One thought I had was to use the ‘normal’ way of uploading media files into Moodle. So uploaded as an individual file.
Then when the course is exported the export block can automatically send the media file to the Oppia server (in the background as there’s already an API for this).
The only real downside I can see to this approach is that then there could only be one media file per activity in Oppia, and it may be tricky to put narrative ‘around’ the media file (currently with the embed approach, multiple media files can be added in a single activity, and as much narrative as is needed)
I would be interested to know if the trade-off here is worth it?