Compress JPG, PNG, and WebP images in your browser. Adjustable quality, instant download, 100% private. No upload required. 100% free and private.
An image compressor reduces the file size of a photo or graphic by re-encoding it with a more efficient compression algorithm or by lowering the quality of the pixel data. Smaller image files load faster on websites, fit inside email attachment limits, and use less storage on phones and cloud drives. This tool compresses JPG, PNG, and WebP images directly in your browser using the HTML Canvas API.
JPG and WebP use lossy compression — they discard fine detail the human eye is unlikely to notice, which is how they reach dramatic size reductions. PNG uses lossless compression — every pixel is preserved exactly, so file sizes are larger but quality is identical to the original. If your image has no transparency, converting PNG to JPG or WebP usually shrinks it by 60–80%.
WebP produces files 25–35% smaller than JPG at the same visual quality, supports both lossy and lossless modes, and handles transparency like PNG. Every modern browser (Chrome, Firefox, Safari, Edge) supports WebP. If your audience uses up-to-date browsers, switching to WebP is the single biggest win for page speed.
This compressor runs entirely in your browser. Your image is decoded, re-encoded, and downloaded without ever being uploaded to a server. There is no tracking, no account, and no file expiry — close the tab and nothing is retained.
Need to change dimensions instead? Try the Image Resizer. Switching formats? Use the JPG to PNG, PNG to JPG, or Image to WebP converters.
Compression quality is the second lever, not the first. A 4000-pixel-wide phone photo displayed inside a 800-pixel column is carrying five times more pixel data than the screen can ever show. Lowering quality on that oversized image degrades what you see while leaving most of the waste intact. Resizing the dimensions to what is actually displayed typically cuts 80–90% of the bytes before compression does anything at all.
Practical target widths: 1600–2000 px for a full-width hero image, 800–1200 px for in-article images, 400 px for card thumbnails, 200 px for avatars. Double those figures only if you are serving high-density (retina) screens with a dedicated 2x asset. Use the Image Resizer for the dimension step, then compress here.
JPG was designed for continuous-tone photographs. It transforms 8x8 pixel blocks into frequency coefficients and throws away the high-frequency detail your eye is least sensitive to. Hard edges — text, UI borders, thin lines, flat colour regions — are pure high-frequency information, so JPG handles them badly: you get visible halos and mosquito noise around letters even at quality 85.
Rules that follow directly from this: keep screenshots, diagrams, charts and logos as PNG or WebP; use JPG for photographs of people, landscapes and products. A screenshot saved as PNG is often smaller than the same screenshot as JPG, because PNG's lossless algorithm compresses large runs of identical colour extremely efficiently.
Compare images at 100% zoom, not fitted to the window — browser downscaling hides artifacts. Look specifically at the places compression fails first: smooth gradients such as sky or studio backdrops (look for banding and blotching), areas of fine repeating texture such as hair, fabric weave and foliage (look for smearing), and high-contrast boundaries (look for coloured fringing). If the image contains a face, check the skin near the jawline and under the eyes, where blocking becomes obvious before it does anywhere else.
A useful workflow is to compress at 85, inspect those four areas, and only drop lower if the file is still too large. Going straight to 50 to "be safe" usually produces a visibly worse image than necessary.
Every lossy save discards information permanently. Compressing an already-compressed JPG re-quantises data that was already approximated, so artifacts compound — this is why images re-shared repeatedly through chat apps end up looking muddy. Always compress from the original file, keep that original somewhere, and if you need three sizes, generate all three from the master rather than shrinking the shrunk copy. Re-saving a PNG is safe because lossless compression discards nothing.
Camera files carry EXIF data — GPS coordinates, timestamp, camera model, lens, and sometimes an embedded thumbnail — which can add tens of kilobytes. Re-encoding through the canvas drops most of it, which shrinks the file and, more importantly, strips the location of where the photo was taken before you publish it. If you specifically want to check what an image currently discloses, inspect it with the EXIF Viewer before sharing. Note the flip side: if you need EXIF preserved for cataloguing or copyright, keep the original file, because compression removes it.
Image compressors are among the most common upload-based web tools, which means photographs of documents, ID cards, medical scans and children routinely get sent to servers whose retention terms nobody reads. This tool decodes and re-encodes through your browser's own canvas, so the image is processed on your device and downloaded from it — no transfer, no queue, no stored copy, and no dependence on a promise about deletion timing. It also means large batches are limited by your device's memory rather than an upload speed cap.
70–80% works for most photos with no visible difference. Use 90%+ for product photos or logos where detail matters. Below 50% you'll see compression artifacts.
PNG is lossless — quality slider has no effect. To shrink PNGs, convert to WebP or JPEG (if the image has no transparency).
No. Compression runs entirely in your browser using the Canvas API. The image file never leaves your device.
Limited only by your browser's memory. Most browsers handle images up to 50–100 MB without issue.
Yes for PNG and WebP. If you convert a transparent PNG to JPG, the transparent area is filled white.
Browse all free tools · Guides and tutorials · PDF tools · Developer tools · Text tools · SEO tools