How to Resize an Image Online in Pixels or Percent

You can resize an image online in your browser: drop a JPG, PNG, WebP or AVIF file, type a new width and height in pixels, or click a percentage shortcut from 25% to 200%. The canvas decoder scales it locally, so nothing is uploaded, and you can lock the aspect ratio or free both sides before downloading.

Two input modes cover most jobs. Pixel mode takes an exact width and height, starting at the original dimensions, and the lock keeps the shape by recalculating the other side when you change one. Percentage mode is one click: 25%, 50%, 75%, 100% or 200%, always calculated from the original file, so 50% of a 4000 x 3000 photo gives 2000 x 1500. You can then choose JPG, PNG, WebP or keep the incoming format, and JPG and WebP add a quality slider from 10 to 100 that starts at 90. Everything runs in the tab, so the file never reaches a server.

The resizer is one of 45 browser-based tools on Aihangsoft, and it does one narrow job: change how many pixels an image has. That job is easily confused with two neighbours. A smaller file at the same dimensions is compression, and a different shape is cropping. This page is about scale.

Image Resizer

Set exact pixels or a quick percentage, lock the proportions, and download. The scaling happens in your tab.

Open the resizer

Resizing is not the same as compressing

Changing the dimensions of an image and compressing it both alter the file size, which is why the two get mixed up, but they act on different things. Resizing changes the pixel count: a 4000 x 3000 photo resized to 2000 x 1500 keeps the same content on a quarter of the pixels. Compressing keeps the pixel count and changes how aggressively the encoder writes those pixels, trading fidelity for bytes. The resizer does a little of both in one pass, because every canvas export re-encodes, but its main lever is width and height, not the quality slider.

That overlap is also why a resize can shrink a file a lot even at high quality: fewer pixels simply means less data to encode. It is not a substitute for a purpose-built compressor, so if you need to hit a specific file size rather than a specific dimension, the guide to compressing an image to 100 KB walks through that workflow.

How to resize an image online

The whole job is five decisions, and doing it takes less time than reading about it. Nothing is submitted anywhere, so you can try any value and simply enter another one.

  1. Add your image. Drop a JPG, PNG, WebP or AVIF file on the drop zone or click to browse. One image is processed at a time. The decoder reads it into the page, and the line under the fields shows the original dimensions in pixels.
  2. Type the new width and height, or click a percentage. Pixel mode accepts any whole number from 1 upward. The quick scale buttons read 25%, 50%, 75%, 100% and 200%, and they always calculate from the original file, so 50% of a 4000 x 3000 photo gives 2000 x 1500.
  3. Decide whether to keep the aspect ratio locked. The lock is on by default: change the width and the height is recalculated to match, keeping faces, products and text in proportion. Unlocking lets both sides move independently, which stretches the picture, so reserve it for a frame whose shape differs from the source.
  4. Pick the output format and quality. Keep the original format, or choose JPG, PNG or WebP. JPG and WebP reveal a quality slider that runs from 10 to 100 and starts at 90. PNG is lossless, so the slider is hidden for it.
  5. Click Resize image, then download. The panel reports the new dimensions, the source size and the output size, and the download name carries the new dimensions, for example photo-1920x1080.jpg. Reset returns the fields to the original size without dropping the file.

Common pixel sizes worth knowing

There is no single correct size, because the right number depends on where the image will live. The values below are widely used working sizes rather than rules published by any one platform, so treat them as starting points and check the requirement at the place you are uploading.

PurposeCommon sizeWhy it is used
Full HD display and video thumbnails1920 x 1080The 16:9 frame most screens and players expect
Square social post1080 x 10801:1, a safe square that crops cleanly on every feed
Portrait social post1080 x 13504:5, taller than square and still narrow enough for a phone
Web hero and blog header1600 x 90016:9, sharp on a laptop without a huge file
Photo attached to a form or email800 x 600Small enough to pass an upload limit and stay readable

Two of those shapes are 16:9 and the others are not, so a phone photo shot in 4:3 will not fit without a change. Resizing alone cannot reshape an image without stretching it, because the tool scales the whole frame. When the target has a different ratio, crop to that ratio first and resize the crop afterwards, as the crop image online guide describes.

What happens when you enlarge an image

Typing a width larger than the source is allowed, and the tool will produce a bigger image, but it cannot add information that was never captured. The canvas interpolates between existing pixels with high smoothing quality, so edges stay clean and the result does not break into blocks, yet it looks softer than the original rather than more detailed.

Interpolation versus reconstruction

Interpolation spreads the pixel data you already have across more space. Reconstruction, which is what an AI upscaler does, guesses plausible detail from patterns learned from other images. For a genuinely sharper picture at a larger size, the AI Image Upscaler is the right tool, as the upscale image online guide explains. The resizer suits an exact pixel count and a modest enlargement, and the safer direction is downward, since halving the resolution discards pixels rather than inventing them.

Which output format should you choose

Format decides what the encoder is allowed to throw away, and it decides whether transparency survives. The resizer offers four choices: keep the original format, JPG, PNG or WebP.

OptionCompressionTransparencyBest for
Keep original formatMatches the source typeMatches the sourceThe default when you do not want to change the file type
JPGLossy, quality slider from 10 to 100Not supported, transparent areas turn whitePhotos and forms that demand maximum compatibility
PNGLossless, no quality sliderPreservedLogos, icons and screenshots with text
WebPLossy, quality slider from 10 to 100PreservedSmaller files for the web while keeping transparency

One behavior catches people out. JPG has no alpha channel, so transparent regions have to become something, and the resizer fills the canvas with solid white before drawing the image on top. A transparent logo exported as JPG therefore arrives on a white background, not a black one. Choose PNG or WebP if the transparency has to survive.

Keep original format is not literal for every input. When the browser reports JPG, PNG or WebP, that type is reused. For an AVIF input, which is readable but not a type the encoder writes, the tool falls back to lossless PNG, so an AVIF photo can come out much larger than it went in.

What the resizer cannot do

Most of these limits follow from how a canvas works.

  • It cannot keep EXIF metadata. Decoding and redrawing on a canvas writes pixels only, so camera model, capture date, orientation and GPS coordinates are all dropped. That is usually welcome for privacy, and it is exactly what the Remove EXIF tool does on purpose. If a GPS tag has to survive, resize a copy and keep the original.
  • It works on one image at a time. The drop zone reads a single file, so fifty photos mean fifty passes. The Image Compressor takes multiple files at once, but the resizer does not.
  • It cannot open formats your browser cannot decode. HEIC files from some phones are the usual case, and they are rejected with a message telling you to convert to JPG or PNG first. Input support covers JPG, PNG, WebP and AVIF in current Chrome, Edge, Firefox and Safari.
  • It does not crop, rotate or straighten. It scales the entire frame, so any reshaping has to happen first, in the Image Cropper or another tool.
  • It cannot guarantee a file size. The quality slider changes size, but the exact byte count depends on the image, and an extremely large target dimension can exceed what the browser will encode at all.
  • It cannot create detail. Enlarging softens the image by design, as covered above, and no setting on this page reverses that.

FAQ

Yes, and the change usually goes in the direction you expect: fewer pixels means less data, so cutting an image to 50% normally produces a smaller file than the original. Three things decide the final byte count. The pixel dimensions set how much data exists to begin with, the output format decides how that data is stored, and the quality slider controls how much of it a lossy encoder throws away. Lowering the quality to 50 at unchanged dimensions will also shrink a file, while enlarging to 200% and saving at quality 100 will make it larger than the source. Every canvas export re-encodes the pixels, so the result is never byte-for-byte identical to the input, and the result panel prints the exact output size before you download.
Yes. Type 1920 in the width field and 1080 in the height field, click Resize image, and the output is that exact pixel grid. One detail matters: with the aspect ratio locked, which is the default, changing one field recalculates the other to match the proportions of your source, so a second number you typed can be overwritten. Unlock the ratio to set both sides independently, and expect the picture to stretch when the new shape differs from the original. For a target like 16:9 the cleaner route is to crop to that ratio first and resize the crop afterwards, because a deliberate crop keeps the subject in proportion while a stretch distorts it.
It can be, in two separate ways. Downscaling resamples the image, which is normally invisible but can soften very fine detail when the reduction is aggressive, and every export re-encodes the pixels, so a lossy format adds a second loss on top of the first. Keeping the quality slider at 90 or above makes that second loss hard to see, and choosing PNG avoids it entirely because PNG is lossless. Enlarging always softens the picture rather than improving it, because interpolation spreads the pixels you already have instead of recovering detail that was never captured. If you need genuine sharpness at a larger size, that is a job for an AI upscaler, not for a resizer.
They become white. JPG has no alpha channel, so transparency cannot be stored in the file at all, and the resizer fills the canvas with solid white before drawing your image on top of it. The practical effect is that a logo or icon cut out on a transparent background arrives as a white rectangle behind the artwork. If the transparency matters, choose PNG or WebP as the output format instead, since both preserve it. If an upload form insists on JPG and the background has to be some other color, flatten the image onto that color before you export, because the white fill is applied by the encoder and cannot be recolored in the resizer.
No. The file is decoded, drawn onto a canvas and encoded again inside your browser tab, so it is never sent to a server, and there is no account, watermark or daily limit. That matters when the image is an ID photo, a signed form or a work document. You can confirm it yourself by opening your browser developer tools, switching to the network panel and resizing an image while it is open, then checking that no request carries the file. The only download involved is the page code itself, which is cached after the first visit. All 45 browser tools on the site follow the same rule.

Resize your image now

Free, private and unlimited. No account needed.

Open Image Resizer