How the automatic size search works
Set a target in kilobytes — or tap one of the 20, 50 and 100 KB presets — and add your images. Each one is drawn onto a canvas inside a Web Worker, a background browser thread that keeps the page responsive while heavy work runs. The compressor then re-encodes the image and measures the result, running a binary search across quality levels: too big, drop quality; comfortably small, raise it. The search settles on the highest quality that still fits under your target.
When even the lowest sensible quality can't get there, the tool scales the pixel dimensions down in steps and searches again. And if a target is genuinely unreachable without making the picture illegible, you get the smallest readable version it produced plus an honest note that the target wasn't met — nothing fails silently.
JPEG or WebP?
JPEG is the default because practically every upload form accepts it. WebP usually encodes the same visual quality in a noticeably smaller file — often 25 to 30 percent less — which means more detail inside the same byte budget. The trade-off is compatibility: a few older portals and email clients reject it. A reasonable rule: WebP for images you control, such as your own site, and JPEG whenever a third-party system will inspect the file.
Compression that stays on your machine
There is no upload step. The file is read, re-encoded and handed back entirely within your browser tab, so no server ever holds a copy of your photo. Re-encoding through a canvas has a useful side effect too: EXIF metadata — GPS coordinates, camera model, capture time — is stripped from the output. One caveat: images already under your target are returned untouched and keep their metadata. Batch processing works the same way; add a folder's worth of images, watch each one's before and after size, and download the results as a single ZIP.