# User Prompt
I. The Most Core Structure (Universal Formula)
Subject + Action + Scene + Lighting + Camera Language + Style + Image Quality + Constraints
Example Template:
A young girl walks slowly by the sea, the breeze ruffles her hair, she smiles at the camera, warm sunset light, 4K HD, cinematic, stable camera movement, smooth and jitter-free footage, clear details.
II. Action Description Techniques (Seedance excels at this)
• Don't just write "dancing" or "walking"; be continuous, natural, and slow-motion.
• Prioritize using: slow, gentle, continuous, natural, smooth, not stiff.
• Avoid: exaggerated, high-speed, complex multi-person interactions, extreme distortions.
Useful Verb Combinations:
• Slowly turn, gently raise hand, step lightly, slightly lower head, sway with the wind.
• Camera slowly pushes in, slightly circles, smoothly tracks.
III. Camera & Movement (Directly enhances cinematic feel)
Seedance 2.0 is very good at recognizing camera movement; directly write:
• Close-up / Medium shot / Wide shot / Extreme close-up
• Slow push-in / Slight pull-out / Smooth dolly / Half-circle orbit
• Static shot, handheld stabilization, no jitter, silky smooth.
Example:
Medium shot, slow push-in, stable tracking, silky smooth footage without stuttering.
IV. Character/Person Stability (Prevents face changes, deformations)
Must add constraint words. Very crucial:
• Clear facial features, stable face, no distortion, no shape-shifting.
• Normal human body structure, natural proportions, actions not stiff.
• Same character, consistent clothing, unchanged hairstyle.
Example:
Character's face is stable and not deformed, body structure is normal, actions are natural and smooth.
V. Image Quality (Must add, improves clarity)
General Image Quality Words (Copy and paste):
• 4K, Ultra HD, rich details, sharp clarity.
• Cinematic texture, natural colors, soft lighting.
• No blur, no ghosting, no flickering, stable image.
VI. Style & Atmosphere (Quickly generates images)
• Healing and fresh, aesthetic and artistic, Japanese fresh, Korean atmospheric.
• Cyberpunk, dark and 고급, retro film, dreamy soft light.
• Minimalist and clean, high-quality texture, realistic scene.
VII. Multi-camera/Storyboard Writing (Seedance 2.0's strength)
Describe using time/shot order, don't be too fragmented:
Opening with a close-up of the face, slowly pulling out to a wide shot, the character walks slowly, the camera follows smoothly, ending with a freeze-frame smile.
VIII. Pitfalls (Absolutely do not write)
• Complex multi-person interactions, fighting, intense running and jumping, exaggerated dancing.
• Vague words: good-looking, beautiful, cool (equivalent to no description).
• Contradictory requirements: ultra-high speed + extreme stability.
• Pornographic, violent, dangerous actions, copyrighted characters.
IX. Universal Templates You Can Directly Copy
Template 1 (Aesthetic Portrait Short Video)
A young girl walks slowly in the woods, the breeze gently blows her hair, a natural smile, warm light and shadow, medium shot, slow push-in, smooth and stable footage, 4K HD, cinematic, clear and undeformed face, normal body structure, rich details.
Template 2 (Atmospheric Landscape)
Sunset by the sea, waves gently lap the shore, camera slowly pans horizontally, warm orange tones, healing and fresh, silky smooth footage, 4K Ultra HD, no flickering, no ghosting.
Template 3 (Image to Video, First Frame Fixed)
Based on the reference image, maintain consistent character appearance and clothing, slow action of raising hand and turning, natural and smooth, not stiff or deformed, stable camera movement, HD details, cinematic texture.
X. Technique Summary
1. Describe actions as slow and continuous.
2. Describe camera movement as stable and simple.
3. Must add stability/no deformation/not stiff.
4. Add image quality + style at the end.
5. Do not write intense, complex, or multi-person actions.seedance2.0 prompts
This is a set of Seedance 2.0 specific, ready-to-use prompt techniques, from basic to advanced, usable by copying to your phone, adapted for text-to-video, image-to-video, multi-shot, camera movement, and character consistency.
Services with a clipboard icon will copy the prompt to your clipboard first.
Version History
Prompt documentation
Related Prompts
A text prompt designed for an LLM or video generation model (like Seedance 2.0/Sora 2) to create a highly dramatic historical scene featuring intense dialogue, emotional conflict, and character revelation. The scene focuses on two characters, Xiao Jingheng and Shen Qingtang, discussing their shared history of reincarnation and failure.
This is a two-step process prompt. First, an initial fantasy image is generated (presumably via Midjourney), and then Nano Banana Pro is instructed to generate a realistic human character based on that fantasy image as a reference. The resulting realistic image is then used as a reference for a video generation model (Seedance 2.0).
Comments (0)
Be the first to comment
to start the conversation.