Skip to content
Initurn

Convert PNG to SVG

Files never leave your device. All processing happens in your browser.

SVG has no transparency, so see-through areas are filled with this colour.

How to convert PNG to SVG

  1. 1

    Drop in a logo, icon or piece of line art.

  2. 2

    The image is traced into coloured paths — expect this to take a moment.

  3. 3

    Download the SVG and tidy it up in a vector editor.

This one needs a warning before anything else: tracing is approximate, and it is the only conversion on this site that is. Every other tool here reproduces your image faithfully. This one looks at a grid of pixels, guesses where the shapes are, and draws new paths around them. The output resembles the input. It does not reproduce it.

Why a real conversion is impossible

A PNG contains no shapes. It contains coloured pixels, and whatever curves and strokes the designer originally drew were thrown away the moment the file was exported. Tracing is reconstruction from evidence, not decoding. When a tool claims to convert a bitmap to vector, this is what it is doing, however it is described.

What traces well

  • Flat logos with a handful of solid colours and clean edges. This is the case tracing was designed for and it works genuinely well.
  • Black-and-white line art, signatures, stamps and simple diagrams.
  • Icons, silhouettes and bold lettering, especially at higher resolutions.
  • Screenshots of simple shapes, though text usually comes out lumpy.

What does not

Photographs. A traced photograph becomes thousands of overlapping coloured blobs — a file many times larger than the original that looks like a bad watercolour of it. There is no setting that fixes this, because the problem is conceptual: photographs have no shape boundaries to find. Anything with gradients, soft shadows, blur or fine texture has the same problem to a lesser degree.

JPEG artifacts are a related trap. If your source has been compressed hard, the tracer faithfully finds the edges of the compression blocks and turns each one into a path. Starting from the cleanest version you have makes a visible difference.

Treat the output as a starting point

Open the result in Inkscape, Illustrator or Figma and expect to do some work. Traced paths carry far more anchor points than a human would draw, curves are approximated by many short segments, and colours are quantised into flat bands. For a logo you are recovering because the original file is lost, that is a reasonable trade. For anything else, redrawing from scratch is usually faster and always cleaner.

Large images are scaled down before tracing, because cost climbs steeply with pixel count and a bigger input mostly produces a bigger file rather than a better one. Tracing runs in a background thread on your own device, so the page stays responsive and nothing is uploaded.

Questions about PNG to SVG

Will the SVG look exactly like my PNG?

No. Tracing approximates the image with coloured paths, so edges shift slightly and colours are quantised into flat bands. On flat logos the difference is small; on anything with gradients or texture it is obvious.

Can I vectorise a photograph?

You can run it, but you should not. A traced photograph becomes thousands of overlapping blobs, producing a file much larger than the original that looks like a poor painting of it. Photographs have no shape boundaries for a tracer to find.

Why is the SVG bigger than the PNG?

Because every colour region becomes a path with many anchor points. Detailed or noisy images produce enormous numbers of paths. Flat artwork with few colours produces small files, which is the case tracing suits.

Can I get my original vector file back this way?

No. The original curves were discarded when the PNG was exported and nothing can recover them. Tracing produces new paths that happen to resemble them, which is genuinely useful when the source file is lost but is not the same thing.

The edges look lumpy. Can I improve them?

Start from the largest, cleanest version of the image you have — an uncompressed PNG rather than a heavily compressed JPG, at higher resolution. Then smooth and simplify the paths in a vector editor afterwards.

Why does it take so long?

Tracing examines every pixel, groups them into colour regions and fits curves around each one. That is far more work than re-encoding an image, and it grows quickly with size, which is why large inputs are scaled down first.

Related converters