Halloween AI Prompts That Actually Work (2026 Guide)

Short answer: A Halloween prompt works when it says what to preserve, what to change, where the scene is, how it is lit, and what must not appear. Specific fabric, colour and lighting details matter more than a long list of spooky adjectives. The eight prompts below follow that order and can be copied into the image or video workflow.

Why most Halloween prompts fail

A prompt like “make me a scary witch” gives the model almost nothing. It does not say whether the face should stay, whether the costume is historical or modern, where the person is standing, what the light looks like, or what should not appear. The model fills the gaps with its own average, which is usually a generic costume and a flat background.

A useful prompt is a short brief. It gives the model a hierarchy: this is the subject, this must stay, this may change, this is the scene, this is the light, these are the failures. The hierarchy is more important than the number of words.

The five-part formula

1. Preserve

Name the identity and the things that must not move. For a person: face, skin tone, facial proportions, pose, number of people. For a pet: breed, fur colour, markings, eye colour, anatomy. For a product: shape, label, logo, material.

2. Change

Name the costume, hair or makeup. Describe fabric, silhouette and colour, not just a character name. “A black velvet coat with a high lace collar and silver buttons” is more useful than “a gothic costume.”

3. Place

Name the scene and the spatial relationship. “A moonlit Victorian garden with a stone path, dark hedges and two distant lanterns.” A scene with a foreground, middle ground and background gives the model depth.

4. Light

Name the direction, colour and softness. “A soft warm key from the left, a cool rim from the back right, deep teal shadows.” Light is what makes an AI image look photographed rather than painted.

5. Avoid

Name the failure modes. “Do not change the face, do not add extra people, do not add a mask, do not add text, logos or watermarks.” This is not a magic spell, but it prevents the most common problems.

Eight prompts you can copy

1. Midnight Braids portrait

Use the uploaded photo as the primary reference. Preserve the subject’s recognizable facial identity, facial proportions, skin tone and natural anatomy. Change the hairstyle into long violet braids with delicate silver ornaments. Replace the clothing with a black-and-purple velvet outfit with lace and metallic details. Place the subject in a moonlit Victorian garden with soft fog and glowing lanterns. Photorealistic editorial photography, cinematic soft lighting. Keep the same number of people. Avoid altered identity, extra limbs, text, logos and watermarks.

2. Cosmic Princess portrait

Edit the uploaded photo into an original celestial Halloween princess portrait. Preserve the subject’s recognizable face, skin texture, body proportions and pose. Replace the outfit with a midnight-blue gown with silver embroidery and softly glowing star patterns. Add a silver moon crown. Place the subject in a moonlit observatory with a starry sky through tall windows. Elegant fantasy fashion photography, cinematic rim lighting, deep blue and violet palette. Avoid altered identity, extra people, text and watermarks.

3. Skeleton Glam makeup

Edit only the makeup in the uploaded portrait. Preserve the face, skin texture, freckles and pose. Add deep plum eyeshadow, a soft smoky edge, a thin silver line along the upper lash line and a berry lip. Keep a luminous but natural finish with visible skin texture. Use soft studio beauty lighting. Do not add a full-face mask, change the face shape, add people, or add text or watermarks.

4. Neon Witch street portrait

Transform the uploaded photo into an original neon witch portrait. Preserve the face, skin tone, body proportions and pose. Replace the outfit with a modern black witch-inspired coat, a wide-brim hat with a metallic band and silver jewellery. Place the subject on a rain-slick city street at night with violet and teal neon reflections, drifting fog and soft practical lights. Cinematic cyberpunk fashion photography. Avoid altered identity, extra limbs, text and watermarks.

5. Pumpkin Pup pet portrait

Use the uploaded pet photo as the primary reference. Preserve the animal’s breed, fur colour, markings, eye colour, facial appearance and natural anatomy. Dress the pet in a soft pumpkin costume with a small green leaf accessory. Keep the costume fitted naturally and do not cover the eyes. Place the pet in a cozy autumn studio with warm lighting and miniature pumpkins. Photorealistic pet portrait, realistic fur texture. Avoid extra animals, unnatural paws, duplicated limbs, text and watermarks.

6. Foggy Porch video

Use the provided image as the exact starting composition. Create a five-second cinematic Halloween shot of a carved pumpkin on a wooden porch at night. Keep the pumpkin shape and carving pattern consistent. Fog drifts slowly from left to right. A warm lantern flickers. The camera performs a very slow push-in. Realistic motion, stable geometry, consistent lighting, no scene cuts, no extra objects, no text, no watermarks.

7. Lantern Glow portrait video

Use the uploaded portrait as the exact reference. Keep the face, costume, hairstyle and body proportions unchanged. The subject remains still. The warm lantern flickers subtly, casting moving highlights on the costume. Soft fog drifts behind the subject. The camera performs a gentle push-in. Maintain realistic skin texture, stable anatomy and consistent lighting. Do not add people, change the costume, cut the scene, or add text or watermarks.

8. Couple Cosmic portrait

Use the uploaded photo of two people as the primary reference. Preserve both subjects’ recognizable faces, skin tones, body proportions and pose. Do not add a third person. Dress the couple in coordinated midnight-blue costumes with silver embroidery and a warm-gold accent. Place them on a stone balcony under a starry sky with soft moonlight and distant lanterns. Photorealistic editorial photography, realistic fabric texture. Avoid face swapping, extra limbs, text and watermarks.

What to change when the result is wrong

If the face drifts, strengthen the preserve line and reduce the number of changes. If the costume is wrong, add fabric and silhouette details. If the image looks flat, add a specific light source and a colour palette. If the model adds extra people, say “keep exactly one person” and “do not add a third person.” If the model adds a mask instead of makeup, say “painted on the skin, no full-face mask.”

Change one variable at a time. If you change the costume, the hair, the background and the lighting at once, you will not know which instruction caused the problem. One variable per attempt is slower to write and faster to finish.

Pets and couples need different protection

A pet prompt must protect species, breed, fur pattern and anatomy. A human costume prompt can change the silhouette; a pet costume prompt cannot. Say “keep four legs, the snout, the ears and the tail” and avoid costumes that cover the eyes.

A couple prompt needs two identity protections and an explicit “do not add a third person.” Keep both faces visible in the source photo. If one face drifts, generate each person separately with the same background and combine the approved stills.

Testing without wasting credits

Generate the shortest duration at the lowest resolution that still lets you judge the result. For a still, one image is enough to see whether the face holds. For a video, five seconds is enough to see whether the motion works. Watch at full size, check the face, hands and edges, and change one instruction before the next attempt.

Do not generate five versions of the same prompt. That is a lottery. Change the preserve line, the costume description or the light, and note which change fixed the problem. A small prompt notebook is more valuable than a larger credit pack.

Frequently asked questions

What is a good Halloween AI prompt?

A prompt that names what to preserve, what to change, where the scene is, how it is lit, and what must not appear. Specific fabric, colour and lighting details matter more than a long list of adjectives.

Can I copy these prompts for free?

Yes. The prompts are published for you to copy and adapt. They are original examples, not a license to reproduce a copyrighted costume or a real person’s likeness. Keep the source photo authorized.

Why does the model change my face?

The model may change the face when the source is low-detail, the prompt asks for too many changes, or the identity instruction is weak. Strengthen the preserve line, reduce the action, and use a sharper source photo.

Can I use the same prompt for a video?

Use the image prompt to create an approved still, then write a separate video prompt with one motion and one camera move. Video prompts need shorter, more physical instructions than still prompts.

Do these prompts work for pets?

Use the pet-specific prompt and change the animal details. Pet prompts must protect breed, fur pattern and anatomy. Do not use a human costume prompt on a pet photo and expect the same result.