input and output as batch of torch images instead of direct .mov/.mp4 files (ComfyUI implementation)

[fafayqa]

I am trying, with the help of ChatGPT, to make implementation of your code for ComfyUI, but unfortunately, we need input to be batch of images (torch ) and output to be also batch of images (torch), so no physical .mov or .mp4 files, so that the result can be passed to other nodes in ComfyUI.

I mean: I installed vidstab to the local ComfyUI python_embeded, but as I said above, problem is, that we need different kind of input (batch of images instead of physical video files).

How can be this done, please?