Why I built Cutaway.

The first time I tried to reverse-engineer an STL file in Fusion 360 was a couple of years ago.

I had just bought a 3D printer, a Creality Ender 3 Pro, and I was excited to try it. I downloaded a few STLs and wanted to modify them. I am kind of a design guy: I enjoy putting my own signature on everything, and I just wanted to make these models visually better.

I loaded the STL into Fusion 360. There is a feature that everyone said would turn your STL into a solid in one click. I was on a trial licence so I could use it. It gave me a broken geometry, just a few faces with the rest hollow. The original part was pretty simple, but the feature could not do the job.

So I decided to model it from scratch. The problem was that I had to follow the mesh, trace along the triangle intersections, and constantly fight performance issues because the STL itself was heavy. Believe it or not, taking a screenshot of the mesh and tracing the screenshot inside Fusion ended up being simpler, just nowhere near accurate.

Over the months I tried the same on more STLs, and it was always the same fight: working directly with the mesh, trying to trace over it, never quite landing on something clean. The mesh fights you. Edges go where the triangulation says they go, not where the part actually is.

The whole time I had this idea in the back of my head: it would be much more useful to cut cross-sections from the mesh, bring those into Fusion, and trace the sections instead of the mesh itself. Cross-sections are different. A flat slice through the model gives you a clean 2D outline. Tracing on top of that feels almost like tracing on paper.

To be honest, not quite paper. Fusion's sketch tool is not built for tracing either, and the only halfway-clean workflow is to convert the imported DXF into construction lines, fix them so you cannot accidentally move them, trace on top, and remove the construction lines as you go. Not exactly user-friendly, but still far better than tracing over a mesh.

Fusion can technically do this part on its own. You drop a construction plane, generate a mesh section on it, then trace and fit curves to that section by hand. But it was clearly bolted on later, built out of whatever tools were already there rather than designed for the job: make a plane, make a section, fit the curves, then repeat the whole thing for the next plane, and the next. It works, it is just slow and tedious enough that it is not worth it once you are past a couple of sections.

So what I ended up doing for a while was creating sections in Blender with boolean cuts, exporting them as DXFs, and bringing them into Fusion. The frustrating part was the next step. When you have multiple sections from the same model, you have to move each one into its correct location, the intersection of the section with the part. That alone was a whole second problem.

I had this idea for a long time: let me cut whatever sections I want without the boolean bullshit, take them into the 3D software of my choice, and trace. And since Fusion supports add-ins, maybe an add-in could place each DXF exactly where it belongs instead of me dragging them around.

I tried building it. It was way above my skills. I have some Python, but doing this on the side of a full-time job was not realistic, at least for me.

Years passed, and AI came into the picture, particularly Claude. I gave it a shot and boom, it actually did the job. It is still early and I have a lot more ideas, but I figured I would release this now and see if people actually fold it into their workflows. If they do, I will keep adding features.

There are already mesh-to-solid tools in Fusion 360 and other apps. They work, but only on basic geometry. None of them can handle the kind of parts you see in the model viewer on the landing page, and those, by the way, are the same models I first downloaded and tried to modify. I might end up showcasing them in a tutorial where I rebuild each one using Cutaway. Honestly, those parts are not that complicated. There is a real gap between what is automatically solvable and what people actually want to recreate.

So the slice app does the first half of the job. Upload an STL, cut sections, export each section as a DXF. You walk away with a clean set of profiles ready to trace.

The Fusion 360 add-in handles the second pain point. When you export sections from any tool and try to bring them into Fusion one by one, you spend most of your time arranging them: pick the right plane, place the DXF, rotate it, nudge it into the correct position. Doing that for ten sections is a nightmare. The add-in skips all of that. One click and every section lands at the correct location and orientation, on its correct plane, automatically.

Those two pieces work. The remaining problem sits inside the 3D software itself: the tracing tools are not built for this. In Fusion you can trace directly on top of the imported DXF in the same sketch, which is messy, full of mis-clicks and overlapping geometry that you keep accidentally selecting. Or you can project the DXF onto a new sketch and trace there, which is slightly better but still not designed for the workflow.

What you actually want is for the DXF to sit in the background in a light, traceable colour, so you can draw over it without ever accidentally grabbing it. That is the next problem worth solving, and it is why Cutaway is going to keep growing.

In the meantime, Cutaway is already usable today. The tracing side in Fusion is not perfect, but it is a long way ahead of wrestling with a raw mesh, and it will only get better as the 3D software catches up. And if it doesn't, well, that is something I can keep building toward myself.