How to use Claude to print Halloween light holders
Claude generated parametric Halloween holders with build123d while Charles McGuinness revised small stencil details, reviewed STL models and printed them.
Original prompt
Shared by the source author. See source notes for context.
For the jack-lantern part, I want to use the image in the inspiration for the stencil
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Charles McGuinness
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author printed the repository's holder examples after stencil revision, parametric modeling and slicer review.
- Human involvement
- McGuinness revised stencil images for printability, reviewed model previews and checked the STL in a slicer before printing.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- McGuinness supplied a stencil image and existing parametric holder examples.
- Claude generated and revised build123d code and showed the model for review.
- The human checked the STL in a slicer and printed the finished holder.
Charles McGuinness used Claude Code and Python's build123d library to produce physical Halloween tea-light holders. The workflow grew from simple electronics cases and heart-pattern holders into decorations with repeated stencil cutouts.
Claude generated parametric code instead of relying on a direct text-to-mesh service. That let later revisions change an opening or other dimension incrementally. McGuinness supplied printer constraints and examples through his cadforge project, then asked the agent to build a holder using an image stored in the part's inspiration folder.
The stencil required separate judgment. ChatGPT produced candidate artwork, but details that looked good on screen disappeared when reduced to a small plastic cutout. McGuinness revised the design across several rounds and checked that internal shapes remained connected rather than becoming unsupported islands.
Claude showed a model preview, generated the STL after review and left a final inspection in the slicer before printing. The post says every example shared in the repository had actually been printed. The result is a reusable parametric workflow and finished decorations, with trial-and-error printability checks still necessary; a plausible digital model alone was not sufficient.
Prompt
Original prompt
Source: Substack · Charles McGuinness
Frequently asked questions
Why use code for the model?
Parametric code allowed incremental dimensional changes rather than full regeneration.
What went wrong with stencil details?
Some details disappeared at the small printed scale, requiring revised artwork.
Were the shared examples printed?
The author states that every repository example had been physically printed.
More details
About this account
Claude generated parametric code instead of relying on a direct text-to-mesh service. That let later revisions change an opening or other dimension incrementally. McGuinness supplied printer constraints and examples through his cadforge project, then asked the agent to build a holder using an image stored in the part's inspiration folder.
Primary account opened. Original prompt excerpt is from the author's demonstrated workflow. Shared repository examples were reported printed. Stencil detail loss and iterative fit remain limitations. Original prompt excerpt retained from the source.
Credit: Charles McGuinness on Substack
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Code · build123d · Python · ChatGPT · 3D printer
3D printing · Parametric models
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.