This tool makes a PDF smaller so it fits under the limit an email, a form or a portal insists on. It works by re-drawing each page as a compressed image at a lower resolution, which is where almost all of the weight in a scanned or image-heavy document lives. That trade is worth knowing about before you use it: the result looks the same on screen but its text is no longer selectable or searchable, so keep the original if you will need to copy from it later. Everything happens in your browser, which matters here more than usual — the documents people need to compress are typically the ones they have just scanned.
No. The PDF is read from your device, each page is rendered on a canvas in the page, and the new document is assembled in memory. Nothing is transmitted.
Scans and photo-heavy documents often drop by 60 to 90 percent. A PDF that is mostly text and vector graphics may barely shrink, or even grow, because converting crisp text into an image is not an efficient trade — for those files, compression is the wrong tool.
No. Each page becomes an image, so text can no longer be selected, copied, searched or read by a screen reader. If those matter, keep the original and send the compressed copy only where size is the binding constraint.
Because the original was already efficient. A text-only PDF stores letters as instructions, which is far more compact than a picture of those letters. If the result is larger, the original did not need compressing.
You can, but you should not. Each pass re-compresses the previous pass's output, so the damage accumulates while the savings shrink. Go back to the original and pick a stronger level instead.