chatgpt-prompts / ChatGPT-Midjourney-Prompt-Generator

This turns ChatGPT into a prompt generating machine that acceps only predefined inputs and responds with only predefined outputs.
452 stars 39 forks source link

The generated prompts describe the setting, not the picture to be drawn #4

Open scrato opened 1 year ago

scrato commented 1 year ago

I tried with [prompt] a scene of a samurai, a chinese opium magnate, an psychologist and an archeologist sitting at the bar

And the results were [p1] In a dimly lit, exotic tavern, each character unfolds a different corner of history. The samurai, a figure of stoic might, sits stoically in his traditional armor, his katana resting against the bar. To his right, the opulent Chinese opium magnate, his eyes hidden behind a haze of smoke emanating from his elaborate opium pipe. On the opposite side of the bar, an intuitive psychologist studies his counterparts, his spectacles reflecting the low glow of the lanterns. Finally, the archeologist, dirt still clinging to his khakis, unfurls a dusty map onto the counter, seeking answers in the parchment's cryptic markings.

[p2] At a crossroads of time and culture, four figures from diverse epochs gather in an anachronistic bar setting. A samurai warrior, his katana gleaming under the neon lights, exudes an aura of martial stoicism. Adjacent to him, the wealthy Chinese opium magnate, surrounded by a cloud of intoxicating opium, rich silks, and intricate jade jewelry. The psychologist, his eyes darting between the characters, pen poised above a notepad, trying to unravel the complexities before him. Amid the vibrant chaos, the archeologist, maps and artifacts spread out before him on the bar, adds a touch of intrigue to the tableau.

[p3] Four worlds collide in an unusual bar scene. In one corner, a samurai, his armor gleaming under the dim, ethereal light, his expression guarded but focused. Sharing the bar, a Chinese opium magnate languidly takes a puff from his pipe, the smoke creating a mysterious aura around him. On the adjacent stool, a thoughtful psychologist, notebook in hand, explores the social dynamics unfolding before him. Finally, an archeologist with a curious glint in his eyes unrolls an ancient map, as if opening a portal to another time.

[MPGM] User input options: Select a prompt with [p1], [p2], or [p3]. Generate a new set of prompts with the [next] command. Provide a new prompt with the [prompt] command. For help, use the [help] command.

This would be fine for writing a story, but not for generating pictures. I'd suggest to tweak the initialization-text so that it's bound to describe the picture, in what art style, which setting, etc. Maybe even give the order that, when given a prompt missing an art style (photography, digital drawing, painting, ...) that it asks the user for this before answering.