Resizing…
Your images never leave this device — resizing runs on the HTML canvas in your own browser.
Drop your images onto the upload area, or click it to browse. You can add up to 20 images at once, each up to 50 MB.
Choose how to resize: exact pixel dimensions, or a percentage of the original. Presets for HD, 4K, square social posts and thumbnails are one tap away.
Leave the aspect ratio locked to keep the original proportions — type a width and the height follows automatically. Unlock it only if you deliberately want to stretch the image.
Pick an output format (keep the original, or convert to JPEG, PNG or WebP) and set the quality for the lossy formats.
Download each image individually, or use Download all as ZIP to get the whole batch in one file.
Set exact pixel dimensions when a platform demands them, scale by percentage when you just want everything smaller, or tap a preset for HD, 4K, square social posts and thumbnails. The aspect-ratio lock keeps proportions correct in every mode.
Queue up to 20 images and apply the same settings to all of them at once. Each result shows its own before-and-after dimensions and file size, and the whole batch downloads as a single ZIP with predictable filenames.
Convert to JPEG, PNG or WebP while you resize, with a quality slider for the lossy formats. WebP typically saves 25-35% over JPEG at the same visual quality, and transparency is handled correctly when converting to and from PNG.
Resizing happens on the HTML canvas inside your own browser. No file is uploaded, no copy is stored on a server, and nothing is transmitted anywhere. Most free resizers upload your photos to process them; this one cannot, because there is no server-side step to upload to. It also keeps working with no connection once the page has loaded.
With the lock on, entering a width computes the matching height from the source proportions, so the image is scaled rather than stretched. The tool also refuses to upscale by default, because enlarging past the original resolution cannot add detail — it only interpolates, producing a softer image at a larger file size.
Twenty images at a time, no watermark, no queue, no account, and no upsell after the third file. The batch is processed locally so there is no per-file cost to us and therefore no reason to meter it. The ZIP is assembled in your browser too.
Every result shows the original and new dimensions alongside the original and new file size, and whether the file got smaller or larger. Converting a photo to PNG or raising JPEG quality can easily produce a bigger file than you started with; the tool tells you when that happens instead of quietly handing you a heavier image.
Resizing an image sounds like a single operation, but three separate decisions are hiding inside it: what dimensions to target, how to compute the new pixels, and what format to save the result in. Get any of them wrong and you end up with a stretched, soft, or unexpectedly enormous file. This guide covers what actually happens when an image is resized and how to choose settings that hold up.
A raster image is a grid of pixels — 4000 across and 3000 down, for example. The aspect ratio is the relationship between those two numbers, in this case 4:3. Resizing to 1000 × 750 preserves that ratio and simply samples the same picture onto a smaller grid. Resizing to 1000 × 1000 does not: the image must be squashed vertically to fit, and every face, circle and straight line in it is distorted.
This is why an aspect-ratio lock matters more than any other control. With it on, you supply one dimension and the other is computed for you. It is also why "fit within 1000 × 1000" and "make it exactly 1000 × 1000" are different requests: the first scales the longest side down until the whole image fits inside the box, leaving the shorter side smaller than the limit; the second forces both dimensions and distorts unless the source happens to be square.
When you shrink an image from 4000 pixels wide to 1000, four source pixels have to become one. The algorithm that decides the value of that one pixel is the resampling method. Nearest-neighbour simply picks one of the four and discards the rest, which is fast and produces hard, aliased edges — appropriate for pixel art, terrible for photographs. Bilinear and bicubic interpolation average the surrounding pixels with increasing sophistication, producing the smooth results expected of photographic downscaling.
Browsers expose this through the canvas image-smoothing settings, and this tool requests the highest-quality option available. In practice that means downscaling a photo here looks equivalent to doing it in a desktop editor for anything short of extreme reduction ratios. If you are shrinking by more than about 4x in one step, editors that offer multi-step downscaling can retain marginally more detail, though the difference is rarely visible at normal viewing sizes.
Enlarging works the same way in reverse: one source pixel has to fill four output pixels, and interpolation invents the values in between by averaging neighbours. No information is added, because none exists in the file. A 400 × 300 photo scaled to 1600 × 1200 has sixteen times as many pixels and exactly the same detail, spread more thinly — which reads as softness or blur.
This is why upscaling is disabled by default here and has to be switched on deliberately. It has legitimate uses: matching a required canvas size, printing at a fixed physical dimension, or preparing an asset for a system that rejects small files. But it never recovers detail, and it always increases file size. AI-based upscalers work differently — they hallucinate plausible detail from a trained model rather than interpolating — which is a genuinely different operation with genuinely different trade-offs, including invented detail that was never in the original.
One conversion trap is worth knowing: JPEG has no alpha channel, so a transparent PNG converted to JPEG must have its transparent areas filled with something. This tool fills them with white, which is usually what people expect. If your image has meaningful transparency, convert to WebP or keep PNG instead.
The quality slider for JPEG and WebP controls how aggressively detail is discarded. The relationship between the number and the file size is steeply non-linear. Dropping from 100 to 90 typically removes more than half the file size while being essentially indistinguishable to the eye. Dropping from 90 to 80 saves a further useful chunk with minor softening in fine texture. Below about 70, artefacts around sharp edges and in smooth gradients become visible, and below 50 most images look obviously degraded.
For most web use, 80-85 is the sweet spot, which is why this tool defaults to 85. Quality has no effect on PNG output, which is lossless by definition — the slider is simply ignored for that format.
Screen images are measured in pixels, and the practical question is how large the image will be displayed. A full-width hero on a typical site needs around 1600-2000 pixels wide to look sharp on high-density displays; a blog body image rarely needs more than 1200; a thumbnail needs 300-400. Serving a 4000-pixel photo into a 600-pixel slot wastes bandwidth on every visitor and is one of the most common causes of poor page-speed scores.
Print is measured in physical size at a given resolution, conventionally 300 dots per inch. A 6 × 4 inch print therefore wants roughly 1800 × 1200 pixels. Below about 150 DPI the result looks visibly soft on paper. Note that the DPI value stored in an image file is only metadata — it does not change the pixels, and changing it without resampling changes nothing about the image itself.
Because this tool redraws your image onto a fresh canvas, the output contains only pixels. EXIF metadata — camera model, exposure settings, timestamps, and any GPS coordinates recorded by a phone — is not carried across. For most people that is a privacy benefit: resized images can be shared without the location where the photo was taken travelling with them. If you specifically need to preserve EXIF, resize with a tool that copies the metadata block explicitly, because canvas-based resizing in any browser tool will drop it.
Shrinking an image is inherently lossy in the sense that pixels are discarded, but the visible result stays sharp if you scale proportionally with the aspect ratio locked and save at a high quality setting (85 or above for JPEG and WebP), or use PNG for graphics. What actually damages quality is stretching, upscaling, and repeatedly re-saving as JPEG.
No. Resizing runs on the HTML canvas inside your browser, so the files never leave your device. There is no upload step, no server-side processing, and nothing is stored. The tool keeps working with no internet connection once the page has loaded.
Upscaling is off by default because enlarging past the original resolution cannot add detail — interpolation only spreads the existing pixels more thinly, giving a softer image in a larger file. Switch on Allow upscaling if you deliberately need a bigger canvas, for example to meet a fixed size requirement.
With the lock on, typing a width computes the matching height from your image's original proportions (and vice versa), so the picture is scaled rather than stretched. Unlocking it lets you force both dimensions independently, which distorts the image unless the new ratio happens to match the old one.
JPEG for photographs where small files matter and transparency is not needed; PNG for screenshots, logos and line art, or when transparency must be preserved; WebP for the web generally, since it is 25-35% smaller than JPEG at comparable quality and every current browser supports it.
Usually because of a format change or a quality increase. Converting a JPEG photo to PNG commonly multiplies its size, since PNG is lossless and poorly suited to photographic content. Raising the quality slider above the source image's original compression level has the same effect. Each result shows the size change so you can spot it.
Up to 20 per batch, each up to 50 MB. All of them use the same resize and format settings, and you can download them individually or together as a single ZIP file assembled in your browser.
No. The image is redrawn onto a clean canvas, so the output contains pixels only — camera settings, timestamps and GPS coordinates are all dropped. This is usually desirable when sharing photos, but if you need EXIF preserved you will need a tool that copies the metadata block explicitly.
What actually degrades an image when you resize it, and how to avoid it. Resampling, aspect ratio, upscaling limits, format choice and quality settings explained.
Read more →