PDF to JPG

Converting a PDF to images renders each selected page pixel by pixel, at the resolution you choose. 96 DPI matches a typical screen, 150 DPI suits most printing, and 300 DPI is print production — and the trade-off is explicit: quadrupling the pixels roughly quadruples the file size and the rendering time. A 50-page document at 300 DPI is on the order of four hundred million pixels, drawn one page at a time.

Choose JPG for pages that are mostly photos or scans, and PNG when the page is text or line art and you want sharp edges. Pick pages in the grid to convert a subset. One page downloads directly; several pages arrive as a ZIP, stored without further compression since the images are already compressed.

The output is images — they have no selectable text. That is the honest meaning of this conversion, and it is why the related tools point the other way: images back into a searchable PDF needs OCR, which runs on your phone in the Android app.

Rendering happens page by page in a worker on this device — deliberately serial, because rendering several large pages at once is how tabs run out of memory on phones. Progress is deterministic and cancelable: the bar counts pages, not vibes, and the cancel button interrupts mid-document without leaving partial files behind. Nothing uploads; the network panel stays empty for the whole run.

The honest limits: output images have no selectable text, no OCR runs here, and very high DPI on very large pages may be automatically reduced to stay inside the browser's memory budget — the result card tells you when that happened. To go the other direction, or to make a scanned PDF searchable again, the Android app handles images-in and OCR on the phone.

The slides case: a speaker wants a deck's key pages as images for a document that will not embed PDFs. Select the four pages that matter, choose 150 DPI, JPG, and download four images sharp enough to project. Nothing else in the deck was rendered, no page numbers were invented, and the original deck is still a PDF for the venue's laptop. The result card names every image after its page number, so slide-14.jpg is findable a month later without opening anything. And the whole run is cancellable mid-document, so a wrong DPI choice costs seconds instead of a full render. One more click back on the grid narrows the export again, because pages unselected cost nothing to render.