You have to choose a shape/subfigure CSS can render from the top toolbar. Then render that on the canvas with a standard drag. You can't freehand, obviously. I

You have to choose a shape/subfigure CSS can render from the top toolbar. Then render that on the canvas with a standard drag. You can't freehand, obviously. I was actually wondering how the tool would handle complexity as CSS has very little in the way of options. This is a neat way to do it.