Seedance 2.0 Prompt Writing Guide

Seedance 2.0 gets better when your prompt behaves like a shot brief, not a bag of adjectives. The most useful mental model is:

1. subject
2. action
3. camera
4. style
5. constraints

That order matters because the model needs to know what the shot is about before it can decide how the shot should move or look.

The working framework

A strong Seedance prompt pattern

Use this as your default pattern:

[subject], [single action], [camera move], [style and lighting], [constraints]

Example:

@Image1 skincare jar, creamy texture smear beside the jar, soft push-in, clean daylight beauty set, premium minimal look, no dirty background no label blur no packaging warp

What good prompts do differently

They describe one shot

Weak prompt:

A creator unboxes a product, talks about it, shows the details, then cuts to lifestyle footage and a final hero shot

Better prompt:

@Image1 creator opens the box on a desk, subtle overhead-to-front angle transition, natural creator lighting, social-first pacing, no hand deformation no box warp

The second prompt is easier to render because it asks for one coherent visual event.

They choose one camera instruction

The quickest way to make Seedance unstable is to ask for:

• orbit
• zoom
• whip pan
• rack focus
• handheld

all in the same shot.

Pick the dominant move first. Add more complexity only after the simple version works.

They protect the fragile parts

If the shot depends on any of these, name them explicitly:

• label readability
• hand anatomy
• face identity
• wheel, watch, or bottle geometry
• reflections and transparent surfaces

The model cannot protect what you never tell it to protect.

How to write by mode

Text to Video

Use text-to-video when you are exploring an idea rather than preserving a specific asset.

Best pattern:

• lead with subject
• keep the action physical
• keep the camera language simple
• define lighting and tone clearly

Example:

Lone runner in a rain-soaked neon alley, side tracking shot, fast controlled pace, reflective puddles, cinematic red and teal contrast, no shaky camera no extra limbs no text watermark

Image to Video

Use image-to-video when the first frame, product silhouette, or composition already exists and you want motion to grow out of it.

Best pattern:

• reference the uploaded image early
• describe one motion layer
• avoid changing the object identity mid-shot

Example:

@Image1 sneaker center frame, slow 180 rotation, soft side rim light, clean commercial shadow floor, no sole deformation no lace chaos no duplicate shoe

Reference to Video

Use reference-to-video when consistency matters more than creative range.

Best pattern:

• define what must stay fixed
• add only one motion instruction after the stability rule
• keep pose and scene change limited

Example:

@Image1 creator identity remains consistent, holds the product at chest level, subtle push-in, natural indoor daylight, social review style, no face drift no finger artifacts no product disappearance

Negative prompts are not optional

Negative prompts are where you stop the most common failure modes before they start.

Use them to block:

• anatomy errors
• text artifacts
• geometry drift
• jitter
• unwanted duplicate objects
• messy transitions

Negative prompt examples by use case

Ecommerce

no logo distortion, no text artifacts, no packaging collapse, no duplicate product

UGC

no extra fingers, no face drift, no lip mismatch, no background warping

Cinematic

no shaky camera, no object melting, no random text, no muddy lighting

Prompt iteration rules that save time

Rule 1: change one major variable per pass

If a result fails, choose one of these to change:

• prompt wording
• reference choice
• camera instruction
• duration
• aspect ratio

Do not change all of them together.

Rule 2: shorten before you complicate

If the shot is unstable, shorten the scope:

• fewer actions
• fewer props
• simpler move
• tighter frame

Shorter prompts are not always better, but narrower shots almost always are.

Rule 3: use the prompt to decide the mode

If your prompt keeps mentioning “same face”, “same bottle”, “same outfit”, or “same shape”, that is usually a signal to switch out of text-only generation.

Prompt examples by goal

Product hero shot

@Image1 beverage bottle, cold condensation across the glass, slow heroic push-in, sharp cool highlights, premium beverage ad tone, no label blur no cap drift no warped glass

Creator review

@Image1 creator speaking directly to camera, subtle hand gesture with product in frame, soft ring-light catchlight, authentic UGC pacing, no finger artifacts no eye drift no shaky framing

Cinematic teaser

Dark warehouse interior, flashlight beam motivates a slow pan reveal, suspended dust, suspense tone, high contrast cinematic lighting, no random text no object melting no flat blacks

The standard to keep

A strong Seedance prompt should make it obvious:

• what the viewer should notice first
• what the camera is doing
• what visual mood the shot lives in
• what failure modes are unacceptable

If the prompt does not answer those four questions, it still needs work.