Understand distance.
Depth Anything estimates relative spatial depth from a single photograph, directly on your device.
An image holds an entire world. Touchscape uses on-device vision models to translate its shapes, depth, and meaning into a tactile-inspired draft you can refine for specialist printing.
Upload a scene. Let the AI discover its layers. Shape its tactile translation, then export a vector prototype.
Experimental vector draft — not validated tactile graphics. Depth is relative, segmentation can misidentify objects, and true tactile production needs specialist review, sizing, and user testing. The Unicode braille legend is a visual representation, not an embossing instruction.
Two neural networks feed a deterministic geometry engine. Every contour in your output comes from image-derived structure, not decorative effects.
Depth Anything estimates relative spatial depth from a single photograph, directly on your device.
SegFormer maps semantic regions, allowing the tactile result to retain meaningful boundaries and a navigable legend.
A custom marching-squares engine extracts vector contours, simplifies lines, adds tactile patterns, and builds SVG.
Photographs, learning materials, and personal memories are overwhelmingly visual. Tactile diagrams are often expensive and labor-intensive to prepare. Touchscape explores a new first step: let AI generate an editable tactile draft, then let people refine it together.
Create a tactile draft ↗