How it works
- You add an image and set either a target width and height in pixels or a percentage scale. The file is decoded once in this tab into an ImageBitmap; nothing is uploaded.
- The bitmap is drawn onto a canvas at the target dimensions using the browser's high-quality resampling. When you are reducing by a large factor, the reduction is done in successive halving steps rather than one jump, because a single large downscale samples too sparsely and produces aliasing on fine detail.
- Aspect ratio is preserved unless you explicitly unlock it. Setting only a width computes the height, so an image cannot be silently squashed by a mistyped number.
- The canvas is encoded back to the source format where the browser supports it — JPEG stays JPEG, PNG stays PNG — so a resize does not quietly change what kind of file you have.
- The result is a blob URL held only by this tab, with the new dimensions and byte size measured from the actual output rather than predicted.
Limitations
Enlarging past the original resolution adds no detail. Scaling a 400-pixel-wide image to 2000 pixels produces a 2000-pixel image containing 400 pixels of information, interpolated. It will look soft, and no setting here changes that — the detail was never captured. Upscaling that genuinely adds plausible detail requires a generative model, which is a different tool with different trade-offs.
Very large images hit a hard browser limit. Canvas dimensions are capped by the browser and by available memory, and the cap is lower on mobile Safari than on desktop Chrome. An image beyond roughly 16000 pixels on a side will fail rather than degrade, which is why the error message tells you to resize in two passes instead of pretending.
Metadata does not survive. EXIF, colour profiles and copyright fields are dropped by the canvas round-trip. Losing the colour profile matters for images destined for print, where the difference between sRGB and Adobe RGB is visible; for web use it is usually the desired outcome.
Resizing a JPEG re-encodes it, so a resize is also a compression pass whether you wanted one or not. Resizing the same file repeatedly compounds that loss in the same way repeated compression does.
Questions
Is there a file size limit?
No cap on file size, but there is a practical cap on dimensions. The decoded image and the target canvas both have to fit in memory, and browsers refuse canvases beyond a certain size — around 16000 pixels on a side on desktop, less on phones.
Can I make a small image bigger?
You can set larger dimensions, but you will not gain detail. The result is an interpolated, softer version of the same information. If a form demands a minimum resolution, this will satisfy the form; it will not make the photograph sharper.
Does the aspect ratio stay the same?
Yes by default. Enter one dimension and the other is computed. You can unlock it deliberately if you need a specific box, and the preview will show the distortion before you download.
Does resizing reduce the quality?
Downscaling loses detail by definition, though a well-resampled reduction usually looks sharper on screen than the original did at that size. The re-encode also applies JPEG compression again, so work from the original rather than resizing a resized copy.
Do I need an account?
No. There is no sign-up, no email step and no usage counter. The tool is paid for by the ads on this page.
When people use it
Specific dimensions demanded by a system are the usual trigger: a profile photo that must be 400 by 400, a document upload that requires at least 600 pixels wide, a marketplace that rejects anything over 2000. These are pass-or-fail numbers, and the only question is hitting them without wrecking the image.
The other use is bandwidth. A phone photograph is four thousand pixels wide and will be displayed in an eight-hundred-pixel column. Serving the full-resolution file wastes most of what it downloads. Resizing to the size it will actually be shown at is a larger saving than any amount of compression on the oversized original.
Photos resized to a system's required dimensions often still have to arrive as one attachment rather than as eleven files. Images to PDF puts them into a single document, one image per page, in the order you set.
How it compares
Doing this in the browser means a photograph is never uploaded to change a number that describes it, and there is no queue and no watermark. What a desktop tool or a build pipeline does better is batches and responsive sets — generating five widths of two hundred images belongs in a script, and tools like ImageMagick or a framework's image component will also pick modern formats per browser, which a single manual resize cannot do.