Image Resizer
Resize images to exact dimensions, a percentage, or to fit inside a box, with an aspect-ratio lock and one-click presets for OG images, Instagram, YouTube thumbnails and more. Handles a whole batch at once and uses progressive downscaling for noticeably sharper results. Everything runs in your browser — nothing is uploaded. Free, no signup.
Add images to resize
Drag and drop, or click to choose. JPG, PNG, WebP or GIF. Batch supported.
Resized images appear here. Nothing is uploaded — everything happens in your browser.
How to Use This Tool
- Add your images — drag and drop or click to browse. The first image seeds the width and height boxes with its own size.
- Choose how to resize — exact dimensions with an optional aspect lock, a straight percentage, or fit inside a bounding box without cropping.
- Or pick a preset — OG image at 1200 by 630, Instagram square, YouTube thumbnail and other common sizes are one click away.
- Download — files are named with their new dimensions, so a batch stays easy to tell apart.
About Image Resizing
Resizing is the single most effective thing you can do to an oversized image, and it is more effective than compression. A 4000 pixel wide photo displayed in a 800 pixel wide column is carrying twenty-five times more pixel data than the page will ever show. No amount of quality tuning fixes that, because the waste is in the dimensions rather than the encoding. Resize first, then compress: doing it in that order routinely takes a multi-megabyte camera file down to something in the tens of kilobytes with no visible difference on screen.
Aspect ratio is where most resizing goes wrong. If you set both width and height to values that do not match the original proportions, something has to give, and the usual result is a stretched or squashed image that looks subtly wrong even to people who cannot say why. The aspect lock here solves that by treating width as authoritative and deriving height from each image own proportions, which matters for a mixed batch. Fit-inside mode solves the same problem differently: it scales the image until it fits entirely within your target box, so nothing is distorted and nothing is cropped, though the result may be smaller than the box in one dimension.
Downscaling and upscaling are not symmetrical operations. Downscaling throws away information, which is straightforward and generally looks good. Upscaling has to invent information that was never captured, and no amount of interpolation can recover detail that does not exist, so an enlarged image always looks softer than a native one at that size. That is why this tool refuses to enlarge by default and warns you when you turn that off. If you genuinely need a bigger image, the right answer is to go back to a higher-resolution original rather than to upscale.
The quality of a downscale depends heavily on how it is done, and this is where a lot of browser-based resizers fall short. Scaling an image directly from 4000 pixels to 400 in a single step forces the renderer to sample very sparsely, which produces aliasing and a crunchy, noisy result. Halving repeatedly until close to the target and only then doing the final step gives the renderer a much better sample at each stage and produces a visibly cleaner image. This tool does that progressive halving automatically whenever the reduction is more than twofold. The one case where you want the opposite is pixel art, where any smoothing destroys the crisp edges, which is what the nearest-neighbour option is for.
A note on DPI, because it confuses almost everyone. For anything on a screen, DPI is meaningless: browsers care only about pixel dimensions, and the DPI field stored in an image file is ignored entirely. It matters only for print, where it describes how densely those pixels get laid down on paper. A 1200 by 630 image is a 1200 by 630 image whether its metadata claims 72 or 300 DPI. If a printer asks for 300 DPI, what they actually need is enough pixels for the physical size, which for a 6 inch wide print means 1800 pixels across.
Pair this with our Image Compressor to shrink the file after resizing, the Image to WebP Converter for modern formats, and the Open Graph Previewer to check how a resized share image actually renders.
Frequently Asked Questions
How do I resize an image without distorting it?
Keep the aspect ratio, which is the relationship between width and height. If you set both dimensions independently to values that do not match the original proportions, the image gets stretched or squashed. The aspect lock in this tool handles that by treating width as the value you control and calculating height from each image own proportions, which matters when a batch contains different shapes. If you need every output to land in exactly the same box regardless of input shape, use fit-inside mode: it scales until the image fits entirely, so nothing is distorted, though it may not fill the box in both directions.
Why does my upscaled image look blurry?
Because upscaling cannot recover detail that was never captured. When you enlarge an image, the software has to invent the new pixels by interpolating between existing ones, and interpolation produces plausible values rather than real detail. The result is always softer than a photograph taken at that resolution. This is a fundamental limit, not a shortcoming of any particular tool, which is why this one refuses to enlarge by default and warns you when you override it. If you need a larger image, go back to a higher-resolution original.
What is the difference between the resampling options?
Smooth resampling averages neighbouring pixels when calculating each output pixel, which produces clean, natural results for photographs and is what you want almost all the time. Nearest-neighbour simply picks the closest source pixel with no blending at all, which produces hard, blocky edges. That sounds worse, and for photos it is, but it is exactly right for pixel art, retro game sprites, QR codes and technical diagrams where crisp boundaries carry the meaning and any smoothing turns them to mush.
What size should an Open Graph image be?
The standard recommendation is 1200 by 630 pixels, which is a 1.91 to 1 ratio, and it is what Facebook, LinkedIn and most other platforms crop toward. Twitter and X accept the same image for a large summary card. Keep the file under about 1 MB, since some crawlers refuse larger ones, and keep important content away from the edges because different platforms crop slightly differently. Also make sure your og:image tag points at an absolute URL rather than a relative path, which is the single most common reason a correctly sized image still fails to appear.
What are the current Instagram image sizes?
Instagram accepts square at 1080 by 1080, portrait at 1080 by 1350, and landscape at 1080 by 566. Portrait takes up the most vertical space in the feed, which is why it generally performs best for reach. Stories and Reels use 1080 by 1920, a 9 to 16 ratio. Whatever you upload, Instagram re-compresses it on their side, so there is little point uploading anything larger than 1080 pixels wide. Platform specifications do change, so treat these as current guidance rather than permanent fact.
Does DPI matter for web images?
No, not at all. Browsers and any other screen display work purely in pixel dimensions, and the DPI value stored inside an image file is ignored completely. An image that is 1200 by 630 pixels behaves identically whether its metadata says 72 DPI or 300. DPI only becomes meaningful in print, where it describes how densely the pixels are placed on paper. When a printer asks for 300 DPI, what they really need is enough pixels for the physical size: a 6 inch wide print at 300 DPI needs 1800 pixels across.
Should I resize before or after compressing?
Resize first, then compress. Resizing removes pixels that the page will never display, which is a much bigger saving than any encoder setting can deliver, and compressing afterwards then works on a smaller image so the quality trade-off is gentler. Doing it the other way round wastes effort, because you carefully tune the compression of pixels you are about to throw away. This tool also lets you set an output quality directly, so for many images the resize step alone gets you where you need to be.
Are my images uploaded anywhere?
No. Files are read locally through the browser FileReader API, drawn to a canvas and re-encoded entirely on your own machine. Nothing is transmitted to any server, nothing is logged, and no copy survives closing the tab. That makes the tool safe for client work, unreleased product photography and anything under NDA. Only your settings, meaning the dimensions, format and quality you chose, are kept in your own browser localStorage so the tool remembers them next time.