The right choice for pages with text
Most PDFs are documents, and most documents are mostly text. That makes PNG the correct default here.
PNG is lossless: every pixel of the rendered page is stored exactly. Letters keep their edges, table rules stay as thin lines, and diagrams have no smearing.
JPG is the wrong tool for that content. Its compression works by discarding high-frequency detail, and a hard black letter against white paper is precisely high-frequency detail. The result is visible haloing around text — obvious at a glance, and impossible to remove afterwards.
Choose JPG only when the pages are dominated by photographs, where the file will be far smaller and the artefacts invisible.
What rasterising costs you
A PNG of a page is a picture of it. Once rendered:
- Text can no longer be selected, copied or searched
- Links stop working
- Enlarging beyond the rendered resolution softens everything
Sharp is not the same as scalable. If you need text that stays crisp at any size, the PDF already is that — convert only when something downstream requires an image.
Expect large files
PNG stores everything exactly, which is the point and also the cost.
A text-heavy page compresses reasonably well, because most of it is white. A page full of photographs can produce a very large PNG — those pages belong in JPG.
Why the zip
A fifty-page document produces fifty images. Rather than fifty download prompts, they are packaged into one archive, assembled in your browser rather than on a server.
Nothing is transmitted
Rendering runs on PDF.js, Mozilla’s PDF engine, inside a Web Worker on your own machine. The Uploaded readout beside the result measures this page’s own outbound network calls directly and reads 0 B on success.