TL;DR
Get decor and gifts delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A maker is using Google’s Gemini image-generation and coding tools to help design 3D-printable parts for a Fallout-inspired Halloween prop with his son. Gemini produced OpenSCAD code that the maker could adjust, but its designs also contained errors, and the finished prop has not yet been built.
A maker is using Google Gemini to develop 3D-printable components for a Fallout-inspired Halloween prop he is building with his son. The project shows how AI-generated OpenSCAD code can help turn an early design idea into adjustable geometry, though the maker says the output required repeated prompting and corrections, and the prop is still in progress.
The project began after the maker’s son chose a costume tied to Fallout 4. A full DIY power-armor costume was not practical on the available schedule, so the maker shifted toward a character with a decorated chest piece and a more improvised, wasteland-survivor look. An attempted order for a motorcycle chest protector was delayed and apparently lost, leading him to consider foam and craft materials for the costume instead.
He has focused on building a distinctive handheld prop. One early modeling task was a truncated cone with a mounting ring. After several iterations, he used Gemini-generated OpenSCAD code to create the shape from two connected truncated cones. The parametric code allowed him to change dimensions and geometry without starting over. A test print was usable for planning the next version, although he reported that the filament was wet and being dried.
The maker also used 3D files for air fittings to check how parts might fit with quarter-inch inner-diameter hoses. He has not decided whether to use inexpensive metal fittings, and says the test helped him consider the options before buying a range of parts. Planned changes to the printed components include making one part larger and adding mounting holes so it can attach to a wooden box that will form the prop’s main body.
AI as a Workshop Design Aid
The project offers a specific example of AI being used not as a finished-product generator but as a starting point for a physical design. For a maker who finds it difficult to model an object that has not been fully worked out mentally, generated code can provide a shape to inspect, print and revise. In this case, editable OpenSCAD parameters gave the maker a way to test dimensions and move toward an assembly.
The limits are just as relevant. Gemini did not consistently follow the maker’s instructions: he said it placed side arms incorrectly, added pill-shaped holes in the wrong areas and introduced chamfers he did not want. The results still needed human judgment, iteration and fabrication work. The account does not establish that AI made the project faster or easier overall; the maker describes it as helping with steps he typically finds tedious, while noting that substantial work remains.
Popular right now
TCL 85-Inch QM6K Series QLED Mini LED 4K HDR Smart Google TV
LG 85-inch QNED75B Mini LED Smart TV Bundle with HDMI and CPS Protection Plan
Ninja Foodi 10-Qt DualZone Smart XL Air Fryer, 6-in-1, DZ550
As an affiliate, we earn on qualifying purchases.
From Costume Idea to Printed Parts
The maker says he is comfortable modeling objects he has worked on before, but finds it harder to design new parts without a settled concept. This project adds another challenge: several components must ultimately fit together. He describes moving from individual pieces to a cohesive assembly as the point where planning becomes difficult, and says he often joins parts with hardware, hot glue or epoxy.
For the prop, he is considering attaching separately designed parts to a small wooden box, roughly tissue-box size. His son wants wires on the outside, while the maker has been exploring a simplified breakdown of a fantasy tool into printable subassemblies. He found the AI’s proposed breakdown unhelpful, though it supported his existing idea of using mounting pins for glued assembly. The pair have discussed making the object a detector, but the son will decide what to call it.
The maker says he may remodel some parts manually in Fusion or similar software, which he expects would make it easier to adjust dimensions and manage relationships between components. For the current deadline, he is leaning toward a rough first version built from separate parts rather than redesigning the entire assembly.
“The AI made a lot of mistakes, and I’m also learning how to better constrain dimensions in my prompts.”
— The maker, in the ToolGuyd report
The Prop Is Still Unfinished
The final appearance, name and function of the prop have not been settled. The maker says he is still trying to help his son choose a concept, and the son may decide what the finished object represents. It is also unclear which fittings will be used, whether the parts will be redesigned in conventional 3D software, and how much of the planned assembly will be completed before Halloween.
No completed prop or final print is reported. The maker says the current output is a test and that a low-density version of another part is being printed. He plans to make a rough version first, then refine it if time allows. The report does not provide a finished build, a confirmed deadline beyond the Halloween goal, or an assessment of how the AI-assisted process compares with his usual workflow.
A Rough Build Before Halloween
The immediate plan is to print and test the revised parts, then mount them to a small wooden box and add other creative elements, potentially including external wires. The maker may print a final-color version if the low-density test works, or revise the design manually in Fusion or similar software for greater control over dimensions and dependencies.
He says the priority is a rough first-version prop that he and his son can build together, with refinement afterward if there is enough time. Whether the printed pieces fit as intended, what the finished prop will be called and whether a more cohesive redesign will be completed remain open.
Key Questions
What is the maker building?
He is working with his son on a custom Fallout-inspired Halloween prop, planned as part of a costume with a wasteland-survivor look. Its final name and function have not been decided.
How is Gemini being used in the project?
The maker used Google Gemini to generate OpenSCAD code for printable shapes and to explore a possible breakdown of a larger tool into subassemblies. He adjusted the code’s parameters and says some generated designs needed correction.
Has the AI-generated design worked?
Some code produced a usable model for a test print, according to the maker. Other results did not follow instructions about part placement and holes, and the project has not yet reached a completed build.
What happens to the parts after printing?
The current plan is to attach separately designed components to a small wooden box, using mounting holes, pins, glue or other fasteners as appropriate. The maker may also revise parts in Fusion or similar 3D software.
When will the prop be finished?
The maker is aiming to have a rough version for Halloween, but gives no confirmed completion date. He says further refinement will depend on how the first assembly works and how much time remains.
Source: rss
