what I mean here is that processing / analyzing a beefy file format like video will take longer than processing text input

same with returning that back to the user as manipulated output (text / code generation is much more rapid than rendering a video)