image-edit

To route an AI image editing request, the skill evaluates your editing intent and selects the most appropriate RunComfy model, such as Nano Banana Edit or Flux Kontext Pro, to execute the edit. This reduces failed iterations and preserves important image details.

To perform mask-driven inpainting, you need a publicly accessible image URL and an optional grayscale mask URL to define the targeted area. The request is then routed to the Z-Image Turbo Inpaint model for precise localized changes.

Yes, you can execute batch identity-preserving background swaps across consistent SKU batches. The skill routes these requests to specialized models that maintain the core subject's identity while replacing the surrounding background seamlessly.

Yes, RunComfy image editing supports multilingual text replacement, allowing you to rewrite image headlines in languages like Japanese. The skill applies model-specific prompting patterns to accurately render the localized text within the image.

Yes, an authenticated RunComfy CLI installation is required to execute image edits. The skill validates your edit requests and submits them through the CLI, allowing you to download the generated results locally.

The best way to compose an image from multiple references is to supply the image URLs and describe the desired composition. The skill routes the request to the appropriate model, applying the correct image limits and aspect ratios for the final output.