Extract PDF pages

Processing limits: 50 MiB and 200 pages; 15 seconds per job. Decoder capacity: 32 MiB per stream and 64 MiB cumulative, with object-count and nesting limits. These are not a bound on total browser memory. Page information and export may remain available when a particular encoding cannot be previewed.

Ordinary links, annotations and AcroForm fields belonging to retained pages are preserved, without flattening forms or rasterizing pages. XFA export is unsupported. A page graph referencing the document catalog or page tree is rejected because copying them could reimport pages outside the output. Editing invalidates digital signatures. Original files are unchanged.

When pages are excluded, links to them lose their destinations. Export is rejected for a form field without page-owned widgets, an annotation owned by both a retained and excluded page, or another structure referring to an excluded page (such as an article thread or tagged structure element). Keeping all pages does not require these refusals. Page deletion is not comprehensive sanitization of hidden or shared data.