Image Resizer

What this image resizer does

Drop images and the first one appears on a canvas with a dashed frame drawn over it: that frame is the result. Type a width or a height — with the proportion lock on (the default), the other number follows and the frame shrinks or grows with it. The quick buttons set the longest side to 1920, 1280, 800 or 400, the numbers most web pages, blogs and thumbnails want. Switch the unit to % to scale by a factor instead. You can also grab the corner of the frame and drag; the numbers update as you drag, and what you see is exactly what the file will be.

Each row shows the original and new pixel dimensions alongside the file sizes, and the output name carries the new dimensions (photo-1280x853.jpg) so batches stay organised. Downscaling uses a multi-step resampler — the image is halved repeatedly before the final pass — which avoids the jagged edges and moiré patterns that single-pass scaling produces on large reductions. Everything runs in the browser; no upload, no queue, no limits.

How to use it

Set the size, then press Resize. In a batch, width × height is a box every image is fitted into: landscape photos are limited by the width, portrait ones by the height, and nothing ever comes out larger than the box — so one setting works for a mixed folder. Click any row in the list to preview that image on the canvas. 'Don't enlarge' (on by default) leaves images that are already smaller than the box untouched. Unlock the proportions only when a platform demands an exact size; the tool then stretches to the box and says so under the canvas. Your settings are remembered on this device.

When the batch is done, download files one by one or as a ZIP, or use the 'Continue with' buttons to send the resized images to the compressor for a final size cut or to the converter for a different format. The three tools share the same list, so the usual web workflow — resize to display size, convert to WebP, compress to target weight — takes three clicks and one file selection. Sideways or upside-down photos go through Rotate image first; product shots that need a clean backdrop go through Remove background before resizing.

Common pitfalls

Enlarging never adds detail. Scaling a 400-pixel image up to 2000 pixels produces a soft, blurry result, because the missing pixels are interpolated rather than recovered; if you need a bigger image, go back to the original or a vector source. Unlocking the proportions ignores the aspect ratio by design and will squash a portrait photo into a landscape box — do it only when a platform genuinely requires a fixed size, and crop first if the proportions differ.

Resizing a PNG screenshot that contains small text makes the text fuzzy; where possible keep screenshots at native size and let the page scale them with CSS. The % unit is relative to each image's own size, so a mixed batch ends up at mixed sizes — use pixels if you want uniform output. And as with all canvas-based tools, EXIF metadata (including the orientation flag) is dropped; the browser applies the orientation before scaling, so photos come out upright, but camera details are not preserved.

Background: how image scaling works

Reducing an image means deciding what colour each new pixel should be from the several source pixels it replaces. Nearest-neighbour picks one and looks jagged; bilinear averages the four closest; bicubic and Lanczos weigh a wider neighbourhood and produce smoother, sharper results. Browsers implement a high-quality mode for canvas drawing, but when the reduction is large — say 6000 pixels down to 600 — a single pass still samples too few source pixels and aliasing appears. The standard fix, used here, is to halve the image repeatedly until it is within a factor of two of the target and then perform the final resample.

Digital cameras and phones now produce 12–50 megapixel images, far larger than any screen: a 4K monitor is only 8 megapixels and a phone screen around 3. Serving an image bigger than its display size costs bandwidth, memory and decoding time for nothing, which is why responsive images (the srcset attribute) and image CDNs exist. Resizing locally before upload achieves the same result with no infrastructure — and, because the work happens in your browser, the photos never leave your device.

Where image resizing is used

Web publishing is the biggest use: matching hero images to their CSS width, generating 2× retina versions, and producing the several sizes a srcset needs. Social platforms want specific dimensions — 1200×630 for link previews, 1080×1080 for square posts, 1500×500 for banners. Email signatures and newsletters need small, fixed-width images. Profile-photo and ID uploads on government, university and job sites often enforce pixel limits such as 600×600 or a maximum of 2 MB. App stores require screenshots at exact device resolutions, and print shops ask for a minimum pixel count for the physical size.

Batch resizing before archiving or sharing a holiday album, shrinking product photos to a shop's grid size, making a set of icons at 16/32/64/128 pixels from one master PNG, and reducing scanned documents to email size are all everyday jobs. Combine with the format converter for WebP output and the compressor to hit a byte target; a resized PNG that contains a QR code should stay PNG so the modules remain sharp. Pages exported with PDF to image can be resized here before going onto a slide or a website.

Frequently asked questions