Image Resizer
Resize by width, height, or DPI with contain, cover, or stretch modes and automatic sharpness preservation.
Resizing batch...
Drop images to scale precisely
Screenshots, hero banners, RAW files, and ZIP bundles all render live previews before export.
Sizing queue
- Add assets to preview width, height, and orientation before resizing.
Canvas settings
Drop files to see live previews.
We purge uploads right after delivering the resized outputs back to you.
Resized files will appear here with preview thumbnails and pixel summaries.
The image resizer changes the pixel dimensions of an image. Give it a target width, a target height, both, or a scaling percentage, and the resizer produces a new image at those dimensions using a high-quality resampling filter that preserves edges and fine detail. Up to 50 files per batch, 40 MB per file. Source format, color profile, EXIF metadata, and animation timing all pass through unchanged.
Three Ways to Specify the New Size
- By width only: set the target width in pixels and leave height blank. The resizer computes the height automatically to keep the original aspect ratio. The most common mode for web work, where the available column width is fixed but image content varies.
- By height only: set the target height and leave width blank. The width is computed to keep the aspect ratio. Useful when you have a row layout with a fixed thumbnail height.
- By both width and height: set both target dimensions. If your numbers do not match the source's aspect ratio, the fit mode (see below) decides how the resizer handles the mismatch.
- By percentage: set a scale percentage (50 halves each dimension and produces a quarter of the original pixel count; 200 doubles each dimension). Useful for proportional changes when you do not have a specific target size in mind.
Fit Modes
When you give the resizer both a width and a height that do not match the source's aspect ratio, three modes decide what to do:
- Contain (default): the resizer scales the image down to fit entirely inside the target box without cropping or distorting. The output's aspect ratio matches the source; the actual output dimensions may be smaller than the target box in one direction. The right choice when you must preserve the entire image and pixel-exact dimensions do not matter.
- Cover: the resizer scales the image so it completely fills the target box, cropping the excess from the edges. The output's dimensions match the target box exactly. The right choice for thumbnails, hero backgrounds, and avatars where the target size is fixed and a centered crop is acceptable.
- Stretch: the resizer forces the image to the exact target dimensions even if it changes the aspect ratio. The image will look squashed or stretched. Rarely the right choice, but useful for layout placeholders or when you need precise dimensions regardless of distortion.
When to Resize
- Before uploading to the web: a 24-megapixel phone photo at 6000 by 4000 pixels is several megabytes. The same photo resized to 1600 pixels wide is more than enough for any modern web hero image and loads in a fraction of the time.
- Hitting platform requirements: LinkedIn cover photos want 1584 by 396, YouTube thumbnails 1280 by 720, Twitter headers 1500 by 500. Resize to exact dimensions instead of letting the platform do its own (usually worse) scaling.
- Catalog and product thumbnail systems: e-commerce sites that need every product photo at a consistent size. Batch-resize a folder of shots to a uniform width and the catalog renders cleanly.
- Sending photos by email or chat: Gmail's 25 MB attachment cap and WhatsApp's compression both make oversized images annoying. Resizing to 2000 pixels on the longest side typically cuts the file size by 80 to 90 percent with no visible difference on a phone screen.
- Print preparation at known DPI: a 4 by 6 inch print at 300 DPI needs a 1200 by 1800 pixel image. Resize to exactly those dimensions before sending to the print shop.
What Happens to Image Quality When You Resize
Downscaling (making the image smaller) is essentially lossless from a perceptual standpoint when done with a good resampling filter. Going from 6000 pixels wide down to 1600 wide throws away pixels, but the remaining ones are computed as weighted averages of their neighbors so the image stays sharp and clean. This is why a resized phone photo looks just as good on a screen as the original did.
Upscaling (making the image larger) is fundamentally limited. The resampler can produce a larger image, but it cannot invent detail that was not in the source. A 200 by 200 photo enlarged to 1000 by 1000 has been stretched 5x in each dimension; the resizer fills in the new pixels by interpolating between the original ones, which produces a smoother but not sharper result. The image will look soft compared to a native 1000 by 1000 capture. For genuine resolution increases on low-resolution sources, you need dedicated AI upscaling tools, not a standard resizer.
Pairing With Other Tools
Resizing is often the first step in a workflow. After resizing, the image compressor shrinks the file further at a controlled quality factor; resizing reduces the pixel count, compression reduces the bytes per pixel. The image converter changes the format afterward if you need WEBP for web delivery or PNG for a transparent variant. To trim the resized image to a specific aspect ratio, use the image cropper either before or after.
Batch Resizing and Privacy
Each resize runs in memory on the server. Files stream to the resizing endpoint, decode into a buffer, run through the resampling filter, and return as base64 inside the JSON response. Nothing is written to disk, indexed, logged, or cached. The buffer is released as soon as the response is sent. Up to 50 files per batch, 40 MB per file. ZIP archives are unpacked server-side and each entry counts against the same 50-file limit. The same target dimensions and fit mode apply to every file in a batch, which is the right behavior for processing many similar photos at once (resizing a folder of product shots to 1200 pixels wide, for example). Outputs are returned individually or repackaged into a single download ZIP for batches above one file. EXIF metadata, ICC color profiles, and animation frame timing pass through unchanged.
FAQ
Downscaling is essentially lossless from a perceptual standpoint. The resizer uses a high-quality resampling filter that weighs many surrounding pixels for each new pixel, which preserves edges and texture far better than simple averaging methods. A 6000-pixel photo resized to 1600 pixels wide looks just as sharp on screen as the original did. Upscaling has natural limits: the resizer can produce a larger image but cannot invent detail that was not in the source, so the result will look softer than a native capture at the same resolution.
Two ways. The simplest is to enter only one dimension (width or height) and leave the other blank; the resizer computes the missing dimension automatically to preserve the source ratio. The second is to use percentage mode, which scales both dimensions by the same factor. If you enter both width and height with values that do not match the source ratio, pick the contain or cover fit mode rather than stretch so the aspect ratio is preserved.
Contain fits the entire image inside the target box, preserving the aspect ratio and leaving the output potentially smaller than the box in one direction (no crop, no distortion). Cover fills the target box exactly by scaling the image up until it covers both dimensions, then cropping the excess (no distortion, but the edges of the source may be cut off). Stretch forces the image to the exact target dimensions even if it changes the aspect ratio (distorts the image). Contain is the safe default; cover is for fixed-size thumbnails and avatars; stretch is rarely the right choice.
Common targets: Instagram feed square posts 1080 by 1080, Instagram portrait posts 1080 by 1350, Instagram Stories 1080 by 1920, YouTube thumbnails 1280 by 720, Twitter (X) headers 1500 by 500, LinkedIn cover photos 1584 by 396, Facebook cover 820 by 312, web hero images 1600 to 2400 pixels wide depending on layout. For general web content images, 1200 to 1600 pixels on the longest side handles most cases without burning bandwidth.
Yes. Every frame of the animated GIF (or animated WEBP) is resized to the same target dimensions, and the original frame timing and loop count are preserved. The output is an animated GIF or animated WEBP at the new size, playing at the same speed and looping the same number of times as the source. Resizing animations is one of the most effective ways to shrink an oversized GIF; halving each dimension typically quarters the file size.
File size scales roughly with the pixel count, which scales as the square of the dimensions. Halving each dimension (50 percent scale) quarters the pixel count, so the file ends up roughly a quarter of the original size. Resizing a 6000 by 4000 photo to 1500 by 1000 cuts the pixel count by 16x and typically cuts the JPG file size by 12 to 15x. For further byte savings beyond what resizing alone gives you, run the resized file through the image compressor with a lower quality factor.
Yes, both are preserved by default. EXIF tags (camera, lens, exposure, GPS, capture time) and ICC color profiles pass through the resize pipeline unchanged. To strip GPS coordinates or other personal metadata before sharing a resized photo, run the result through the converter with the strip-metadata option enabled.
Yes. Up to 50 files per batch, 40 MB per file. The same target dimensions and fit mode apply to every file in the batch, which is exactly what you want when resizing a folder of product photos to a uniform 1200 pixels wide, or a set of social media graphics to 1080 by 1080. ZIP archives are unpacked server-side and each entry counts against the same 50-file limit. Outputs are returned individually or repackaged into a single download ZIP.
No. Files stream to the resizing endpoint, decode into a memory buffer, run through the resampling filter, and return in the response. Nothing is written to disk, indexed, logged, or cached. The buffer is released as soon as the response is sent. The tool requires no registration and does not track which images you have resized.
Free with no registration. No rate limits, no watermarks added to outputs, no premium tier with extra features held back. The same applies to all imgdeal tools, including cropping, rotating and flipping, format conversion, and compression.