Resize Images

Both get lumped under "making the photo smaller", but they do different things. Resizing reduces the number of pixels: turn a 4000×3000 photo into 1080×810 and it drops from 12 million pixels to about 870 thousand, appearing that much smaller on screen. Compressing keeps the pixel count and raises the compression ratio to lighten the file, so the on-screen size does not change.

So pick the tool by the problem. If the photo displays too large — say it overflows a page layout — use this tool. If the file is too heavy — an attachment over the 5 MB cap, for instance — use the image compressor, which can auto-adjust quality to hit a target size in KB. Of course, fewer pixels also means a smaller file, so when a file is massively oversized, shrinking its dimensions here first gives a far cleaner result than cranking up compression alone.

Instagram mostly uses 1080×1080 for squares, 1080×1350 for portrait, and 1080×566 for landscape. Portrait fills the most screen, so many accounts now default to 1080×1350. YouTube thumbnails are standardized at 1280×720 — upload larger and the platform scales it down for you.

For blog posts, content columns usually run around 800–900px wide, so a longest side of 1200–1600px looks sharp without weighing the page down. A handy rule of thumb for high-DPI screens is roughly twice your content width.

Keep in mind that platforms change these numbers all the time. Instagram's square standard climbed from 640 to 1080 over a few years, and recommended ratios keep shifting. For anything important, check the service's current guidelines. When in doubt, go a little bigger — you can always shrink later, but enlarging cannot be undone.

Scale an 800×600 photo up to 1600×1200 and the pixel count quadruples — but the information for those new pixels does not exist in the original. All software can do is blend neighboring colors and invent plausible values, and the result is a soft, mushy image. Detail that was never captured cannot appear.

That is why "Never enlarge beyond the original" is on by default here: if your target is larger than the original, the image is only scaled up to its original size, ratio intact. If you genuinely need a bigger file, do not upscale — go find a bigger original. Dig up the camera file, or for screenshots, raise the display scaling and capture again. For print, work out the required dimensions at 300dpi and source an original to match.

Changing the aspect ratio always cuts something. Crop a 16:9 landscape shot to a 1:1 square and big slices of the left and right disappear; force a 9:16 portrait shot into 16:9 and the top and bottom vanish entirely. This tool keeps the largest possible area of the original that fits the target ratio, minimizing the loss — but if the ratios differ, some trimming is unavoidable. Once you add a photo, the preview above darkens the parts that will be cut so you can check before downloading, and dragging the bright rectangle lets you fine-tune the position beyond what the nine anchor buttons offer. In a batch, the position you set on the first photo is applied proportionally to the rest.

The default anchor is center, which works fine for landscapes and objects. But portraits often do better with the top anchor. Faces usually sit in the upper part of the frame, so center-cropping a portrait shot to a square tends to slice off foreheads while keeping the floor at someone's feet. Move the anchor to the top and the face survives. For group shots or full-body photos, try a few anchors and judge by the result sizes shown in the table.

The result is 1px off from what I set

With Keep aspect ratio or longest-side mode, the shorter side is rounded to the nearest whole pixel. Set the longest side of a 4032×3024 photo to 1920 and the short side comes out exactly 1440.0 — but a 4032×3016 original gives 1436.19, which rounds to 1436. Pixels cannot be split, so this is unavoidable. If both dimensions must be exact, switch to exact-size mode, turn off Keep aspect ratio, and type both values. If those differ from the original ratio the image will look stretched, so it is better to enable Crop to ratio alongside.

My portrait photo comes out sideways

Phones always store photos in one fixed orientation and record which way is up in the EXIF metadata. Programs that ignore that data show the photo lying on its side. This tool honors EXIF orientation and processes the image exactly as you see it, so it will not tip over. If a file was already saved sideways somewhere else, rotate it in an image editor first, then bring it here.

Converting a transparent PNG to JPG gave me a black background

JPG cannot store transparency, so transparent areas must be filled with some color. This tool fills them with white. A black fill is the classic symptom of software that saved without filling at all. To keep the transparency, set the output format to PNG or WebP. To fill with a color of your choice, the image converter has a background color picker.

It gets slow with many images

All the work happens in your own browser. Nothing is sent to a server — which keeps private photos safe — but speed depends on your device. Camera originals over 4000px take a moment just to open. To keep memory in check the tool processes one image at a time and frees each as it finishes, so even dozens of images will not freeze the browser. Just let the progress bar do its thing.