Making every piece behave

SpinPack starts with a simple challenge: rotate the pieces, fit them into the box, and leave no gaps. Add a few layers of depth, though, and a shape that looked perfect suddenly needs another rethink. Recent development has focused on the machinery behind those puzzles. I’ve been improving how the 3D renderer keeps and reuses its models between puzzles, and tightening up how it updates their positions, rotations and sizes. The latest fix makes related model updates happen together. That covers the puzzle pieces, box outline, landing guides and placement markers—all the visual clues you rely on when deciding where a piece belongs. It’s quiet work, but a spatial puzzle needs its display to be dependable. I’ve also updated the app’s underlying Expo dependencies and clarified the Android release workflow, with separate production and internal testing builds. Alongside that work, SpinPack has starter puzzles, deeper challenges, optional visual aids and hints for those moments when one awkward piece has you stumped. You can also unlock puzzle sharing and send someone else the same challenge. I’d love to hear where you get stuck: judging depth, finding the right rotation, or figuring out which piece to place first. Those moments are useful clues for making SpinPack better.