How it works
We render each page with pdfjs-dist to a canvas, snapshot the canvas as a PNG, and bundle every page into a single ZIP using JSZip. Everything happens in your browser — no upload.
FAQ
Does this extract the original embedded JPEGs? No — it renders each page to a PNG at the resolution you choose. A true XObject extraction is fragile across PDF producers (older PDFs use stencil masks, indexed color spaces, JBIG2, etc., that are awkward to round-trip back to a standalone image). Rendering covers every visual case the user usually wants. If you need a single embedded JPEG, opening the PDF in Acrobat → Tools → "Save as Image" still works.
Why is the output PNG and not JPEG? PNG preserves any transparency in the rendered page (e.g. invoices on a white background round-trip cleanly). PNGs are larger than JPEGs but for ZIPs of 10-20 page documents the difference is negligible.
How big is the output? Roughly 200-500 KB per page at 2× scale for a typical letter-sized text PDF; multiply by the page count. At 4× the file is 4× larger. If you only want screen-quality, 1× or 2× is plenty.