Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.comfy.org/llms.txt

Use this file to discover all available pages before exploring further.

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

Overview

This node generates or edits images by sending a text prompt to Google’s Vertex AI API. It uses a specific Gemini model to create new images or modify existing ones based on your instructions.

Inputs

ParameterData TypeRequiredRangeDescription
promptSTRINGYesN/AText prompt describing the image to generate or the edits to apply. Include any constraints, styles, or details the model should follow.
modelCOMBOYes"Nano Banana 2 (Gemini 3.1 Flash Image)"Selects the Gemini model to use for image generation. Currently only one option is available. This parameter includes additional sub-parameters for resolution, aspect ratio, thinking level, and image input.
seedINTYes0 to 18446744073709551615When the seed is fixed to a specific value, the model makes a best effort to provide the same response for repeated requests. Deterministic output isn’t guaranteed. Also, changing the model or parameter settings, such as the temperature, can cause variations in the response even when you use the same seed value. By default, a random seed value is used. (default: 42)
response_modalitiesCOMBOYes"IMAGE"
"IMAGE+TEXT"
Determines the format of the response. Choose “IMAGE” to receive only an image, or “IMAGE+TEXT” to receive both an image and a text description. (default: “IMAGE”)
system_promptSTRINGNoN/AFoundational instructions that dictate an AI’s behavior. This is an advanced parameter. (default: A pre-defined system prompt instructing the model to always produce an image)
Note on model parameter: The model parameter is a dynamic combo that includes additional sub-parameters for resolution, aspect ratio, thinking level, and image input. These sub-parameters are defined within the model selection and are not listed as separate inputs in this table. Note on image input: You can provide up to 14 images as input to the model. These images are passed through the model parameter’s image sub-field and are used for editing or as visual context for generation.

Outputs

Output NameData TypeDescription
IMAGEIMAGEThe generated or edited image.
STRINGSTRINGA text description or caption generated by the model.
thought_imageIMAGEFirst image from the model’s thinking process. Only available with thinking_level HIGH and IMAGE+TEXT modality.

Source fingerprint (SHA-256): 0b9af4e937874f2e192f3dc0b67f8e769b37d6595fbddb7a326b1f13e3e444d3