Alibaba introduced a free tool Qwen-Image-Layered, which uses AI to create images with automatic separation into multiple transparent RGBA layers. The model independently separates the background, objects, and text during image generation. This allows users to quickly change the background, move characters, or edit text without disrupting the overall composition of the image.
Qwen-Image-Layered is trained on PSD files and understands the structure of layered images. By default, it creates four layers, but the number can be adjusted according to user needs. This approach makes editing each element simpler and more precise, and changes in one layer do not affect other parts of the image.
The tool is already available in open access. It can be used for basic editing operations, such as replacing text on posters or quickly updating backgrounds. Users note that this approach is especially convenient for working with illustrations and digital art, where it is often necessary to change individual elements without losing the quality of other parts.
Some users note that the current version of Qwen-Image-Layered works slowly when increasing the number of layers, and the image quality may be lower than expected. However, even now the tool can be useful for quick edits, experimenting with layers, and automating routine tasks in graphic editors.
