Zombie Dog Memory Rescue VIDEO Breakdown (Prompts Included)
AI Generated Video
Recreating this Zombie Dog Memory Rescue video is straightforward with Pollo AI. Follow the workflow below, prepare the same references, and paste the prompts to rebuild the viral AI video style quickly.
Step 1
Open pollo.ai, select Image Generation, paste the image prompt below and click Generate. Set GPT Image 2 as the default image model. Adjust the aspect ratio according to your video needs; multiple sizes including 16:9 and 9:16 are supported.
Image Generation Screen
For the first character photo（image①）, you need to upload your personal photo and use Image-to-Image to generate the character portrait. For the German Shepherd image（image②/③）, Text-to-Image works fine; you can also generate the first German Shepherd photo with Text-to-Image, then use Image-to-Image to create the second one.
Image Prompt- 1:
hyperrealistic studio photograph,full body shot, head to toe, an 28-year-old man standing straight facing camera, slim and wiry, tired posture with shoulders slightly forward, brown leather flight cap with ear flaps, amber-tinted brass goggles pushed up on her forehead, faded olive-green hooded military jacket with dark leather shoulder panels and ammo loops on the chest strap worn open and half slipping off her left shoulder, dark olive sleeveless top, wide grey utility belt with a heavy metal buckle and pouches, baggy grey-blue cargo pants ripped and patched and dirty tucked into her boots, strapped knee pad on the left leg, laced black-brown combat boots with extra straps and worn soles, thigh holster with a pistol on her right thigh, pale hard pouch strapped to her left thigh, fingerless tactical glove on her right hand, scratched mechanical wristwatch on her left wrist, olive canvas rucksack with a rolled bedroll and a worn dirty pink plush bunny hanging off it, small yellow bone charm on her belt, holding a bolt-action hunting rifle with a scratched dark red painted stock wrapped in leather cord and grey tape, a makeshift dark red hooked spear slung across her back, muted desaturated palette of olive green grey-blue brown leather and black, pure white seamless cyclorama background, soft even diffused studio lighting, flat neutral light, subtle contact shadow under the boots, shot on Sony A7R V, 50mm lens, f/5.6, sharp from head to toe, visible fabric weave and stitching, scuffed leather, dust and dirt on the trousers, ultra sharp detail, 8k, photorealistic, documentary realism.
Reference Image
AI Generated Image
Image Prompt- 2:
full body,a large male black-and-tan German Shepherd, 5 years old, 35kg, classic saddle-back markings with a black saddle over the back and rich golden-tan on the legs chest and face, dark facial mask, large erect ears, clear amber-brown intelligent eyes, wet black nose, thick glossy well-groomed coat, bushy sabre tail, wearing an olive-drab MOLLE tactical harness with two side pouches and a top grab handle, a round metal dog tag on the collar, a dark red knitted scarf wrapped around his neck, a small yellow bone charm clipped to the harness, holding a worn yellow-green tennis ball in his mouth, alert and happy, neutral studio lighting, plain grey background, photorealistic, highly detailed
AI Generated Image
Image Prompt- 3:
full body,the same large male black-and-tan German Shepherd but infected and diseased, milky white clouded eyes with almost no visible pupil, burst blood vessels around the eyes, matted clumped fur with bald patches showing grey skin on the flank and neck, dark almost black veins visible under the skin of the muzzle and around the mouth, dried blood on the lower jaw and chest, strands of saliva, ragged hoarse breathing, slight muscle tremor in the shoulder, wearing the same olive-drab MOLLE tactical harness now torn on one side with one pouch hanging loose, the dark red knitted scarf faded dirty and half unwrapped, a small clean yellow bone charm still clipped to the harness and glinting, sitting upright and calm like a trained dog, restrained realistic decay, not a monster, no glowing eyes, no exaggerated fangs, no gore, neutral studio lighting, plain grey background, photorealistic, highly detailed
AI Generated Image
Create Zombie Dog Memory Rescue Video on Pollo AI
Upload your references and prompt to generate a polished zombie dog memory rescue short.
Start Creating Free
Step 2
Once your image is ready, go to Video Generation. Upload your prepared image or pick assets from your Creations library. Select reference-to-video mode and paste your video prompt. You can also upload a video for motion reference.
Note: Add @your-image for image reference and @your-video for video reference inside the prompt. Pick any of these models: Seedance 2.5, Seedance 2.0, MiniMax H3 or Wan 3.0. Click Generate after confirming all settings.
Video Generation Screen
Video Prompt
Create a short post-apocalyptic thriller with emotional depth. The central characters are a woman and the same dog. The main plot involves the woman being chased down an abandoned street by the infected dog. Just as the dog is about to pounce on her, a motion-matched crossfade transitions to a warm memory from the past—a healthy dog leaping into her arms—before crossfading back to the present, where the woman recognizes the dog and reaches out to comfort it. The entire video should avoid hard cuts, white flashes, or sudden cuts to black.
Reference Material Requirements:Image 1: Used to establish the woman's facial features, hairstyle, body type, clothing, and overall identity in the present-day sequence; no background image to be used. Image 2: Used to establish the dog's breed, body type, coat density, ears, tail, facial structure, and affectionate demeanor while healthy. Image 3: Used to establish the same dog's coat texture, gaze, stains, aggressive posture, and menacing demeanor while infected. Image 4: Used to depict the woman's facial features, hairstyle, body type, clothing, and the gentle act of hugging the dog in the warm memory sequence; no background is to be included. Images 1 and 4 together define two visual states of the same woman; two different women must not be generated. Images 2 and 3 together define two states of the same dog; two different dogs must not be generated.
Subjects and Relationships:The entire film features only one woman and one dog. The woman is always the same person defined by both Image 1 and Image 4: the "real-world street" segment prioritizes the state shown in Image 1, while the "warm memory" segment prioritizes the state shown in Image 4. The dog is always the same dog defined by both Image 2 and Image 3: the "real-world street" segment uses the infected state, while the "memory" segment uses the healthy state. The relationship between the woman and the dog evolves from flight and fear to recognition, approach, embrace, and comfort.
Sequential Action Script:The opening scene is a desolate, abandoned city street in cold, gray tones, with a damp ground, damaged vehicles in the distance, dust clouds, and dilapidated buildings. The woman runs, panting, glancing back frequently, her face etched with fear and disbelief. The infected dog chases her from deep within the street, crouching low before suddenly accelerating with ferocious movements. Its appearance adopts the infected state shown in Image 3, while retaining the basic breed and body structure of Image 2. The woman takes cover beside a wrecked vehicle as the infected dog lunges at her. The dog's body leaves the ground, its front paws reaching toward the woman's shoulder, its mouth agape as it growls low, its entire body carrying a powerful forward momentum. The woman raises her arm in defense, leaning backward, nearly knocked to the ground. The camera zooms in on the dog's front paws, the woman's arm, and the dog's shoulders and back; the car shakes, and dust flies, creating a tense moment without showing exaggerated or gory imagery.
Just as the dog is about to knock the woman down—in that split second in midair—use a motion-matched similar dissolve transition to cut to the past. Do not use a hard cut. Maintaining the dog's forward momentum, outstretched front paws, and the woman's leaning-back posture, the scene begins to fade softly: the dog's disheveled fur gradually becomes transparent and transforms into the healthy, soft fur shown in Image 2; the woman's raised, defensive arm fades into her past self with arms wide open to catch the dog; the cold, gray edges of the car fade into the edges of a tree trunk in the park; and the dust and mist fade into warm, golden sunlight and tree shadows. Throughout the transition, the dog continues to lunge toward the woman's arms along the same trajectory; the landing points of its front paws, the height of its body, and the woman's backward movement to catch it remain consistent.
In this heartwarming memory, the woman—in the gentle state depicted in Image 4—stands beside a tree in the sunlight, spreading her arms wide to catch the healthy dog. The dog's front paws rest on the woman's shoulders, its body carrying the momentum of leaping forward into her embrace; the woman takes a half-step back with the flow of the moment, smiling as she pulls her arms tighter to hold it close. The grass is bright, the tree shadows soft, and sunlight filters through the leaves—the scene is warm, quiet, and reminiscent of a cherished old memory. The camera gently traces a half-circle around the woman's arms and the dog's shoulders and back, maintaining the speed and direction of the embrace.
The memory sequence continues to show the embrace, then gently fades back to the street in the present. The outline of the woman holding the healthy dog, the position of the dog's front paws on her shoulders, and the placement of her arms all align with the woman's posture as she holds the infected dog in the present. The warm golden sunlight, grass, and tree shadows gradually fade, giving way to the cold, gray street, damaged vehicles, and dust-filled haze. As the transition completes, the woman is already standing beside the car, holding the infected dog by its neck and shoulders. Her body is pressed against the car by its weight, but she no longer pushes it away. Her expression shifts from fear to choked emotion and heartache, and she slowly tightens her arms. The infected dog's aggressive momentum is defused by the embrace; its front paws rest on the woman's shoulders, and its body gradually calms down.
Finally, the shot cuts to a close-up. The woman cradles the infected dog's face in both hands, gently guiding its gaze toward her, and slowly brings her forehead close to the dog's. The dog still retains the infected appearance shown in Image 3, but its gaze briefly softens. The woman's eyes are moist, and her lips tremble slightly, though she says nothing. The final shot lingers on the woman and the infected dog facing each other closely, with the dilapidated street still in the background.
Cinematography and Style:The overall style is cinematic and realistic. The reality sequences use cool gray-blue tones, handheld tracking shots, shallow depth of field, low-angle chase shots, and tense close-ups; the memory sequences use warm golden sunlight, stable and soft shots, bright grass, and tree shadows. The most crucial transition occurs at the moment the infected dog is about to pounce on the woman, achieved through a similar dissolve: the dog's forward momentum, the landing point of its front paws, its body height, the woman's backward lean to brace herself, and the movement of her arms are perfectly matched between the reality and memory sequences. The dissolve is smooth and natural, avoiding hard cuts or unrelated scene changes.
Sound:Reality sequences: The woman's rapid but clear breathing; the brief sound of her shoes hitting damp ground; the infected dog's low, dry growl; the scraping sound of its claws against the ground; and a single, low, muffled metallic thud as the car is struck. The dog's growl is a short, low-frequency, chest-driven threat—without any wet, bubbly sound effects. When fading into the memory sequence: The real-world sounds transition into a gentle breeze, the rustling of leaves, the soft panting of a healthy dog, and the rustling of fabric. The sound changes should feel as though the air has become lighter and the space more open, rather than as if entering water. When fading back to reality: The sounds of leaves and the warm breeze gradually fade, and the desolate ambient sounds of the cold, gray street slowly return; only a very faint layer of the infected dog's low guttural growl remains, without masking the woman's breathing. The ending leaves only the woman's suppressed breathing, the dog's now-fainter snoring, and the low ambient sounds of the abandoned street in the distance—the overall atmosphere is quiet and restrained.
Maintain Consistency:Keep the number of women at one and the number of dogs at one. Image 1 and Image 4 depict different states of the same woman; do not generate a second woman. Image 2 is used only for the healthy memory state, and Image 3 is used only for the infected reality state; both states belong to the same dog—do not generate two dogs. Maintain consistency in the woman's identity, the dog's breed and structure, the vehicle's position, the street's direction, and the contrast between warm and cool color tones.
Final Summary
Create a short, cinematic zombie dog memory rescue sequence using the prepared reference assets. Keep the subject identity, first-frame setup, motion timing, camera style, and continuity rules consistent.Want to bring your Zombie Dog Memory Rescue video to life in minutes? Upload your references, paste the prompt, and create your own cinematic AI short with Pollo AI.