Reglyph reads a screenshot, exported PNG, or flattened PDF and recovers what's actually inside it — text, shapes, layout — then rebuilds it as clean vector geometry you can open and edit.
Most "image to vector" tools just draw outlines around pixels. Reglyph reconstructs what the image was made of, so the file behaves like it was designed, not scanned.
Text blocks return as actual type layers with a matched font, size, and alignment — not a raster patch pretending to be text.
Frames, containers, and spacing come back as a real grid with parent-child relationships, so resizing one thing doesn't break everything else.
Flat icons and simple illustrations are traced back into paths with proper anchor points — ready to recolor, scale, or restyle.
Four stages, run automatically on every upload.
Drop in a screenshot, exported PNG, or flattened PDF page — anything where the original file is gone.
A vision model reads the image and infers what's underneath it: text blocks, containers, shapes, and stacking order.
That structure is rebuilt as real vector geometry — paths, anchors, and live text, not a pixel-perfect trace.
Download a clean SVG, or hand it straight to your design tool of choice.
If you've ever had to rebuild a layout by eye from a screenshot, this is for you.