What image compression actually does
Image compression works by finding and removing information in a way that's least noticeable to the human eye. For JPG-style compression, this typically means simplifying subtle color gradients and slightly blurring fine detail that the eye is less sensitive to, especially in areas of the image that aren't the main subject.
This is fundamentally a trade: lower the "quality" setting and the file gets smaller, but at some point the simplifications become visible as blockiness, blurring, or color banding. The goal of good compression isn't to hit the smallest possible file size. It's to find the point where the file is meaningfully smaller but the visible quality loss is still acceptable for how the image will be used.
Finding the right quality setting for your use case
The "right" compression level depends entirely on where the image is going. A photo displayed at full size on a high-resolution monitor needs more quality headroom than a thumbnail that will be shown at 150 pixels wide. A product photo customers will zoom into needs more care than a background texture nobody will look at closely.
As a rough starting point: 80-90% quality is usually visually indistinguishable from the original for most photos, while cutting file size substantially. Below 60%, artifacts typically start becoming noticeable, especially in images with text, sharp edges, or large areas of flat color. There's no universal "correct" number. The only real test is looking at the compressed result at the size and context it'll actually be viewed in.
How to compress an image with this tool
- Drag and drop your image, or click to browse and select a file.
- Use the quality slider to adjust the compression level. The preview updates so you can see the visual result at each setting.
- Watch the file size readout update live as you adjust the slider, so you can balance visual quality against the size you need.
- Once you're happy with the result, download the compressed file.
Everything happens locally in your browser via canvas re-encoding. Your image is never uploaded anywhere, which matters if you're compressing anything you'd rather not send to a third-party server.
Common mistakes when compressing images
The most common mistake is compressing an image multiple times in a row across different tools or sessions. Each round of lossy compression compounds the quality loss, so an image that's been compressed three separate times will look noticeably worse than one compressed once at a lower quality setting. If you need to compress further, it's better to start over from the original, uncompressed source if you still have it.
Another common issue is compressing an image that's already been resized down aggressively. Very small images have little detail left to compress, so aggressive compression on a tiny image tends to show artifacts much faster than the same setting on a larger one. If both resizing and compressing are needed, resizing first usually gives a cleaner result.