Wan 2.1 for portrait diffuser work
Wan 2.1 is the first generation of the family, best known for motion consistency. This page is the 30-second text to video route for a diffuser, in a portrait treatment.
What to set
- Duration: 30 seconds. Longer runs cost more credits, and the credit cost is shown next to the form before you submit, so the price is known before anything runs.
- Task: text to video. Describe the shot in words; the model builds motion from the prompt.
- Aspect: 16:9 - landscape, the default for anything that plays on a screen.
- Style: portrait. Put the style words first in the prompt - they weigh more than modifiers placed at the end.
A prompt that works for a diffuser
portrait, a diffuser moving slowly, natural light, shallow depth of field, 30-second shot, no text
Treat the first render as a test rather than a final asset: watch the motion, then change one thing at a time. The live controls on the page are the source of truth for what the model accepts, not this text.