Image Compressor

No. This is a no-upload image compressor — every step happens in your browser through the Canvas API, and no file is transmitted, logged or stored anywhere.

You can verify it in about ten seconds: open developer tools, switch to the Network tab, and compress an image. No request appears. Disconnect from the internet after the page loads and the tool still works perfectly. That makes it safe for client work, ID documents and anything you would not want sitting on someone else's server.

You can drop in JPG/JPEG, PNG, WebP, AVIF, GIF and BMP files. Output keeps the original format for JPEG, PNG and WebP; anything else (GIF, BMP, AVIF) is written out as JPEG or PNG depending on whether it has transparency, since those are the formats every browser can reliably encode.

Animated GIFs are a special case — only the first frame survives, because canvas-based processing cannot preserve animation. If you need to shrink an animation, convert it to a video format instead.

For website images, 75–85 is the sweet spot. That range typically removes half to three-quarters of the file size while staying visually indistinguishable from the original at normal viewing distance. Below about 60 you start to see blocky artefacts in smooth areas like skies and skin.

For anything destined for print, or images that will be edited further, stay at 90 or above. And remember that quality is relative to content: a soft-focus portrait survives aggressive compression far better than a screenshot of a spreadsheet. Use the side-by-side preview above and lower the slider until you can actually see the difference — then go back up one notch.

Not unless you ask it to. With Max width left at 0, pixel dimensions are preserved exactly and only the encoding changes.

Set a value and any image wider than that limit is scaled down proportionally, so aspect ratio never changes and narrower images are left alone. This matters more than people expect: a 4000px photo displayed in a 800px column wastes bandwidth on every single page view. Resizing to the width you actually display usually saves more bytes than the quality slider does.

By default each image keeps its own format — a PNG stays a PNG. Tick Force JPEG and everything is written out as JPEG instead.

This is genuinely useful when a PNG contains a photograph rather than a logo or screenshot: photographic PNGs are often several times larger than an equivalent JPEG at the same visual quality. The catch is transparency. JPEG has no alpha channel, so transparent areas are composited onto a white background rather than turning black as they do in some tools. If you need the transparent background, leave this option off.

The queue holds up to 60 images per session with a 30 MB cap per file. Those limits exist to keep your browser tab responsive, not because of any server quota — there is no account, no daily allowance and no watermark.

Files are processed one after another so the page stays usable, and each row shows its own before and after size. If you have more than 60, clear the queue and run a second pass. Very large files (20 MB and up) may pause the tab briefly during encoding; that is normal.

At quality 80 and above, almost never on photographs — which is exactly why the side-by-side preview exists. Compare the two panes at full size before you commit to a setting.

Where loss does show up is in images with hard edges: logos, text overlays, diagrams, QR codes and pixel art. Lossy encoding produces faint halos around sharp contrast boundaries, and those artefacts compound every time you re-save. Two rules keep you safe. Always compress from the original master rather than from an already-compressed copy, and if the source is a graphic rather than a photo, keep it as PNG at a high quality setting or skip compression entirely.