WanReferenceVideoApi - ComfyUI Built-in Node Documentation

Complete documentation for the WanReferenceVideoApi node in ComfyUI. Learn its inputs, outputs, parameters and usage.

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

The Wan Reference to Video node uses the visual appearance and voice from one or more input reference videos, along with a text prompt, to generate a new video. It maintains consistency with the characters from the reference material while creating new content based on your description.

Inputs

Parameter	Data Type	Required	Range	Description
model	COMBO	Yes	"wan2.6-r2v"	The specific AI model to use for video generation.
prompt	STRING	Yes	-	A description of the elements and visual features for the new video. Supports English and Chinese. Use identifiers like character1 and character2 to refer to the characters from the reference videos.
negative_prompt	STRING	No	-	A description of elements or features to avoid in the generated video.
reference_videos	AUTOGROW	Yes	-	A list of video inputs used as references for character appearance and voice. You must provide at least one video. Each video can be assigned a name like character1, character2, or character3.
size	COMBO	Yes	"720p: 1:1 (960x960)" "720p: 16:9 (1280x720)" "720p: 9:16 (720x1280)" "720p: 4:3 (1088x832)" "720p: 3:4 (832x1088)" "1080p: 1:1 (1440x1440)" "1080p: 16:9 (1920x1080)" "1080p: 9:16 (1080x1920)" "1080p: 4:3 (1632x1248)" "1080p: 3:4 (1248x1632)"	The resolution and aspect ratio for the output video.
duration	INT	Yes	5 to 10	The length of the generated video in seconds. The value must be a multiple of 5 (default: 5).
seed	INT	No	0 to 2147483647	A random seed value for reproducible results. A value of 0 will generate a random seed.
shot_type	COMBO	Yes	"single" "multi"	Specifies whether the generated video is a single continuous shot or contains multiple shots with cuts.
watermark	BOOLEAN	No	-	When enabled, an AI-generated watermark is added to the final video (default: False).

Constraints:

• Each video provided in reference_videos must be between 2 and 30 seconds in duration.
• The duration parameter is limited to specific values (5 or 10 seconds).

Outputs

Output Name	Data Type	Description
output	VIDEO	The newly generated video file.