Winter fortress: a Genjutsu prompt
Download our clip, add your selfie or avatar, and become the girl racing through the snow toward a glowing winter fortress.
This is a scene we created in Higgsfield: a woman stands in a snowy field, a reindeer walks past her, and she turns and runs toward a white fortress lit up for the holidays, with fireworks bursting over the domes. With Higgsfield Genjutsu you can replace her with yourself or your avatar, while everything else stays the same: the reindeer, the snowy field, the fortress, the fireworks and the camera movement. Below is everything you need: the clip, the prompt and the settings.
Download the clip
15 s · 1080×1920 · MP4
Download this clip and upload it to Genjutsu as your source video. It provides the motion, framing and scene; only the person will change.
Prepare your photos
Genjutsu will replace the woman in the clip with the person in your photos. You can use a selfie or an image of your AI avatar.
- Upload 3 to 4 photos: one from the front, one in profile and one from behind.
- In the first half of the clip she is seen in profile, and in the second half she runs away from the camera, so the profile and back photos matter most.
- She wears a fur hat, so your hair is visible mostly below it: make sure the length and color of your hair are clear in the photos.
- Use daylight, no filters, no sunglasses. Only you in the photo, no other people.
Open Genjutsu in Higgsfield
- Go to higgsfield.ai and log in.
- Open Video → Create Video and select Higgsfield Genjutsu in the model picker.
- Choose a mode: Object Swap if you want to put yourself into our exact scene, or Motion Transfer if you want to keep only our movement and build your own scene (see block 05).
- Upload the downloaded clip as the source video.
- Upload your photos as references.
- Copy the prompt below and paste it into the description field.
- Choose the resolution (1080p recommended) and click Generate.
Prompt: Object Swap
Replace the woman in the video with the person from the reference images. Keep the person's exact face, identity, skin tone, hair color and hair length from the references in every frame, including the profile shots and the shots from behind. Keep the same leopard print fur hat, brown leather bomber jacket, yellow scarf, floral maxi skirt and black boots. Preserve the original motion, the turn and the run, camera movement, framing, lighting and timing exactly as in the source video. Do not change the reindeer, the snowy field, the pine trees, the white fortress with domes and holiday lights, the Christmas tree, the river or the fireworks. Photorealistic, cinematic, cold winter dusk light. Do not add any text.
Prompt: Motion Transfer
Object Swap changes only the person and keeps our whole scene. Motion Transfer takes only the movement from our clip, the turn and the run, and rebuilds everything else from your references: you, your outfit and your location. Upload your photos and, if you want, a photo of a place or an outfit, then replace the part in brackets with your own scene.
Transfer the exact motion, the turn and the run, camera movement and timing from the source video to the person in the reference images. Keep the person's exact face, identity, hair and outfit from the references in every frame, including the profile shots and the shots from behind. Rebuild the scene as [a snowy street of an old European city at dusk, with holiday lights and a big Christmas tree at the end of the street]. Keep a reindeer walking past at the start and fireworks bursting in the sky at the end. Photorealistic, cinematic, cold winter dusk light. Do not add any text.
Tips
- Not sure which mode to pick? Start with Object Swap: it is the safest way to get a result that looks like our clip. Use Motion Transfer when you want your own location or a completely different look.
- Want a different outfit? In the prompt, replace "Keep the same leopard print fur hat, brown leather bomber jacket, yellow scarf, floral maxi skirt and black boots" with a description of your outfit, or add a photo of the outfit as an extra reference.
- If you look different while running away from the camera, add another photo from behind and generate again.
- The exact credit cost is shown before you generate. Genjutsu Lite is a cheaper and faster option, with less control.
You can use this clip for personal projects and social media posts. Only use your own face or the face of someone who has given you permission. Do not use it for impersonation, misleading or offensive content. When you publish the result, label it as AI-generated. Alveon is not responsible for content created by users. If you post your result, we'd love to see it: tag us @alveonai.
Clip and prompt: Alveon Studio. Tool: Higgsfield Genjutsu.














