About this tool
Resizing changes how many pixels wide and tall an image is. Photos from a modern phone are often over 4000x3000 pixels, and posting one of those straight to a blog or a shop listing means visitors download several megabytes to see something that renders at a fraction of that size. The picture looks identical; the page just takes longer to load and eats more mobile data. Scaling images down to the size they will actually be displayed at is one of the cheapest ways to make a page feel faster.
Going the other way deserves caution. Shrinking discards information, which is always possible. Enlarging invents information that was never captured, which is not. Blowing a 300x200 image up to 3000x2000 gives you more pixels but not more detail. Keep the original and resize a copy. This tool works in either pixels or percent, locks the aspect ratio for you, and uses bicubic interpolation so downscaled edges stay smooth instead of going jagged. Your image is never uploaded. Editing and saving both happen inside your browser, so you can open files you would not want to hand to a server — photos with faces in them, internal documents, anything sensitive.
How to use
- Click "Pick an image & start" to open a file. You can also drag a file onto the window.
- Choose your unit: pixels when you have an exact target, percent when you just want "half the size".
- Enter a width or a height. With aspect-ratio lock on, the other value follows automatically.
- Apply, and the canvas updates immediately. Press Ctrl+Z if you want to try a different number.
- Go to File > Export and save as PNG, JPEG, or WebP.
Frequently asked questions
- Does resizing down hurt image quality?
- At the new size, barely at all. Downscaling merges pixels together, so information is discarded, but the image is displayed proportionally smaller and the loss is hard to see. The problem appears if you later enlarge that smaller file — the discarded detail does not come back.
- Can I make a small image bigger?
- You can add pixels, but not sharpness. Missing detail is estimated from neighbouring pixels, so edges go soft. Up to about 1.5x is usually acceptable; beyond that it shows. If you need a large image for print, start from a large original.
- What happens if I turn off aspect-ratio lock?
- The picture stretches or squashes. It is most obvious on faces and circular logos. If you have to hit an exact size that does not match the original ratio, scale proportionally first and then crop away the excess — that looks natural, stretching does not.
- What size should I use for the web?
- For photos in article bodies, 1200-1600 pixels wide covers most cases. High-density screens mean it is sensible to supply roughly twice the display width. Small thumbnails work at 400-600 pixels, and full-width background images at around 1920.
- Does resizing also reduce file size?
- Yes, and it is usually the biggest win available. File size roughly tracks pixel count, so halving both width and height leaves a quarter of the pixels. That is why resizing should come before compression — and doing both together goes further still.
- What happens to EXIF data like date and GPS?
- It is not carried over. Exporting redraws the edited canvas into a new file, so the original metadata does not come with it. If you are worried about publishing photos that quietly reveal where they were taken, that works in your favour. If you need the capture data, keep the original separately.
- Is transparency preserved?
- With PNG and WebP, yes. JPEG has no way to store transparency, so transparent areas are filled with white on save. Use PNG or WebP for logos and icons that need to sit on any background.
- Can I resize many images at once?
- Yes — the editor has Automate > Batch, where you can select multiple files and process them together. Useful for product shots or profile pictures that all have to match one spec.
- What is bicubic interpolation?
- It is one way of deciding the colour of pixels that did not exist before. Copying the single nearest pixel produces hard, stair-stepped edges; bicubic looks at the surrounding 16 pixels and blends between them. That suits photographs. It is the wrong choice for pixel art, where the hard edges are the point.