Extracting Pages Without Re-Processing the Whole Document
Splitting a PDF means creating a new document containing only the page objects you select, along with whatever fonts, images, and resources those specific pages actually reference — everything unrelated to the selected range is left out entirely, which is why a 5-page extract from a 500-page document stays small rather than carrying the weight of the full original (see Complete Guide to PDF Processing for the underlying structural explanation).
Specifying page ranges like "1-3, 5, 8" — mixing individual pages and ranges in one request — is standard convention for exactly this kind of extraction, since real-world needs (pulling out a specific chapter plus one appendix page) rarely map to a single clean contiguous range.
Doing this entirely client-side matters specifically for sensitive documents: extracting just a few pages from a confidential contract or report via an online tool that uploads to a server means the entire document passes through that server, not just the pages you actually wanted — a real, easily overlooked exposure that processing locally in the browser avoids completely.