Copy, paste, generate
Halloween AI Prompts
Copy-ready Halloween AI prompts for selfies, costumes, pets, couples and short videos, with negative prompts and a working browser generator.
A Halloween prompt is not a list of spooky adjectives. It is a small production brief: what must stay the same, what may change, where the subject stands, how the light behaves and what the model must avoid. The same structure works for a selfie, a pet, a couple or a five-second video.
The prompts below are deliberately original. They describe wardrobe, material and light instead of naming a film, studio or living artist. Copy one as a starting point, then change the wardrobe and setting slots for your own idea.
What changes
Identity: every people prompt starts with what must remain recognizable.
Wardrobe: fabric, trim and accessories are described as visual traits.
Scene: background, time of day and light are specified separately.
Constraints: negative prompts list the failures that matter most.
Prompt recipe
This is model input, so it stays in English. Change the wardrobe, scene and light lines; keep the identity and anatomy lines.
Use the uploaded photo as the primary reference. Preserve the subject's recognizable facial identity, skin tone, approximate age, body proportions and pose. Transform the subject into an original Halloween portrait with a clearly described costume, scene and light. Keep the same number of people as the uploaded photo. Use photorealistic editorial photography with realistic fabric and skin texture. Avoid altered identity, extra limbs, duplicated fingers, distorted anatomy, text, logos and watermarks.
Negative prompt
Named film character, studio logo, celebrity likeness, cartoon, plastic skin, extra people, extra fingers, warped hands, distorted face, text, watermark.
Image-to-video prompt
Use the provided image as the exact first frame. Add one small movement: a slow head turn, a blink, fabric moving in a breeze or fog drifting through the scene. Keep the face, costume and lighting consistent. One smooth camera move, no cuts, no extra characters, no text or watermarks.
More copy-ready prompts
Halloween selfie
Edit the uploaded selfie into an original Halloween portrait. Keep the face recognizable and the skin natural. Add smoky eye makeup, a black lace collar and a single silver ear cuff. Place the subject in a dim room with one warm lamp and deep teal shadows. Photorealistic, close-up editorial portrait, realistic skin texture, no text or watermark.
Costume portrait
Use the uploaded photo as reference. Dress the subject in an original velvet Halloween costume with a structured shoulder, lace cuffs and a small metallic brooch. Keep the pose and body proportions. Foggy garden at night, warm lantern light, cool moonlight rim. Photorealistic fashion editorial, realistic fabric, no named character, no text.
Pet costume
Use the uploaded pet photo as the primary reference. Preserve the breed, fur color, markings, eye color and natural anatomy. Add a soft pumpkin costume with a green leaf collar that does not cover the eyes. Cozy autumn studio, warm light, miniature pumpkins. Photorealistic pet portrait, realistic fur, no extra animals, no text.
Couple portrait
Use both uploaded photos as references. Keep each person's face and identity consistent and do not swap faces. Dress one subject as a modern cowboy and the other as an ethereal angel in an original costume. Golden-hour desert road, warm dust haze, cool twilight sky. Photorealistic cinematic couple portrait, exactly two people, no extra people, no text.
Halloween video
Use the uploaded image as the exact starting frame. Five-second cinematic clip: the subject slowly turns their head toward the camera, fog drifts through the background, a breeze moves loose hair, one smooth push-in. Stable identity and anatomy, consistent lighting, no cuts, no extra characters, no text or watermarks.
Try it in the browser
The generator below is already set to the right mode and prompt. Upload a photo, check the cost, then generate.
How to get a clean result
Pick the job first
Decide whether you need a selfie edit, a full costume, a pet portrait, a couple photo or a video. Each job needs different constraints.
Change one slot at a time
Keep the identity and anatomy sentences. Edit the wardrobe, scene and light separately so you can see what caused a change.
Save the prompt that worked
Once a look holds together, keep the exact prompt and reuse it with a new photo. Consistency comes from repetition, not more adjectives.
Best for
- First-time users who do not know how to prompt
- Creators building a reusable Halloween template
- Comparing photo, pet and couple workflows
- Finding a starting point for a short video
Limits to plan around
- A prompt cannot guarantee a real person's likeness or a licensed character.
- Longer prompts are not automatically better; the identity and anatomy lines matter most.
- Prompt text is model input, so it stays in English even when the interface is localized.
- Negative prompts reduce common failures but do not eliminate them.
Questions
Do I need a different prompt for every model?
Start with the same structure and adjust the motion or camera language if a model ignores it. The identity, wardrobe, scene and constraint slots transfer well.
Why are the prompts in English?
The prompt is model input, not interface copy. English is the most consistently supported prompt language across the models, so translating it can change the output.
Can I sell the images I generate?
Check the plan and the acceptable use policy first. Do not use a real person's likeness, a copyrighted character or a brand asset without permission.
What should I change first if the result looks wrong?
Remove accessories and secondary scene details. A simpler costume and a single light source usually fix more than adding another adjective.
Generate this look
The generator below is already set to the right mode and prompt. Upload a photo, check the cost, then generate.