Seedance 2.0 Showcase: Real Outputs, Ready Prompts
Browse curated Seedance 2.0 video examples with the exact prompts and reference inputs behind each result. Copy, remix, and build on what already works — new entries added every week.
Core Capabilities
Four inputs. One directed video.
Seedance 2.0 accepts images, reference video clips, audio, and text — combine any of them to generate a controllable, cinematic video up to 15 seconds.
Reference images — pixel-level fidelity
Upload up to 9 images to lock faces, outfits, and environments. Seedance 2.0 reproduces every detail precisely, so every generated video looks unmistakably like your reference.
- Pixel-accurate character and scene reproduction
- Style transfer from any reference photo
- Up to 9 images combined in a single generation
Reference videos — copy any camera movement
Feed a clip and Seedance 2.0 replicates the dolly, tracking shot, or whip-pan exactly. Show it a movement once and it returns it faithfully every time.
- Clone camera motion from any reference clip
- Match movement rhythm and pacing
- Transfer transitions and creative effects
Extend & continue — full scenes, not fragments
Chain clips together with seamless transitions and consistent character, lighting, and scene continuity. Turn a single generation into a complete multi-shot sequence.
- Extend any clip up to 15 seconds total
- Smooth shot-to-shot continuity
- Prompt-guided multi-shot sequences
AI video editing — modify existing footage
Swap characters, remove objects, or add new scene elements to footage you already have. Seedance 2.0 functions as an editor, not just a generator — with built-in sound and music generation.
- Replace characters in existing video
- Add or remove objects from any scene
- Built-in audio: sound effects and music
Real Outputs with Copyable Prompts
Every example shows the exact prompt and reference materials used. Copy a prompt, change one variable, and iterate from there. New examples added every week.
A detailed, cinematic prompt designed to generate a high-action fight scene between two men, featuring ultra-high-speed punches with afterimages, dramatic camera movements, slow-motion effects, and a powerful finishing blow, originally written by an LLM (Chappy).
An extremely detailed, multi-segment prompt for generating a 15-second cinematic, realistic video of a street basketball scene. It describes the setting (LA outdoor court, golden hour), the characters (a tall, muscular Black man and a petite, athletic blonde woman), the action (the man's confident warm-up, his aggressive drive, and the woman's precise, slow-motion block), and specific camera work, sound effects, and post-production requirements.
A creative prompt for generating a video featuring a steampunk robot working as a Michelin three-star chef in a futuristic kitchen. The robot uses mechanical arms to precisely stir-fry Kung Pao Chicken, with ingredients automatically flying into the wok, while alien nobles applaud the meal.
A detailed, multi-scene Seedance 2.0 video generation prompt designed to create a humorous, high-energy live stream video in the 'Northeast MC Style' (Kuaishou/Douyin style), featuring a male host selling products with intense DJ music and visual effects.
Three highly detailed, multi-scene prompts for Seedance 2.0, each generating a 15-second cinematic sequence of the Chinese goddess Chang'e flying to the moon, utilizing different traditional Chinese art styles: Version 1 (Ink Wash Scroll), Version 2 (Fine-line Heavy Color), and Version 3 (Freehand Ink Splash). Each prompt specifies camera movements, scene progression, sound design, and style constraints.
A user shares a sophisticated 9-panel grid prompt template for Seedance 2.0, designed to generate a cohesive 10-20 second cinematic sequence with a clear emotional arc (setup → build → turn → payoff) using a reference image for the actor. This template is designed for advanced storyboarding and character consistency.
A detailed, multi-scene prompt for Seedance 2.0 to generate a 15-second cinematic trailer in the style of Zelda: Breath of the Wild, focusing on an ancient ruin exploration. The prompt specifies a blend of cel-shading and watercolor, cold blue-green tones, camera movements, and sound design for the awakening of a hero in a temple.
A detailed, multi-scene prompt for Seedance 2.0 designed to generate a 15-second cinematic trailer in the style of Zelda: Breath of the Wild, using a blend of cel-shading and watercolor aesthetics. The prompt specifies camera movements, scene progression (dawn, gliding, combat), and sound design.
A highly detailed, multi-shot prompt for Seedance 2.0 focusing on complex camera work and scene transitions involving a high-speed train (Shinkansen) passing through a rural landscape. The prompt uses advanced cinematic techniques like 'Ascending Reveal,' 'Vertigo Chase,' 'Spatial Transition,' and 'Elegant Orbit' to create a six-part sequence.
A detailed prompt for the first shot (0-3s) of an action sequence, specifying a low-angle wide shot of a matte black Charger, emphasizing kinetic energy, golden hour lighting, high contrast, and handheld shake.
A prompt for generating an extreme close-up video of Anya Forger's mouth, focusing on dramatic lighting, steady breathing, and subtle micro-expressions as she delivers a line with absolute conviction.
A prompt used to generate a fight scene video between Brad Pitt and Tom Cruise, with their dialogue revolving around the Epstein files.
A complex reasoning prompt for Seedance 2.0 and Grok Imagine, asking for a man deciding whether to walk or drive to a car wash 50 meters away, followed by him executing the decision.
A detailed system prompt designed for the Seedance 2.0 model to convert simple user ideas into highly structured, cinematic JSON prompts, including parameters for shot composition, lighting, character details, scene atmosphere, visual beats, physics rules, and audio elements.
A prompt describing a showdown where a scarred woman defeats swordsmen on a glass floor above a zen garden, focusing on her nimble movements and slicing action.
A detailed prompt for a top-down shot of a woman (Maki Zenin) defeating swordsmen with powerful attacks, focusing on camera movement and a jumpcut sequence with blood.
A detailed cinematic prompt describing a scene in the Japanese countryside where the camera races alongside a bullet train at impossible speed.
The first part of a detailed, multi-shot prompt for a Hollywood professional racing movie scene, focusing on the veteran driver's close-up shot during a rainy night race.
A detailed cinematic prompt for generating a scene from Stranger Things where the main party members simultaneously unleash unique superpowers in a ruined lab, utilizing specific filmic elements like grain, flares, and volumetric mist.
A dialogue-based prompt for Seedance 2.0 creating an alternate, humorous ending to Lord of the Rings where Sam suggests taking the Eagles to Mount Doom, followed by Frodo agreeing and the subsequent destruction of Sauron's tower.
A detailed, cinematic prompt designed to generate a high-action fight scene between two men, featuring ultra-high-speed punches with afterimages, dramatic camera movements, slow-motion effects, and a powerful finishing blow, originally written by an LLM (Chappy).
An extremely detailed, multi-segment prompt for generating a 15-second cinematic, realistic video of a street basketball scene. It describes the setting (LA outdoor court, golden hour), the characters (a tall, muscular Black man and a petite, athletic blonde woman), the action (the man's confident warm-up, his aggressive drive, and the woman's precise, slow-motion block), and specific camera work, sound effects, and post-production requirements.
A creative prompt for generating a video featuring a steampunk robot working as a Michelin three-star chef in a futuristic kitchen. The robot uses mechanical arms to precisely stir-fry Kung Pao Chicken, with ingredients automatically flying into the wok, while alien nobles applaud the meal.
A detailed, multi-scene Seedance 2.0 video generation prompt designed to create a humorous, high-energy live stream video in the 'Northeast MC Style' (Kuaishou/Douyin style), featuring a male host selling products with intense DJ music and visual effects.
Three highly detailed, multi-scene prompts for Seedance 2.0, each generating a 15-second cinematic sequence of the Chinese goddess Chang'e flying to the moon, utilizing different traditional Chinese art styles: Version 1 (Ink Wash Scroll), Version 2 (Fine-line Heavy Color), and Version 3 (Freehand Ink Splash). Each prompt specifies camera movements, scene progression, sound design, and style constraints.
A user shares a sophisticated 9-panel grid prompt template for Seedance 2.0, designed to generate a cohesive 10-20 second cinematic sequence with a clear emotional arc (setup → build → turn → payoff) using a reference image for the actor. This template is designed for advanced storyboarding and character consistency.
A detailed, multi-scene prompt for Seedance 2.0 to generate a 15-second cinematic trailer in the style of Zelda: Breath of the Wild, focusing on an ancient ruin exploration. The prompt specifies a blend of cel-shading and watercolor, cold blue-green tones, camera movements, and sound design for the awakening of a hero in a temple.
A detailed, multi-scene prompt for Seedance 2.0 designed to generate a 15-second cinematic trailer in the style of Zelda: Breath of the Wild, using a blend of cel-shading and watercolor aesthetics. The prompt specifies camera movements, scene progression (dawn, gliding, combat), and sound design.
A highly detailed, multi-shot prompt for Seedance 2.0 focusing on complex camera work and scene transitions involving a high-speed train (Shinkansen) passing through a rural landscape. The prompt uses advanced cinematic techniques like 'Ascending Reveal,' 'Vertigo Chase,' 'Spatial Transition,' and 'Elegant Orbit' to create a six-part sequence.
A detailed prompt for the first shot (0-3s) of an action sequence, specifying a low-angle wide shot of a matte black Charger, emphasizing kinetic energy, golden hour lighting, high contrast, and handheld shake.
A prompt for generating an extreme close-up video of Anya Forger's mouth, focusing on dramatic lighting, steady breathing, and subtle micro-expressions as she delivers a line with absolute conviction.
A prompt used to generate a fight scene video between Brad Pitt and Tom Cruise, with their dialogue revolving around the Epstein files.
A complex reasoning prompt for Seedance 2.0 and Grok Imagine, asking for a man deciding whether to walk or drive to a car wash 50 meters away, followed by him executing the decision.
A detailed system prompt designed for the Seedance 2.0 model to convert simple user ideas into highly structured, cinematic JSON prompts, including parameters for shot composition, lighting, character details, scene atmosphere, visual beats, physics rules, and audio elements.
A prompt describing a showdown where a scarred woman defeats swordsmen on a glass floor above a zen garden, focusing on her nimble movements and slicing action.
A detailed prompt for a top-down shot of a woman (Maki Zenin) defeating swordsmen with powerful attacks, focusing on camera movement and a jumpcut sequence with blood.
A detailed cinematic prompt describing a scene in the Japanese countryside where the camera races alongside a bullet train at impossible speed.
The first part of a detailed, multi-shot prompt for a Hollywood professional racing movie scene, focusing on the veteran driver's close-up shot during a rainy night race.
A detailed cinematic prompt for generating a scene from Stranger Things where the main party members simultaneously unleash unique superpowers in a ruined lab, utilizing specific filmic elements like grain, flares, and volumetric mist.
A dialogue-based prompt for Seedance 2.0 creating an alternate, humorous ending to Lord of the Rings where Sam suggests taking the Eagles to Mount Doom, followed by Frodo agreeing and the subsequent destruction of Sauron's tower.
How to write effective Seedance 2.0 prompts
Three components. Consistent output every time.
Define the subject and action
One clear sentence: who or what is in frame and what they are doing. Use motion verbs — slow, sharp, float, pivot. Keep it under 20 words.
Specify the camera
Name the shot type and movement: close-up, wide, tracking, dolly-in, handheld sway. The more specific, the more consistent the output.
Lock the visual style
One style phrase is enough — cinematic grain, neon haze, soft rim light. Change only one variable per iteration so you can isolate what drives each result.
Seedance 2.0 FAQ
Stay updated as Seedance 2.0 launches
New showcase examples are published every week. API generation access opens when Seedance 2.0 goes live — follow the showcase to be ready on day one.
Browse the full showcase now. Direct generation access follows when the API opens.