The Moment I Knew Something Was Wrong Two weeks ago, I pulled a 12-hour print off my build plate. The model was beautiful. And 35% of the filament I paid for was sitting in the trash — twisted, snapped, and sanded-off support structures that had done their job and become instant waste. I'd been printing for years. I knew supports were "necessary evil." But then I started noticing something: every slicer on the market — Cura, PrusaSlicer, OrcaSlicer, Bambu Studio — uses the same algorithm to decide where supports go. A simple angle threshold. If a face is steeper than 45°, slap supports everywhere beneath it. No per-geometry intelligence. No optimization. Just brute force. So I built a tool that finally asks the right questions. Here's how it works under the hood. Architecture at 10,000 Feet I wanted a CLI tool that: Takes a standard STL file Analyzes the actual geometry — not just surface angle Generates optimized supports per-region, not per-uniform-rule Outputs STL you can drop into any slicer No database.…