Portrait breathes
The person in the photo blinks, breathes, and looks from the window to the camera with a slight smile. Hair moves in a gentle breeze. Locked-off camera, quiet room tone.
Wan 3.0 Prime Prompts: The Guide and 30 Tested Prompts
Wan 3.0 Prime starts from pictures, so the best prompts describe change, not appearance. This guide shows how to animate a photo, how to direct reference images, and how to fill a thirty-second take, with 30 prompts to copy.
The Wan 3.0 Prime prompt formula
Write the parts in this order. Each one answers a question the model otherwise guesses.
- 01
Who or what
“The woman from the reference images”
- 02
What changes
“turns from the window and walks to the door”
- 03
Camera
“slow follow at shoulder height”
- 04
Light
“late afternoon sun through blinds”
- 05
Sound
“footsteps on wooden floor, a distant radio”
Describe motion, not the picture
The start image already shows the subject, the clothes and the room. Repeating that in the prompt wastes words and can pull the model away from your photo. Spend the prompt on what happens: who moves, how, and what the camera does.
Use present-tense verbs and a clear order of events.
Vague
A beautiful woman with long brown hair in a green dress standing in a garden.
Directed
She turns toward the camera, brushes her hair back and walks slowly along the garden path. The camera drifts backward with her. Birdsong and gravel underfoot.
Directing reference images
References work best when they agree with each other: several angles of the same person, or the same product in consistent light. Mixing different people in one reference set blurs the result.
Name the subject by role in the prompt (“the man from the references”, “the bottle in the reference images”) so the model knows the references are the subject rather than the background.
Filling thirty seconds
Give a thirty-second clip three events in order. Tie each to a location or an action so the camera has somewhere to go. For a single moment, choose 5 or 10 seconds instead; it costs less and stays tighter.
Vague
A man walking in a city. 30 seconds.
Directed
A man walks out of a subway exit, buys a coffee from a street cart, then crosses a busy square toward an office tower. Tracking shot from the side. City noise, a cart bell, pigeons.
Resolution strategy
Draft at 480p to check motion and timing, then render the keeper at 1080p with the same prompt and images. The price table on the model page shows how much this saves on long takes.
30 Wan 3.0 Prime prompts by use case
Copy a prompt, or open it in AI Video Pro with the engine already selected. Swap the subject for yours and keep the structure.
Photo to video
Travel snapshot
The waves in the photo roll in and the palm leaves sway; the woman on the sand shades her eyes and looks out to sea. Slow push in. Surf and wind.
Old photo, gentle motion
The couple in the old photograph turn to each other and laugh, the grain and warm tones of the print preserved. Very slow push in. Soft room tone.
Pet portrait
The dog in the photo tilts its head, ears perk up, then it trots toward the camera wagging its tail. Low camera at the dog’s eye level. Paws on grass, a happy bark.
City window
Rain begins to streak down the window in the photo while the city lights beyond blur and flicker. The camera stays fixed. Rain against glass, faint traffic.
Consistent characters
Morning routine
The woman from the references wakes by a sunny window, stretches, and walks to the kitchen to make tea. Handheld follow. Kettle, birds outside.
3–6 reference angles of one person work best.
Office arrival
The man from the references walks through a glass lobby, nods to the receptionist, and steps into an elevator as the doors close. Tracking shot. Footsteps and a soft elevator chime.
Chef in motion
The chef from the reference images tosses vegetables in a wok over high flame, then plates them with a flourish. Medium shot, warm kitchen light. Sizzle and clatter.
Runner at dawn
The woman from the references runs along a waterfront path at dawn, slows down, and stops to catch her breath at the railing. Side tracking shot. Footsteps, breathing, gulls.
Musician on stage
The guitarist from the references plays a slow chord progression under a single spotlight, eyes closed, then looks up to the crowd. Slow orbit. Acoustic guitar and applause.
Products
Hero orbit
The camera orbits the product from the references on a stone pedestal, morning light raking across its surface. Clean neutral background. Low ambient hum.
In the hand
A hand picks up the product from the reference images, turns it to show the label, and sets it down on a marble counter. Close-up, soft daylight. A gentle clink.
Lifestyle table
The product from the references sits on a breakfast table as sunlight moves across it; a cup of coffee is set down beside it. Slow push in. Cafe ambience.
Floating reveal
The product from the references rises slowly out of soft mist and turns to face the camera, ending centered. Dark background, rim light. A low whoosh.
Shelf pull
On a well-lit shop shelf, a hand reaches in and pulls the product from the references to the front. Eye-level shot, shallow depth of field. Store ambience.
Places and landscapes
Canopy rise
From a misty forest floor the camera rises through the trees and breaks above the canopy onto a mountain view. Birdsong fading into wind.
Harbor evening
Fishing boats rock gently in a small harbor at sunset, lights turning on along the quay one by one. Slow pan left to right. Water lapping, distant voices.
Desert wind
Wind lifts ribbons of sand off the crest of a dune under a bright sky; the camera tracks along the ridge. Steady wind.
Snowfall street
Snow begins to fall on a quiet old-town street, lamps glowing, a single cyclist passing through. Locked-off camera. Muffled footsteps and a bicycle bell.
Rooftop garden
A rooftop garden in the city at golden hour, plants moving in the breeze, the skyline behind. Slow dolly forward between the planters. Wind and distant traffic.
Illustration and styles
Painted fox
The illustrated fox trots across the snowy meadow, leaves paw prints, then stops and looks back. Keep the painted style of the image. Soft wind, crunching snow.
Comic panel
The comic-style hero in the image turns, cape lifting in the wind, and looks out over the city. Keep the ink lines and flat colours. Wind and distant sirens.
Watercolor harbor
Boats in the watercolor painting drift and water ripples, colours bleeding softly as they move. Keep the paper texture. Gentle waves.
3D character wave
The stylized 3D character in the image waves at the camera, then does a small happy hop. Keep the soft studio render look. Cheerful ambient sound.
Poster in motion
The travel poster comes to life: clouds drift, the train in the image moves across the bridge, the sun glints. Keep the flat retro style. Train sound in the distance.
Long takes and endings
Night market walk (30s)
A woman in a yellow raincoat walks through a night market: first past a noodle stall, then under paper lanterns, then out onto a quiet street. Tracking shot. Rain, sizzling pans, chatter fading.
Room reveal to last frame
The camera moves through the empty room in the start image and ends on the furnished room in the last-frame image, light warming as it goes. Soft footsteps.
Sunrise timelapse (30s)
Locked-off shot of a valley before dawn; the sky brightens, mist lifts from the fields and the first sunlight crosses the hills. Rising birdsong.
Loop-ready
Gentle waves roll up the beach and pull back, ending in the same position as the first frame so the clip loops. Locked-off camera, surf sound.
Extend: keep going
Continue the same scene and motion at the same pace. Same lighting, same camera, no cuts.
Use in the Extend tab with Wan 3.0 Prime selected.
Frequently asked questions
What makes a good Wan 3.0 Prime prompt?
Describe what changes in the image, not what is already in it: who moves, how, what the camera does, and two or three sounds. Keep it to one to four sentences.
How should I prompt with reference images?
Use references that agree (several angles of one subject) and name the subject by role in the prompt, for example “the woman from the reference images”.
Why does my Wan clip ignore my photo?
Usually the prompt describes a different person or scene. Remove appearance details and describe only the motion.
Can I use these prompts on other engines?
Yes. The structure works across AI Video Pro. Reference prompts need an engine that accepts references, and 30-second takes need Wan 3.0 Prime or Seedance 2.5.
Are there content rules?
Yes. GoCrazyAI is for adults (18+) only. Prompts that break the community guidelines are blocked automatically, and uploads are checked before generation.
Put a prompt to work
Wan 3.0 Prime opens preselected in AI Video Pro. The credit cost shows before you generate.
Last updated: