There are two completely different ways to get pictures out of a PDF, and most people pick the wrong one. You can extract the images the document actually contains, or you can render each page and photograph the result. The first gives you the original files at full resolution. The second gives you a screenshot. This guide covers when to use which, and how to do the first one properly.
How to extract images from a PDF in under a minute
- Open ihatepdf.cv/extract-images — no account, no email, nothing to install.
- Drop your PDF in, or click to select it. The file is read inside your browser and is never uploaded.
- Review what was found. Every embedded image appears with its true pixel dimensions, file size, and the pages it is used on.
- Filter out the noise. Images under 64 pixels are hidden by default — those are almost always spacers and rules, not pictures.
- Save what you want — click any single image, or download the whole selection as a ZIP.
Why extraction beats screenshotting, by a lot
A PDF does not store a picture of each page. It stores a list of drawing instructions, and any photographs placed on those pages are embedded as separate image files inside the document. Crucially, an image is stored once at whatever resolution it was embedded, then drawn at whatever size the layout calls for.
That gap is usually enormous. A product photo occupying a two-inch box on the page might be a 3000-pixel original underneath. A logo displayed at 40 pixels might be a 1200-pixel master. When you screenshot the page or export it as an image, you capture the scaled-down rendering and throw the rest away. When you extract, you get the file that is actually in there.
This is why extraction is the right tool for a scanned document in particular. A scanned page is typically one large image per page, captured at 300 DPI or better. Extracting hands you that scan exactly as the scanner produced it.
Extract images vs PDF to JPG — which do you want?
Use Extract Images when you want the photographs, logos, charts or scans that someone placed into a document. Use PDF to JPG when you want a picture of the page itself — text, layout, background and all — for a thumbnail, a preview, or a slide.
A quick test: if you would be annoyed to see the surrounding text in the result, you want extraction. If the layout is the point, you want conversion.
Why you sometimes get twelve strips instead of one photo
Open a one-page PDF with a single photograph and you may get back a dozen files. This is normal and it is not a bug in the extractor. Word processors, design tools and print drivers routinely slice a large image into horizontal bands or square tiles before embedding it, because smaller chunks compress better and stream faster over a slow connection.
The PDF genuinely contains those pieces as independent images, so an honest extractor returns them as it finds them. Reassembling them is not reliably possible — the tool cannot always know their arrangement. If you need the image as it appears on the page, that is the case where PDF to JPG is the better answer.
And why you sometimes get fewer images than you can see
The reverse also happens. A logo repeated in the header of every page is stored exactly once and referenced from each page. A fifty-page document therefore yields a single logo file, credited to all fifty pages, rather than fifty identical copies. That is the correct result and it is what you want — but it surprises people who counted the logos on screen.
What about transparency?
PDFs store transparency as a separate greyscale image called a soft mask, held alongside the picture it applies to. A naive extractor returns the two separately, leaving you with a cut-out logo on an opaque black or white rectangle and a mysterious greyscale file you have no use for. Folding the mask back into the alpha channel of the exported PNG is what keeps a transparent background transparent, and it is worth checking that whatever tool you use actually does it.
The privacy problem with every other extractor
Nearly every online image extractor works the same way: you upload the PDF, their server opens it, and you download the results, with a note reassuring you that the file is deleted afterwards. That is a reasonable engineering choice and a poor privacy one. For a contract, a medical scan, an unpublished manuscript, a client deliverable or anything under NDA, the document has already left your control by the time the deletion promise applies.
The ihatepdf extractor decodes images with WebAssembly inside your own browser tab, using the same codec work that powers the compressor. There is no upload, so there is nothing to delete, log or breach. Once the page has loaded you can disconnect from the internet entirely and the extraction still completes.
Which formats come back
An image stored in the PDF as a JPEG is written out byte-for-byte identical to the file that was embedded — nothing is re-encoded, so there is no second round of compression damage. Everything else — Flate, CCITT fax, indexed-palette, CMYK, 1-bit stencils, greyscale scans — is decoded and saved as PNG, which is lossless. Two rare formats, JPEG 2000 and JBIG2, are not supported and are reported as skipped rather than saved incorrectly.
If the PDF is locked
An encrypted PDF has encrypted image data, so nothing can be read out of it while it is locked. Run it through Unlock PDF or Remove Password first, then extract. A good extractor should tell you the file is encrypted rather than reporting "no images found", which is the same output for a completely different problem.
Frequently asked questions
Are the extracted images the same quality as the originals?
Yes. JPEGs come out byte-for-byte identical to what was embedded, and everything else is written as lossless PNG. Nothing is re-compressed at any point.
Why are the extracted images bigger than they looked in the document?
Because you were looking at them scaled down. PDFs store an image at its embedded resolution and draw it at whatever size the layout needs, so the stored original is frequently several times larger than the on-page rendering.
Can I extract images from a password-protected PDF?
Not while it is locked — the image data itself is encrypted. Remove the password first, then extract.
Does it find images inside stamps and signatures?
Yes, if the tool walks annotation appearance streams as well as page resources. That is where stamps, signature images and form field graphics live, and a page-only walk misses them entirely.
Does this work on a phone?
Yes, in any modern mobile browser on Android or iPhone. Very large documents are limited by device memory rather than by any server rule.
Can I extract images from several PDFs at once?
One document at a time. Keeping it to a single file is what allows accurate page attribution and duplicate detection within that document.
Is there a file size limit?
No server limit, because there is no server. The practical ceiling is your device memory.