Why Pack3D exists.
One engineer, a friend's print farm, and a packing engine that beats nesting by hand.
It started with a friend's print farm.
He was losing hours every week arranging parts on plates by hand, and it was exactly the kind of problem my background in mathematics and software was built for. So I wrote a program to do it for him.
The first working version packed around 20% more parts per plate than a skilled human nester, depending on the shape. That program is the engine behind Pack3D today.
What the engine actually does.
Pack3D works from the full 3D geometry of each part, not the box around it. It tries rotations nobody would attempt by hand, tests millions of positions, and keeps the layout that fits the most parts on the plate.
You upload STL or 3MF files and choose your printer, clearance, and print scale. Pack3D hands back a packed 3MF that is ready to slice. A Bambu Studio or OrcaSlicer project opens back up on the plate it came from.
Where the ideas come from.
The engine is never finished. Some ideas come from the packing and optimization literature, which I read steadily. Others are my own: operators and heuristics I design for problems the published work does not cover.
Both are held to the same test. I build the idea, run it against the current engine on real parts, and keep it only if it measurably wins. Most do not. If it does not make plates denser, it does not ship.
What I bring to it.
By day I work as a software engineer, building financial software for banks. I hold an M.Sc. Eng. in Mathematical Modelling and Computation from DTU, the Technical University of Denmark.
Bank software and plate packing have more in common than they sound. In both, a mathematical model has to be correct, fast, and trustworthy enough that someone runs their business on its output. Pack3D is held to the same standard.
