Convert PDF pages to high quality JPG or PNG images

A PDF page has no fixed pixel size. It has a physical size, and the dpi setting decides how many pixels are drawn per inch. A US Letter page (8.5 by 11 inches) comes out at 816 by 1056 pixels at 96 dpi, 1275 by 1650 at 150 dpi and 2550 by 3300 at 300 dpi. An A4 page at 300 dpi lands near 2480 by 3508, give or take a pixel, because fractions of a pixel are dropped.

Rendering uses pdf.js in print intent, so text, vector drawings and embedded images are all drawn from the page description at the size you asked for. Text and line art stay crisp at any setting. A photograph inside the PDF can only be as sharp as the file that was embedded; 300 dpi adds no detail that was never there. Pages are drawn on white, so transparent areas come out white in both formats.

Screen (96 dpi) is for slides and web pages viewed at normal size: small files, fast to load, soft if anyone zooms in or shows them full-screen on a sharp display. Standard (150 dpi) is the default and safe middle, holding up on most displays, including a deck projected at 1080p. When unsure, start here.

Print (300 dpi) is for anything going through a printer, or that will be cropped into or enlarged later. It draws four times as many pixels as 150 dpi, so files are much bigger and each page takes longer. With more than 30 pages loaded and Print selected, the tool warns that the job will use a lot of memory and suggests converting a smaller range at a time.

Browsers also cap canvas size, so a render above 16 million pixels is scaled down to fit. Letter and A4 at 300 dpi sit around 8 million, well inside it; an A3 poster at 300 dpi comes back slightly smaller than a true 300 dpi render.

The cards put it plainly: JPG gives smaller files and suits scans and photos, PNG is lossless and suits diagrams and text. The difference is how each treats a hard edge: JPG softens the boundary between black and white, which is what small text and thin rules are made of, while PNG stores every pixel as rendered at the cost of size on a photographic page. A text page or line drawing headed for the web is usually better as PNG, often not much larger since flat areas of one color compress well; a scan or a photograph-heavy page is better as JPG.

Choosing JPG reveals a JPG quality (%) field set to 85, accepted between 30 and 100. Raise it toward 100 for print work. Check a low value on a page with small text, where artifacts show soonest. The field disappears for PNG, which has no quality setting.

Pages render one after another in your browser's memory, and every finished image stays there until you download, which is why a few hundred pages at 300 dpi is a bad idea on a phone. Under Convert, switch to Chosen pages and work in batches, using commas and ranges such as 1-20 and the word last for the final page; the button counts the selection, reading Convert 20 pages.

Files are named after the PDF with the page number appended, padded to the width of the page count: a 12-page file gives report-page-01.jpg to report-page-12.jpg, a 5-page file gives report-page-1.jpg. The padding keeps them in page order when sorted by name. One page gives a Download button carrying the file name; several give Download all as ZIP plus a per-file Download link.

If the PDF came from a low-resolution scan, every page is already a picture and the tool is re-rendering that picture larger, so 300 dpi gives a bigger file with the same fuzziness. A PDF put through Compress with Flatten is the same, since flattening replaces each page with an image of itself. Try selecting a line of text in a reader: if nothing highlights, the page is a picture, and dpi only sets how large a copy you get.

The images carry no text layer either, so a PDF rebuilt from them with JPG to PDF cannot be searched or copied from. Keep the original for anything that must stay readable as text, or run OCR PDF over the rebuilt file to put a text layer back.