万圣节 AI 提示词与 Prompt Builder

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Create long violet braids, an elegant black-and-purple velvet costume and delicate silver hair accessories. Keep the outfit wearable and fully clothed.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Create an original midnight-blue celestial gown with silver embroidery, subtle star details and a delicate moon crown. Keep the clothing realistic and fully covering.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Dress the subject in a believable wearable robot costume with articulated shoulder panels and subtle blue light accents. Keep the original face visible; do not replace it with a robot head.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Create an original forest-witch costume with a tall soft hat, layered moss-green fabric, a dark cloak and a simple wooden staff. Use realistic materials and a grounded editorial composition.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Create a classic, non-graphic vampire costume with a tailored black cape, burgundy fabric and a small antique brooch. Keep the face recognizable and the expression natural; no blood or injury.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Create an original eerie unicorn mascot costume using an ivory textured hood, a short silver horn and worn theatrical fabric. Keep the human face fully visible, with no franchise logos, no copied character design and no gore.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Apply tasteful, non-graphic skeleton-inspired face makeup. Preserve the subject's facial geometry, hairstyle, body, existing clothes and pose. Change makeup and lighting only; do not replace the outfit.

Preserve the original background and setting; change makeup and lighting only.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded pet photo as the primary reference. Preserve its breed, fur color, distinctive markings, eye color, facial appearance and natural anatomy. Keep the same number of animals.

Dress the pet in a soft pumpkin-inspired costume with a small green leaf accessory. Keep eyes and ears unobstructed, with natural paws and a comfortable-looking fit.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded pet photo as the primary reference. Preserve its breed, fur color, distinctive markings, eye color, facial appearance and natural anatomy. Keep the same number of animals.

Dress the pet in a soft, wearable dinosaur costume with a short padded tail and subtle luminous trim. Keep its face visible and all limbs anatomically natural.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded pet photo as the primary reference. Preserve its breed, fur color, distinctive markings, eye color, facial appearance and natural anatomy. Keep the same number of animals.

Dress the pet in a soft tan baked-potato-inspired outfit with a fabric foil-like wrap detail. Keep the costume clear of the eyes, nose and paws; do not turn the animal into food.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded group photo or separate portraits as identity references. Preserve each person's own recognizable face, age, skin tone and body proportions. Keep the same number of people as the references. Never swap identities between people or invent extra people. Assign each person a clear reference role before editing.

Assign the first person a warm western cowboy costume and the second person a pale angel costume with lightweight fabric wings. Preserve which person wears which outfit. Keep both people fully clothed and naturally positioned together.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded group photo or separate portraits as identity references. Preserve each person's own recognizable face, age, skin tone and body proportions. Keep the same number of people as the references. Never swap identities between people or invent extra people. Assign each person a clear reference role before editing.

Create coordinated, age-appropriate moon-and-star costumes for the whole group, using soft navy capes, silver star accessories and comfortable clothing. Keep all children and adults their original ages and sizes; keep the mood friendly rather than frightening.

Set the scene in a moonlit autumn garden with a little ground fog and warm lantern light.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Use the uploaded photo as the primary reference. Preserve the subject's recognizable face, facial proportions, skin tone, approximate age and natural anatomy. Keep the same number of people.

Keep the people, faces, hairstyles, clothing, pose and framing unchanged. Replace only the background with the selected setting, adding carved pumpkins, warm lanterns and dry leaves. Match the original light direction and contact shadows.

Use a refined autumn photography studio with realistic soft lighting, contact shadows and an uncluttered backdrop.

Use realistic editorial photography, believable fabric and fur, natural texture and a clear full composition.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.

Locked tripod and fixed lens. Keep the original background landmarks, framing and camera distance unchanged. Show the entire person, head and both feet throughout. All motion happens in place; no zoom, reframing or movement toward the lens.

A brisk five-second full-body werewolf transformation. After one brief ordinary-person beat, the knees flex, shoulders and ribcage broaden, the back arches, the arms lengthen into powerful forearms and clawed hands, and the thighs and calves reshape into a strong bent-hock stance beneath the trousers. The compact muzzle, ears, jaw and neck fur change together with the body. Keep the ORIGINAL TROUSERS visibly present for every frame: the same opaque fabric and colour remain over the waist, hips, seat, groin, thighs and calves, stretching and reshaping around the new leg contours while staying securely worn. The trousers never vanish, tear open, slide down or turn into fur. Fur appears only on the neck, exposed arms and hands, and ankles below the trouser cuffs; clawed paws remain visible below those cuffs. Keep the original shirt opaque and intact, adapting to the broader torso. Finish the whole-body transformation by the third second, then hold a low, forceful stance with one short claw gesture. Tactile practical-creature detail, clear full-body silhouette, no exposed intimate anatomy.

Timing for the five-second output: keep the ordinary-person beat and full-body change at the same brisk natural speed, completed by 3.0 seconds, with the short claw gesture finished by 4.0 seconds. From 4.0 to 5.0 seconds, hold the final full-body stance with only subtle breathing and fur settling. The original opaque shirt and trousers remain securely worn. No new action, slow motion, forward travel or camera change.

A brisk five-second practical-effects Halloween transformation, one continuous shot. Use the uploaded ordinary adult portrait as the exact opening appearance. Keep the same opaque shirt, shoulders, body proportions and real background. A locked camera, fixed focal length and original framing throughout; all movement is head and shoulders at the original distance.

Begin the action immediately: the person turns their head and shoulders a little to one side with a confident expression. A tight sweep of glowing orange embers passes sideways across the head. As the embers clear, the entire head has become a three-dimensional, naturally proportioned real carved pumpkin, replacing the hair and face rather than painting a mask onto human facial skin. The pumpkin occupies the original head-and-hair volume, with a compact stem, deep organic rind grooves, asymmetric narrow carved eyes and a sharp crooked jagged mouth. Real pumpkin flesh is visible along the thick cut edges. The hollow interior contains flickering amber fire; it is not a smiley face or a plastic costume mask.

By the second second the transformation is complete. The pumpkin-headed person snaps their head back to face the viewer, makes one short sharp sideways head tilt, and turns one shoulder with a confident, menacing Halloween pose. At that movement the fire flares once through the carved openings, throwing a brief warm glow onto the unchanged shirt and a few embers sideways; then it settles to a low flicker for a very short final beat. Clear anticipation, a decisive reveal, and one energetic payoff instead of a static face-filter hold. Keep the pumpkin head attached to the neck and the whole shirt fully opaque and intact. No camera push, walking toward the lens, giant head, removal of clothing, gore, text or logos.

Timing for the five-second output: finish the transformation by 2.0 seconds and the head tilt, shoulder turn and single fire flare by 4.0 seconds at brisk natural speed. From 4.0 to 5.0 seconds, let the small embers fade and the interior fire settle to a low flicker; the head and shoulders stay at their original distance. No second flare, new action, slow motion or camera change.

One continuous five-second wide master shot, locked tripod, fixed focal length. The opening frame matches the reference photograph: the same adult person standing full length in the same setting and lighting as the reference, with the same visible clothes, shoes, ground, background landmarks and camera distance. Keep the complete person and the ground around their feet inside this same fixed wide composition for the entire shot. The background landmarks never enlarge or shift. No zoom, push-in, dolly, reframing, cuts or close-up. Preserve the person's identity and original opaque outer clothing and complete trousers, fully opaque and intact throughout.

Play this exact action sequence in order, as a fast horror-to-comedy reversal. Begin with the person alone and the background empty: the ghost is NOT already present in the opening frame.

0.0-1.3 seconds: In the empty background over the person's right shoulder, low charcoal mist rapidly twists upward into a ragged dark plume. A single gaunt, hollow, threatening face emerges inside a hood-shaped crown of dense black smoke. The person notices movement and visibly turns their head and shoulders back toward the newly forming face. Show the mist growing from empty space AND the person's turn before any other interaction.

1.3-2.5 seconds: Seeing the looming face, the person visibly recoils in fear, eyes wide, stumbles one short step sideways and falls onto their bottom on the visible ground beside their original standing spot. Show the full fall in the fixed wide shot: knees bend, hips drop, one hand braces on the ground, ending clearly seated on the ground with bent knees. A quick harmless startled tumble, not an injury, not a jump or dance. Their original top and trousers remain fully on. The person looks up fearfully at the ghost.

2.5-4.0 seconds: The smoky apparition slides sideways to the seated person's side at the same camera depth and unexpectedly breaks into a playful dance instead of attacking: two quick, exaggerated side-to-side shuffle steps and two rhythmic shoulder bounces, its opaque dark lower plume swishing left then right. Make the dance clearly readable, energetic and funny against the person's frightened seated reaction. The person stays seated and watches in confused surprise; only the ghost dances. Keep the hollow smoke face and ragged black-mist outline recognizable. The apparition is dense opaque smoke or a completely opaque hooded silhouette; no exposed human anatomy or unclothed double. Both characters remain fully in view. At 4.0 seconds the ghost is doing the last dance bounce beside the seated person.

Brisk natural-speed action, no slow motion. The payoff is the contrast: frightening emergence, human turns and falls in fear, then the ghost dances. Preserve that causal order, the original setting and lighting and the fixed full-body wide framing. No contact with the lens, no fighting, no blood, no text or logos.

From 4.0 to 5.0 seconds, the ghost finishes its last dance bounce and settles beside the seated person with only a little sideways smoke drift. The person remains seated, looking up in confused surprise; no second fall or additional interaction. Keep both characters fully visible at the same camera depth. Preserve the original 0.0-1.3 / 1.3-2.5 / 2.5-4.0 second action boundaries and brisk natural speed; do not stretch the emergence, turn, fall or dance across five seconds.

A lively eight-second Halloween dance duet in one fixed wide shot. Start from the uploaded adult portrait and keep the same recognizable person, original background, lighting, fully opaque original top and complete original trousers throughout. Keep head, hands and shoes inside the frame. Locked tripod, fixed lens and camera distance. No zoom, forward steps, cuts or exposed human anatomy.

A friendly white GHOST is the second DANCER. It has a rounded white head and pear-shaped floating body, big black oval eyes, a smiling mouth, two long flexible white arms with rounded hands, and a curved tail instead of legs. The ghost has solid three-dimensional volume with soft edges, not a flat light or a robe worn by a person. It appears in the empty space to the viewer's left of the adult, about chest-to-knee height. Leave a small clear gap between the two dancers.

0-0.6s: the ghost pops into view and immediately catches the same dance beat as the adult.
0.6-7.6s: BOTH partners perform a continuous, energetic step-touch groove together. Repeat this short phrase: lean LEFT and sweep the right arm across the body; bend DOWN in a pronounced bounce; lean RIGHT and sweep the left arm across the body; stretch UP with both elbows lifted. Repeat the phrase at least twice. The adult adds relaxed side steps, knee dips and shoulder bounces, keeping the natural human dancing of a casual party. The ghost copies the same beat with BIG whole-body bends, a squash-and-stretch bounce, alternating arm sweeps and a tail flick. Its whole pear-shaped torso visibly changes shape and direction on each beat. Keep the ghost's dance compact beside the person; its arms and body dance rather than flying around the person. BOTH move simultaneously, with no static floating ghost or held ending.
7.6-8s: both finish with one last bounce and arm accent, still dancing as the clip ends. Playful confident energy at normal speed. Preserve original clothes and fully covered human body. No additional people, text or logos.

Use the approved image as the scene and identity reference. Keep its architecture and overall framing. Keep every existing person, animal, outfit and recognizable face consistent throughout the clip. Preserve the original number of subjects and natural anatomy.

The subject turns once toward the camera as a nearby lantern switches on and a thin bank of fog crosses the ground. Give the action a clear beginning and ending in one continuous shot. Keep the camera fixed.

Keep motion coherent and the original composition readable. Changes in light must follow the described event with natural illumination and shadows.

Avoid extra limbs, duplicated subjects, distorted anatomy, text, logos and watermarks.