Writing
How to Compress Images for the Web
A practical guide to quality, dimensions, formats, and the trade-offs that affect page speed.
Writing
A practical guide to quality, dimensions, formats, and the trade-offs that affect page speed.
Images are usually the largest files on a page. Compressing them well improves load time, mobile data use, and Core Web Vitals without making the page look worse.
A 4000-pixel photo displayed at 800 pixels wastes bandwidth. Resize to the largest size the layout actually needs, then compress. A two-column layout rarely needs more than 1200 pixels; a full-width hero may need 1600–2000 pixels. Retina screens benefit from up to twice the CSS size, but not from a 6000-pixel original.
This site's Image Compressor re-encodes to JPEG at the original dimensions. Use the Image Resizer first if the dimensions are too large.
The quality slider trades file size for detail. For photos, 70–80% is often visually indistinguishable from the original at normal viewing size. Text, sharp edges, and gradients need more quality; a screenshot at 60% can look muddy. Always compare at 100% before publishing.
JPEG does not support transparency. A transparent PNG saved as JPEG may show a black, white, or coloured background. Keep PNG or WebP when the alpha channel matters.
Check the compressed file size, not just the preview. A page with ten images at 300 KB each is 3 MB before any scripts or fonts load. Set a budget: for example, 150–250 KB per content image and under 100 KB for thumbnails.
Each JPEG re-encode loses a little quality. Keep the original, compress once for the web, and regenerate from the original if you need a different size. Repeatedly editing and saving a JPEG adds artifacts that no later setting can remove.
The compressor runs in your browser, so the image is not uploaded. That matters for screenshots, client work, and personal photos.