Add export 3D printable STL option

[nicemit]

Feature Request

Problem

While the editor can export STL files, they are not production-ready for 3D printing. Current exports often have issues like:

• Non-manifold geometry (holes, inverted normals)
• Mesh integrity problems that cause printer slicing errors
• Improper orientation for print beds
• Suboptimal scaling for physical constraints

This limits users who want to create models, 3D print them, and showcase the physical results.

Related Issues

This builds on the broader import/export discussion in #145, but focuses specifically on making exports print-ready.

Solution

Add a dedicated "Export 3D Printable STL" option that generates valid, production-ready STL files with:

• Automatic mesh validation and repair (closing holes, fixing normals)
• Orientation and scale optimizations for printing
• Print bed placement suggestions
• Export validation to verify printability

Benefits

• Users can create models in the editor and directly 3D print them
• Removes the need for post-processing in slicer software
• Builds confidence in the tool for makers and designers
• Increases engagement and use cases for the platform
• Helps users showcase their work through physical prints

Interest in Contributing

I'm interested in contributing this feature to the project! I'm opening this discussion to gather feedback and inputs from the maintainers and community on:

• The technical approach and architecture
• Integration points with the current codebase
• Any existing utilities or libraries already in use for mesh validation/processing
• Best practices for implementing this feature within your development workflow

Looking forward to your thoughts and guidance on moving this forward!

[Aymericr]

Thanks for offering to contribute. The current STL command deliberately exports the cleaned scene geometry but does not claim printability, so this remains a useful separate track.

A production-friendly first PR should stay narrower than automatic repair: add a pure mesh-validation pass on the cloned export scene, report non-manifold edges, degenerate triangles, open boundaries, and final bounds/scale, and cover it with small wall, opening, slab, and intentionally broken fixtures. That gives us measurable acceptance criteria without silently changing architecture. Orientation, unit/scale selection, and opt-in repair can then follow once the validator proves what each transformation fixes. Please avoid mutating the live Three.js scene; the integration point is the prepared export clone used by ExportManager.

[nicemit]

Thanks for the detailed explanation and for pointing me in the right direction!

I completely understand the reasoning behind keeping this as a separate export path. I was thinking along the same lines, a dedicated "Export 3D Printable STL" option that performs the repair and printability checks without modifying the existing export pipeline or the live mesh.

I'll continue exploring the codebase and work on this feature. If I run into any questions or need clarification on the architecture or implementation details, I'll follow up on this discussion (or reach out through any other preferred channel).

Thanks again!

[Aymericr]

This shipped — #701 and #703, both merged 2026-08-21 — and it landed close to the validation-first shape I described above, so it's worth being precise about what you actually get.

There's one new action in Settings, Export 3D print files, alongside the unchanged GLB/STL/OBJ exports. It produces 3MF rather than STL: millimetre units, Z-up, split into per-level parts, architectural structure only, at 1:100. 3MF because it carries units and orientation in the file instead of relying on slicer conventions, which removes two of the four problems in your original list outright.

On the other two: shells are compiled with Manifold in a worker, and preflight reports disconnected components, winding, thickness, and unsupported geometry. Anything that fails blocks the download and shows you the diagnostic rather than handing you a file that dies in the slicer. That's validation, not automatic repair — the split you and I discussed above. Repair is still a reasonable follow-up now that the validator exists to prove what each transformation fixes.

The classic STL export is untouched and still makes no printability claim, which was the honest starting point of your report.

Closing as delivered. Two threads left open if you're still interested in contributing, and both are more valuable now that there's a baseline to measure against: opt-in mesh repair, and exposing the print-STL path (it exists internally as a format but isn't surfaced in the UI, since 3MF is the better default). Thanks for the original write-up — the framing of "exports exist but aren't print-ready" is what made this a separate track instead of a tweak.

[nicemit]

Thanks for the detailed update.
I’ve pulled the latest main and confirmed that the print-export work is already in place, including the printable 3MF export and its preflight diagnostics.
I’ll avoid duplicating that work. I’d like to take the smaller remaining UI task first: expose the existing internal print-stl export in Settings alongside the current 3MF option, while keeping 3MF as the default/recommended format. It will use the same existing print export profile and validation flow.
I’ll keep that as a narrow PR. Once that is complete, I can look separately at the opt-in mesh-repair follow-up.