This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHubThe ByteDance Image Edit node allows you to modify images using ByteDance’s AI models through an API. You provide an input image and a text prompt describing the desired changes, and the node processes the image according to your instructions. The node handles the API communication automatically and returns the edited image.
Inputs
| Parameter | Data Type | Input Type | Default | Range | Description |
|---|---|---|---|---|---|
model | MODEL | COMBO | seededit_3 | Image2ImageModelName options | Model name |
image | IMAGE | IMAGE | - | - | The base image to edit |
prompt | STRING | STRING | "" | - | Instruction to edit image |
seed | INT | INT | 0 | 0-2147483647 | Seed to use for generation |
guidance_scale | FLOAT | FLOAT | 5.5 | 1.0-10.0 | Higher value makes the image follow the prompt more closely |
watermark | BOOLEAN | BOOLEAN | True | - | Whether to add an “AI generated” watermark to the image |
Outputs
| Output Name | Data Type | Description |
|---|---|---|
IMAGE | IMAGE | The edited image returned from the ByteDance API |