JPG to PDF

It should not be. Phones often store a portrait photo as landscape pixels plus an EXIF orientation tag that tells viewers how to turn it. JPG to PDF reads that tag and turns the photo on the page the way your browser shows it, sized to its upright shape, without re-encoding the JPEG. If a page is still sideways, the camera saved the wrong tag (or none): open the PDF in Rotate PDF, which also runs in your browser, and turn just that page.

Not with Max file size at No limit. Then each JPEG goes in as it is, not decompressed and saved again (a turned phone photo only has its rotation tag reset to upright), so every pixel of the original is still there even when the photo is displayed smaller to fit the page, and the PDF is as heavy as the photos. With a Max file size that the PDF is over, yes, on purpose: the photos are saved again with fewer pixels and stronger compression until the PDF fits, and the panel says the images were saved again.

With Max file size at No limit, yes. Because each JPEG goes into the PDF as it is, apart from a rotation tag reset to upright, any metadata stored inside it travels too, including camera details and GPS coordinates if your camera recorded them. Strip it from the photos first if the PDF will be shared widely. A photo saved again to fit a Max file size keeps only its pixels: its camera details and location are not in the PDF.

Yes, all four, mixed with JPGs in the same PDF, each on its own page. A PNG goes in as it is, and any transparent areas come out white. A WebP, GIF or BMP is decoded by your browser and saved again: as a high-quality JPEG when the picture has no transparency, as a PNG when it does. That is a re-encode, so a WebP photo can come out slightly larger or softer than the file you had, and an animated GIF gives only its first frame. Each file is judged by its contents, not its name, so a WebP saved as .jpg still goes in. A file of any other kind is turned away when you add it with "Choose PNG, JPG, WebP, GIF or BMP images."

Yes, with Max file size: choose 100 KB, 200 KB, 500 KB or 1 MB, or Custom and a whole number of KB. The PDF is made as usual first and kept as it is if it already fits. If it is bigger, your browser saves the images again as JPEGs, step by step at a lower quality and pixel size, measures each PDF, and keeps the sharpest one that fits. If none fits, you get the smallest it made and a line saying it is above your limit. Here 1 KB is 1,024 bytes; a portal that counts 1,000 needs a slightly lower number. A transparent PNG saved again this way has its see-through areas filled with white, the colour of the page.

The tool sets no count of its own, only the 25 MB ceiling per image. Every photo you add is held in the tab's memory until the PDF is built, so a very large batch is bounded by what your device can hold.

Not fully. The page and its processing engine load over the network, and parts of the engine are fetched the first time a step needs them. The processing itself runs on your device.

Up to 25 MB per file.

Each image goes on its own page, centred. Unless you change the page setup, that page is US Letter (8.5 × 11 in) portrait with a 0.5 in margin. Larger images are scaled down to fit; smaller ones are never enlarged. Fit to image makes each page the size of its image instead, plus the margin. To change it, pick another option before you create the PDF. Page size: Fit to image, A4, US Letter, US Legal, A3 or A5. Orientation: Auto, Portrait or Landscape (Auto turns the page for a wide image). Margin: None, Small (0.25 in) or Large (0.5 in).

Yes. Under Output, choose One PDF per image (ZIP): each image becomes its own one-page PDF, named after the image (holiday.jpg becomes holiday.pdf), and when there is more than one image they download together as one ZIP.

Yes — each image becomes one page, in the order shown. Use “Move up” and “Move down” to change it before you create the PDF.

They show as white: each page is filled white before the image is drawn on it.