Image to PDF

Image to PDF builds a new document from the pictures you add, one page per image, in a single pass. Everything you select goes into the same PDF rather than being converted file by file, so a folder of scans comes back as one document. Order is either the order the files arrived in or file name A to Z, sorted naturally, which puts 2 before 10 instead of the other way round.

Page size decides the shape of the result. Fit the page to each image never resamples the picture: at the default margin of zero the page is exactly the picture, and raising the margin grows the sheet around it instead of shrinking what is drawn. The catch is that one image pixel becomes one PDF point, so a 4000-pixel-wide photo produces a page around 1.4 metres across — fine on a screen, useless at a printer. Choosing A4, US Letter, A3 or A5 instead scales each picture to fit inside the page and centres it, and Match each image turns the sheet to landscape when the picture is wider than it is tall. The page background colour applies either way, and it is what shows through wherever a PNG is transparent.

Format support is narrower than the file picker suggests, and how a picture gets in matters as much as whether it does. A JPEG is the one format copied in untouched: its compressed bytes become the page, so nothing is decoded and nothing is re-encoded. A PNG is not copied — pdf-lib decodes it to pixels and deflates them again as full-colour RGB, which keeps every 8-bit sample but rebuilds the page data, reduces a 16-bit PNG to 8 bits per channel, and leaves a greyscale or palette screenshot noticeably heavier than the file it came from. Anything else the browser can display — WebP, GIF, BMP, AVIF — is painted onto a canvas and handed on as a PNG, so it takes that same route, and an animated GIF or WebP contributes only its first frame. HEIC and TIFF are refused outright, with a message naming the file that stopped the run, because no browser can decode either one and the browser is all this tool has.

The whole conversion runs in this tab using pdf-lib, reading the images from disk into memory, and no file is uploaded. That boundary also explains what is missing. There is no text recognition, so a PDF built from scans is a stack of pictures and nothing in it is searchable. No picture is ever downsampled on the way in, so the document weighs at least what the images weighed and frequently a little more. A JPEG's EXIF orientation tag is ignored, so a phone photo that relies on that tag to sit upright arrives on its side. And an animated PNG stops the run outright, with an error from pdf-lib that does not say which file caused it.

Runs on your device. Your file is processed directly in this browser.