It would be very handy when I select a text files or images on Finder. Just launch a Command to process that file.
This will turn those 2-3 steps into only 1 step. This will save a lot of time. No need to open Word/PDF to get the texts, or open Preview to open image and take an Image Analysis with screenshot.
It's like when I have a bunch of pdf/txt/doc, I can directly summarize it right from Finder.
Please don't limit vision model to only external LLM, since some people run locally Qwen2-VL with vision capable.