Turn the pages of a PDF into PNG or JPG images at whatever resolution you need. Rendering uses pdf.js — the same engine Firefox uses to display PDFs — so pages come out looking the way a browser would draw them, fonts and vectors included, rather than as a rough approximation. Pick a scale factor and the output resolution follows: 2× a standard A4 page is around 1,240 by 1,754 pixels, which is enough for most print and presentation use. The scale is checked against what the browser can actually allocate before rendering starts, because a canvas past the limit does not error — it comes back blank, and a tool that hands you an empty PNG without saying so is worse than one that gives you a smaller image. Page ranges, multiple PDFs and a ZIP download are all supported, and nothing is uploaded.
| Scale | Approx. pixels | Roughly equivalent to |
|---|---|---|
| 1× | 595 × 842 | Screen preview |
| 2× | 1190 × 1684 | Retina screen, small print |
| 3× | 1785 × 2526 | ~225 DPI |
| 4× | 2380 × 3368 | ~300 DPI, print quality |
| 6× | 3570 × 5052 | Large-format print |
Scale multiplies the page's natural size, which for a standard A4 page is 595 by 842 points. At 2× you get roughly 1,190 by 1,684 pixels, at 3× about 1,785 by 2,526. As a rule of thumb, 2× is fine on screen and 4× is comparable to a 300 DPI scan.
PNG for anything with text, diagrams or flat colour — it is lossless and keeps edges sharp. JPG for photographic pages where the file size matters more than perfect edges. PNG also supports transparency; JPG does not, so transparent areas are filled white automatically rather than coming out black.
Because the page at that scale would exceed what the browser can allocate — roughly 16,384 pixels on either axis, and less on Safari. Past that limit a canvas does not raise an error, it just renders blank, so the scale is reduced and you are told, rather than handed an empty image.
Yes. Enter a range like 1-3,7,10-12. Leave it empty to convert everything. Ranges are clamped to the document, and a reversed range like 5-3 is read as 3-5 rather than producing nothing.
No — the output is an image, so the text becomes pixels. If you need the words, use the PDF Text Extractor instead, which pulls the actual text layer out.
No. The file is read and rendered entirely in your browser. Contracts, statements and scans are never transmitted, logged, or stored.
Each page is rendered with pdf.js, the same engine Firefox uses to display PDFs, so what you get is what a browser would draw rather than an approximation. A viewport is built at the requested scale, a canvas is sized to it, and the page is painted and then encoded with canvas.toBlob. The scale factor is clamped before rendering: browsers refuse canvases beyond roughly 16,384 pixels on either axis and beyond a total area limit that is lower still on Safari, and exceeding either does not throw — the canvas comes back silently blank, which is far worse than a smaller image. So the maximum safe scale is computed from the page's own dimensions and the request is reduced with an explanation rather than producing an empty PNG. JPEG output fills white first, since JPEG has no alpha channel and an unfilled transparent region encodes as black. Pages render sequentially with progress rather than in parallel, because a dozen simultaneous high-resolution canvases is how a tab runs out of memory. Filenames are zero-padded to the document's page count so a ZIP of a hundred pages sorts correctly.