Plain Upscaler

plainupscaler.com

AI Halloween Video Prompt Guide

A still is the starting point, not the finish line. These prompts add small, controlled motion to a costume portrait without changing the person.

Most bad AI Halloween videos fail for the same reason: the first prompt asks for too much. A head turn, a hair movement, a light flicker, and a camera push-in are enough for five seconds. The model does not need a full scene change.

Each recipe below is written for image-to-video, not text-to-video. Generate the portrait first, keep the face and costume consistent, and add one or two motions. Then upscale the finished clip if it needs a cleaner export.

The video itself is generated by the model you choose, such as Wan 2.6 or Kling 2.6. Plain Upscaler can clean and enlarge the finished file; it does not animate the still on its own.

See all 10 Halloween looks Upscale the finished clip

1

Start with a still you like

Use the image prompt on a look page first. Do not animate a portrait whose face or costume edges you have not checked.

2

Add one motion

Pick a head turn, a blink, a light flicker, or a slow push-in. One clear motion is easier to control than four at once.

3

Upscale the finished clip

If the clip is soft or noisy, run it through the video upscaler and compare before and after.

Adult model in an original haunted unicorn costume with a plush horn and violet cape in a foggy moonlit orchard

01

Haunted Unicorn

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic Halloween portrait animation. Keep the subject’s face, costume, hairstyle, and body proportions consistent with the reference. The subject slowly turns their head toward the camera with a subtle mysterious expression. A gentle breeze moves the plush hood and the edge of the violet cape. Soft fog drifts through the moonlit orchard. The camera performs a slow, smooth push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, replace the costume, add characters, or introduce sudden scene cuts. No text or watermarks.

Adult model with long violet braids in a black and purple velvet gothic outfit in a moonlit Victorian garden

02

Midnight Braids

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic Halloween portrait animation. Keep the subject’s face, braids, costume, and body proportions consistent with the reference. The subject slowly turns their head toward the camera with a calm, mysterious expression. A gentle breeze moves loose braid strands and the edge of the velvet outfit. Soft fog drifts through the moonlit garden and the lanterns flicker subtly. The camera performs a slow, smooth push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, replace the costume, add characters, or introduce sudden scene cuts. No text or watermarks.

Adult model in a midnight-blue celestial gown with glowing star patterns and a silver moon crown in an observatory

03

Cosmic Princess

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic fantasy portrait animation. Keep the subject’s face, crown, gown, and body proportions consistent with the reference. The subject slowly turns toward the camera while the star patterns on the gown shimmer softly. A gentle draft moves the outer layer of the gown. Distant stars drift slowly outside the observatory windows. The camera performs a slow, elegant push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, replace the costume, add characters, or introduce sudden scene cuts. No text or watermarks.

Adult model wearing a handmade robot costume with matte foam armor panels and blue LED accents under studio lighting

04

Halloween Robot

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic Halloween portrait animation. Keep the subject’s face, robot costume, and body proportions consistent with the reference. The subject slowly turns their head toward the camera while the blue LED accents pulse gently. A subtle studio light sweep crosses the armor panels. The camera performs a slow, smooth push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, replace the face with a robot head, add characters, or introduce sudden scene cuts. No text or watermarks.

Small dog wearing a soft pumpkin Halloween costume in a cozy autumn studio with miniature pumpkins

05

Pumpkin Pet

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic pet portrait animation. Keep the animal’s breed, fur markings, costume, and body proportions consistent with the reference. The pet blinks slowly and tilts its head slightly toward the camera. The green leaf accessory moves a little. Warm studio light shifts softly and dust motes drift in the background. The camera performs a very slow, gentle push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the animal’s species or markings, add animals, or introduce sudden scene cuts. No text or watermarks.

Two adults in coordinated cowboy and angel Halloween costumes on a desert plain at dusk

06

Cowboy & Angel

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic couple portrait animation. Keep both people’s faces, costumes, and body proportions consistent with the reference, and do not swap their identities. Both subjects turn slightly toward the camera. A warm evening breeze moves the angel’s feathers and the cowboy’s fringe. Dust drifts slowly across the desert plain. The camera performs a slow, smooth push-in. Maintain realistic motion, stable anatomy, and consistent lighting on both faces. Do not change either person’s identity, replace the costumes, add characters, or introduce sudden scene cuts. No text or watermarks.

Adult model dressed as a forest witch with moss-green robes, an antler headpiece, and a lantern in a misty forest

07

Forest Witch

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic forest witch portrait animation. Keep the subject’s face, robes, antler headpiece, and body proportions consistent with the reference. The subject slowly raises their eyes toward the camera. The lantern flame flickers and the mist drifts between the trees. A few dried flowers and loose strands of hair move in a gentle draft. The camera performs a slow, smooth push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, replace the costume, add characters, or introduce sudden scene cuts. No text or watermarks.

Adult model with elegant Day of the Dead inspired skeleton makeup and marigold flowers in candlelight

08

Skeleton Makeup

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic beauty portrait animation. Keep the subject’s face, skeleton makeup, outfit, and body proportions consistent with the reference. The subject slowly opens their eyes and turns slightly toward the camera. Candle flames flicker gently and soft shadows move across the altar. Marigold petals drift slightly in the warm air. The camera performs a slow, smooth push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, remove the makeup, add characters, or introduce sudden scene cuts. No text or watermarks.

Adult explorer in a yellow raincoat in an eerie yellow-wallpapered corridor with fluorescent lights

09

Backrooms Portrait

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic liminal-horror portrait animation. Keep the subject’s face, raincoat, and body proportions consistent with the reference. The subject stands still and slowly looks over their shoulder toward the camera. The fluorescent lights flicker once, and the camera drifts forward at walking pace. Dust hangs in the air and the damp carpet texture stays stable. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person’s identity, add monsters or characters, or introduce sudden scene cuts. No text or watermarks.

Small dog wearing a glow-in-the-dark dinosaur costume with green luminous spikes in a backyard with Halloween lights

10

Glow Dinosaur Pet

Generate the portrait first, then use this prompt with an image-to-video model such as Wan 2.6 or Kling 2.6. Keep the first clip to five seconds and ask for small motion.

Image-to-video prompt

Use the provided image as the exact visual reference and starting composition. Create a continuous five-second cinematic pet portrait animation. Keep the animal’s breed, fur markings, costume, and body proportions consistent with the reference. The pet looks toward the camera and wags its tail slightly. The green luminous spikes pulse softly, and warm string lights shimmer in the background. A few autumn leaves drift past. The camera performs a very slow, gentle push-in. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the animal’s species or markings, add animals, or introduce sudden scene cuts. No text or watermarks.

Upscale the finished clip

The clip is generated by the video model you choose. Plain Upscaler's video upscaler can clean and enlarge the finished file; it does not animate the still by itself.

Upscale the finished clip

AI Halloween Video Prompt Guide: FAQ

A still is the starting point, not the finish line. These prompts add small, controlled motion to a costume portrait without changing the person.