What resizing actually does
Resizing an image changes its pixel dimensions. A photo that's 4032×3024 pixels coming straight from your phone gets re-rendered as, say, 1200×900 pixels. The image still shows the same scene, but the file is smaller and the picture displays at a smaller size when viewed at native zoom. The Image Resizer does this one job, runs entirely in your browser, and accepts PNG, JPG, and WebP up to 50 MB.
This is the right tool for the most common "my photo is too big" problem: phone cameras produce images at 8-12 megapixels by default, which is overkill for anything that ends up on a screen. A laptop screen is at most 4 megapixels (a 4K display). A phone screen is 2-3. Sending someone a 12 MP photo of your dog so they can look at it on Instagram is wasting bytes — and inflating the file size by maybe 10×.
How this compares to iLoveIMG, ResizeImage.net, and Convertio
The competitors all do roughly the same thing — upload your image, server resizes it, download the result. The differences are in the limits and the trade-offs.
iLoveIMG is the polished one. Bulk resize works on the free tier, file caps at 25 MB, but they pace you with daily limits and aggressive upsells. Their resize quality is fine — server-side ImageMagick with sensible defaults. Trade: your image uploads to their server, sits there for retention, comes back.
ResizeImage.net is a single-page tool with the same upload-then-resize flow. It supports a few extras like rotation and basic cropping. Free tier is generous on file count but the file size cap (around 10 MB) frustrates anyone resizing modern phone photos that often exceed 10 MB straight out of the camera.
Convertio has the same kitchen-sink model as in their image compressor — many formats, server-side processing, 100 MB daily total on the free tier before they push you to subscribe.
The trade we picked: we use the browser's native canvas API with high-quality smoothing turned on. No upload, no file count, no daily cap, no signup. The resize happens in 1-3 seconds depending on size. The quality is what the browser produces — for downscaling (the 95% case) that's effectively identical to ImageMagick. For upscaling, we flag it as a bad idea and tell you what to use instead.
How to use the Image Resizer
Drop or pick your image. The tool reads the original dimensions and file size and shows them at the top.
- Drop or pick a PNG, JPG, or WebP. Up to 50 MB. The tool displays the source dimensions (e.g.
4032 × 3024) and file format right away. - Pick a mode. "By pixels" takes exact target dimensions; "By percent" scales by a percentage (50% means half-size in each direction, which is a quarter of the total pixel count).
- Type your target dimensions. With the aspect-ratio lock on (default), changing the width auto-updates the height to match the original ratio. Uncheck the lock if you need exact non-proportional dimensions — for example a 1200×630 social-media card.
- Pick output quality for JPG and WebP. 92% (the default) is the sweet spot for re-encoded photos. PNG ignores this slider because PNG is lossless.
- Click Resize. The output is named with the new dimensions baked in (
photo.jpgbecomesphoto-1920x1080.jpg).
The resize runs through canvas with high-quality smoothing — the browser's bilinear-or-better interpolation algorithm. For typical downscaling (phone photo to web size), the output is indistinguishable from what a desktop tool like Photoshop or ImageMagick would produce. No bytes leave the tab; check the network panel if you want to verify.
Pixels vs percent: which mode to use
Both modes do the same math, just with different inputs.
By pixels is the right mode when you have a hard target — a specific upload size, a known display dimension, a CMS that requires 1200×630 cards. Type the exact numbers and let aspect-ratio lock fill in the other.
By percent is the right mode when you don't care about exact numbers, you just want it smaller. "Half-size" usually does what you want — a 4032×3024 photo becomes 2016×1512, which is still plenty for any screen, and the file shrinks to roughly a quarter of the original size.
The aspect-ratio lock matters more than the mode. With it on, your image keeps its proportions and doesn't get squished. With it off, you can force any combination of width and height — useful for fitting a 4:3 source into a 16:9 frame, painful if you forget to turn it back on for the next image.
A worked example: shrinking a phone photo for email
Real case: you took a photo on an iPhone 14 — 4032 × 3024 pixels, 4.8 MB JPG out of the camera. You want to email it to a relative. Gmail's attachment cap is 25 MB, but the relative's email provider chokes at 10 MB and their internet is slow. You want the photo to be under 1 MB and still look fine.
| Target size | Dimensions | File size (92% JPG) | Reduction | Notes |
|---|---|---|---|---|
| Source | 4032 × 3024 | 4.8 MB | — | iPhone full-resolution JPG |
| 75% scale | 3024 × 2268 | 2.9 MB | 40% | Still full-screen sharp on a 4K monitor |
| 50% scale | 2016 × 1512 | 1.4 MB | 71% | Sharp on every phone and laptop screen; under most strict email caps |
| 33% scale | 1330 × 998 | 680 KB | 86% | The right answer for email-to-Grandma |
| 25% scale | 1008 × 756 | 410 KB | 91% | Web-thumbnail size; fine for sharing, looks small on big screens |
For the email case, 33% scale (1330 × 998 px, 680 KB) is the right answer. It's under 1 MB so it goes through any email provider. At 1330 px wide it still fills most phone screens edge-to-edge and looks sharp on laptops. The 4.8 MB original would have been a slow upload on the relative's connection and downloaded forever on theirs.
Notice how much more impact resizing has than compression alone. The same photo, compressed at 75% JPG quality but kept at full resolution, would still be around 2.5 MB. Cutting the dimensions does more for file size than cutting quality — by a wide margin. Both work even better together: resize first, then compress with the Image Compressor.
About upscaling (and why our tool flags it)
Upscaling — making an image larger than its source — is supported by the tool but warned about, on purpose. Browsers do bilinear or bicubic interpolation when stretching a canvas. That means each new pixel is computed as a weighted average of nearby source pixels. It's fast, deterministic, and produces a result that's mathematically correct. It's also slightly blurry.
The problem: doubling a 1000×1000 photo to 2000×2000 doesn't add real detail. Where the original had a single sharp pixel, the doubled version has a 2×2 block of similar pixels. Real upscaling — preserving sharpness, hallucinating plausible detail in the new pixel space — needs an AI model trained on millions of image pairs. Tools that do this well include Topaz Gigapixel, ESRGAN-based open-source upscalers, and Upscayl (free, runs locally).
For most use cases — making images smaller, not larger — none of this matters. The Image Resizer is built for that 95% case. If you specifically need to enlarge a low-resolution image and have it stay sharp, use an AI upscaler. If you just need it bigger and don't care about a slight softness (printing a 1000×1000 image at 4×4 inches on a low-DPI printer, for instance), our tool is fine.
Compressing vs resizing: which one to pick
These two operations are often confused. They do different things and have different best uses.
| Operation | What it changes | What stays the same | Best for |
|---|---|---|---|
| Resize | Pixel dimensions (1920×1080 → 960×540) | Quality per pixel | Image displayed at smaller size; need to reduce both file size and display size |
| Compress | Encoding quality (file gets smaller, image looks roughly the same) | Pixel dimensions | Need a smaller file, want the image to still display at original size |
| Both (recommended) | Dimensions and quality | — | Maximum file-size reduction with predictable results |
The combo gives the biggest wins. A 4032×3024 phone photo at 4.8 MB, resized to 1920×1440 and compressed at 75% JPG, lands at around 380 KB — a 92% reduction with no visible loss at typical viewing distance. The same photo compressed alone (kept at full resolution, 75% JPG) is still around 2 MB. The same photo resized alone (1920×1440, 92% JPG default) is around 850 KB. Resize plus compress beats either one alone.
Related image tools
Image Resizer fits inside a broader image toolset:
- Image Compressor — change file quality without touching dimensions. Pair with resize for maximum size reduction.
- Image Format Converter — change between PNG, JPG, and WebP. Converting a PNG photograph to WebP typically saves more than resizing the PNG.
- Image Cropper — cut out a section of an image. Often better than resize when you only need part of the picture.
- PNG to WebP — specifically for the PNG-to-WebP conversion case.
- HEIC to JPG — for the iPhone-photo-format problem before you resize.
Frequently asked questions
Does the aspect-ratio lock keep my image from getting distorted?
Yes. With the lock on (default), changing the width auto-updates the height to keep the original aspect ratio — your image won't get squished or stretched. Uncheck the lock to set both dimensions independently, which is useful for fitting an image into an exact non-proportional space like a 1200×630 social-media card, even if the source is 4:3.
Is my image really not uploaded?
Correct. Decoding uses the browser's native image loader; resizing uses canvas with high-quality smoothing; encoding uses canvas.toBlob(). Zero outbound requests during the resize. Check your browser's network tab if you want to verify.
Why does the warning appear when I upscale?
Because the browser uses bilinear interpolation for canvas resizing. Doubling the resolution of a 1000×1000 photo to 2000×2000 doesn't add real detail — it averages neighboring pixels, which produces a soft, slightly blurry result. For real upscaling (preserving sharpness, adding plausible detail), you need an AI-based upscaler (Topaz, ESRGAN-based tools, Upscayl). For most use cases — making images smaller, not larger — this tool is exactly right.
What output formats do you support?
The same format as the input. PNG stays PNG (with transparency preserved). JPG stays JPG. WebP stays WebP. To change formats and resize, run through this tool, then through our Image Format Converter — or vice versa.
What's the maximum size?
50 MB per image file. Output dimensions are capped at 16000×16000 px — anything larger may exceed canvas size limits on some browsers. For most use cases (phone photos, screenshots, web images), the cap is far above what you'd need.
Does the output preserve EXIF metadata?
Mostly no — canvas re-encoding strips most EXIF. Camera model, GPS coordinates, exposure data are lost. For users sharing photos online, this is usually a feature (it strips GPS coords you didn't want to share). For archival use cases where EXIF matters, use a tool that preserves it — ImageMagick or ExifTool can resize without re-encoding.
How does this differ from compressing?
Resizing changes the pixel dimensions (1920×1080 → 960×540). Compressing keeps the dimensions but changes the encoding quality (1920×1080 → 1920×1080, smaller file). Both shrink the file; resizing also visually shrinks the image when displayed at native size. For maximum size reduction: resize first, then compress.
Does this work for batch resizing?
Not in v1 — one image at a time. Multi-image batch resize is on the roadmap. For now, run images through individually or use a desktop tool (XnConvert, ImageMagick) for big batches.