-vtt https://www.youtube.com/watch?v=XXXXXXXXXXX --debugIf the live stream will eventually complete (e.g. a live stream of an event that lasts for a few hours), and you're able to wait, then the commands above will work for that recording once it has completed.
However, if you need to get captions from a live stream in progress, you can't yet use youtube-vtt or saveVttFiles(). I do have experimental code that will work to get captions for a live stream, which can be found here. I'm still trying to figure out the best way to integrate this with the main code base, so let me know if you have any suggestions. I'm still wondering what the main use case for consuming captions from a live stream would be.