About PNG2PDF

You add pictures — one screenshot or fifty scans, up to 50 MB each — and the tool writes them into a PDF, one picture per page, in the order you arranged them. Pages can take the size of the picture itself (at 96 DPI, so a 1240-pixel-wide screenshot becomes a 930-point page) or a print format: A4 at 210 × 297 mm, US Letter at 8.5 × 11 inches, with a margin of 0, 6, 13 or 19 mm and the picture centred inside it.

PNG is what it is built around, but the browser's decoder is what does the reading, so JPG, WebP, GIF, BMP and AVIF files work in the same batch. You can rotate a page in 90° steps, reorder pages, sort them the way a person would read file names (scan-2 before scan-10), and choose between one combined PDF and a separate PDF per picture zipped together.

There is no conversion service behind this page. Four browser APIs do the whole job:

1. Decode. createImageBitmap turns the file into pixels and applies any EXIF orientation, so a photo taken sideways arrives upright.
2. Draw. The bitmap goes onto a canvas, which is where your chosen rotation is baked in and where pictures beyond 16 megapixels are scaled down — that ceiling is the browser's, not ours.
3. Compress. In lossless mode the canvas is read back band by band and its RGB samples are pushed through CompressionStream, the same deflate a PNG uses, so the pixels that land in the PDF are the pixels you started with. An alpha channel is compressed separately and attached as the image's soft mask. Compact mode skips all that and asks the canvas for a quality-85 JPEG.
4. Assemble. The PDF is written by this site's own code: image XObjects, one content stream per page positioning the picture on its media box, the page tree, the cross-reference table and the trailer. The result is a standard PDF 1.7 file that any reader opens.

Reading the canvas in horizontal bands rather than in one go is what keeps a 16-megapixel picture from needing 48 MB of pixel data in memory at once — it matters on phones, where that allocation is the difference between a PDF and a crashed tab.

Nothing is uploaded, because there is nowhere to upload to. The decode, draw, compress and assemble steps above all run inside your browser tab, and PNG2PDF operates no conversion server — no picture, no filename and no pixel data is transmitted, logged or stored by this site.

That also means the tool has no queue and no waiting: a screenshot is a PDF in the time it takes to press one button, even on a connection that would take a minute to upload it. Disconnect after the page loads and everything still works. When you remove a picture, close the tab or press Clear, the in-memory copies are released; nothing is written to local storage, so a later visit starts empty.

What this claim covers is our own behaviour with your files, which is to convert them and hand them straight back. It says nothing about browser extensions, a corporate proxy or the device itself — those are outside any web page's control, including this one.