LumaImageNode2 - ComfyUI Built-in Node Documentation - ComfyUI

This node generates images from text descriptions using the Luma UNI-1 model. It takes a text prompt and optional settings like aspect ratio and style, then sends the request to the Luma API to create an image. Two model variants are available: uni-1 and uni-1-max.

Inputs

Common Inputs

Parameter	Description	Data Type	Required	Range
`model`	Model to use for generation. Selecting a model reveals additional settings for that model.	DYNAMIC_COMBO	Yes	`"uni-1"` `"uni-1-max"`
`prompt`	Text description of the desired image. 1–6000 characters. (default: "")	STRING	Yes	1 to 6000 characters
`seed`	Seed controls whether the node should re-run; results are non-deterministic regardless of seed. (default: 0)	INT	Yes	0 to 2147483647

uni-1 and uni-1-max Inputs

Shared by the uni-1 and uni-1-max model options. These settings appear when either model is selected.

Parameter	Description	Data Type	Required	Range
`aspect_ratio`	Output image aspect ratio. `"auto"` lets the model pick based on the prompt. (default: `"auto"`)	COMBO	Yes	`"auto"` `"3:1"` `"2:1"` `"16:9"` `"3:2"` `"1:1"` `"2:3"` `"9:16"` `"1:2"` `"1:3"`
`style`	Style preset. `"auto"` picks based on the prompt; `"manga"` applies a manga/anime aesthetic and requires a portrait aspect ratio (2:3, 9:16, 1:2, 1:3). (default: `"auto"`)	COMBO	Yes	`"auto"` `"manga"`
`web_search`	Search the web for visual references before generating. (default: False)	BOOLEAN	Yes	True / False

Reference Inputs

Parameter	Description	Data Type	Required	Range
`image_ref`	Growable slot: connect 1 to 9 items (e.g. `image_1` to `image_9`). Up to 9 reference images for style/content guidance.	IMAGE	No	Up to 9 images

Note: If style is set to "manga", the aspect_ratio must be either "auto" or one of the portrait ratios "2:3", "9:16", "1:2", "1:3". Using any other ratio with "manga" style will cause an error. The maximum number of reference images is 9 for both uni-1 and uni-1-max.

Outputs

Output Name	Description	Data Type
`image`	The generated image as a tensor.	IMAGE

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

Source fingerprint (SHA-256): 27254fe4627fd340426a68f651cab4513ffb6668cafc0accd17f2c442f7d3125