Prerequisites, six steps from upload to download, the failures that account for most bad meshes, and what to try once it works.
Convert an image to STLThis page covers how to convert image to STL from first file to finished mesh: what the converter can read, what it needs from you, and how to tell a good result from a bad one.
The workflow is the same whatever converter you use: shape the input, set the depth, generate the mesh, then check it before printing.
Converters trace edges, not objects, so a high-contrast subject on a plain background gives a clean outline. Gradients, soft shadows and textured backdrops turn into ragged walls and blobby fill. Flat logos, icons, silhouettes, stencils and solid lettering all convert well; a busy photograph usually does not.
Trim the file to the subject so the converter does not waste resolution on empty background. Raise contrast, delete the background, and close small holes inside the shape so you do not get stray islands in the mesh. Export as PNG rather than JPEG if you can, because JPEG compression leaves speckle along hard edges.
Most tools expose a threshold or brightness cutoff, an invert switch, a thickness or depth, and a smoothing value. Move the threshold until the preview shows the shape you actually want, and invert it if the converter picked the background instead of the subject. Depth is the real height of the part in millimetres, and 2mm to 4mm is a sensible starting range for a flat sign.
Upload the file, up to 20MB here, press convert and wait for the mesh to build. Large images with fine detail take longer, because every pixel edge becomes geometry. The download is an STL, which stores geometry only, so there is no colour and no texture in the file.
Open the STL in a slicer before you print anything. Check that it is one closed solid, that no hairline spikes hang off the outline, and that the dimensions in millimetres match what you planned. Most slicers flag non-manifold errors and offer a repair pass you can run before slicing.
Pick a layer height from the detail you need: 0.2mm is fine for a flat sign or coaster, while 0.12mm or lower suits fine relief and small lettering. Add a brim if the part is tall and thin, and use two or three perimeters so thin outlines do not print as gaps. Scale in the slicer if your plate is smaller than the exported model.
Almost every bad result comes from one of these six causes, and each has a straightforward correction.
| Symptom | Cause | Fix |
|---|---|---|
| Nothing happens after upload | Image is above the 20MB upload limit | Resize the image or save a smaller JPEG |
| The STL is a flat rectangle | Whole frame converted as one relief block | Switch to silhouette or threshold mode and re-convert |
| Fine detail turns into blobs | Source resolution too low and threshold too aggressive | Start from a larger image and lower the threshold |
| The part is solid with no holes | The shape interior was filled instead of the background | Invert the mask, or fill the background rather than the subject |
| Slicer reports a non-manifold mesh | Stray pixels and thin spikes create self-intersections | Smooth the outline, delete specks, run the slicer repair tool |
| The print comes out far too large | Model built at pixel scale, not millimetres | Set the width in mm before converting, or scale in the slicer |
Work through these in order and you will fix most files in a couple of attempts. If two or three passes still give a messy outline, the problem is in the source image rather than the converter: rebuild the graphic with hard edges, a transparent background and a single solid fill, then run it again.
Three things worth trying when a straight silhouette conversion is no longer enough.
Instead of a flat extrusion, feed the converter a grayscale copy where darker areas end up deeper and lighter areas stay shallow. The result is a relief panel that follows the tonal range of the image. How well it reads depends entirely on how much contrast the source has.
If you convert a set of icons, letters or tokens one at a time, write down the depth and base height you used for each. Identical numbers in give identical parts out, so the pieces sit flush against each other when you print them side by side.
STL stores geometry only, so a two-colour print means two files. Export the raised shape and the base as separate meshes, then assign each one to a different extruder or a filament change in the slicer.
Load your JPG, PNG, SVG or HEIC, set the threshold and a depth in millimetres, then download the mesh. There is nothing to install, and the same page runs on a desktop, a laptop, a Chromebook or a phone.
Yes, as long as the image holds shape information the converter can trace. A silhouette becomes an extruded outline, and a grayscale image can become a height map where brightness sets the depth. A photograph with no clean edges will not turn into a recognisable object on its own.
The pages that answer the next question after this one.
Skip the complex modeling software. Upload your image now and get a print-ready STL file instantly.
Start Converting Free