The SVG engine
Prompt to SVG, or image to SVG. Real editable paths, not a bitmap in an SVG wrapper.
What it does
Two inputs, one engine, one output shape.
- Prompt to SVG. Describe a shape in plain words. You get vector markup.
- Image to SVG. Upload a PNG or JPG. You get vector markup.
Both return the same thing: an SVG document made of real paths, plus the files people actually ask for afterwards (PNG, PDF, EPS, DXF, an icon pack, and a simplified version that still reads at 16px).
The studio, the REST API, the MCP tools and the x402 endpoints all give the same file at the same price, whichever way you come in.
What makes it different
Most converters wrap a bitmap in an <svg> tag, or leave the seams showing.
What you get here instead:
| Common problem elsewhere | What you get with AMNT |
|---|---|
| Wide art squashed into a square | Your picture keeps its own shape |
| A pale ring of the old background round the outline | Clean edges, no ring |
| Thin lines of background showing between shapes | Shapes that meet with no gaps |
| Thin grey slivers between colours | Solid, clean colours |
| Wobbly curves | Smooth curves |
| 200 unnamed paths in no useful order | Layers, grouped by colour or by object, each one named and movable |
The result opens in Figma as objects you can click, move and recolour.
A photograph will never be pixel-exact as vector. Flat art, logos, icons and illustration come out clean. Photos come back as an illustration look, and that is the honest description of what you are buying.