Извлечение изображений

Each image comes out in the format it's stored in inside the PDF, most often JPEG or PNG, with no re-encoding. Files are numbered in the order they're found, for example image_1.jpg and image_2.png, and Download All packs them into extracted-images.zip.

It only pulls out embedded raster images. If the graphics are vector drawings, charts, or text drawn as paths, there's nothing to extract, and you'd need to render the page as a picture with PDF to PNG instead.

You get the image data exactly as it's embedded, at its stored resolution rather than the size it appears on the page. If whoever made the PDF downsampled the images before embedding them, that lower resolution is all that exists in the file.

A scanned PDF stores each page as one large image, so you'll get one full-page picture per page rather than the individual photos or figures on it. Crop what you need from those page images afterward.

Images are collected page by page, so a logo or header graphic that's reused on many pages is extracted once for every page it appears on. Keep the first copy and discard the rest.

Upload as many PDFs as you want and click Extract once. Images from all the files are numbered continuously and shown in a single grid, and Download All puts everything into one ZIP.

There are no page or selection options; every embedded image on every page is extracted. You can still save only the ones you want from the preview grid, since each thumbnail has its own download button.

No. The PDF is opened by a PyMuPDF engine running as WebAssembly inside your browser, and the images are pulled out of the file in memory on your device. The preview grid, the individual downloads, and the ZIP are all generated locally.