Actions75
- AI Actions
- Artifact Actions
- Audit Log Actions
- Chat Actions
- Compliance Pattern Actions
- Dataset Actions
- Dataset Row Actions
- Miscellaneous Actions
- Project Actions
- Reference Actions
- Space Actions
AI → Edit Image
AI-generatedSummary
Edits each input image using a text prompt and the selected image model, then stores the generated image file in the specified binary property.
Inputs
- Input Image Binary Property (required) — Name of the incoming binary property containing the source image (PNG, JPEG, or WebP). Defaults to
data. - Prompt (required) — Text describing the edit to apply; up to 32,000 characters.
- Put Output File in Field (required) — Binary property name for the edited image output; defaults to
data. - Additional Options — Optionally choose input fidelity (high or low), image model, number of images (1–10), output format (JPEG, PNG, or WebP), quality, and output size or aspect ratio. Defaults are low fidelity, GPT Image 1.5, one image, WebP, auto quality, and auto size. Input fidelity is unsupported by GPT Image 1 Mini.
Output shape
Input items with edited image files in the configured binary property.
The operation generates the requested number of images per input. The supplied source pack only exposes the operation's function signatures, so it does not show whether output items retain other input fields or how multiple generated files are named.
Examples
Example 1: Edit an image from the incoming item's `data` binary property and save the result to `editedImage`.
Set the prompt to describe the desired change; optionally set the model, output format, size, and number of images.