Model guide
Vidu AI Prompt Guide
Learn practical Vidu AI prompt patterns for AI video recipes, camera motion, style, lighting, and constraints.
Practical starting point
Vidu recipes should separate the subject, action, scene, and visual style so each part stays readable.
Advertisement
Model prompt notes
- Concise subject-action-scene format with style and motion separated.
- Vidu output depends on the current model version and input mode.
- Best for: stylized clips
- Best for: anime-inspired shots
- Best for: simple motion scenes
Prompt structure
- Use labeled subject, action, scene, motion, look, lighting, and negative lines.
- Keep style words grouped together instead of mixing them into the action sentence.
- Use a clear motion line for camera or subject movement.
- End with format and negative constraints.
Workflow tips
- Use Vidu drafts when a stylized scene needs clean separation between look and movement.
- For anime or character prompts, test face and hand stability early.
- Keep background text and sign detail controlled.
- If the style drifts, remove competing style words before adding new ones.
What to avoid
- Combining too many art styles in one prompt.
- Long camera descriptions that hide the main character action.
- Detailed background signs or symbols that the model may not preserve cleanly.
Vidu prompt FAQ
Why separate style and motion for Vidu prompts?
Separating style and motion makes it easier to adjust the look without accidentally changing the action or camera behavior.
What recipes are good starting points for Vidu?
Anime scenes, stylized clips, and simple motion scenes are good starting points because the prompt can keep look and action compact.
How do I reduce character artifacts?
Keep the character description concise and add negative constraints for hands, eyes, face shape, and extra limbs.
Related prompt pages
Independent prompt drafting aid. Verify final prompts inside the current model interface.