Free Online Image Compressor – Reduce Image File Size Without Losing Quality

Compress JPG, PNG, WebP, AVIF, GIF and more directly in your browser. No upload to server, no sign-up, completely free.

What Compression Actually Throws Away

Lossy compression does not shrink an image by making it smaller. It shrinks it by deciding which information you will not miss and discarding that information permanently.

Two properties of human vision make this possible. We notice changes in brightness far more than changes in colour, so colour information can be stored at lower resolution than brightness with almost no visible effect. And we are poor at detecting fine detail in busy areas, so texture can be simplified where a lot is already happening.

This is why compression works so well on photographs and so badly on screenshots. A photograph is mostly busy texture and gradual colour; a screenshot is mostly flat areas and hard edges, which is exactly what these shortcuts handle worst.

Quality Settings Have Sharply Diminishing Returns

The relationship between the quality setting and the file size is not linear, and the useful range is narrower than the slider suggests.

The step from 100 to 90 usually removes a large share of the file size for no perceptible change. The step from 90 to 80 removes rather less and is still generally invisible. Below about 70 each further reduction costs visible quality for progressively smaller savings, which is the point at which resizing becomes the better tool.

Resize Before You Compress

The most common reason an image is too large is not that it is insufficiently compressed but that it contains far more pixels than will ever be displayed.

A photograph straight from a phone might be 4000 pixels wide. Displayed in a 800-pixel column, three quarters of that width is discarded by the browser at render time — after the visitor has downloaded all of it. Resizing to roughly twice the display width, which covers high-density screens, typically removes far more file size than any quality setting can, and it costs nothing visible.

The order that works is resize first, then compress. Compressing a 4000-pixel image aggressively produces a file that is both large and visibly degraded, which is the worst of both.

Our image resizer handles the first step, and the image converter covers choosing the format, which also affects size substantially.

Compression Damage Accumulates

Each time a lossy image is saved, compression is applied again — to a file that already contains the artefacts of the previous pass. The algorithm treats those artefacts as real detail and compresses them along with everything else.

Practically, this means an image edited and re-saved several times degrades noticeably even if every individual save used a high quality setting. It also means a file received from someone else has already been compressed an unknown number of times, so compressing it again starts from a worse position than the file size suggests.

Keep an untouched original wherever possible, and produce compressed versions from it rather than from the last compressed version.

What Compression Cannot Fix

• An image with the wrong dimensions. Compression reduces file size at the same pixel count. If the pixel count is the problem, resizing is the answer.
• Text and screenshots. Hard edges are what lossy compression handles worst. Use a lossless format instead of a lower quality setting.
• Too many images on one page. Twenty well-compressed photographs still add up. Lazy loading and showing fewer images at once do more than another 10% off each file.
• An image that was already low quality. Compression cannot recover detail, and applying it again only removes more.

Judging the Result Properly

Compare the compressed image against the original at the size it will actually be displayed, not zoomed in. Artefacts that are obvious at 400% magnification are frequently invisible at normal size, and optimising against a magnified view produces needlessly large files.

Where an image will be viewed closely — product photography, artwork, anything a reader will zoom into — apply the opposite reasoning and keep the quality high. The right setting depends on how the image will be looked at, which is a judgement no default can make for you.

Compression here runs entirely in your browser, so nothing is uploaded and the original file never leaves your device.