Halloween 2026 · Portrait templates

Halloween AI portraits from the photo you already have

Halloween AI portraits put you in a costume you never had to buy, dress up in or shoot. Upload your own portrait to the AI Photo Editor, paste one of the template prompts below, and the model rebuilds the costume, the set and the lighting around the face you uploaded.

Open the AI Photo Editor

Five Halloween portrait templates for your own photo

Every example below started from the same fictional man in a grey T-shirt, generated for this page, so you can see how much each template changes and what stays put. Press Use this look to open the AI Photo Editor with that prompt filled in, then upload your portrait.

Fictional bearded man as a vampire count in a black velvet cape with crimson lining and a garnet pin, candelabra in a stone hall behind him

Candlelit Castle Count

Velvet cape with a crimson lining, a garnet pin and two kinds of light on the face: candle on one side, moonlight on the other.

Image editing prompt

Turn the uploaded portrait into a candlelit Halloween vampire count portrait. Preserve the person’s face exactly: same facial features, face shape, skin tone, beard or lack of beard, eye colour, age and expression; keep one person only. Change the outfit to a high-collared black velvet cape lined with deep crimson satin over a dark brocade waistcoat and a white cravat with an antique garnet pin; slick the hair back neatly. Setting: a stone castle hall at night, tall candelabra behind the shoulder, warm candle glow on one side of the face and cool moonlight on the other. Style: a realistic oil-portrait look photographed with an 85mm lens, rich shadows, subtle film grain. Avoid fangs dripping blood, gore, changed identity, extra fingers, text, logos or watermarks.

Animate this look (image to video)

Keep the uploaded vampire portrait exactly as it is. The candle flames flicker softly, the person slowly raises their eyes to the lens and gives a slight knowing smile, the cape collar shifts in a draught. Slow push-in, camera otherwise still, one continuous 8-second shot, no cuts, no text.

Use this look

Fictional man in a charcoal witch hat and moss-green cloak holding a glowing brass lantern on a moonlit heath with bare trees

Heath Lantern Witch

A weathered witch hat, a moss-green hooded cloak and a brass lantern under a low moon. The face keeps its natural skin tone: no green paint, no warts.

Image editing prompt

Place the person from the uploaded photo in a moonlit Halloween witch portrait. Keep the same face, facial features, skin tone, hairline, facial hair, age and expression; one person only. Dress them in a weathered wide-brimmed charcoal witch hat with a dried-heather band, a hooded moss-green wool cloak over a dark linen shirt, and a small brass lantern held at chest height. Setting: a misty heath at blue hour with crooked hawthorn trees and a large pale moon low on the horizon, lantern light warming the face from below. Style: cinematic photograph, shallow depth of field, cool blue and amber palette. Avoid green skin, warts, changing the face, broomstick clutter, text, logos or watermarks.

Use this look

Fictional smiling man in a pith helmet and khaki shirt with stained linen wraps around his shoulders in a torch-lit sandstone tomb

Lost-Tomb Explorer in Linen Wraps

Half adventurer, half mummy: tea-stained linen over a 1920s expedition shirt, with the whole face left uncovered so you still look like you.

Image editing prompt

Restyle the uploaded portrait as an adventurous Halloween explorer half-wrapped in mummy linen. Preserve the person’s real face, skin tone, eye colour, beard, age and smile exactly and leave the whole face uncovered. Wrap the shoulders, arms and neck loosely in aged, tea-stained linen bandages over a khaki 1920s expedition shirt with a leather satchel strap; add a canvas pith helmet. Setting: a torch-lit sandstone tomb corridor with carved, unreadable hieroglyph-like reliefs and drifting dust in the torch beam. Style: realistic adventure-film still, warm torchlight, deep shadows, 35mm grain. Avoid covering the eyes or mouth, creepy decay, corpses, readable text, logos or watermarks.

Animate this look (image to video)

Keep the explorer portrait and the face unchanged. Dust drifts through the torch beam, the torch flame flutters, a loose linen strip lifts slightly in a draught and the person glances down the corridor and back to the camera with a grin. Slow handheld drift, one continuous 8-second shot, no cuts, no text.

Use this look

Fictional man with black and white skull face paint, black tuxedo, bow tie and white rose boutonniere in a ballroom with gold lights

Black-Tie Skull Paint

A crisp black-and-white skull painted over your own bone structure, worn with a tuxedo and a white rose in a gold-lit ballroom.

Image editing prompt

Edit the uploaded portrait into a black-tie Halloween skeleton look. Keep the person’s identity: same face shape, features, skin tone visible at the edges, eye colour, beard shape, age and pose; one person. Paint a crisp black-and-white skull design over the face as theatrical makeup, following the real cheekbones, nose and jaw, with the eyes clearly open and readable. Dress them in a fitted black tuxedo with satin lapels, a white wing-collar shirt and a black bow tie, plus a small white rose boutonniere. Setting: a dim art-deco ballroom with gold bokeh lights. Style: glossy editorial portrait, flash with soft fill. Avoid real bones, gore, changing the face underneath the paint, text, logos or watermarks.

Use this look

Fictional man dressed as a scarecrow in a straw hat, red plaid shirt and denim overalls in a pumpkin patch with glowing jack-o-lanterns

Golden-Hour Pumpkin Patch Scarecrow

The friendliest option for family cards: plaid flannel, patched overalls, straw at the cuffs and glowing jack-o’-lanterns behind you.

Image editing prompt

Turn the uploaded photo into a cosy harvest scarecrow Halloween portrait. Preserve the same face, skin tone, facial hair, eye colour, age and friendly expression; keep the face clean except for two small painted rosy circles on the cheeks; one person only. Outfit: a faded red plaid flannel shirt with straw poking from the cuffs and collar, patched denim overalls, a frayed straw hat and a burlap neckerchief. Setting: a pumpkin patch at golden hour with stacked hay bales, carved jack-o’-lanterns glowing softly and a few crows on a wooden fence in the distance. Style: warm, natural lifestyle photograph, 50mm lens, gentle backlight. Avoid stitched mouths, horror faces, changed identity, text, logos or watermarks.

Animate this look (image to video)

Keep the scarecrow portrait and the person’s face unchanged. A light breeze moves the straw at the cuffs and the hat brim, the jack-o’-lanterns flicker, a crow hops along the fence, and the person laughs and tips the hat. Gentle slow push-in, one continuous 8-second shot, no cuts, no text.

Use this look

Examples were made from one fictional AI-generated portrait with an open image model (FLUX.2 klein) during testing, not with Remaker AI’s paid models. Your result depends on your photo and the model you select.

In short

You upload one portrait of yourself (or of an adult who agreed); nothing is pasted onto another person’s body.
Each template prompt says what to keep (face, skin tone, beard, age) and what to change (outfit, set, light).
Nano Banana 2 edits run at 30 credits per 1K image before plan discounts; the editor shows the cost before you generate.
A finished still can become an 8-second Veo 3.1 clip in the video tool with the “Animate this look” prompt.

Your photo is the starting point

The AI Photo Editor starts from the portrait you upload, and the prompt describes everything that should change around you: the costume, the set and the light. Because the costume is redrawn on your own photo, posture and proportions stay yours.

That is why each template above starts with a “preserve” sentence and ends with an “avoid” sentence. The middle of the prompt is the costume, the set and the light. If a result drifts, it is usually because the face part of the prompt was shortened or removed.

Pick a portrait the editor can keep

The model can only preserve what it can see. A good source photo does more for likeness than any extra adjective in the prompt.

Face the camera or turn slightly; both eyes, the mouth and the jawline should be visible.
Use daylight or a bright room. Heavy beauty filters and strong colour casts get baked into the costume version.
Crop to head and shoulders. A tiny face in a group photo gives the model too little to hold on to.
Glasses, beards and hairstyles are kept unless you ask to change them; say so in the prompt if you want them gone.
Use your own photo or one an adult has agreed to. Do not edit photos of children, public figures or anyone who has not said yes.

Make your Halloween portrait step by step

Copy a template Open “Image editing prompt” on the look you want and press Copy prompt.
Open the AI Photo Editor Upload your portrait, paste the prompt and choose a portrait ratio such as 4:5 or 3:4 for social posts.
Adjust one detail at a time Swap a colour, prop or setting in the middle of the prompt. Leave the preserve and avoid sentences as they are.
Generate and compare Check the eyes, skin tone and hairline against your original before you download. Run it again if the face drifted.

From still to short clip

Three templates include an “Animate this look” prompt. Download your still, open the Veo 3.1 tool in image-to-video mode, upload the still as the first frame and paste that prompt. Clips are 8 seconds with native audio. Keep motion small: candle flicker, a glance, a breeze. Big movements are where faces start to change.

Video costs more than a still: image-to-video clips run 100 to 200 credits depending on the Veo 3.1 route, so get the still right first.

Frequently asked questions

What does the editor change, and what does it keep?

You upload your own portrait and the editor redraws the costume, set and lighting around it. Each template asks it to keep your face, skin tone, age and facial hair, so the result is you in costume rather than a different person.

Will it still look like me?

Usually, if the photo is clear and you keep the “preserve” part of the prompt. Likeness can still drift, especially with face paint or dramatic lighting. Compare the eyes, skin tone and hairline with your original and generate again if something changed.

How much does one costume portrait cost?

With Nano Banana 2 it is 30 credits at 1K, 40 at 2K and 50 at 4K before plan discounts; Nano Banana is 5 credits at 1K. The editor shows the exact cost for your plan before you generate.

Can I use the templates with a group or couple photo?

The templates are written for one person. For two people, upload a photo of both, change “one person only” to “keep both people and both faces exactly as they are” and describe both outfits. Do not combine faces from different people.

Can I post the result?

Yes, if it is your own photo or you have the person’s permission. Say that the image is AI-edited where the platform asks, and do not use it to impersonate anyone.

Updated 2026-10-20. Remaker AI (remaker-ai.net) is an independent studio and is not affiliated with Google, OpenAI, ByteDance, Black Forest Labs, Alibaba or xAI. Model names are marks of their respective owners. Prices are shown before every generation.