Google Vertex AI icon

Google Vertex AI

Use Google Vertex AI (Gemini) for text, media analysis, and image generation/editing

Actions11

Image → Edit Image

AI-generated

Summary

Edits one or more input images from the incoming item's binary data using a text prompt and a Gemini image ("Nano Banana") model on Vertex AI, returning the edited image(s) as binary output.

Inputs

  • Image Model (required) — Gemini image model: 2.5 Flash Image (default), 3 Pro Image, 3 Pro Image Preview, or 3.1 Flash Image.
  • Images (required) — One or more entries, each naming a binary property on the incoming item (default "data"). All are sent to the model in order. Blank names are ignored. If none remain, the operation errors with "At least one input image is required".
  • Prompt (required) — Instruction describing the change to make, e.g. "add a red hat".
  • Options > Aspect Ratio — Output aspect ratio such as 1:1, 16:9 or 9:16. Leave blank for the model default.
  • Options > Image Size — Output resolution of 1K, 2K or 4K. The description says 2K and 4K are supported by the Gemini 3 image models.
  • Options > Put Output In Field — Binary field name for the result (default "edited"). Additional images are stored as edited1, edited2, and so on.

Output shape

One output item per input item. Edited images are in binary fields (edited, edited1, …) with filenames like edited.png. json contains {text} if the model returned any text, otherwise it is empty.

Duplicate images in the response are dropped. If the model returns no image, for example because the prompt was blocked, the item contains json {text} plus blockReason when present, and no binary. Requests use the Google Vertex AI service-account credential's project and region. With Continue On Fail, errors are returned as json {error}.

Examples

Example 1: Change the background of an uploaded product photo

Images: binaryField=data; Prompt: "Replace the background with a plain white studio backdrop"; Image Model: gemini-2.5-flash-image.

Example 2: Combine two images into one 16:9 composite at high resolution

Images: add two entries (data, logo); Prompt: "Place the logo on the shirt"; Image Model: gemini-3-pro-image; Options: Aspect Ratio 16:9, Image Size 2K.

Links

Discussion