-
stitching [0], [1], and [2]
[0]https://github.com/kkroening/ffmpeg-python#quickstart
[1]https://kkroening.github.io/ffmpeg-python/#ffmpeg.zoompan
[2]https://ffmpeg.org/ffmpeg-filters.html#Examples-133
-
CodeRabbit
CodeRabbit: AI Code Reviews for Developers. Revolutionize your code reviews with AI. CodeRabbit offers PR summaries, code walkthroughs, 1-click suggestions, and AST-based analysis. Boost productivity and code quality across all major languages with each PR.
-
- be able to run on cloud hardware as a render service
Imagine creating a promotional video, then automatically swapping in each client's logo and re-rendering it for each one! The possibilities here are endless.
DaVinci Resolve can't be beat for #1 and seems to support #3 well, but its API seems designed more for plugins than for automated authoring. For instance, the documentation at https://forum.blackmagicdesign.com/viewtopic.php?f=21&t=9927... seems to give just read-only access to a lot of things at the TimelineClip level. And https://github.com/pedrolabonia/pydavinci has exactly one contributor.
By contrast, Shotcut seems to be built on, and by, https://mltframework.org/ with a close correspondence between the runtime API surface and the file system representation of a file, and it's fully open-source. Very possibly the sweet spot we've been looking for!
-
Also give https://kdenlive.org/ a try. I settled on it as a high-power, steep learning curve solution when I found myself hitting the wall with alternatives. It's clunky but it does everything I need.
Like Shotcut, it's all MLT underneath the hood. You can click "Generate script" on the render dialog, and get a .mlt file.
Another lead for you: https://github.com/AcademySoftwareFoundation/OpenTimelineIO