Image Resizer

How to resize an image

1. Add your images. Drag them onto the box above or click to browse. Batches are fine.
2. Choose a mode. By width or by height sets one dimension and derives the other. By percentage scales both. Fit inside a box keeps everything within limits you set. Exact size sets both dimensions.
3. Enter your numbers, or use one of the presets for common sizes.
4. Resize and download. Each row shows the original and new dimensions with the size change. Save individually or download the batch as a ZIP.

Nothing is uploaded. Everything happens in your browser.

When you would need this

Meeting a required size. Passport photo uploads, competition entries, marketplace listings and profile pictures often specify exact pixel dimensions and reject anything else.

Preparing images for a website. Serving a 4000-pixel photo into a 600-pixel column wastes most of the bytes a visitor downloads. Resizing to what is actually displayed is the single most effective image optimisation there is.

Reducing file size properly. Resizing removes pixels rather than degrading the ones that remain, so it reduces file size dramatically without introducing compression artefacts.

Normalising a mixed batch. Photos from different cameras and phones arrive at wildly different sizes. Fitting them all inside one box makes them consistent without cropping anything.

Fitting an email or upload cap. Halving the dimensions of a photo cuts roughly three quarters of its file size, which usually solves the problem in one step.

Common sizes worth knowing

Most resizing tasks are one of a small number of targets. These are the ones that come up repeatedly.

The principle behind all of them is the same: an image only needs enough pixels to cover the space it occupies on screen, doubled for high-density displays. Everything beyond that is downloaded and discarded by the browser.

Why downscaling looks good and upscaling does not

The asymmetry is worth understanding, because it explains a limit no tool can work around. Downscaling combines several source pixels into one output pixel — the information needed to decide each output value genuinely exists, so the result is an accurate summary of the original.

Upscaling has the opposite problem. Doubling an image's width means three quarters of the output pixels correspond to nothing that was ever captured, and their values have to be estimated from neighbours. Estimation is why enlargements look soft: the invented pixels are averages, and averages have no edges. Sharpening afterwards adds contrast at boundaries but cannot restore texture that was never recorded.

Machine-learning upscalers improve on this by predicting plausible detail from training data rather than averaging — but they are inventing a convincing guess, not recovering the original. For anything where accuracy matters, such as a document or a product photograph, starting from a larger original is the only real answer.

Tips for resizing well

Downscale freely, upscale reluctantly. Removing pixels looks fine; inventing them does not. Enlarging produces softness or artificial-looking detail because the extra pixels are estimated from their neighbours. Keep the "never enlarge" option on for batches, so a small image mixed in with large ones is left alone instead of being blown up.

Resize before compressing, not after. Both reduce file size, but they compound in the right order. Resize to the dimensions you actually need first, then apply compression to what remains. Compressing a huge image aggressively and then resizing it means you kept the artefacts and threw away the pixels.

Use fit-inside-a-box for mixed orientations. If a batch contains both portrait and landscape images, setting a width forces the portraits to be very tall and vice versa. Fitting inside a box handles both correctly, because the constraint applies to whichever dimension is larger.

Double your dimensions for high-density screens. Phones and modern laptops pack two or three physical pixels into each CSS pixel. An image displayed at 600 pixels wide looks soft on those screens unless it is around 1200 pixels. Beyond roughly double there is no visible benefit.

Do not stretch to fit a shape. If an image must fill a differently-shaped space, crop it rather than stretching. Cropping loses content at the edges, which viewers rarely notice; stretching makes faces and circles visibly wrong, which they notice immediately.

Check the smallest result in a batch. When resizing many images at once, the settings that suit the largest may reduce a smaller one further than you intended. The dimensions shown on each row make this easy to spot before downloading.

Limits and requirements

Enlarging cannot add real detail. No resizing method recovers information that was never captured. This tool uses the browser's high-quality smoothing, which is good, but the limit is physical rather than algorithmic.

Aspect ratio is preserved except in exact mode. Every other mode maintains proportions automatically. Exact mode will distort unless you leave the keep-aspect-ratio option on.

Metadata is dropped. Re-encoding does not carry EXIF across, so timestamps, camera settings and GPS coordinates are removed.

Memory is the constraint. Each image is decoded to raw pixels at roughly four bytes per pixel — a 50-megapixel photo needs around 200MB while being processed. Large batches of high-resolution images are best split into smaller groups on a phone.

Animated GIFs lose animation. Only the first frame is resized. The output is a still image.

Very small results. Dimensions are never allowed to round down to zero, so scaling a tiny image by a very small percentage yields at least one pixel rather than an error.

Privacy: how this tool handles your images

Your images never leave your device. When you add a file, your browser decodes it into memory, draws it at the new size, re-encodes it, and hands the result back. There is no upload, no server-side processing and no copy on our infrastructure — because none is involved.

The images people resize are frequently personal: profile pictures, family photos, pictures of a home, photographed documents being prepared for an application. Free resizing sites handle this by uploading the full-resolution original — complete with the GPS coordinates embedded by the camera — to a server you know nothing about, and it is entirely unnecessary given what browsers can now do locally.

You can verify this rather than trusting it. Open your browser's developer tools, switch to the Network tab, and resize an image — nothing carries your file. Or load this page, disconnect from the internet, and resize anyway. It still works.