A round knob above the chosen box turns it (Shift: 5 degree steps); a Turn
field under the canvas types the angle and a double-click levels it. The
mask follows the turned box exactly, bboxes and elements keep the straight
box around it, and a turned text box adds how it is tilted to its
description. A box that is not turned saves and outputs exactly as before.
The arrow keys move a box one pixel (Shift: ten), and the grid starts off
when a picture is wired in. Handles, tags and outlines keep one size on the
screen at any zoom, and over a picture the boxes are see-through. Each box's
tag says whether it is text or an object, a chip per box picks the box to
edit, and the type buttons read as that box's own setting. A gear sets the
grid size. With a picture shown, the canvas keeps the picture's own shape
even when width and height are wired.
Draw boxes on a canvas and say what goes in each one: a thing, or words to
write in the picture. Wire elements to MOSS-Image Layout Prompt and the
picture puts each thing where its box is.
It gives what ComfyUI's own Create Bounding Boxes gives (preview, bboxes,
elements; same types, same order, same values from the same boxes), so it
can take that node's place, and its canvas also draws in the classic node
look. The boxes are kept as parts of the canvas in a hidden socketless
string, so a wired size never moves them.
ausboss