How to crop and resize images in bulk
Drop your images into the box above, or click Choose files. JPG, PNG, WebP and AVIF work, up to 500 at a time, and videos can go in the same batch.
Pick the size: 512×512, 768×768 or 1024×1024 for AI training, 1080×1350 for Instagram, 1920×1080 for video, or your own.
Every image shows a frame of that shape. Drag a frame to keep a face or subject in view; double-click to centre it again.
Choose the format and quality, click Crop and resize, and download them all as a ZIP.
Why crop images to one size?
Training an image model (Stable Diffusion, LoRA, DreamBooth) or a classifier needs every picture at the same size, usually a square like 512×512 or 1024×1024. Stretching photos to fit distorts them; cropping each one keeps faces and objects in proportion.
Galleries, product listings, thumbnails and social posts look tidy when every image has the same shape. Doing it one by one in an editor takes ages; here the whole set is done at once.
Crop, then resize
Each image is first cropped to the target's shape, as large as it fits, then scaled to the exact pixel size. Nothing is stretched or padded. Images smaller than the target are enlarged to fit, which can look soft; the result says so.
Resize image keeps each picture's shape and only changes its size. Crop image lets you draw any frame on one picture. This tool does both for many pictures at once, with one shape for all.
Is it safe?
Yes. Bulk Crop & Resize runs entirely in your browser. Your files are read and converted by your own device using built-in browser technology, so they are never uploaded to a server and nobody else can see them.
You can check this yourself: open your browser's developer tools, switch to the Network tab and convert a file. No upload requests are made.
Images are cropped and resized in your browser; they never leave your device. Location and camera details (EXIF) aren't copied into the new files.
Files stay in this tab's memory only until you remove them or close the page.
Frequently Asked Questions
What size should images be for AI training?
Use the size your model trains at: 512×512 for Stable Diffusion 1.5, 768×768 for 2.x and 1024×1024 for SDXL and most newer models. Check your training tool's settings if unsure.
Can I choose which part of each image is kept?
Yes. Every image shows its frame; drag it, or select it and use the arrow keys, to keep a face or subject inside. Double-click to centre it, or use Centre all.
What happens to images smaller than the target size?
They're enlarged to fit, so every file comes out at exactly the size you picked. Enlarged images can look soft; for training, it's better to leave out images much smaller than the target.
Which format should I choose?
Keep the original format, or choose PNG for lossless files (common for training) and JPG or WebP for smaller ones. Quality applies to JPG and WebP.
Can I crop photos and videos together?
Yes. Add both to the same batch, for a video model's training set or an Instagram carousel: every file comes out at the same size, photos in the format you choose and videos as MP4.
How many images can I do at once?
Up to 500 per batch. Download them as a ZIP, then add the next batch.