Skip to main content

Bring a still to life with Wan 3.0 Pro, Alibaba's Wan 3.0 delivered at 1080p, 2K or 4K: the opening frame becomes 2 to 30 seconds with generated audio, an optional last frame fixes the ending, and a sharp source matters at the larger tiers.

From $0.144/s20% below the reference list price

Workbench

Input

Image to video
278/20000
Optional here, because the frame already carries the look. Use it for motion and camera only. at least 1 characters · at most 20000 characters
1/1

Drop files or media links, paste, or browse · 1 file · Max 10 MB · JPG · PNG · WEBP · GIF

  • 1
1 / 1 files
The Pro master begins on image_url, so inspect this image before enlargement; accepted files are JPEG, PNG without transparency, BMP or WEBP, 240–8000 px per side, no wider than 8:1 and at most 20 MB. at least 1 characters
0/1

Drop files or media links, paste, or browse · 1 file · Max 10 MB · JPG · PNG · WEBP · GIF

0 / 1 files
Control the Pro shot’s landing composition with this optional exact final frame; image-to-video requires the opening image too; JPEG, opaque PNG, BMP or WEBP, 240–8000 px per side, 8:1 maximum, 20 MB maximum. at least 1 characters
affects price
For Pro delivery, choose 1080p, 2k or 4k (default 1080p); above native 1080p the file is upscaled, so inspect lettering and texture at full size; the resolution tier changes the price.
For a Pro master you plan to crop, specify 16:9, 4:3, 1:1, 3:4 or 9:16; leaving adaptive, the default, lets framing follow the source and scene.
Durationaffects price
Set the Pro shot length before choosing delivery resolution: 2–30 whole seconds, default 5; -1 reserves the 30-second ceiling then settles on actual output. no lower than 2 · no higher than 30
Generate audio
Pro includes a generated audio track by default; false suits a separately scored master, and in video extension it changes only the new segment, leaving the original audio intact.
Prompt expansion
Pro rewrites prompts by default; turn expansion off when wording is final, but leave it true or unset for reference_file_url or reference_link_url because those inputs are read through expansion and false is rejected.
Seed
Keep a Pro comparison on the same random starting point with 0–2147483647; compare the actual files because repeated seeds do not ensure pixel-identical output. no lower than 0 · no higher than 2147483647
Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run
Example result

Related models and next tasks

Missing the model you need?

One frame in, motion out, upscaled to 4K

Turn a finished character portrait or an approved photograph into a moving shot. The image establishes the starting look; the prompt directs a turn, a glance, a camera move or the sound around it. Choose 2–30 seconds and 1080p / 2k / 4k.

Open the playground
One frame in, motion out, upscaled to 4K
Starting material
One opening image; optional closing image
Length
2–30 s · 5 s default
Delivery
1080p / 2k / 4k · 1080p default
Sound
Generated audio is on by default; switch it off for a separate soundtrack.
Starting material
Choose the right starting pointUse an opening image when the face, outfit and composition are already chosen. Add a closing image when a reveal, final pose or loop needs a defined destination.
Length
Plan the cost before extending the shotA fixed-length clip costs the selected per-second rate multiplied by its length. Find the movement in a five-second draft, then select the delivery resolution and check the live quote before submitting.

What to create with Wan 3.0 Pro

Choose the format for the actual destination

Choose the format for the actual destination

Start at 1080p to choose the movement. The 2k and 4k tiers enlarge that render for delivery; select them when the placement needs those dimensions and inspect the source render before enlarging it.

  • A couple and a fogged mirror

    Steam, reflections and two nearby faces give you useful details to inspect. Watch the mirror-clearing gesture in the middle of the clip.

Useful ways to apply Wan 3.0 Pro

  • Build a character teaser around one clear beat

    Character creators

    Build a character teaser around one clear beat

    Direct an adult character’s glance, approach or small gesture. One camera move and a clear ending give a short NSFW teaser an understandable arc.

  • Give fabric, light and posture a visible role

    Fashion and editorial

    Give fabric, light and posture a visible role

    Choose the garment and lighting before writing motion. A turn toward the window, a curtain moving or a hand lifting a cup gives the viewer something specific to watch.

  • Make atmosphere part of the shot

    Sound-led scenes

    Make atmosphere part of the shot

    Name the sound source: rain on glass, room tone, footsteps or a brief line. Leave generated audio on for an audiovisual draft, or switch it off when you will add a track in editing.

  • Keep a reusable visual anchor

    Series and campaigns

    Keep a reusable visual anchor

    Keep the successful still, prompt and settings together. Use that image for subsequent image-driven shots, or consistent references when the location changes. Review the middle of each clip as well as its first frame.

Make your first Wan 3.0 Pro clip

  1. Choose a clear opening and a purposeful ending

    Crop the opening still for the final placement. Add a closing still only when its pose or composition matters; keep the lighting and camera relationship understandable between the two.

  2. Direct the change

    Describe one action, its pace and one camera move. “She turns toward the lens as the camera slowly moves closer” gives the shot a clear instruction. Keep the face visible if identity is the main selling point.

  3. Choose five seconds and read the quote

    Begin with a short take at a suitable draft resolution. After the direction works, increase length or output size one at a time and review the new quote.

  4. Keep the take that works

    Watch the opening, middle and ending for face, hands, garment and set continuity. Save the inputs with the clip. Hold seed while comparing one edit; use a new seed when exploring a different take.

Wan 3.0 Pro: saved request and prompt ideas

  • Saved example: A couple and a fogged mirror

    The saved example A couple and a fogged mirror requested duration_seconds=5; resolution="1080p". Open Examples to compare its actual input and video, then try one change with your own material.

    He draws her closer and rests his chin on her shoulder; she tips her head back against him, laughs and wipes a clear patch in the fogged mirror with her fingertips so they can see themselves. Steam rolls up past the brass sconce. Slow push in, cinematic, shallow depth of field.
  • A packshot returning to its hero angle

    Using the same product image as both image_url and last_image_url keeps the edit's start and end frames identical while the prompt directs only the motion in between.

    Camera slowly orbits the perfume bottle from the source image, catching light off the glass, then returns to the exact starting angle and composition, soft studio light, subtle steam rising in the background
  • A portrait with camera-only direction

    With appearance already fixed in the source image, the prompt describes nothing but camera movement and a single gesture, leaving wardrobe and lighting untouched.

    Slow push-in on the woman from the source image as she lowers her chin and looks up at camera, hair catching a slight breeze, everything else in the frame held still
  • A quiet two-frame close

    Pinning both the opening and closing frame keeps a private, close moment from drifting into a different pose than the one already approved for the shot.

    The couple from the source image lean closer together under low lamp light, foreheads nearly touching, breathing visible, then settle into the exact framing shown in the closing reference

Choosing and using Wan 3.0 Pro

A prompt pattern you can use

Write: starting state → one action → camera movement → ending → sound. Give a longer take a visible sequence, not more adjectives. Set duration and resolution in the controls, then use the prompt to direct what the viewer should see.

All Wan models22

Video models18