Seedance 2.0 Prompt Guide
Master the art of prompt writing for image-to-video and text-to-video generation. Learn how to create compelling videos with precise motion, camera, and narrative control.
Table of Contents
Text to Video
Image to Video Tips
Basic Prompts
Prompt = [Subject] + [Motion], [Background] + [Motion], [Camera] + [Motion]
Keep it simple and direct
Use simple words and sentence structures. The model will expand your prompt based on its understanding of the image to generate the expected video.
Negative prompts don't work
The model does not respond to negative prompts. Focus on describing what you want, not what you don't want.
Focus on motion, not static elements
Since the image already provides the scene, minimize descriptions of static parts. Focus on what moves: subject motion, background changes, and camera movement.
Highlight distinctive features
When the subject has prominent features, add them to better identify the subject — e.g., "the old man", "the woman wearing sunglasses". Use clear degree adverbs for motion: "fast", "dramatically".
Follow the input image
Write prompts based on the actual image content. Clearly specify the subject and desired action or camera movement. Don't contradict the image content or basic parameters.
Multi-Action Prompts
The model responds well to multi-shot actions, supporting sequential actions in a timeline and different actions for multiple subjects.
Prompt = Subject 1 + Movement 1 + Movement 2
Prompt = Subject 1 + Movement 1, Subject 2 + Movement 2 ...
Simply list the actions one by one. The model will expand the prompt and generate the expected video based on its understanding.
Camera Movement
Describe camera changes in natural language. Supported movements include: orbit, aerial, zoom, pan, follow, handheld, and shot transitions.
For consistent multi-shot prompts, describe the connection between shots.
Use "shot transition" to connect changes between shots.
If the scene changes after a transition, describe the new scene.
When using camera movement prompts, select "non-fixed camera" in basic parameters.
Camera movement prompts also work for text-to-video scenarios.
Degree Adverbs
Use degree adverbs to emphasize the frequency, intensity of actions, or subject features in the video.
Key Principles
Clarity: The model cannot infer motion degree from the reference image alone, so you must specify it explicitly in the prompt.
Exaggerate appropriately: Slightly exaggerating the degree enhances expressiveness. E.g., change "man roaring" to "man roaring wildly", or "wings flapping" to "wings flapping powerfully".
Text to Video Tips
Basic Prompts
Prompt = [Subject] + [Movement] + [Scene] + [Lens], [Style]
Subject + Movement + Scene are the core elements. The model will expand the prompt and generate the expected video.
Tips for sequential actions, camera movement, and degree adverbs from the image-to-video section also apply here. Negative prompts are not supported.
How to Write Better Descriptions
Detailed Character Description
Pay attention to the character's appearance, clothing, and posture. The more specific, the better the result.
Environment Details
Describe natural environments (mountains, deserts, waterfalls) or built environments (studios, bathrooms) to emphasize visual and sensory experience.
Emotion & Dynamics
Depict the character's emotional state combined with environmental dynamics to create rich narratives.
Atmosphere & Lighting
Use lighting descriptions to set the mood: dusk, dawn, dim light, warm glow, dramatic shadows, golden hour, etc.