Seedance 2.0 Cinematic AI Video Prompt Professional Edition

Problem

Seedance 2.0 prompts often fail because of unstable prompt structure: too many abstract terms, conflicting instructions, overloaded motion, unclear foreground/midground/background, and physically implausible camera or lighting cues. This leads to face drift, messy camerawork, audio-video mismatch, garbled text, or lip-sync failure.

How It Works

The skill turns prompt engineering into executable checks:
- Four-layer assembly: builds prompts through L1 technical layer, L2 spatial layer, L3 lighting layer, and L4 motion layer, covering aspect ratio, frame rate, duration, depth, light sources, and a single motion strategy.
- Templates and diagnostics: includes T1-T8 scenario templates for e-commerce, sci-fi, food, sports, Chinese classical aesthetics, and more; the diagnostic table maps symptoms such as face drift, flickering light, or blur to fixes.
- Anti-pattern filtering: detects abstract wording, conflicting instructions, three or more motion terms, and physics violations, while reducing false positives from compound terms.
- Boundary warnings: flags high-risk cases such as duration, lip sync, complex text, fluids, and high-speed motion, leaving room for post-production or voiceover fallback.

Scope

It is best used as a Seedance 2.0 prompt generation and review tool, converting scene descriptions into structured, reviewable prompts. Complex text, lip sync, fluid physics, and fast motion remain model-limited, so use reference assets, simplify motion, and reserve retries or post-production.