SliceFoundry's image to STL tool turns a single reference picture into a printable 3D model. Upload a photo, drawing, screenshot, or product image, optionally add a description, and the AI builds a 3D model that matches the silhouette, proportions, and major features of the image — then exports it as a binary STL ready for any FDM slicer.
Use it when you have a clear visual idea of what you want, but no CAD experience, no 3D scan, and no Thingiverse model that fits. The image carries the intent, the description fine-tunes it, and SliceFoundry handles the geometry.
This isn't photogrammetry — you don't need dozens of photos around an object, calibration, or a turntable. Instead, SliceFoundry treats the image as the primary visual target for an AI model that emits OpenSCAD-style geometry. The AI matches the image's silhouette, part-to-part proportions, and clearly visible features, while you provide any extra constraints in plain language (target dimensions, mounting holes, wall thickness, what it attaches to). The result is a single printable STL you can preview, iterate on, and download.
3D scanning and photogrammetry both reconstruct existing objects from many images and produce a mesh that matches the original closely. They're great when you have the physical part in hand and want a near-exact digital copy.
Image-to-STL solves a different problem: you have a visual reference of what you want (a photo, a sketch, a screenshot, a product image) and you want a printable, parametric, easily-editable 3D model that captures the same intent. You can't 3D scan a sketch. You can't photogram a single product photo. Image-to-STL fills that gap.
Once SliceFoundry generates the STL, you can either slice it directly or generate a matching slicer profile for your specific printer and filament. The export is a clean, FDM-friendly STL — manifold geometry, sensible wall thickness, and printable proportions — not a polygonal mesh exported from a photo-realistic renderer.
If your first print struggles, use SliceFoundry's Fix My Print assistant to diagnose the issue and correct your slicer settings, then iterate on the model itself if the geometry needs an adjustment.
Yes — for the right kinds of objects. Image-to-STL works best on functional, geometric parts with clear silhouettes and a small number of distinguishable features. It won't reproduce a photo-realistic statue from a snapshot, but it will turn a product photo of a bracket, a knob, a holder, or a 3D-printed-style prop into a printable starting point you can refine.
No. Image-to-STL works from a single image. The AI infers the underlying 3D shape from the visible silhouette and clear features. If a critical feature is hidden in your photo, mention it in the description.
JPG and PNG are the most common, and both work. Any standard web image format the browser can display is fine. SliceFoundry normalises the upload before sending it to the AI, so you don't have to resize or crop manually.
Yes. Sketches, line drawings, and digital concepts work as long as the major shapes are recognisable. Add a description to clarify proportions and features that aren't obvious from a 2D drawing.
Text to STL generates a model purely from a written description. Image-to-STL anchors the design in a visual reference first, then layers the description on top. If your idea is easier to draw or photograph than describe, image-to-STL is usually the better fit.
Anything an FDM 3D printer can produce — replacement parts, brackets, mounts, enclosures, knobs, holders, props, ornaments, jigs, and fixtures. The STL is slicer-agnostic and works in Cura, OrcaSlicer, PrusaSlicer, Bambu Studio, SuperSlicer, and any other modern slicer.
Open SliceFoundry, upload your reference image, and convert it to a printable STL →