A fixed 20 pixel border looks completely different depending on what it is wrapped around. On a 400 pixel phone screenshot it would swallow a tenth of the image. On a 6000 pixel camera scan it would be almost invisible. Every measurement in this tool, the border thickness, the corner radius, the polaroid bottom strip, the shadow blur and offset, is stored as a percentage of the photo's shorter side rather than a raw pixel count. A 5 percent border reads as roughly the same visual weight whether the source image is tiny or huge, which is the whole point of a resizable frame.
This also means the live preview and the final download agree with each other exactly. The preview renders against a smaller, resized copy of your photo purely for speed while you are dragging sliders, but the percentage math is applied identically in both places, so nothing shifts or changes proportion between what you see on screen and what actually downloads.
Why the polaroid bottom strip is not just a caption bar
Instant film cameras store their developing chemicals in a small pod sealed inside the bottom of the film pack. When the photo ejects from the camera, rollers spread that chemical pod across the frame to develop the image, and the extra material needed for that pod is what pushes the bottom border so much wider than the top and sides. It was never a design decision, it was a manufacturing constraint that became the signature look. This tool keeps that asymmetry intentionally, since a symmetrical border with a caption glued onto the bottom reads as a border with text on it, not as a polaroid.
Why rotation and shadow both need a bigger canvas
A photo with a border is just a rectangle. Rotate that rectangle even a few degrees and its corners swing outward, so the smallest canvas that can contain it without clipping any corner is now larger than the original rectangle in both directions. Add a shadow on top of that and the canvas needs to grow again, this time to leave room for the shadow's blur radius and its offset in whichever direction it is nudged. That extra space is exactly why a rotated, shadowed polaroid exports at noticeably larger dimensions than the source photo it was built from, and it is also why the canvas background setting matters: the resize tool can bring the final file back down to a specific size once the frame and shadow are baked in, if you need an exact dimension afterward.