Wan 3.0 Prime Reference to Video

Name: Wan 3.0 Prime Reference to Video
Author: Nodespell Team

Official

Connect at least one reference, then describe how to use it. Reference chips identify ordered images, videos and audio independently. A document or pu

Nodespell AI

AI / Video / Alibaba

Connect at least one reference, then describe how to use it. Reference chips identify ordered images, videos and audio independently. A document or public webpage can supply context alongside media; provide only one of these and keep Enhance Prompt on. Use Video Edit to change a source clip, or Extend Video to add footage before or after it.

Prime offers faster generation with the same controls and higher cost than Wan 3.0.

Reference video seconds and output seconds both contribute to cost. The combined duration cannot exceed 30 seconds. This model is in preview. Results can vary between runs.

Inputs (6)

Prompt

String

Describe the scene, movement and sound. Maximum 20,000 characters after resolving reference labels.

Multi InputMin: 0Max: 100

Reference Images

String

Up to 10 ordered JPEG, PNG, BMP or WebP images, each at most 20 MB and 240–8000 pixels per side. Transparent PNGs are not supported. Refer to them as @Image1, @Image2, and so on.

Multi InputMin: 0Max: 10

Reference Videos

String

Up to 5 ordered MP4 or MOV clips, each 1–15 seconds, at least 16 fps and at most 100 MB. Combined video length at most 15 seconds. Input video length contributes to cost. Refer to clips as @Video1, @Video2, and so on.

Multi InputMin: 0Max: 5

Reference Audio

String

Up to 5 ordered WAV or MP3 clips, each 1–15 seconds and at most 15 MB; combined audio length at most 15 seconds. Refer to them as @Audio1, @Audio2, and so on.

Multi InputMin: 0Max: 5

Document

String

Choose or upload one document asset, at most 100 MB and 50 pages where applicable. External document URLs are not accepted. Cannot be combined with Webpage URL. Requires Enhance Prompt.

Min: 0Max: 1

Webpage URL

String

One publicly accessible webpage that does not require login. Cannot be combined with Document. Requires Enhance Prompt.

Min: 0Max: 1

Parameters (11)

Prompt

String

Describe the scene, movement and sound. Maximum 20,000 characters after resolving reference labels.

Default:

Document

String

Choose or upload one document asset. External document URLs are not accepted. Use either a document or a webpage.

Default:

Webpage URL

String

Optional public webpage URL. Requires Enhance Prompt; cannot be combined with a document.

Default:

Resolution

String

Higher resolution increases cost.

Default: 720p

Duration

String

Output length in seconds. Input video plus output must fit 30 seconds. Auto may use the full remaining duration; its estimate is an upper bound.

Default: 5

Aspect Ratio

String

Auto chooses framing from the inputs and instruction.

Default: auto

Generate Audio

Boolean

Include generated dialogue, music or sound effects. Turning audio off does not lower the price.

Default: true

Enhance Prompt

Boolean

Expand the prompt to add detail. Required for document or webpage references; may increase generation time.

Default: true

Advanced

Object

Seed and watermark controls.

Default: false

Seed

Number

Use -1 for a random seed, or 0–2147483647 for a fixed seed. A fixed seed does not guarantee identical results.

Default: -1

Watermark

Boolean

Add a visible watermark to the generated video.

Default: false

Outputs (1)

Video

Inferred

Generated video with embedded audio when enabled.

Nodespell Team

Creator profile

Type

Node

Status

Official

Package

Nodespell AI

Keywords

WanWan3Wan 3 0Reference To VideoPrimeFast

Use in Workflow