Bildstorleksändring

Resize images by pixels, percent, or social preset

This image resizer gives you three ways to set a new size. Use exact pixels when you know the width and height you need — type one value with the aspect-ratio lock on and the other calculates automatically. Use percentage to scale proportionally to the original size, so 50% halves a photo and 25% makes a quick thumbnail. Or use a preset to jump straight to a platform size like Instagram Post (1080x1080) or YouTube Thumbnail (1280x720) without guessing at numbers.

A live preview card shows each result with its new dimensions and estimated file size before you download. Switch between PNG, JPEG, and WebP output at any time — the tool handles transparency and background fill correctly for each format. Works on any photo type: JPG, PNG, WebP, GIF.

True bulk and batch image resizing

Most free resizers handle one picture at a time. This is a proper bulk image resizer: drop in a whole folder of photos, set dimensions or percentage once, and every image resizes with the same settings instantly. Each appears as its own card with a Download button, or grab the entire batch as a single ZIP archive in one click.

It's the right tool for product photos, blog images, app icons, email attachments, or any situation where you need a consistent size across many files. Competitors like BulkImageResize.com and Bulk Resize Photos send all your files to a remote server. This tool processes everything locally — a hundred images is just as private and fast as resizing one.

Resize without losing quality — the honest answer

To resize an image without losing quality, scale down, not up. Downscaling discards only the pixels you don't need and stays crisp using high-quality interpolation. Export as PNG or WebP for lossless results; with JPEG, keep quality high (90% or above) to avoid visible compression artifacts.

Enlarging is a different story. You cannot genuinely enlarge an image without losing quality using any non-AI method — stretching existing pixels over a larger canvas makes them soft and blocky. There's no fake 'AI upscale' here. If you need a larger image, start from the biggest original you have. Tools like Topaz Gigapixel AI can do AI upscaling at around $99/year, but that's a different product that runs server-side. For everything else — resizing down, targeting specific dimensions, or batch processing — this tool handles it.

Fit modes, social presets, and target file size

When the proportions of your photo don't match your target dimensions, fit modes handle the gap. Contain fits the entire image inside the box and can pad the edges with a background color to reach the exact pixel count. Cover fills the box completely and crops the overflow — ideal for square avatars and social media tiles. Stretch forces the exact width and height when you specifically need those dimensions regardless of distortion.

Built-in presets cover Instagram, Facebook, X/Twitter, LinkedIn, Pinterest, YouTube, and common profile picture and passport-photo sizes. The target file size mode resizes an image to a specific number of kilobytes — the right tool for government forms, school portals, and job applications with strict file size limits like 20 KB, 50 KB, or 100 KB.

Competitor comparison — free online image resizers

iLoveIMG offers a resize tool that works well, but it uploads your images to their servers and limits free users to files under 5 MB per batch. TinyWow also processes files server-side and has added a sign-in requirement as of 2024. Adobe Express has a solid resize interface but requires a free Adobe account and uploads to Adobe's cloud. Simple Image Resizer is fast but server-side, and its interface hasn't been updated in years. Squoosh by Google is browser-based and excellent for compression, but handles only one image at a time.

UtiloKit Image Resizer does everything in your browser: no upload, no account, no watermark, no file size cap, no daily limit. It adds features those tools don't have — true batch processing with ZIP download, a target-KB mode, custom fit modes, and social media presets. If you've been using iLoveIMG for quick resizes and want something that doesn't send your photos to a server, this is a direct replacement.

Output formats and privacy — runs entirely in your browser

Every resize happens on an HTML canvas inside your own browser. Your images are never uploaded to a server, never stored anywhere, and never seen by anyone but you — which is exactly what you want for personal photos, ID documents, medical scans, or confidential work files. It's an online tool that behaves like a local desktop app.

For output format: choose PNG for lossless quality and preserved transparency, JPEG with adjustable quality (0-100%) for smaller files, or WebP for the best balance of quality and file size. The tool works in Chrome, Edge, Firefox, Safari, and all major mobile browsers. No install, no account, no file size cap — resize as many images as you need, any time.

Image resampling algorithms: what happens to pixels when you resize

Every resize operation involves a mathematical decision: when the output canvas has a different number of pixels than the input, how does software calculate each new pixel's color value? The simplest method is nearest-neighbor interpolation, which assigns each output pixel the value of the closest input pixel. It's computationally instant and preserves hard edges in pixel art, but produces a visibly blocky 'staircase' effect when upscaling photographs. Bilinear interpolation improves on this by computing a weighted average of the four nearest input pixels, producing smoother gradients — but the smoothing can make enlarged images look soft or blurry at high magnification.

Bicubic interpolation uses a 16-pixel neighborhood and fits a cubic polynomial curve through the surrounding values, producing sharper results than bilinear with less noise. It's been Photoshop's default resampling method for decades and is the standard for professional photo resizing. Lanczos resampling (also called Lanczos3) goes further still: it applies a sinc-based filter across an even wider pixel neighborhood to minimize ringing and aliasing artifacts. Lanczos is widely considered the highest-quality algorithm for downscaling photographs, though it is computationally heavier. Modern image CDNs like Cloudflare Images and imgix use Lanczos by default for exactly this reason.

At the frontier is AI-based upscaling, where convolutional neural networks — trained on millions of image pairs — generate plausible fine detail rather than interpolating existing pixels. Tools like Topaz Gigapixel AI, Adobe Photoshop's Super Resolution (powered by Sensei), and the open-source ESRGAN model can upscale an image by 4× while synthesizing realistic texture that looks convincing to the eye. This is categorically different from mathematical interpolation: the network is inventing detail based on learned patterns, not computing it from the source pixels. For portraits and landscapes the results can be extraordinary; for images with text or precise geometry, artifacts can still appear.

Resolution, PPI, and the screen pixel density landscape

The phrase 'save at 72 PPI for the web' has been repeated so often it has become dogma — and it is almost entirely obsolete. The number originates from the original 1984 Macintosh display, which had a physical pixel density of 72 dots per inch, so a 72 PPI image mapped one-to-one with screen pixels. Modern displays bear no resemblance to that figure. A standard 1080p desktop monitor at 24 inches runs at roughly 92 PPI. A 4K monitor at 27 inches is around 163 PPI. An iPhone 16 Pro's Super Retina XDR display is 460 PPI. A Samsung Galaxy S24 Ultra reaches 505 PPI. The 72 PPI figure in an image's metadata has zero effect on how a browser renders it — what matters is the pixel count in the file, not the embedded PPI value.

Device Pixel Ratio (DPR) is the genuinely important concept for screen-bound images. DPR is the ratio of a device's physical pixels to its CSS pixels. On a DPR=2 screen (all modern iPhones, most Android flagships, Retina MacBooks), one CSS pixel is rendered using a 2×2 grid of four physical pixels. A 400×400 CSS image on a DPR=2 screen will look sharp only if the image file is 800×800 — the browser scales it down by half, filling each CSS pixel with four physical ones. On a DPR=3 screen (iPhone 15 Pro, Pixel 8 Pro), a crisp image needs 3× the CSS dimensions in actual pixels.

The HTML srcset and sizes attributes exist precisely to serve the right resolution to each device without wasting bandwidth. A markup pattern like <img src="photo.jpg" srcset="photo-2x.jpg 2x, photo-3x.jpg 3x" alt="..."> lets the browser automatically fetch the 2× version for Retina screens and the 3× version for high-DPR phones, while older desktop browsers fall back to the 1× original. When you use this image resizer to export multiple sizes of the same photo, you are preparing exactly the assets this pattern needs.

Platform image size requirements: social media, e-commerce, and web

Every major platform enforces its own optimal dimensions, and uploading the wrong size wastes quality — either the platform downscales your image (discarding resolution you provided) or upscales it (stretching pixels and revealing compression artifacts). For Instagram: square posts render at 1080×1080 px, portrait posts at 1080×1350 px (the highest-reach aspect ratio at 4:5), and landscape posts at 1080×608 px. Stories and Reels are 1080×1920 px (9:16 vertical). For Facebook: cover photos are 820×312 px on desktop; shared link images render at 1200×630 px. X/Twitter headers are 1500×500 px. LinkedIn profile photos require at least 400×400 px, and banners are 1584×396 px. YouTube thumbnails are 1280×720 px (16:9), with a 2 MB file-size cap.

For e-commerce product photography, the requirements are driven by zoom features. Amazon's main product image requires a white or pure-white background with the product filling at least 85% of the frame; the minimum for zoom to activate is 1000×1000 px, and 2000×2000 px is the recommended size to ensure crisp zoom at all viewport sizes. Etsy recommends a minimum of 2000 px on the longest side. Shopify's image guidelines suggest 2048×2048 px for square product images, enabling their built-in photo zoom without any additional CDN configuration. Uploading 800×800 px images to any of these platforms disables zoom and makes products look underproduced compared to competitors who supply high-resolution assets.

For web performance, the sizing philosophy reverses: use the smallest file that still looks sharp at your target display size. Full-width hero images rarely need to exceed 1920×1080 px, and on most websites 1600 px wide is sufficient. Blog post inline images are rarely displayed wider than 800 px. Avatar and thumbnail images in grids are typically 200–400 px wide. Favicons follow a specific ladder: supply a 512×512 PNG for the web app manifest, which browsers and operating systems scale down; the traditional ICO file embeds 32×32 and 16×16 variants. Matching your image dimensions to the actual display size — rather than uploading the raw camera file and letting the browser scale it — is one of the highest-leverage improvements available for Google Core Web Vitals and overall page speed.

The megapixel myth: when resolution matters and when it doesn't

Camera manufacturers have long used megapixel count as a headline specification, and consumers have learned to equate more megapixels with better image quality. This relationship is real but limited. Sensor size, pixel pitch, and lens quality matter far more than raw megapixel count. A 50 MP image from a 1/2.5-inch smartphone sensor — where each photosite is tiny and captures little light — is frequently lower quality in terms of dynamic range, color accuracy, and noise than a 12 MP image from a full-frame mirrorless camera where each photosite is much larger and captures far more photons per exposure. When you view a 50 MP phone image and a 12 MP camera image at the same output size, the 12 MP image often looks cleaner, especially in any scene that isn't flooded with light.

High resolution is genuinely useful in specific contexts. Cropping flexibility is the most practical: a 50 MP image lets you crop to 25% of the frame and still have a 12.5 MP image — enough for most uses. Large-format print demands high pixel counts: an A4 sheet printed at 300 DPI (the standard for photographic quality printing) needs 2490×3510 pixels — about 8.7 MP. A poster at 24×36 inches at 300 DPI needs roughly 65 MP. But billboard printing is typically done at 72 DPI or even lower because viewing distance is 10+ feet, making even a 10 MP image more than sufficient. For most web publishing, images displayed at 800 px wide on a DPR=2 screen need only 1600×1067 px — about 1.7 MP.

The practical cost of excess resolution is concrete. Storage and bandwidth scale quadratically with linear resolution: double the megapixels and you roughly double the file size (for the same content and quality). A 50 MP RAW file from a Sony A7R V is around 100 MB; a single second of 4K RAW video from an Arri camera is around 25 MB per frame at 24fps. For workflows that involve resizing images down to web or social dimensions, starting from an overly large file simply adds processing time and storage overhead with no quality benefit at the output stage. The right resolution strategy is to capture at the resolution your largest output demands, and export at the resolution each delivery channel actually needs — which is precisely the workflow a batch image resizer supports.