Scott Sykora, the developer behind PartForge, spends a lot of his free time making mythical playthings for his 6-year-old daughter.
"She's in her dragon era, so she's constantly asking me to 3D print more dragon toys for her," Sykora says.
His 10-year-old son, Auggie, has his own requests, including – recently – a gag pen that squirts water instead of writing.
"For that, we found a little hand-sanitizer pump mechanism and then used Partforge to build a fake pen and water reservoir around it."
It's the kind of project that captures what makes 3D printing so appealing: the ability to dream up something specific, then actually make it.
But there's a catch. While 3D printers have made fabrication more accessible, designing custom parts remains a major hurdle. Traditional computer-aided design (CAD) software is powerful, but notoriously difficult to learn.
Sykora knows this firsthand. And his frustration with that learning curve eventually led him to create PartForge, an AI-powered platform that lets users describe what they want in plain language, generate a customizable 3D model, and refine it until it's ready to print.
From Toy Dragons to Robotic Arms
After watching AI transform knowledge work and digital creation, Sykora became curious about its potential in the physical world, particularly robotics. What he discovered surprised him: a thriving open-source community of hobbyists designing and building sophisticated machines at home.
There were robotic arms, bipedal walking robots, intricate mechanical drives, and adorable little robotic creatures. Many of the designs were freely shared, allowing other makers to reproduce, modify, and improve them.
And Sykora wanted in.
"The AI embodiment isn't quite there for a useful robot butler, but it's getting close," he says. "I want to be making in the space when it does cross that threshold." His motivation goes beyond the technology itself.
"I want this era to be driven by makers tinkering at home and not some giant corporation. Making and sharing things myself seems like the best way to help that happen."
The Problem With CAD
There was one major obstacle between Sykora and his robotics ambitions: CAD software. Ironically, he already knew how to use it. After college, Sykora worked as a freelance CAD modeler, helping inventors visualize their ideas with SolidWorks. He'd also used 3D Studio Max and Blender for various digital projects.
But it had been years since he'd worked with CAD, and he was rusty. "Each of those tools is amazing and incredibly powerful in its own right, but each took months to learn enough to be productive in them," he explains. "And now, after not using them, I'm back at square one."
The interfaces were unfamiliar. The muscle memory was gone. And the simple mechanical shapes he needed for his robotic arm hardly justified spending months relearning complex software.
He began wondering why he couldn't just describe the part he needed and have AI build it. That question became PartForge.
Rather than manually constructing a model through menus, sketches, and constraints, users can explain what they want in ordinary language. PartForge generates an adjustable 3D design, complete with parameters that can be changed as the project evolves. Users can also upload photos or sketches, giving the AI a visual starting point.
The Aha Moment
For Sykora, the platform's potential became clear during a robotic arm build. He'd ordered the wrong size bearings, a small mistake that would ordinarily require manually reworking the CAD model to accommodate the new dimensions.
Instead, he told PartForge what had happened and specified the bearings he'd purchased. The AI looked up the bearing specifications, found the correct measurements, and adjusted the design automatically.
"It even added a little control later to switch it back, no questions asked." That was the moment PartForge began to feel like something more than an easier way to generate 3D models. It could help solve real engineering problems while keeping designs flexible enough to accommodate changes.
A Shared Workshop
Partforge's vision goes beyond individual projects.
Every published design, or forge, can join a shared library of customizable creations. Other makers can build on those designs, adapt their dimensions, or incorporate existing mechanisms into entirely new projects, subject to licensing permissions. Crucially, the AI can also draw on that growing library, building on work that other users and agents have already completed.
Sykora recently created an articulated steampunk spider with print-in-place joints that allow its legs to be repositioned.
The spider is a fun project in its own right. But the joint mechanism is now available in Partforge's library, where it could become the foundation for other articulated creatures, toys, or mechanical designs.
"I'm really excited where other makers take that and what that system compounds into," he says.
It's a vision of collaborative creation in which every new project potentially makes the next one easier.
And perhaps that's the most exciting part of all. For Sykora, the future of making things isn't just about robots or mechanical designs. It's about giving anyone with an idea – including a couple of kids dreaming up dragons and mischievous inventions – the tools to bring it to life.