PNG Compressor

Reduce PNG image file size without losing transparency.

PNG is a lossless format: every pixel you put in comes back out exactly. That is why PNG files are large, and why compressing one works differently from compressing a JPEG — the size comes down by storing the same image more cleverly, not by discarding detail.

Where the size actually goes

A 24-bit PNG reserves room for over sixteen million possible colours per pixel, plus an alpha channel for transparency. Most PNGs never use anything close to that range. A screenshot of an interface, a logo, or a flat illustration might use a few hundred distinct colours in total, while still being stored as though it might use millions.

  • Colour quantization rebuilds the image from a smaller palette — often 256 colours — and stores an index per pixel instead of a full colour value.
  • On flat-colour artwork the visible difference is usually nil, while the file can drop by 60–80%.
  • On photographs, quantization is visible as banding in smooth gradients, which is why photographs belong in JPEG or WebP.

Transparency survives

The alpha channel is preserved through compression, including partial transparency such as soft shadows and anti-aliased edges. This is the single reason to keep an asset as PNG rather than converting it to JPEG: a logo that needs to sit on any background colour cannot be a JPEG at any file size, because JPEG has no way to record "this pixel is see-through".

PNG-8, PNG-24, and which one you have

PNG-24 stores full colour per pixel and is what most tools export by default. PNG-8 stores a palette index per pixel and is capped at 256 colours. Compressing a PNG usually means converting PNG-24 to PNG-8 with a palette chosen to fit the image. For screenshots, icons, and flat graphics that conversion is close to free; for anything with a photographic gradient it is not.

  • Screenshots, UI captures, diagrams → compress hard, the loss is invisible.
  • Logos and icons with few colours → compress hard, keep the transparency.
  • Photographs saved as PNG → convert to JPEG or WebP instead of compressing.

If the PNG is a photograph, change format instead

A photograph stored as PNG is often five to ten times larger than the same photograph as a high-quality JPEG, because a lossless format has to record every sensor-noise variation faithfully. If your PNG is a photo and does not need transparency, converting it will save far more than compressing it — the image converter handles that directly.

Frequently asked questions

Will my transparent background survive compression?
Yes. The alpha channel is preserved, including partial transparency such as soft shadows and anti-aliased edges, so a compressed logo still sits cleanly on any background.
Is PNG compression lossless?
PNG as a format is lossless, but meaningful size reductions usually come from reducing the colour palette, which is a lossy step. On flat-colour graphics the result is visually identical; on photographic gradients it can introduce banding.
Why is my PNG so much larger than a JPEG of the same photo?
Because PNG stores every pixel exactly, including the invisible sensor noise in a photograph, while JPEG is allowed to discard detail your eye will not miss. For photographs that difference is commonly five to ten times. PNG wins on flat artwork and on anything needing transparency.
What is the difference between PNG-8 and PNG-24?
PNG-24 stores a full colour value for every pixel. PNG-8 stores an index into a palette of at most 256 colours, which is much more compact. Compressing a PNG generally means converting it from PNG-24 to a well-chosen PNG-8 palette.
Can I compress a PNG that has already been compressed?
You can, but there is usually little left to gain — once the palette has been reduced, reducing it again mostly costs quality rather than bytes. If the file is still too large, resizing its dimensions is the more effective next step.

What to do next

This tool is free. When you need more than file housekeeping, these pick up where it stops.

Bulk Image Upscale Compressed a photo that was already small? Upscaling rebuilds detail instead of stretching the pixels you have left.Background Removal Cutting the subject out drops everything behind it — often a bigger saving than compression, with a cleaner result.AI Image Editor Fix the image itself — lighting, distractions, unwanted objects — instead of only making the file smaller.