Create 3D-printing models with AI
Build or revise a dimensioned model, retaining editable sources alongside exported geometry, useful previews and an explicit account of what was checked.
Create 3D-printing models with AI around explicit dimensions and intended features. The print-3d skill chooses between CadQuery, Blender and OpenSCAD for mechanical parts, procedural forms or suitable existing models. Editable sources remain part of the handover, so later changes do not depend on reconstructing a design from its final mesh.
Checks concern the actual exported geometry, and previews should show that exported model. Units, dimensions and important holes or mating surfaces need comparison with the request. A closed mesh does not establish strength, wall thickness, physical fit, slicing success or print quality. The workflow does not operate a physical printer.
What it does
- Preserve editable CAD or scene sources.
- Compare exported dimensions with the design brief.
- Inspect features from useful preview angles.
- Retain an original mesh while making revisions.
- Deliver the requested STEP or mesh format.
- Record assumptions about units and fit clearances.
Install and use
- Give a file- and command-capable agent the complete package and an authorised workspace. Use Python 3.11 or newer in an isolated environment; the source tests Python 3.12. Prepare CadQuery, Blender or OpenSCAD according to the task. No mandatory MCP server or paid API is required.
- For example, request a cable holder with specified dimensions and hole positions, retaining its source and a view that exposes the holes.
- Agree units, features and output formats. Keep generated exports and preview scenes separate from editable originals, and approve any intended replacement of existing outputs.
- Review the geometry report and actual export previews. Critical fits need separate physical calibration or a test piece; report unperformed checks rather than treating them as passes.
FAQ
Source
Made by genhoi. Original skill: print-3d, published on GitHub under the MIT licence.
View on GitHub