XRMM / Creative guides
Wan 3.0 Prompt Guide.
Shape every shot.
A practical Wan 3.0 prompting guide for subjects, lighting, motion, and style. Keep the formula simple. Add the details that matter.
Maintained by XRMM · Updated
Try your prompt01 / The foundation
Wan 3.0 prompt formulas
Describe what the viewer should see and what should move. More words do not automatically make a better video. Start with a clear subject and one action.
Start with the basics
Subject + Scene + Motion
Name the person or object, place it in a setting, and describe what happens. Include the direction and speed of the action when they matter.
A ceramic cup sits beside a window. Steam rises slowly from the coffee while rain runs down the glass.
Add visual direction
Subject + Scene + Motion + Lighting + Style
Add a light source, framing, or material. Choose compatible details and remove instructions that compete with each other.
A ceramic cup sits beside a rain-covered window. Steam rises slowly. Soft side light, warm natural colors, a close-up with a fixed camera.
For image-to-video
Motion + Camera Movement
Your image already establishes the subject, setting, and appearance. Describe what changes and what should stay consistent. Use a fixed camera when only the subject should move.
Preserve the cup and composition from the starting image. Steam rises gently. The camera slowly moves closer to the cup. One continuous shot.
These short prompts are starting templates, not verified output demonstrations. Try image-to-video mode.
02 / Lighting & composition
Set the light. Shape the scene.
Choose a light source, contrast, and time of day. Describe visible lighting rather than simply asking for a cinematic look.
Shape every shot.
Read the full prompt
Edge light, low contrast, medium close-up, daylight, left-side composition, clean single shot, warm tones, soft light, sunny day, side light, daytime. A young girl sits in a field of tall grass with two fluffy little donkeys behind her. The girl, about eleven or twelve, wears ethnic clothing, black hair in two braids, and an innocent smile. She sits cross-legged, gently touching wildflowers. The donkeys are sturdy, upright-eared, curiously looking at the camera. Sunlight falls on the field, creating a warm, natural atmosphere.
Artificial Light
Read the full prompt
Edge light, medium close-up, artificial light, side light, low saturation, warm tones, right-side composition, over-the-shoulder shot. The camera focuses on a foreign girl in a blue and white checkered shirt, with defined features and bright eyes. A few strands of hair fall from her neat ponytail, adding softness. She tilts her head, lips slightly parted, as if listening. The background is a dimly lit room with a vintage table lamp providing soft side light, illuminating her outline. A black wooden door is faintly visible, adding mystery and depth.
Moonlight
Read the full prompt
Moonlight. A young woman stands in an old-fashioned room with mottled tile walls and an old wooden door. She has short black hair, slender eyebrows, and deep blue eyes that look profound in the faint moonlight, with a focused, contemplative expression. She wears a black top with a white collar, hands down, body slightly turned, gazing off-camera with a tight-lipped, emotional look. Indoor light comes from moonlight through the window, casting soft shadows on her face and enhancing the scene's tranquility and tension. The camera focuses on her expression, creating an intimate, storytelling frame.
Soft Light
Read the full prompt
Sunset, soft light, side light, edge light, warm tones, low saturation, center composition, medium close-up, eye-level shot. A foreign man and woman stand beside a yellow taxi in a half-body shot. The man wears a beige trench coat, hands in pockets, smiling and leaning forward. The woman wears a green beret, matching dress, and lace gloves, elegantly holding the man's arm. The background shows vintage buildings, two green street lamps emit warm halos, and a policeman stands in the distance. Light comes from the right, creating soft edge light and a warm, natural atmosphere.
Hard Light
Read the full prompt
Hard light, side light, medium shot, low saturation, high contrast, medium focal length. Eye-level shot of a foreign man in close-up, wearing a striped shirt, sitting at a wooden table with folders and a red telephone. He crosses his arms, head resting on the wall, eyes closed, mouth slightly open as if speaking. There's a blurred figure in the foreground. The background is a dark wooden wall with paintings and decorations, creating a quiet, contemplative mood.
Top Light
Read the full prompt
Top light, clean single shot, telephoto, close-up. Eye-level shot of a foreign woman's face with brown hair and eyebrows, green eyes, slightly wide nose, mouth slightly open showing teeth. She has two wounds on her forehead with blood seeping out, and a few drops of blood on her left cheek. She looks to the upper right, mouth slightly open then closes. The background is a blurred red wall, overall dark tone.
Daytime
Read the full prompt
Sunrise, daytime, soft light, side light, edge light, medium close-up, center composition, warm tones. A white woman sits on a white bed wearing a blue and white checkered cotton top with messy, fluffy golden-brown wavy hair, looking slightly tired with soft eyes toward the camera as if conversing. Sunlight pours through beige curtains, creating soft light and shadow on the sheets and walls. White iron railings surround the window, creating a peaceful, cozy indoor atmosphere. The camera focuses on her upper body with a blurred background to highlight the subtle interaction between emotion and environment.
Night
Read the full prompt
Practical light, high saturation, mixed light, night, top light, soft light, wide shot, symmetrical composition, wide angle. Eye-level wide shot of two foreigners sitting at a table talking in dim lighting. The man on the left wears a green sweater and white pants, looking focused at the woman on the right. The woman wears a striped shirt and blue jeans, holding a book or folder. Their surroundings have a vintage feel with old-fashioned furniture, chandeliers, and curtains. The table has various items placed on it.
03 / Motion & camera
Give every movement a direction.
Separate subject motion from camera motion. For a Wan camera dolly prompt, specify a slow move toward or away from the subject. A fixed camera, tracking shot, and dolly are different instructions.
Street Dance
Read the full prompt
A group of energetic, stylish hip-hop dancers perform street dance on a wide stage illuminated by vibrant neon lights. Their figures are surrounded by side lighting, forming halos. A wide cinematic shot captures their synchronized steps, lively energy, and youthful expressions. Fast-paced camera work matches the music beat, showcasing complex footwork, explosive moves, and team coordination.
Running
Read the full prompt
A sprinter with a strained face from exertion, facial muscles tight, jaw clenched, wearing a thin athletic vest and shorts, professional spikes, sprints at incredible speed at the finish line of a 100-meter track! Body leans forward, head thrusts out, arms swing with maximum amplitude and frequency, one leg pushes off explosively, the other strides forward. Chest breaks the finish line. The background is a blurred track and audience, timer shows the key time, and the finish tape is clearly visible.
Anger
Read the full prompt
A battle-hardened warrior, face and body covered in mud and scars, wearing heavy leather armor and fur. In a rainstorm, he stands alone on a muddy battlefield, staring ahead in anger.
Fear
Read the full prompt
In a dark room, only the faint light of a phone screen illuminates a young woman's face. Her pupils are dilated in extreme fear, lips slightly parted but unable to make a sound. Cold sweat drips from her forehead, down her temple, slowly across her stiff cheek.
Camera Push-in
Read the full prompt
In a dimly lit room, a detective sits at a desk and opens a file folder. As he sees the contents, the camera pushes in for a close-up of his pupils contracting by 30%, and his expression becomes tense.
Camera Pull-out
Read the full prompt
A soft, round animated character with a curious expression wakes up to find their bed is a giant golden corn kernel. The camera pulls out to reveal the room is a huge corn silo, echoing with sound. Corn kernels are piled into towering walls, and a beam of warm sunlight streams through a high window, casting long shadows on the floor.
04 / Style & color
Choose a consistent visual language.
Use materials, color, and composition to describe the style. Start with compatible directions, then adjust one at a time.
Felt Style
Read the full prompt
Golden hour sunlight, soft lighting, medium-wide angle, low perspective, warm tones. A whimsical village sits at the base of a giant hollow tree trunk, surrounded by lush greenery and mushrooms. Little felt goblins with wool hats drive a miniature wooden train along mossy tracks. The train, pulled by a small engine, carries acorns and berries in tiny carts. Sunlight filters through the leaves above, casting soft shadows and highlighting the rich textures of bark and foliage.
3D Cartoon
Read the full prompt
3D cartoon style, a surreal dream where everything is made of corn. The protagonists ride a corn train through giant corn cobs and kernels. The scene is bathed in warm golden light, enhancing the dreamlike quality. Characters wear rustic clothing and show wonder and curiosity as they travel through this whimsical world. The corn train moves smoothly, its wheels made of perfectly shaped kernels, creating a playful and enchanting atmosphere.
Tilt-shift Photography
Read the full prompt
Tilt-shift photography, high-angle view of a city corner, focusing on a bustling intersection where vehicles and pedestrians weave together, forming a dynamic urban rhythm. Traffic lights flash alternately, guiding the flow. Building edges are sharp, colors vivid, as if a delicate model. Neon signs and shopfronts line the streets, adding a magical miniature world feel. The whole scene is rendered to highlight the unique visual impact of tilt-shift, like observing a busy yet orderly mini urban ecosystem.
Time-lapse
Read the full prompt
Time-lapse, dusk, sunset, edge light, medium-wide shot, side light, warm tones, soft light, moderate contrast, center composition. In the shot, an anthropomorphic metal robot walks on a busy city street. Its shell has a silvery sheen, body structure is precise, and joints have fine mechanical textures. The robot is about two meters tall, with an oval head and two faintly glowing blue sensors as eyes. Its arms swing naturally, steps steady, and leg joints emit a slight hydraulic sound. On both sides of the street are tall glass buildings, the afterglow of the setting sun casts soft rim light on the buildings and robot. Pedestrians move in the blurred background, adding to the urban rhythm. The ground is covered with dark asphalt, some areas reflecting the sky and buildings.
05 / Troubleshooting
Wan 3.0 prompt FAQs
Can I turn off prompt enhancement in Wan 3.0?
Yes. Use the Prompt enhancement switch at the top right of the prompt box in the XRMM workspace. Turn it on to enable enhancement or off to disable it before generating. It is off by default and available in text, image, and reference modes.
Your choice applies to the next generation request. Use the switch rather than writing “disable enhancement” in the prompt. Enhancement does not guarantee that every instruction will be followed.
Why is Wan 3.0 not following my prompt?
If Wan 3.0 is generating something completely different, check for conflicting instructions, mismatched reference inputs, or too many actions in one shot. These are troubleshooting starting points, not a diagnosis of every result.
- Reduce the prompt to one subject, one action, and one camera movement.
- Remove conflicts such as “fixed camera” together with “camera orbits the subject.”
- For image-to-video, describe motion that fits the starting image. For reference mode, check that each media label points to the intended asset.
- Choose enough time for the action in the duration control. Describing a longer sequence in text does not change the output setting.
- Change one instruction at a time and compare outputs with the same input and settings.
For example, replace “spin around the product, zoom in, cut to a new room, and keep the camera still” with “slowly move the camera toward the product in one continuous shot.” Review the new result before adding complexity.
Return to the prompt formulasAbout this guide
XRMM maintains this guide for users of its video workspace. It covers general prompt-writing methods, rather than an official model specification or a benchmark. Example videos are illustrative; exact results depend on the inputs, settings, and generation.
For current Wan 3.0 model inputs and parameters, consult Alibaba Cloud. The API documentation and XRMM workspace controls are separate.
To evaluate a prompt, save the input, prompt, settings, and output together. Change one instruction, compare the result, and record what improved or failed.