Thanks for making a pull request to jupyterlab!
To try out this branch on binder, follow this link:

Currently the two extensions have a very similar structure and approach, but it probably still makes sense to keep them separate. So they can be disabled separately, but also later to be able to add more features specific to audio or video, such as the ones mentioned in the top comment:

• Add toolbars items to perform actions on the files: trim, edit...
• Captions for videos via <track>
• Waveform visualizer for audio files

OK I think the basic functionality should be in place. All CI is now all passing ✅

Marking as ready for review to encourage folks to try it, for example on Binder: https://mybinder.org/v2/gh/jtpio/jupyterlab/audio-video-extension?urlpath=lab

Maybe it's in a good enough so that it can be available in an early JupyterLab 4.5 alpha release, to end users can try it too. Happy to iterate more on it if some things need to be fixed until the JupyterLab 4.5 final release.

This works well with tabs, should it also work as output renderer?

from IPython.display import display
from IPython.core.formatters import DisplayFormatter

video_path = './big_buck_bunny.mp4'
with open(video_path, 'rb') as f:
    video_data = f.read()

mime_bundle = {'video/mp4': video_data}

# Display the MIME bundle
display(mime_bundle, raw=True)

Yes it would likely make sense to support this too. For large files the amount of data to transfer from the kernel can be quite big, but we can also imagine this being useful for smaller files.

Yes it would likely make sense to support this too. For large files the amount of data to transfer from the kernel can be quite big, but we can also imagine this being useful for smaller files.

I'll have a look next week. Unless folks are ok if we do that in a separate follow-up PR?