714+ curated video prompts from 400+ creators worldwide. Browse, search, and copy the best Seedance 2.5 prompts to spark your next creation.
Enter a prompt and generate your video with Vibbit AI
Community examples span multiple Seedance versions. Adapt the original prompts for Seedance 2.5.
A poetic prompt for Seedance 2.0 to generate a Chinese-style dreamscape video, featuring blue and white porcelain transforming into a Pegasus and a paper crane flying through a window, set against a snowy, oriental aesthetic, offering an escape from a cyberpunk city.
A detailed video generation prompt for Seedance/Sora, describing a man working in a lab with multiple camera shots, specifically excluding music and talking, and requesting only ambient sound.
A detailed description of a scene and cinematography for a modern dance performance. The setting is a dimly lit indoor home environment (modern kitchen/living room). The performer executes a highly rhythmic modern dance, conveying confidence and allure. The prompt specifies a fixed camera position, moderate depth of field, and single-sided lighting.
A prompt for Seedance 2.0 generating a fantasy video of a fox spirit transforming into a beautiful woman in a moonlit bamboo forest, emphasizing the visual style of ink painting and CG illustration, and focusing on the transformation process and the character's subsequent movements.
A highly detailed, two-shot cinematic prompt for Seedance 2.0, set in a Hollywood Sci-fi style. It describes a violent corkscrew maneuver inside a spaceship, focusing on a couple (Darrow and Victra) bracing themselves, sharing a desperate kiss mid-rotation, and culminating in Victra slamming Darrow's visor down as the ship takes impact and the camera completes a 360-degree rotation.
A highly detailed 15-second vertical video prompt for Seedance 2.0, simulating a first-person perspective (POV) selfie video of a Taiwanese girl sneaking out late at night to buy snacks at a 7-Eleven in Taipei, including specific lighting, camera shake, dialogue, and scene transitions.
A detailed prompt for the Seedance 2.0 I2V (Image-to-Video) model, instructing the generation of a cinematic, full-color Japanese action anime scene. The scene features a character (referenced by an image) engaged in dynamic, fast-paced, close-quarters combat against an army of humanoid robots using a samurai sword. The prompt emphasizes dynamic cuts, superhuman abilities, ignoring physics, and a dark, cataclysmic robot uprising aesthetic.
A highly detailed, multi-shot prompt for Seedance 2.0 to generate a 15-second video of a young girl's morning routine, styled as a casual, low-quality iPhone vlog, emphasizing her soft, sweet demeanor and specific actions like stretching and saying 'Good morning'.
An extremely detailed, multi-shot prompt for Seedance 2.0 to generate a 15-second cinematic food documentary short film, mimicking the style of 'A Bite of China' (舌尖上的中国), focusing on a Northern Chinese New Year's Eve dinner preparation, including specific camera movements, sound design, and narrative voiceover.
A detailed, five-shot prompt for Seedance 2.0 to generate a high-energy, stylized fight scene between two martial artists in the distinct national comic style of 'Wushan Wuxing' (Fog Hill of Five Elements), emphasizing intense action, dynamic camera work, and powerful visual effects like ink flow and speed lines.
This is a detailed prompt designed for generating a continuous shot video in an ultra-cinematic, photorealistic sci-fi style. It specifies a dense, futuristic megacity setting with noir cyberpunk elements, emphasizing high contrast, metallic reflections, neon accents, and volumetric rain haze. The core instruction is for ONE continuous shot with aggressive camera movement.
A simple prompt for Seedance 2.0 to generate a rhythmic video featuring the transformation of the Twelve Flower Goddesses, utilizing quick turns, costume changes, special effects, and push/pull camera movements synchronized to a beat.
A highly detailed, multi-shot cinematic prompt for Seedance 2.0 describing an explorer's journey through a vast, silent snowscape, culminating in the discovery of a hidden, warm valley, utilizing specific camera cuts and emotional cues.
A humorous video generation prompt for Seedance 2.0 describing a scene at a red light where a ragdoll cat on a bicycle attempts to snatch a fish from a nearby basket, is startled by a car horn, and then glares and swipes angrily at the camera (POV shot).
A long, detailed, two-part stop-motion narrative prompt for Seedance 2.0 featuring an Uzbek woman puppet. The first part describes a quiet tea-drinking scene, and the second describes a vigorous cooking scene (kneading dough and baking samsas in a tandoor), using [cut] markers for scene changes and specifying materials and camera angles.
A detailed, multi-step video prompt for Seedance 2.0, generating a hyper-realistic, pure and innocent style video of a girl in a cozy living room setting from the first-person perspective of her boyfriend, including specific dialogue and actions for a playful interaction.
A detailed, continuous cinematic shot prompt for Seedance 2.0, describing a high-energy action sequence featuring a Japanese schoolgirl sprinting through the Shibuya crosswalk, incorporating specific audio cues, camera movements (low push, side-roll, explosive jump, rotation), and a dramatic mid-air music cut.
A detailed 15-second narrative prompt for Seedance 2.0, focusing on a pure love story between a girl and a boy in a cinematic, realistic style, using specific camera movements, lighting, and sound effects across three distinct scenes (0-4s, 5-10s, 11-15s).
A highly detailed, multi-shot cinematic prompt for Seedance 2.0 designed to emulate the style of 90s Hong Kong Art Cinema, featuring a melancholic scene in a rainy phone booth with specific instructions for visual effects, lighting, and emotional performance.
A multi-shot video generation prompt for Sora 2 or Seedance 2.0, describing Harry S. Truman getting out of a car and walking through a crowd of photographers, focusing on camera cuts and ambient sound.
A highly detailed, cinematic video generation prompt designed to showcase the effortless, silent birth of a complete Tesla ecosystem, blending magical engineering with minimalist design. It features a mysterious, glowing crate unfolding to reveal a Cybertruck and assemble a sleek showroom around it, complete with precise technical specifications, camera movements, audio mixing instructions, and quality control tags.
A detailed prompt for generating an epic fantasy aerial video sequence, focusing on the visual awakening of a desolate land, using specific color palettes, extreme contrast lighting, and dynamic camera movements to convey a sense of grandeur and power.
An extremely detailed, cinematic prompt for Seedance 2.0 describing the aggressive transformation of an old American-style bus into a grumpy, evil Cybertronian mechanical giant caterpillar, including specific details about the mechanical components, the driver's reaction, camera movement, and desired visual style.
A highly detailed, multi-stage prompt for Seedance 2.0 to generate a continuous, first-person POV video of a character skateboarding and then snowboarding, transitioning seamlessly through the four seasons (Spring, Summer, Autumn, Winter) with corresponding changes in clothing, scenery, and action, maintaining a 2D anime style and cinematic quality.
Use the search bar to find prompts by keyword, or filter by category (Cinematic, Dance, Fashion, Narrative, etc.). Click any prompt card to preview the generated video, view the full prompt text, and copy it for your own use.
Yes! Every prompt in the gallery is fully editable. Copy any prompt and modify details like camera movements, scene descriptions, lighting, or subject actions. The Seedance 2.5 prompt formula (Subject → Action → Camera → Scene → Style) helps you structure effective customizations.
Click 'Copy Prompt' on any prompt card, then paste it into Vibbit's Seedance 2.5 workspace. You can also click 'Try in Vibbit' to launch directly with the prompt pre-filled. Generation typically takes 30-60 seconds for a 5-second clip.