The Bounding Box Grid node in ComfyUI raised a new technique to layout and position elements within the composition in a visual and more Cartesian precise way. It’s equivalent to the Edit from Qwen, Flux and others well known generative models.
This study will refer the use of Bounding Box in a MOSS Image workflow.
Tip 1: if your Bounding Box node is greyed out and you can not draw the rectangles inside the grid area, you need to turn on the Node 2.0 on ComfyUI. You cna do that by clicking the ComfyUI logo and then on Node 2.0.
Tip 2: you have to manually enter the width and height of your output otherwise the box calculation will be a miss. If you change the aspect ratio, you have to change here the dimensions as well.
The workflow requires three new nodes to be setup. Double-click the empty area and choose the Create Bounding Box node, the MOSS-Image Layout Prompt node and the Prompt as Text node. From the first node you will connect the output Elements to the input Element, and the output Prompt to the input Source. To finish, you have to connect the output String to the input of your Positive Prompting, in this case would be Text to Image (MOSS Image).
Once everything is setup, you can start prompting and drawing the rectangles inside the Bounding Box node. Check some examples of how the Bounding Box node can affect the composition.
