Rainy Night Chase Cinematic Video
A prompt for a 30-second hyper-realistic video of a tense chase on rainy city streets, featuring wet road reflections and dynamic camera movement.
8,010 prompts
A prompt for a 30-second hyper-realistic video of a tense chase on rainy city streets, featuring wet road reflections and dynamic camera movement.
Detailed prompt for a 30-second cinematic 3D animation of a girl and her wolf cub exploring a magical wilderness during golden hour.
A prompt for a 20-second cinematic video featuring a Korean woman making a dramatic entrance and performing martial arts in a subway station.
A detailed prompt for generating a 15-second cinematic martial arts fight video, including specific camera angles, character consistency, and physics-based choreography.
Prompts for generating fresh and bright beauty commercial videos featuring a Southeast Asian woman demonstrating a skincare routine with specific products.
A comprehensive prompt for Seedance 2.0 Mini to generate a 15-second Xianxia short film titled "Xie Bai," featuring complex character interaction and moral tension.
A prompt for a 30-second hyper-realistic video showing a Korean girl's morning routine with cinematic lighting and vertical composition.
Detailed prompt for Seedance 2.0 to create a realistic documentary-style clip of a man dropping giant soap into a volcano, causing bubbles to flood a helicopter.
A detailed prompt for Seedance 2.0 to generate a cinematic music video with specific characters, styles, and lyric instructions.
A highly detailed prompt for generating a 15-second cinematic Chinese Xianxia short film using Seedance 2.0 Mini, including instructions for character consistency and Arri Alexa visual aesthetics.
A detailed music video storyboard prompt depicting an Asian woman in a tattered wedding dress among rainy ruins, used for model comparison.
A detailed prompt for a fantasy action video depicting a clawed Chinese female character battling enemies in a palace, emphasizing cinematic combat and mystical effects.
Detailed prompt for generating hyper-realistic drone footage from path images using Seedance, including camera movement and subject preservation instructions.
A humorous prompt for a 12-second realistic smartphone video of a cat riding a goose being chased by a dog, including physics and comedic timing instructions.
A detailed prompt for Seedance 2.0 Mini to generate a 15-second comedic Chinese Xianxia short film about a generational hierarchy realization during a duel.
A cinematic video prompt showing a small bear interacting with and driving a vintage car in a rural setting.
A prompt for generating epic Three Kingdoms battlefield videos with Hollywood blockbuster aesthetics, featuring hyper-realistic armor and dynamic camera movements.
A prompt for creating hyper-realistic cinematic racing scenes focusing on a woman surrounded by a pit crew with high-contrast lighting.
A detailed prompt for a 15-second 3D animation of a curly-haired girl trying to fix her bun, only to fail hilariously.
A prompt for Seedance 2.0 to generate a 28-second cinematic vlog following a Korean girl through various scenes with character consistency.
A detailed prompt for a 30-second 2D hand-drawn Chinese fantasy animation of a monkey swordsman fighting a hydra with specific VFX and choreography.
A prompt for generating a cinematic, realistic lifestyle video documenting a peaceful day of a young Korean college student.
A detailed video prompt for Seedance 2.0 Mini depicting a surreal showdown between an ancient immortal cultivator and a modern job interview.
A prompt for Seedance that uses timecodes to control character blinking and hesitation movements in a video.
Last reviewed August 17, 2026 · Editorial synthesis of current Topview samples and public prompting guidance
Quick answer
A useful Seedance 2.0 prompt assigns a clear job to every input, describes one primary on-screen action, gives the camera one readable instruction, and protects the details that must not drift. For image-to-video, let the source frame define identity, composition, wardrobe, packaging, and color; use the text mainly to direct motion, timing, camera behavior, sound, and precise continuity constraints.
[Output format] + [source or reference role] + [subject] + [one primary action] + [camera] + [scene and light] + [optional sound cue] + [continuity locks] + [specific exclusions]
State the intended frame and use case, such as a 9:16 creator ad, 16:9 cinematic shot, product macro, or medium talking-head shot.
When an input is available, say exactly what it controls: first-frame composition, character identity, product design, motion reference, or audio timing.
Identify the person, product, prop, or environment with only the details needed to keep it recognizable. Do not overwrite a specific source image with a conflicting description.
Choose one dominant change that can be read from start to finish, such as opening a bottle, turning toward a window, or lifting a product to camera.
Use a shot size plus one compatible movement: static medium close-up, slow push-in, lateral track, gentle handheld follow, or controlled orbit.
Name the setting, time of day, key light, and one visual treatment that supports the intended story or brand.
Tie a concrete sound to an event, such as a cap click, fabric rustle, footsteps, room tone, or one spoken line. Omit audio direction when it is not needed.
List the few non-negotiable attributes: same face, same outfit, unchanged bottle silhouette, readable label, stable background, or consistent screen direction.
Prevent likely failure modes with concrete limits such as no new text, no extra people, no product redesign, no camera cut, or no background replacement.
Treat the source image as the visual contract. Prompt the change over time instead of redescribing everything already visible.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Animate a portrait without identity drift | Use the source image as the exact identity and composition. Add one small performance beat and one subtle camera or environmental movement. Lock face, hair, and wardrobe. | Use the uploaded portrait as the exact identity and first-frame composition. She turns her eyes toward the window and blinks once as a light breeze moves two loose strands of hair. Static medium close-up with a very slow push-in; preserve her face, hairstyle, cream jacket, and warm window light; no new accessories or background change. |
| Animate a product still | Assign the image to product identity, then direct one physical event, one clean camera move, and the product details that must remain unchanged. | Use the source image as the exact bottle design and opening composition. The cap turns once and lifts with a soft click while a narrow light sweep crosses the glass. Slow 15-degree lateral camera slide; keep the bottle silhouette, cap, label placement, colors, and marble surface unchanged; no extra text or props. |
| Bring an environment to life | Keep the scene layout fixed and animate two related ambient elements. Use restrained parallax or a locked camera. | Use the source frame as the exact cafe layout. Steam rises from the cup and rain trails move down the window while reflections shift softly on the table. Locked wide shot with subtle depth parallax; preserve furniture positions, signage, color grade, and empty seats; no people enter the frame. |
| Move from a defined start to a defined finish | State what the first frame controls, describe one continuous transition, and define the final composition without adding intermediate redesigns. | Begin from the supplied closed compact case and preserve its exact materials and logo. The lid opens in one smooth motion as the camera pushes from a medium product shot to a centered macro view of the interior. End with the case fully open and the logo facing camera; no cuts, no added objects, no shape change. |
The action tells the viewer what changes; the camera tells the viewer how to experience it. Keep both directions compatible and easy to stage.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Premium product reveal | One product action + slow, controlled camera move + lighting event. | The watch rotates a quarter turn on a black stone pedestal as a cool rim light reveals the brushed metal edge. Slow clockwise orbit at product height, shallow depth of field, stable horizon. |
| Natural creator performance | One conversational gesture + lightly imperfect handheld framing + no competing camera move. | She raises the serum beside her cheek, looks into the phone camera, and delivers one short recommendation. Natural arm-length selfie framing with gentle handheld sway; no zoom, no orbit, no cutaway. |
| Cinematic subject movement | Give the subject a clear route and let the camera follow from a named position. | A cyclist crosses the wet bridge from left to right while neon reflections trail beneath the wheels. Low side-tracking shot matching the cyclist speed, then a subtle deceleration at the far end; keep screen direction consistent. |
| Short multi-beat scene | Use two or three time boxes only when the clip genuinely needs multiple beats. Give each interval one action and a compatible shot. | 0-3s: macro view as the coffee bag tears open. 3-7s: top-down shot as grounds fall into the filter. 7-10s: slow push-in as the finished cup releases steam. Keep the same counter, bag design, cup, and morning light across all beats. |
Commercial prompts need recognizable product facts, while creator-style prompts also need believable framing, pacing, and interaction.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Ecommerce hero clip | [Format] + [exact product] + [single reveal] + [clean camera] + [surface and light] + [geometry and label locks]. | 9:16 ecommerce hero shot of the referenced silver earbuds case on clear acrylic. The case opens once as the camera makes a slow centered push-in. Bright high-key studio light with soft reflections; preserve the hinge, proportions, finish, and logo placement; no extra copy or floating parts. |
| UGC hook with product in hand | [Creator framing] + [fast but plausible hook] + [one product interaction] + [product visibility] + [natural setting] + [identity and object locks]. | Vertical phone-camera video in a bright bathroom. The creator leans into frame, says, "This is the step I stopped skipping," then lifts the referenced moisturizer beside her face and turns the label toward camera. Natural handheld movement and room tone; keep her face, robe, jar shape, lid color, and label layout consistent. |
| Product-in-use demonstration | [Starting position] + [one complete use action] + [detail shot] + [tactile sound] + [no redesign]. | Begin with the exact referenced lipstick closed in her hand. She removes the cap, twists the bullet up once, and applies one clean stroke. Tight over-shoulder mirror shot with a brief macro insert and a soft cap click; preserve packaging geometry, shade, hand count, and mirror background. |
| Creator testimonial without visual clutter |
Use constraints to protect specific assets and spatial logic. A short prioritized list is more actionable than a long generic negative prompt.
| Goal | Prompt pattern | Example wording |
|---|---|---|
| Lock character identity | Preserve the same [face markers], [hair], [wardrobe], and [body proportions] from first frame to last; exclude the most likely identity changes. | Keep the same face, short black bob, green earrings, navy blazer, and body proportions throughout. No face replacement, hairstyle change, wardrobe change, duplicate person, or age shift. |
| Lock product design | Preserve [silhouette], [materials], [color], [label/logo placement], and [part count]; prevent additions and deformations. | Keep the exact rectangular bottle, amber glass, black pump, cream label, and centered wordmark. No warped label, extra cap, duplicated bottle, color shift, invented text, or package redesign. |
| Lock scene geography | Define the subject route and fixed anchors; preserve screen direction and background placement. | The runner moves left to right along the same platform. Keep the red bench behind her, tracks on frame right, and station clock in the background. No axis flip, teleportation, new doorway, or background replacement. |
| Simplify a fragile shot | Protect the highest-value detail, reduce motion around it, and exclude only failure modes that would make the clip unusable. | Prioritize a readable front label. The bottle remains still while only the camera slides slowly right and condensation moves. No fast spin, liquid splash across the label, glare over the text, or camera cut. |
Make this image into an amazing cinematic luxury product ad. Add lots of movement, dramatic camera angles, a transformation, beautiful lighting, text, music, and viral UGC energy.
Use the uploaded image as the exact identity and first-frame composition of the amber skincare bottle. The bottle remains centered and still while two condensation drops move down the glass and a warm light sweep travels left to right. Slow 10-degree camera push-in at label height, soft bathroom ambience, 9:16 vertical framing. Preserve the bottle silhouette, black pump, cream label, wordmark placement, marble surface, and warm color palette. No product rotation, package redesign, extra text, added hands, background change, or camera cut.
The weak prompt stacks taste words and conflicting jobs without defining what the source image controls. The stronger version separates identity from motion, gives the product and camera one action each, connects sound and light to physical events, and protects the details that determine whether the clip remains usable.
Assign the source image to identity, design, or composition, then describe only the changes you want. If the person already wears a blue jacket, do not introduce a different wardrobe in text unless a wardrobe change is the actual task.
Choose the single event the viewer must notice. Move secondary ideas into a later clip or a short, clearly time-boxed sequence.
Use one dominant camera path per beat. A slow orbit, fast handheld chase, static macro, and aerial pullback should not all compete inside the same brief moment.
Translate words such as premium or cinematic into framing, light, material, motion, and pacing that can appear on screen.
Name what each referenced image, video, or audio input should control. Do not assume that a list of assets communicates which one supplies identity, motion, framing, or timing.
Prioritize exclusions tied to the shot: identity drift for portraits, geometry and label drift for products, axis changes for action, or unwanted text for ads.
Reduce action or camera complexity first. Then change one instruction at a time so the next result reveals which direction had an effect.
Start with the output format and input role, then specify the subject, one primary action, one camera instruction, the scene and light, optional sound, continuity locks, and a few concrete exclusions. The exact length can vary; clarity and compatible instructions matter more than filling every field.
A text-to-video prompt must establish the subject and visual world. An image-to-video prompt can let the source frame carry those details and spend more words on motion, timing, camera behavior, sound, and what must stay unchanged. Avoid making the text fight the image.
Use a clear source image, explicitly assign it to identity or product design, keep the movement manageable, and name the highest-value attributes to preserve. For a person, that may be face, hair, wardrobe, and proportions; for a product, silhouette, materials, colors, part count, and label placement.
Common building blocks include a locked shot, slow push-in, lateral track, restrained orbit, low follow, top-down view, or natural handheld framing. Pair the move with a shot size and subject action, and avoid stacking camera directions that imply incompatible paths.
Use them when a predictable failure would make the clip unusable. Product ads often need protection against shape, label, and text drift; character shots may need identity and wardrobe locks. Keep the list specific to the scene instead of adding every possible defect.
Use time boxes when the clip needs a short sequence, tutorial, or ad progression. Give each interval one readable action and preserve shared subjects, props, lighting, and screen direction across the cuts. For a single visual idea, one continuous shot is usually easier to direct.
Describe believable creator framing, a quick spoken or visual hook, one product interaction, and how the label or key feature faces the viewer. Preserve both creator identity and product design, and leave clean space for captions if they will be added later rather than asking the generation to invent marketing text.
Use short, physical cues linked to visible events: a cap click, shoe impact, fabric movement, room tone, or one line of dialogue. If sound is not important to the task or the selected workflow, it is reasonable to omit it and focus the prompt on the visuals.
How this guide was built
This guide is an editorial synthesis of recurring structures in Topview's public Seedance 2.0 prompt library, including reference-led edits, first-frame animation, product ads, creator-style videos, cinematic shots, camera scripts, continuity locks, and audio cues. Public Seedance 2.0 guides and model documentation were reviewed to identify the questions users commonly need answered; all formulas, examples, comparisons, and troubleshooting language here were written specifically for Topview. The patterns are starting points rather than guarantees, because results vary with the selected workflow, source assets, settings, and generation run.
| [Medium close-up] + [one spoken benefit] + [one supporting gesture] + [safe caption area] + [no invented claims or interface elements]. |
| 9:16 medium close-up at a tidy desk. The creator points once to the referenced desk lamp and says, "I can change the mood without rearranging the room." Gentle handheld phone framing, practical evening light, clear space above the desk for later captions; no on-screen text, new controls, or background people. |