Wan 2.1 Video Generation

Wan 2.1 Prompting Guide with Video Effects

When using Wan 2.1 with LoRAs, or what you call video effects, prompt quality comes from two layers working together: the base prompt defines the video, and the effect nudges the output toward a specific learned look, motion pattern, or transformation.

Best Overall Formula

Effect trigger + subject + environment + action + camera + lighting + style + constraints

Why This Mode Feels Different

With video effects, the prompt describes the shot while the effect pushes it toward a learned behavior, distortion, transformation, or style lane.

Single Biggest Rule

Prompt the video first, then let the effect bend it.

Best Order to Write

Trigger + who + where + what happens + camera + look + stability priorities

Prompt Anatomy

1) Trigger

Use the required effect trigger exactly if the effect depends on exact wording.

2) Subject

Define the character, object, outfit, or visual anchor that the effect will act on.

3) Action

Describe the main motion clearly so the effect has a readable base shot to work with.

4) Control

Use camera, lighting, mood, and stability cues to keep the effect from overwhelming the scene.

Text-to-Video

For text-to-video, define the whole shot clearly so the effect has a strong structure to work on.

sq41sh squish effect, a young woman in a yellow raincoat standing on a city sidewalk at night, her body and umbrella comically compressing and stretching as cars pass by, medium shot, slight handheld camera, wet pavement reflections, neon storefronts, cinematic lighting, smooth motion, detailed textures, consistent anatomy, coherent background

Image-to-Video

For image-to-video, let the source image anchor the look and use the prompt mainly for motion, camera, atmosphere, and effect behavior.

[effect trigger], preserve the subject’s face, outfit, and pose from the reference image while applying the learned transformation effect during the movement. The subject slowly turns toward camera as the effect appears naturally through the motion. Gentle push-in, stable background, clean anatomy, consistent identity, no text or distortion.

Motion Guidance

Style Effects

  • alter look and finish
  • change texture or rendering
  • push toward a stylized lane
  • benefit from stable motion
  • need clear subject identity

Motion Effects

  • stretching
  • bouncing
  • squashing
  • warping
  • signature learned movement

Transformation Effects

  • require start-state clarity
  • need end behavior described
  • work best with one main action
  • benefit from stable framing
  • need strong preservation cues

Camera Movement

Clear camera language helps separate the shot from the effect itself.

  • close-up
  • medium shot
  • full-body shot
  • wide shot
  • slow push-in
  • tracking shot
  • fixed camera
  • handheld camera

Effect Control Language

preserve identity, stable background, smooth motion, coherent anatomy, consistent proportions, clean transformation, detailed textures, no text, no watermark, no flicker

Useful Negatives

  • no blurry frames
  • no distorted anatomy
  • no malformed hands
  • no extra limbs
  • no flickering background
  • no watermark
  • no subtitles
  • no static frame

Strong Master Template

[VIDEO_EFFECT_TRIGGER],
[subject] in [environment],
performing [one clear action],
with [specific motion behavior],
shot as [framing + camera movement],
[lighting + mood + visual style],
preserve [identity / anatomy / background / temporal stability],
[constraints / negatives]

Bottom Line

Write the shot clearly first. Then use the effect to push the result, not to replace the need for a good prompt.