AI video prompts
Real prompts with the videos they made. Copy one, or press Generate: Create Video opens with the model, settings and files already set, so you only add your photo.
7 prompts · 0 guides
All prompts
Charming Smirk 😏
Replace the face of the main male character in the source video with my face from . The attached collage shows the same person from multiple angles; use all views to accurately preserve my identity. Match my facial proportions, eyes, nose, lips, jawline, skin tone, and distinctive facial features. My face must remain recognizable and consistent throughout the entire video, including during head turns, speech, and fast movements. Preserve the original character’s body, clothing, actions, head movements, gaze direction, facial expressions, and lip-sync timing. Keep the original camera movement, framing, background, other people, editing, and audio unchanged. Use Image 1 only as a facial identity reference; do not copy the suit or background from the photos. Blend my face naturally into the scene with matching lighting, shadows, perspective, and motion blur. Preserve realistic skin texture and natural occlusions when objects or hands pass in front of the face. No identity drift, flickering, warped features, beauty filters, or visible face-swap seams.
Beach Gallop 🐎
Replace the female rider in the video with the person from @[Image 1]. Keep their face, facial features, skin tone, hairstyle and body proportions exactly as in the photo, with consistent identity in every frame. Keep everything else identical to the original video: the dapple-grey horse with the red saddle pad and black tack, the full-speed gallop along the wet shoreline, the golden-hour beach, soft blue sky, rolling waves, water splashes and reflections on the wet sand. The person from @[Image 1] rides confidently in the same pose: leaning forward in the saddle, hands on the reins, body moving naturally with the horse's rhythm, hair and clothes blowing in the wind. Dress them in a fitted white top and light riding pants with black boots, matching the original outfit. Match the original camera tracking, motion, lighting and colour grading. Natural shadows and wind on the new rider, realistic skin texture, no face distortion, no flicker, smooth continuous motion, photorealistic, cinematic quality.
Party Entrance 😎
Replace only the main foreground man wearing the black T-shirt with the large K18 chest graphic with MARCEL, the golden cat shown in the reference image. Track this exact man continuously throughout the entire source video, including every front view, profile, turn, back view, partial occlusion, and the final frames. The reference sheet shows three views of ONE cat. Generate only ONE Marcel. IDENTITY AND APPEARANCE: Match the reference cat closely: round feline face, golden shaded fur, lighter cream muzzle and chest, green eyes, pink nose, detailed whiskers, small triangular ears, compact feline body, rounded paws, and a furry tail with a dark tip. Preserve the reference silver Cuban-link chain and the pendant reading “MARCEL”. Keep his identity, fur markings, proportions, and accessories consistent across all frames. Replace the man's entire body and clothing with Marcel's furry feline body. Marcel wears only the reference chain and pendant. No black T-shirt, no trousers, no K18 graphic, no human skin or human hands. SOURCE MOTION IS THE CHOREOGRAPHY: Transfer the source man's performance frame by frame. Preserve the exact timing, rhythm, sequence, screen position, direction of travel, foot placement, body sway, shoulder movement, head turns, arm gestures, pauses, and acceleration. Marcel performs upright on his hind legs. Map the man's arms to Marcel's front legs and paws, and the man's legs to Marcel's hind legs, while retaining believable feline anatomy. Adapt limb reach to the reference cat's proportions without changing the timing or intention of any gesture. Preserve the source approach toward the camera, the right-hand-to-mouth gesture as a right-front-paw-to-muzzle gesture, the lowering of that paw, the close frontal performance, the sideways turn, the complete turn away from the camera, and the final movement back toward the guests. Preserve all visible handheld objects, their trajectories, and their contact with the corresponding paws. FACIAL PERFORMANCE: Transfer the timing and emotional intent of the man's visible expressions into realistic feline facial movement: gaze direction, blinks, eyelid tension, muzzle movement, mouth opening and closing, and head tilts. Keep Marcel's recognizable cat face throughout. No human lips, human teeth, or exaggerated cartoon expressions. SCENE INTEGRATION: Keep the original guests, their faces and clothing, furniture, decorations, room, camera movement, framing, duration, and audio unchanged. Preserve correct foreground and background occlusions. Match the moving pink, purple, green, and warm event lighting on Marcel's fur, eyes, chain, and pendant. Add physically consistent floor contact shadows and natural motion blur. Keep his paws grounded without sliding. The chain, pendant, fur, and tail must react naturally to his movement without obscuring the source gestures. Maintain consistent scale and perspective as he approaches and moves away from the camera. Reconstruct any background revealed by replacing the man's silhouette. Photorealistic live-action integration, detailed natural fur, stable anatomy, coherent lighting, and consistent identity throughout. No additional cats, duplicated limbs, human fingers, residual human face, body morphing, flickering fur, floating paws, sliding feet, warped jewelry, new cuts, slow motion, or invented choreography. Do not leave the original man visible in profile or back-facing frames.
Elevator Meet-Cute
STRICT SOURCE CAMERA AND ENVIRONMENT LOCK: Replace only the two actors. Preserve the supplied video's exact elevator wall and door colors, materials, reflections, exposure, white balance, saturation and lighting in EVERY shot. Do not recolor or redesign the elevator. Keep exact source camera position, lens perspective, field of view, camera distance, framing, headroom, crop and subject screen size at every frame. No zooming in or out, no wider or tighter shots, no reframing, no new camera movement. Preserve every original cut at its exact time. Reference photos define ONLY the people and their clothes, never the framing, lighting, background, warm color grade, grain, borders or printed text. Render the replacements under the source video's existing light. Ignore all letters and borders in the reference photographs. Recreate the ENTIRE supplied elevator video with the TWO fictional people in the reference photographs performing the exact source acting and movements. PERMANENT CAST: PHOTO1 replaces ONLY the original man on the LEFT wearing headphones, grey sweater, white collar and striped tie. PHOTO2 replaces ONLY the original woman on the RIGHT with dark bangs and white sleeveless blouse. Track each same person throughout every shot, close-up, turn and partial occlusion. The woman exits at the exact original moment. The final shot contains ONLY the PHOTO1 man; do not reinsert the woman or transfer her identity to him. FULL REPLACEMENT: Use each reference's complete recognizable face, skull shape, hairline, hairstyle, hair length and color, ears, facial hair or lack thereof, skin tone, neck, shoulders, torso, arms, hands, natural body build and proportions. Rebuild entire heads and bodies, not just faces. Remove the original man's hairstyle and original woman's bangs and hair, replacing them fully with their assigned reference hairstyles in front, profile and rear views. No inherited source head shapes or bodies. Preserve normal human proportions and each person's own identity through expression changes. STRICT SOURCE CAMERA AND ENVIRONMENT LOCK: Replace only the two actors. Preserve the supplied video's exact elevator wall and door colors, materials, reflections, exposure, white balance, saturation and lighting in EVERY shot. Do not recolor or redesign the elevator. Keep exact source camera position, lens perspective, field of view, camera distance, framing, headroom, crop and subject screen size at every frame. No zooming in or out, no wider or tighter shots, no reframing, no new camera movement. Preserve every original cut at its exact time. Reference photos define ONLY the people and their clothes, never the framing, lighting, background, warm color grade, grain, borders or printed text. Render the replacements under the source video's existing light. Ignore all letters and borders in the reference photographs. Recreate the ENTIRE supplied elevator video with the TWO fictional people in the reference photographs performing the exact source acting and movements. PERMANENT CAST: PHOTO1 replaces ONLY the original man on the LEFT wearing headphones, grey sweater, white collar and striped tie. PHOTO2 replaces ONLY the original woman on the RIGHT with dark bangs and white sleeveless blouse. Track each same person throughout every shot, close-up, turn and partial occlusion. The woman exits at the exact original moment. The final shot contains ONLY the PHOTO1 man; do not reinsert the woman or transfer her identity to him. FULL REPLACEMENT: Use each reference's complete recognizable face, skull shape, hairline, hairstyle, hair length and color, ears, facial hair or lack thereof, skin tone, neck, shoulders, torso, arms, hands, natural body build and proportions. Rebuild entire heads and bodies, not just faces. Remove the original man's hairstyle and original woman's bangs and hair, replacing them fully with their assigned reference hairstyles in front, profile and rear views. No inherited source head shapes or bodies. Preserve normal human proportions and each person's own identity through expression changes. CLOTHING: PHOTO1 man wears his exact reference muted navy-blue short-sleeved camp-collar button-up shirt, charcoal trousers and dark shoes. Preserve his reference black parted hairstyle, moustache and light chin stubble. PHOTO2 woman wears her exact reference cream cardigan and cream top, dark indigo wide-leg jeans and black loafers, small earrings. Preserve her shoulder-length layered dark hair and reference facial features. Match each reference's actual clothing, colors, fit and accessories, without inheriting source clothing. Retain original headphones and shoulder bag as external props on the man's track at their original times. ACTION AND LIP SYNC: Transfer precise poses, hand gestures, gaze, emotional reactions, head turns, timing and walking to the reference people's natural bodies. Each person visibly performs only the exact words spoken or sung by their corresponding original role, with syllable-accurate lips/jaw synchronization to the unchanged source audio. Preserve complete phrases, pauses and silent listening; offscreen speech stays offscreen. Natural mouth anatomy, no random chewing, protruding tongue or frozen portrait expressions. SCENE: Preserve the original elevator, doors, lighting, shadows, camera positions, movement, framing, all cuts, original soundtrack and complete source duration from first to last frame. Do not import photo backgrounds or static poses. Photorealistic live-action, natural skin and hair, stable outfits and identities, correct prop contact and occlusion. No extra people, duplicated identities, swapped faces, merged bodies, extra limbs, captions or new actions. SCENE: Preserve the original elevator, doors, lighting, shadows, camera positions, movement, framing, all cuts, original soundtrack and complete source duration from first to last frame. Do not import photo backgrounds or static poses. Photorealistic live-action, natural skin and hair, stable outfits and identities, correct prop contact and occlusion. No extra people, duplicated identities, swapped faces, merged bodies, extra limbs, captions or new actions.
Storm Ai
1. CENTER CHARACTER: Replace ONLY the single main character standing in the exact center of the group with the character from @Image1 . The center position, body placement, scale, pose, movement, and screen position must remain consistent with the original video. 2. SURROUNDING CHARACTERS: Replace EVERY OTHER PERSON surrounding the center character with the character from @Image2 . This includes ALL visible people around the center character: front row, left and right sides, middle rows, and background. Do NOT leave any of the original surrounding people unchanged. Every surrounding person becomes the same @Image2 character, repeated as identical copies. 3. IDENTITY LOCK: Keep each replacement character's face, hairstyle, skin tone, body proportions, outfit, and color palette exactly as shown in the reference images, in every single frame. No morphing, no identity drift, no blending between image_1 and image_2. 4. MOTION AND TIMING: Preserve the original choreography frame by frame. Each replaced character follows the exact same body movement, limb motion, head turn, facing direction, walking speed, and timing as the original person they replace. Do not add, remove, or re-time any motion. 5. CAMERA AND FRAMING: Keep the original camera angle, camera movement, focal length, depth of field, and shot duration unchanged. Do not re-frame, crop, zoom, or stabilize. 6. ENVIRONMENT: Keep the background, set, props, floor, and all non-human elements 100% identical to the original video. Only the people are replaced. 7. LIGHTING INTEGRATION: Match the original lighting direction, intensity, color temperature, shadows, and reflections onto the new characters. Contact shadows and ground contact must stay physically correct, so the characters look natively filmed in the scene, not pasted on.
Mafia Boss
Butterfly
Transfer the exact body movements, choreography, gestures, timing, and motion from the source video to the person in the reference image. Preserve the person’s exact face, hairstyle, skin tone, body proportions, clothing, and identity. Keep the movement natural and realistic. Maintain consistent facial features and clothing throughout the entire video.
How to write an AI video prompt
A good AI video prompt says who or what is in the shot, what happens, how the camera moves and how it should look. Seedance follows plain sentences best: one subject, one action and one camera move per shot.
Write it like a note to a camera crew. Put the subject first, then the action and its pace, then the camera, then the light and style. When the video uses your photos, name them with @ tags (like @Face) so the model knows which file is which.
| Subject | Who or what is in the shot: “a woman in a red saree” |
|---|---|
| Action | What happens, and how fast: “spins slowly” |
| Camera | “slow push in”, “drone shot following from behind” |
| Light and style | “golden hour, cinematic, soft film grain” |
| Files | @Face for your photo; @Style or @Music for files a prompt comes with |
How to use a prompt
- 01
Pick a prompt
Watch the sample video to see what it makes. Copy the prompt, or keep reading the guide for tips.
- 02
Press Generate
Create Video opens with the prompt, model, aspect ratio, length, resolution and any files already set. Add your photo if the prompt asks for one.
- 03
Generate and download
Change anything you like, then generate. Most videos are ready in a few minutes.
Frequently asked questions
Can I copy these prompts?
Yes. Every prompt has a Copy button, and Copy all on each guide copies the whole list at once.
What does Generate do?
It opens Create Video with the prompt, the model, the aspect ratio, the length, the resolution and any files the prompt needs already set. You add your own photo where a prompt asks for one, change anything you like, and generate.
Which model are these prompts for?
Most are written for ByteDance Seedance 2.5, which makes 4–30 s videos at 480p, 720p, 1080p. Each prompt shows its model, and Generate picks it for you.
What do tags like @Face or @Music mean?
They name the files a video uses. @Face is the photo you add; tags like @Style or @Music are files that already come with the prompt. The model reads each tag as that file.
How much does a video cost?
Each Generate button shows its price in credits. Plans start at $7 / ₹499 and credit packs at $4 / ₹299. Failed videos are refunded automatically.
Write your own prompt
Describe the shot, add a photo if you like, and generate with Seedance 2.5.