1987 Dark Fantasy AI Prompt — Rumpelstiltskin & Midjourney

Four prompts to copy, then make your own

Pick the shot you actually need. The dancer prompt leaves the feet visible; the reaction prompt gives the face room; the empty barn separates the setting from the cast. Replace a subject or garment before changing every style detail at once. These recipes have not been tested across every image model, so expect to adjust the result.

The prompts below are in English, including on the Chinese page, so you can copy the same starting text. The final --ar parameter is Midjourney syntax. Remove it when a different tool asks for aspect ratio in its own settings.

Sources: Midjourney — Aspect Ratio

What does “1987 dark fantasy” mean in a prompt?

It is shorthand for an imagined period look, not a film title or a guarantee of authenticity. A date on its own leaves most of the picture undecided. Give the model rough timber or stone, practical-looking costumes, a limited palette and a light source that belongs in the room. Soft lens detail and a little grain can finish the image; they cannot rescue a scene that still looks like a modern studio.

For the Rumpelstiltskin joke, keep some warmth. A brightly readable little dancer against darker hay is easier to understand than a murky horror scene. “Dark fantasy” does not have to mean black shadows everywhere. If the first result is too polished, ask for worn fabric and uneven set materials before piling on more filters.

Using a Rumpelstiltskin Midjourney prompt with references

Start with the still image, and choose your reference according to what you want to carry over. Midjourney describes Style Reference as a way to borrow a visual look, such as texture, lighting and color; it does not copy a person’s identity. An Image Prompt can influence the new picture’s content and composition. Treat those as different jobs.

On the web, put the reference into the appropriate image slot rather than trying to explain a URL inside the sentence. In Discord, Style Reference uses --sref followed by an image URL. Use a reference you are allowed to use. A style reference is useful for keeping the barn and reaction shot in the same visual world, but it does not make them a matched video or lock a face across every frame.

Sources: Midjourney — Style Reference · Midjourney — Creating on Web

A still-image prompt is only half the dance

The dancer standing on tiptoe is a pose. To make it a video, you still need an action sequence: lift the heels, take small sideways steps, pause with the arms out, turn toward the spectator, then cut to the reaction. Describe the camera as well as the movement. Simply adding “dancing” to a portrait prompt leaves the model to invent most of the performance.

If your goal is to put your own face in this particular meme, you can skip the separate Midjourney step. Rumpel’s photo tool supplies the scene stills and starting storyboard. Upload your portrait, choose dancer, spectator or both, and review the credit cost. For your own longer scene, the shot-by-shot guide is the better next read.

Make a photo dance with the Rumpelstiltskin AI generator · Write a Rumpelstiltskin video storyboard

Is this the prompt that made the original meme?

No. These are our own example prompts for a related visual style. Know Your Meme documents the AI clip’s origin, but that does not establish the original model settings or production prompt. A convincing old-film look is not evidence that an image came from 1987, either.

Keep the distinction clear when sharing a result: a new AI fantasy scene, a personal meme remix and footage from the real 1987 film are different things. The first two can stand on their own without inventing a lost-movie credit.

Sources: Know Your Meme — AI Rumpelstiltskin

Frequently asked questions