What this tool does
It puts your PNG images into a PDF, in the order you choose, without touching the image data.
That last part is the whole point. The usual way to build a PDF from images in a browser is to draw each one onto a canvas and re-encode it. It works, and it quietly costs you a generation of quality every time. Here the PNG bytes are embedded exactly as they arrived, so the picture inside the PDF is bit-for-bit the picture you dropped in.
Why that matters more than it sounds
Most people converting images to PDF are going to compress the PDF afterwards — for an application portal, a job form or an email limit. If the conversion step has already softened the image, the compression step is working on a degraded original, and you lose twice.
Doing it losslessly here means that when you go on to compress the result, the compressor is working with everything the original had.
Page size: what to pick
Fit the image makes each page exactly the size of its picture. No white borders, no letterboxing. This is right for screen reading, and for a single image you want to look deliberate.
A4 (595 x 842 pt) or US Letter (612 x 792 pt) centres the image on a standard page with a ~10 mm margin, scaled to fit and never stretched. Pick this when the document will be printed or submitted somewhere that expects a page size.
Nothing is ever stretched to fill the page. A stretched ID photo is a rejected application.
PNG is often the wrong format for the job
An inconvenient thing worth knowing: if your "PNG" is really a photograph — a phone picture saved as PNG, or a screenshot of one — you are carrying three to five times more bytes than you need for no visible benefit.
PNG stores every pixel exactly, which is perfect for text, screenshots of interfaces, logos and diagrams with flat colour, and wasteful for anything with smooth gradients. If the resulting PDF is uncomfortably large, that is usually the reason, and converting the source images to JPG first will fix it far better than compressing the PDF afterwards.