What changes when you convert PNG to PDF
A PNG carries an alpha channel and a PDF page does not, so the first thing that happens in this conversion is a decision about transparency. Transparent regions are composited over white, because a page has to be some colour and white is the one that prints correctly. If your PNG has a transparent background and you were relying on it, the PDF will show white there.
Each image becomes its own page, sized to that image's aspect ratio. Nothing is cropped: an image larger than the page box is scaled down to fit, with the ratio intact, so a wide screenshot gets a wide page rather than losing its edges. When you convert several images at once, the order you set in the staging list is the page order, top to bottom.
PNG is a lossless format, which means the file going in is usually large and the PDF coming out is too. There is no compression pass on the way in, deliberately — screenshots of text are exactly the content that JPEG artefacts damage most, and a screenshot that has become unreadable is worse than one that is a few megabytes. If the result is too large for an upload form, compress it afterwards as a separate, visible decision.
The output contains pictures of whatever the PNGs showed. There is no OCR step, so text in a screenshot is not selectable, not searchable and not readable by a screen reader in the resulting PDF. Converting that PDF to Word later will give you a Word file with an image in it. If searchability matters, the text needs to come from a text source rather than a screenshot.
The whole assembly happens in your browser. The images are read into memory in the tab, embedded by pdf-lib, and the finished document is handed back as a blob URL that exists only until you close the page. No request is made, which matters when the PNGs are screenshots of internal dashboards or documents.
This page is the PNG-to-PDF case specifically. The general tool, with its full set of limitations and questions, is Images to PDF.