Online AI image editor with layers
Build on a finite canvas with photos, generated accessories, editable text and shapes in separate layers. The assistant analyzes the whole composition, edits across layers and adds or refines elements in stages, checking the actual composite after each saved step.
Open toolUpdated
- What you provide
- Uploaded images or layered files, existing private generated images, text and editing instructions.
- What you get
- Editable private projects, transparent PNG, JPEG, WebP, SVG, PDF, PSD and Compositor-compatible packages.
How to use it
- Create a canvas and set its width and height, or open an image, PSD, Sparkles project or zipped Compositor project.
- Add images from your library, text and shapes. Use selections, masks and layer properties to refine the composition.
- Tell the assistant the poster or retouching goal. It can edit all unlocked layers; selected layers only indicate focus. It saves layout, generated elements and refinement as separate versions, then inspects the composed image.
- Save and inspect the export. Changing canvas bounds crops or extends the working area; scaling the image also resizes its contents.
Common questions
Does resizing the canvas stretch my photo?
Choose canvas bounds to crop or extend the area without scaling the layers. Choose scale entire image when the content should change size too.
Can generated accessories stay on separate layers?
Yes. Ask for separate image layers and a transparent background where needed. They can then be moved, masked and adjusted independently.
Will the assistant overwrite simultaneous manual edits?
Edits use the frozen project version. If a newer saved version exists, the assistant result is kept separately and can be opened as a copy.