Compress PNG

In the default Lossless mode, yes: the image is decoded to raw pixels and re-encoded as PNG, so every pixel value and every alpha value survives exactly. That mode discards nothing. If you switch the output format to JPEG or WebP with a quality setting, you are no longer doing lossless compression, and the tool labels it accordingly rather than pretending otherwise.

Yes, for PNG and WebP output — both formats carry a full alpha channel and it is copied across untouched. JPEG output cannot store transparency at all, so if you choose it, transparent pixels must be filled. The Flatten transparency option decides whether they become white or black; leaving it off and exporting to JPEG is what produces the black box people see around badly converted logos.

It depends entirely on the content. Screenshots, diagrams, logos and anything with large flat areas often drop 20–60% in lossless mode. A photograph saved as PNG will barely move, because there is almost no repetition for a lossless algorithm to exploit — for a photograph the real win is converting it to WebP or JPEG, which routinely cuts it by 80% or more.

Usually because it is a photograph in the wrong container. PNG stores every pixel exactly, so a 12-megapixel photo has to record twelve million distinct color values with no shortcuts. Screenshots are also larger than people expect when they are taken on a high-density display, since a screenshot of a 1440-point-wide window on a 2× screen is really 2880 pixels wide.

For the web, usually yes. Lossless WebP is typically 25–35% smaller than PNG for the same image with the same alpha channel and no visual difference whatsoever. Every browser released in the last several years supports it. Keep PNG when the file has to be opened by older desktop software, or when it is a source asset you will edit again.

For screenshots, almost always. A screenshot captured at 2880 pixels wide and displayed in an 800-pixel column is carrying more than three times the pixels it needs, and no encoder setting will fix that. Cap the dimensions first, then compress.

No. The file is read into memory by your browser, decoded onto a canvas, re-encoded, and handed straight back to you as a download. There is no server involved in the processing at any point, and no copy of your file exists outside this browser tab.

No. This tool does not perform color quantization or palette reduction, so it will not introduce the banding you sometimes see from tools that convert to 8-bit PNG. That means the savings in lossless mode are more modest than a quantizing tool would claim, but the output is genuinely identical to the input.

With the Strip metadata option on, text chunks, creation timestamps and embedded comments are not carried into the output, because the file is rebuilt from raw pixels. Screenshots in particular can carry the application name and window title, which is worth removing before publishing.

Yes. Add as many as you like — they are queued and processed in parallel across Web Workers, so the interface stays responsive. Each file gets its own row in the results table, and Download all packages the finished batch into a single ZIP.

In lossless mode, no — not at any zoom level, because no pixel has changed. If you convert to JPEG or WebP at a quality setting, then yes: sharp edges and text will soften slightly, which is exactly why lossless mode is the default here.

Because the original was already encoded by a very well-optimized tool, and a browser's PNG encoder does not always beat it. When that happens the tool reports the increase honestly instead of hiding it — and the sensible response is either to keep the original or to switch the output to WebP.