Image Compressor

What this tool does

The compressor reads an image from your device, redraws it on a canvas and asks your browser to encode a new file. You can lower the quality, and the longest side is capped at 1,200 pixels. The original and compressed versions are shown side by side with their sizes.

Nothing is uploaded. The image is read through the browser's File API, processed in the page and downloaded from an object URL.

How to use it

1. Choose a JPEG, PNG, WebP, GIF or another raster image your browser can open. SVG files are not supported.
2. Set the quality. The default is 80%.
3. Select Compress image. The result appears next to the original.
4. Download the compressed file, or move the quality slider and release it to try again.

Quality, format and file size

Quality is passed to the browser's image encoder. It is not a percentage of the original file size, and the same value can produce different results in different browsers. A lower value usually makes a smaller file, but it can introduce visible artifacts, especially around text and sharp edges.

This tool asks for WebP output when the browser supports it and falls back to JPEG otherwise. WebP often produces a smaller file at a comparable quality, but JPEG may be more widely accepted by older software. The result panel notes which format was used.

Compression is not guaranteed to shrink every file. An image that is already heavily optimised may grow when it is re-encoded. If that happens, keep the original or try a lower quality. For a wider performance checklist, read how to speed up your website.

Dimensions, transparency and animation

Images larger than 1,200 pixels on the longest side are scaled down proportionally. Smaller images are not enlarged. WebP can preserve transparency; the JPEG fallback does not, so transparent areas are filled with white.

Animated GIFs are treated as still images: only the first frame is redrawn and compressed. The tool does not create an animated WebP or keep the animation.

Metadata and privacy

The browser redraws the pixels into a new image, so most embedded metadata from the original file, such as camera details, is not carried over. This is not a dedicated metadata scrubber, and you should not rely on it for privacy-sensitive files without checking the result.

Processing happens in this browser tab. The image is not sent to NelcByte. Large images use more memory, and an older or memory-constrained device may fail to decode or encode a very large file.