ByteDanceSeedreamNode - ComfyUI Built-in Node Documentation

Complete documentation for the ByteDanceSeedreamNode 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 ByteDance Seedream 4 node provides unified text-to-image generation and precise single-sentence editing capabilities at up to 4K resolution. It can create new images from text prompts or edit existing images using text instructions. The node supports both single image generation and sequential generation of multiple related images.

Inputs

Parameter	Data Type	Input Type	Default	Range	Description
model	MODEL	COMBO	”seedream-4-0-250828”	[“seedream-4-0-250828”]	Model name
prompt	STRING	STRING	""	-	Text prompt for creating or editing an image.
image	IMAGE	IMAGE	-	-	Input image(s) for image-to-image generation. List of 1-10 images for single or multi-reference generation.
size_preset	STRING	COMBO	First preset from RECOMMENDED_PRESETS_SEEDREAM_4	All labels from RECOMMENDED_PRESETS_SEEDREAM_4	Pick a recommended size. Select Custom to use the width and height below.
width	INT	INT	2048	1024-4096 (step 64)	Custom width for image. Value is working only if size_preset is set to Custom
height	INT	INT	2048	1024-4096 (step 64)	Custom height for image. Value is working only if size_preset is set to Custom
sequential_image_generation	STRING	COMBO	”disabled”	[“disabled”, “auto”]	Group image generation mode. ‘disabled’ generates a single image. ‘auto’ lets the model decide whether to generate multiple related images (e.g., story scenes, character variations).
max_images	INT	INT	1	1-15	Maximum number of images to generate when sequential_image_generation=‘auto’. Total images (input + generated) cannot exceed 15.
seed	INT	INT	0	0-2147483647	Seed to use for generation.
watermark	BOOLEAN	BOOLEAN	True	-	Whether to add an “AI generated” watermark to the image.
fail_on_partial	BOOLEAN	BOOLEAN	True	-	If enabled, abort execution if any requested images are missing or return an error.

Outputs

Output Name	Data Type	Description
IMAGE	IMAGE	Generated image(s) based on the input parameters and prompt