PNG to SVG.
Drop a PNG and get a real vector SVG, not a bitmap wrapped in SVG code. Colours and transparency are kept. Free, nothing is uploaded, no sign-up.
Reading image
How to convert a PNG to SVG
- Drop your PNG. Logos, icons and flat illustrations give the cleanest result. Photos turn into posterised blobs, because a vector needs flat colour regions.
- Watch it trace. The colours are found, the shapes are traced into curves with sharp corners kept, and the SVG is compared with your PNG until the match stops improving. Transparent pixels stay transparent.
- Download. You get a plain SVG with one path per colour. Open it in Figma, Illustrator, Inkscape or a browser. Use Colours to cut the palette down and Detail to trade smoothness for fidelity.
Questions
Is it really free?
Yes. No account, no watermark, no download limit. The site is paid for by ads.
Are my images uploaded?
No. Conversion runs on your device, inside this page. Your image never leaves your browser, so it also works with confidential files.
Does a transparent PNG stay transparent?
Yes. Transparent areas stay empty and semi-transparent edges get a matching opacity. No white box is added behind the shapes.
Is this a real vector or a PNG embedded in SVG?
A real vector. Some online converters only wrap the PNG in an SVG tag, which cannot be edited or scaled cleanly. Every shape here is a path.
Why does my PNG come out blocky or with too few colours?
Raise Detail and Colours. Photos and heavy gradients are the hard case; a logo with flat colours is the easy one.
I have a JPG, not a PNG.
Use the JPG to SVG converter. For a vinyl or paper cutter see SVG for Cricut.
What does the match score mean?
It compares the rendered SVG with your image on colour, transparency and edges. 100% would be pixel identical. Flat artwork usually lands above 99%; photos trace into many shapes and score lower, because vectors are not made for photos.
What do the settings do?
Type picks what kind of image it is, which sets sensible defaults. Colours sets how many colours the image is split into (Auto picks them from the image and adds more where needed). Detail sets how closely curves follow the pixels: higher keeps small features, lower gives smoother shapes and smaller files.