Can I make a comic frame by textually describing what I want?

I am like to write short stories. For reasons I am not aware of, a few people like to read them.

It would be cool to a nice cover picture for each story, now that I am making eBooks for them (epub, pdf, mobi)

I dont have any friends or friends of friends who could draw it (Which is far from wanting to do it).

I dont have any money to hire a proper person for it.

I am thinking of something along the lines of an extremely versatile version of the police "sketch a suspect" (whatever that is actually called) You describe a face, and the computer generates pictures in an iterative process that hopefully may lead to a "photo" of the suspect.

Is there any such tool out there where I can describe what I want in a picture and have the computer generate it?

Something like:

Frame 1:

Takes place in a bar: (small, futuristic, sparsely populated) 1) Rhinoceros (Gray), with human head (Black eyes, smoking, stubble)

2) Humanoid StarWars Greedo (laser rifle, angry). 3) Humanoid Bartender, Robot, red eyes.

I have not spent a lot of time figuring out what the structure should be but really anything that lets you declare textually what you want in an image.

I imagine the first version will look wrong but by refining it by declaring more details and have the program create different suggestions the process could produce a useful output.

I do know I can create diagrams of many types now by just textually describing it

This would be taking the concept to a whole different level, and I presume it would require a gigantic assortment of possible assets

Does this exist?