Instead of layers and masks you just write what to change: “remove the people behind me”, “make the background an office”, “make me look younger”. The AI changes only what you named and tries to keep the face, the pose and the rest of the frame.
Background, clothes, lighting, unwanted objects and people. A short request works too: the “Improve” button turns it into a precise instruction.
Upload more than one photo and build a shared frame, for example two people in the same picture.
A finished picture can be animated at once, upscaled or turned into a series of variations.
The models try to keep the face and pose, but the result depends on the request. If the face drifts, say in the description that it must not change.
Ordinary JPEG and PNG, and iPhone photos in HEIC: they are converted to JPEG automatically on upload.
The interface is open without an account; you need to sign up to run something.