About this tool
PDFs don't store pictures as separate JPG or PNG files — every image is embedded inside the document as an internal object, encoded however made sense when the PDF was built. This tool digs into that structure directly in your browser: it walks every page looking for embedded raster images, hands back photos that were stored as JPEG byte-for-byte identical to the original file (no re-compression, no quality loss), and reconstructs a real, valid PNG file around any raw uncompressed pixel data it finds so that becomes downloadable too. Drop in a PDF, click "Extract images", and every picture it can recover appears as its own thumbnail with a one-click download.
Nothing is uploaded — the PDF is parsed locally with pdf-lib, walking each page's internal `/Resources/XObject` dictionary and telling actual raster images apart from Form XObjects (reusable page fragments that aren't photos at all, such as pages embedded by an n-up or booklet layout), so only real pictures are ever extracted. A few exotic internal encodings — indexed color palettes, CMYK, 16-bit channels, or JBIG2/CCITT fax compression typically used for scanned black-and-white pages — aren't repackaged into a downloadable file by this tool. Rather than guess at pixels it can't correctly decode, it reports that image's exact dimensions honestly and flags it as an unsupported encoding instead of producing a broken file.
Frequently asked questions
Why does it say "unsupported encoding" for some images?
PDFs can store images using several different internal encodings. This tool fully supports the two most common ones (JPEG-encoded images, extracted as-is, and raw/Flate-encoded RGB bitmaps, repackaged into standard PNG files). Less common encodings (indexed color palettes, CMYK, JBIG2 fax-style compression, or JPEG2000) are detected — you'll still see their dimensions and know they exist — but aren't currently re-packaged into a downloadable file, to avoid guessing at pixel data it can't decode reliably.
Will this find images that were used as a page background or watermark?
Yes — the tool scans every page's embedded raster image resources regardless of how they're visually used (a full-page background image, a small logo, a photo, or a scanned watermark bitmap are all detected the same way). The one thing it deliberately skips is TEXT-BASED or vector-drawn watermarks (like the ones our own PDF Watermark tool creates), since those are drawn as font glyphs or vector paths, not raster images — there's nothing to "extract" as an image file in that case.