Turning pixel art into a 3D model

The short version: draw the sprite, paint depth onto it, export a GLB. pixemboss does this entirely in the browser - there is nothing to install and no account is needed to get to a finished file.

Draw the shape first, worry about depth later

Start on the 2D canvas with the Pencil, Line, Rectangle, Ellipse and Fill tools. Work at a small size - 16×16 or 32×32 is plenty, and a smaller grid means fewer, larger faces in the finished mesh.

A flat, readable silhouette matters more than detail. The extrusion step turns every filled pixel into a column of geometry, so a shape that reads clearly in 2D reads clearly in 3D.

Paint depth onto the pixels

Switch to Depth paint and brush the areas that should stand further forward. Each pixel carries its own depth value, and that value becomes the thickness of its column when the drawing is extruded.

The Auto depth panel can seed the whole map for you - From brightness uses each pixel's luminance, From color index uses its palette slot, and From noise roughens a flat surface. Seed with one of those, then brush corrections by hand.

Export a model

Open Export and pick a target. The targets are presets over the same pipeline, so you can also start from one and adjust the advanced controls:

Colour modes decide how the model is textured

Under the advanced controls, Color mode changes what an engine receives. Vertex colors bakes the palette into the mesh and needs no texture at all - the flattest, lightest option. Palette texture writes a tiny palette image, which is what you want if you plan to swap colours later. Unwrapped UVs gives you a real UV layout to paint on.