Choose the right compression for your assets
Reduce asset sizes for websites, product listings and ads. Default lossless mode preserves pixel content and format; choose lossy quality or target size for stronger reduction and compare the result.
How to use
Add images by choosing, dropping or pasting, then inspect formats, dimensions and sizes.
Choose lossless optimization, or set quality/target size, output format and JPG fill.
Compare actual size, quality and target status; download images individually or as ZIP.
Specifications and data handling
Up to 50 images, 20 MB each and 100 MB total; 16 MP and 8192 px per side. Outputs are limited to 32 MB each and 150 MB total. Local processing without uploads or persistence; leaving releases images.
Lossless mode preserves format, alpha and animation. Quality/size re-encoding exports only the default image or first frame and removes original metadata. Keeping PNG still uses the lossless path.
Use cases
Optimize site banners, product images and ad bundles; try target sizes for upload limits; choose lossless for alpha, pixels or animation.
FAQ
How does lossless differ from quality compression?
PNG lossless optimization restructures compression while preserving decoded pixels; JPG only clears optional text comments. Quality/size modes re-encode JPG / WebP and may lose detail. High quality is still not lossless.
Why do some files stay the same size?
Files may already be optimized or lack safely removable data. Lossless never outputs a larger file. Smart/original output can keep originals when not smaller, with an explicit label.
Can every image meet the target KB?
No. At fixed dimensions, quality 10–95 is searched for a higher-quality result within target. If even the lowest quality is too large, a smaller candidate is returned and marked unmet. PNG kept in its format is lossless only. Try reducing resolution.
Are transparency, animation and orientation preserved?
Lossless preserves them; optional comment removal keeps EXIF orientation and color profiles. Lossy exports static frames: WebP keeps alpha, JPG uses a fill. Source orientation is applied by the decoder; original metadata is not retained.