The file never leaves your device. The file is opened in your browser and the images taken out. Anything stored as JPEG is returned byte for byte without re-compression, so nothing is lost.
Rather than capturing the page, this pulls out the images stored inside a PDF, HWP or HWPX file at their original quality. Nothing is re-compressed, so nothing is lost. Nothing is uploaded either.
The file never leaves your device. The file is opened in your browser and the images taken out. Anything stored as JPEG is returned byte for byte without re-compression, so nothing is lost.
Pulls the images embedded in a PDF, HWP or HWPX back out at their original size. This recovers the pictures stored inside the file rather than screenshotting the page.
A screenshot only keeps what your display shows. Documents usually still hold the picture that was inserted, often far larger — which is how you get an original back after losing it.
Images inside HWP and HWPX come out as well, without needing the Hangul word processor installed. Useful on a Mac or Chromebook that received a Korean document.
Opening the document and extracting from it both happen in the browser, so internal or contractual documents stay put.
A PDF from a scanner holds one large JPEG per page. Extracting therefore gives you each whole page as a picture — a direct way to turn a scan into JPGs. A PDF exported from Word or Hangul, by contrast, stores text as text, so only the inserted pictures come out.
Formats that cannot be lifted out as they are, such as JPEG 2000, fax compression or CMYK, take a second route: the PDF rendering library is loaded and the image is decoded to pixels and saved as PNG. What still fails, Windows metafiles that are drawing commands rather than pictures, and embedded objects such as tables and equations are not extracted; the reason is listed instead. Tiny images used as icons or masks are left off the list so that hundreds of them do not pour out.
Hangul documents quite often contain a file named .jpg that is actually a PNG. The format is read from the first few bytes and the file is saved with the correct extension, so nothing you extract fails to open. JPEGs are returned byte for byte, without recompression, so no quality is lost.
The name must end in .pdf, .hwp or .hwpx, but which parser opens the file is decided by its first few bytes, not its name. Hangul sometimes saves HWPX under a .hwp name, and files with the extension simply changed to .pdf circulate too. Such files open as long as the content is one of the three; if it is none of them, the tool says so.
In a PDF, any image whose width times height is under 400 pixels is not counted at all. Bullet symbols, fragments of underline and transparency masks are that size, and a document can hold hundreds of them, burying the photographs. They are left off the skipped list as well. If you really need a tiny icon, zoom in and take a screenshot.
The document's graphics are shapes or text rather than pictures. Tables, arrows and text boxes are drawing instructions, so there is no image to extract.
It was already shrunk when it was inserted. Nothing can recover detail the file does not contain.
That is a different job — use Save as PDF from the print dialog, or a screenshot.
Logos in headers and background patterns are stored as images too. Use the previews to pick the ones you want.
Downloading many files in a row prompts the browser for permission. Allow it and they save in turn, named after the original with a 01, 02 sequence appended.
Under each preview are the pixel dimensions, format and file size. Save only the ones worth keeping.
HWPX is a compressed folder with the pictures stored as files; HWP holds them inside one container. Either way they come out on the same screen.
Before opening an HWP the flags in its header are checked; if it is password-protected or saved as a distribution (DRM) copy, the tool says so and stops. Remove the password by saving again from Hangul; a distribution copy has to be re-issued as a normal document. PDFs get no such check, so make an unprotected copy of a locked PDF first.