Also available in: Español · Português · Français · العربية
PNG to WebP converter
Convert a PNG to WebP in your browser, and find out whether the result is actually worth keeping.
What is a PNG to WebP converter?
PNG and WebP are both image formats that a browser can display, and WebP is the newer of the two. PNG stores pixels exactly, so what you saved is what you get back forever. WebP can do that too, but it can also compress the way JPEG does, throwing away detail the eye is unlikely to miss in exchange for a much smaller file.
A converter re-encodes the picture from one to the other. This one does it in your browser using the same canvas that any web page has, which means your image is never uploaded and there is no queue, no size limit imposed by a server and no copy of your work anywhere but on your own machine.
The usual reason people convert is to make pages faster, and the usual advice is that WebP is simply smaller. That is true often enough to be worth trying and false often enough to be worth checking, which is the part this tool is built around.
How to use it
- Choose a PNG. The file is read locally and its header is parsed, so the tool knows the real dimensions and whether the image is truecolour, greyscale, a palette or carries transparency. That is checked from the bytes rather than from the file name, so a mislabelled file is caught straight away.
- Pick a quality and convert. Lower quality means a smaller file and more visible damage. There is no lossless setting and the reason is explained below, so treat every conversion here as a trade rather than a format change.
- Read the comparison before you download. The result panel shows both sizes and says which file is smaller. When the WebP came out bigger, the download button is not offered and the tool tells you to keep your PNG, because at that point the conversion has cost you pixels and bytes at once.
Why there is no lossless option
WebP has a lossless mode, and it is very good. A browser cannot reach it. The HTML standard gives the canvas one quality argument, says it applies to an image format that supports variable quality, and specifies no mechanism at all for requesting lossless encoding. There is no flag to set and no workaround, so every converter that runs in a browser tab, this one included, is turning a lossless source into a lossy copy.
That is worth stating plainly because the expectation runs the other way. PNG is lossless, so converting it to another modern format sounds like a change of container rather than a loss of information. It is not. Even at the highest quality this tool offers, most of the pixels in a typical image come back slightly different from the ones you put in.
If you need lossless WebP, it needs a command-line encoder or a build step rather than a web page. That is a real limitation of the approach rather than of this particular tool, and it is the one thing worth knowing before you replace an original with a converted copy.
When WebP comes out bigger
Lossy compression is cheap on smooth gradients and expensive on hard edges. PNG works the other way round: it looks for runs and repetition, so large areas of flat colour cost it almost nothing. Put those two facts together and the outcome depends entirely on what the picture is of.
On photographs, WebP wins comfortably and it is not close. On flat graphics, screenshots, diagrams and logos with transparency, it can lose, and it can lose badly. We built four images that isolate each case and measured them: a synthetic photograph went from 562 KB of PNG to 191 KB of WebP at quality 90, while a synthetic screenshot went from 5 KB of PNG to 68 KB of WebP. That screenshot figure is a property of the example rather than a rate you should expect, but the direction is a property of the formats and it will show up on your own images.
The same four images tell the other half of the story. Lossless WebP beat PNG on every single one, including the screenshot, where it was smaller by a factor of forty. So WebP really is the better format for this content, and the browser simply cannot produce the version that proves it.
This is why the tool compares rather than predicts. The header gives a hint before you convert, and it is only a hint: a palette PNG or one with an alpha channel is the kind that tends to go badly. The number that matters is measured afterwards, on your image, and if it says the PNG was smaller then the PNG was smaller.
Honest limits
The conversion is done by your browser's own WebP encoder, so the exact bytes depend on which browser you are using and which version. Two people converting the same PNG at the same quality can get slightly different files. The comparison is still valid, because it measures whatever your browser actually produced.
Animated PNGs are flattened to their first frame. The canvas holds a single still image and has no notion of a sequence, so anything that moves does not survive.
Colour management is the subtler loss. An ICC profile attached to your PNG does not travel through the canvas, so a file with a wide-gamut or otherwise unusual profile can come out looking different even before the lossy compression has done anything. For ordinary sRGB images this makes no visible difference, which covers almost everything on the web.
There is also a ceiling on how large an image the browser will hold, and it is lower on mobile than on desktop. A very large PNG can fail to convert rather than convert badly, and the tool says so rather than handing you a truncated result.
Why is it free?
Because your own computer does the work. The file is read locally, decoded by your browser, re-encoded by the same canvas any page can use, and handed back to you. There is no upload, no server paying for bandwidth and no copy of your image anywhere else.
So there is no account, no sign-up, no watermark and no cap beyond what your browser can hold. The engine and the 70 checks that verify it sit in the repository beside the page, along with the script that produced the measurements quoted above.