Tiny Online Tools
Extracting images, not screenshots of them
Most "PDF to image" tools render each page to a bitmap. That gives you the page — text, background and all — at whatever resolution you chose. It is not the same thing as getting the photograph the author placed on it.
The images in a PDF are separate objects in the file, and in the common case they are a complete JPEG sitting there compressed exactly as it was imported. This tool walks the object graph, finds them, and writes the JPEG straight out: same bytes, same quality, no second round of lossy compression.
What comes out
DCTDecode(JPEG) andJPXDecode(JPEG 2000) — the original file, byte for byte, after the transport filters are unwrapped. Nothing is re-encoded.FlateDecode,LZWDecode,RunLengthDecoderaw bitmaps — decoded and written as PNG, honouring the colour space (grey, RGB, CMYK, ICC-based, Indexed with its palette), the bit depth (1, 2, 4, 8 or 16 bits per component), the/Decodearray and any/SMaskalpha channel./ImageMaskstencils — 1-bit masks come out as a transparent PNG.JBIG2DecodeandCCITTFaxDecode— these are not standalone image files and a browser cannot decode them. They are listed with their size, filter and pages and marked as not extractable, rather than quietly dropped.
Effective DPI is the number that matters
A 600 × 400 bitmap is 600 DPI in a one-inch box, 150 DPI across four inches and 75 DPI across eight. The intrinsic pixel size on its own tells you nothing about print quality. This tool parses the page's content stream, tracks the transformation matrix through it, and reports for every image the size it is drawn at and the DPI that results — flagged amber below 300 and red below 150.
That makes it the fastest way to answer two very practical questions: why is this PDF 40 MB? (an image at 1200 DPI that nobody needs) and why did this print badly? (a 90 DPI logo blown up across the cover).
De-duplicated
One image object drawn on twelve pages is one image used on twelve pages, not twelve copies. The card lists every page that uses it.
Filtering
Set a minimum width and height to skip the 1 × 1 spacer images that word processors scatter through their output, and optionally include images that are present in the file's resources but never actually painted.
Privacy
The PDF is opened and unpacked entirely in your browser. Nothing is uploaded and no network request is made while it works.