Maia RouterInstall

Summary

Edit an image by applying AI-driven modifications based on a textual instruction prompt using one of multiple supported AI models, submitting either a binary image file or an image URL as input.

Inputs

• Model (required) — Select the AI model to perform the image edit: GPT Image 1, Gemini Nano Banana, or Gemini Nano Banana Pro.
• Prompt (required) — Provide the textual instruction describing the desired edit to apply to the image.
• Input Data Mode (required) — Choose the input mode for the image: either provide the image as binary data (from previous workflow node) or as a publicly accessible image URL.
• Binary Property (required) — When 'Binary File' is chosen, specify the name of the binary property containing the image data to edit.
• Image URL (required) — When 'Image URL' is selected, provide the URL of the image to edit.
• Aspect Ratio — For Gemini models, specify the desired aspect ratio of the output edited image (e.g., 1:1 square, 16:9 landscape).
• Quality — For the GPT Image 1 model, select the quality level for output image generation (auto, high, medium, low).
• N (Number of Images) — For GPT Image 1 model, define how many edited images to generate (1 to 10).
• Size — For GPT Image 1 model, specify the resolution of the edited image from 256x256 to 1024x1024 pixels.

Output shape

a list of JSON objects representing the edited image(s), each including base64-encoded image data or URLs depending on the model used

Output data structure depends on the AI model; multiple images can be returned when parameter 'N' is greater than one (GPT Image 1). Errors are propagated per item.