Image Compressor

Image Compressor uses the browser-image-compression library, which orchestrates a Canvas re-encode loop inside a dedicated Web Worker. Your file is decoded into a bitmap, drawn onto an OffscreenCanvas (or a regular Canvas on Safari, which still lacks full OffscreenCanvas support outside Workers in older versions), and re-encoded with the browser's JPEG, WebP, or PNG encoder. The worker keeps the encode off the main thread so the page never freezes, even on 50 MB camera RAW exports.

Quality is controlled through two complementary modes. Quality mode passes a 0-1 quality factor straight to the encoder; lower values give smaller files at the cost of more visible JPEG artefacts in flat colour areas. Target size mode runs a binary search across quality values, encoding the image multiple times until it lands within roughly 5 percent of the size you asked for, which is how you can hit a strict 200 KB or 1 MB ceiling without manual fiddling.

JPEG and WebP are lossy and benefit most from compression: JPEG drops to 30-50 percent of the original size at quality 0.8 with no visible difference, and WebP at the same setting is usually another 25-35 percent smaller still. PNG is lossless and only benefits from sub-pixel-accurate downscaling and zlib re-encoding, which trims the size modestly but never as dramatically as JPEG or WebP.

Resizing during compression is supported through a maxWidthOrHeight option. The library uses Pica-style resampling under the hood, which gives sharper down-scaling than the default Canvas drawImage path. Combine this with a quality drop to hit very small file sizes for hero images that load above the fold.

EXIF metadata handling is configurable. By default the compressor preserves the orientation tag (so phone photos display the right way up) but strips GPS coordinates, camera serial, and timestamp data, which keeps files small and avoids accidentally publishing your location with a photo. A toggle preserves all metadata if you need it for archival.

Files up to about 100 MB can be compressed in a single pass on a desktop browser; mobile Safari has a tighter per-tab memory limit and may struggle above 30-40 MP. For very large originals the right play is to resize first, since pixel count drives memory more than file size does.

Everything happens locally. The file lives as a Blob in your tab, the worker executes JavaScript bundled with the page, and the encoded output is downloaded directly. There is no upload step and no server-side compression queue, which is why the tool can promise that your images never touch external infrastructure.