Question
What is a mask in AI image editing?
A mask is a map that marks which part of an image the AI may change and which part it should keep. Usually the editable area is painted white or made transparent, and the rest is protected. The model fills the marked area to match the prompt and the surrounding image.
Conventions differ by tool. Hugging Face's Diffusers documentation uses white for the area to fill and black for the area to keep. OpenAI's image edit endpoint instead expects a mask with an alpha channel, "of the same format and size" as the image, and says its masking is "entirely prompt-based", meaning the mask guides the edit rather than locking pixels.
Many newer editors let you skip drawing a mask: you describe the area in words ("the sky", "the mug on the left") and the model works out where it is.
ImageQuarry works by description: you select a version and say what to change and what to keep, and the change arrives as a new version beside the old one.
Source: Hugging Face Diffusers inpainting documentation, OpenAI image generation guide and the ImageQuarry homepage, checked October 20, 2026.
Related questions
Last reviewed October 20, 2026 by ImageQuarry
Create an image. Edit it with words. Keep refining.
Describe the area instead of drawing it
Select a version, name the part to change, and keep the earlier version for comparison.
Start creatingAlready have an account? Sign in