Why we don't just delete metadata but never copy it
Many tools clear the fields in the document properties and call it done. A PDF can hold information in many other places, though: XMP metadata, annotations, forms, embedded files or earlier versions appended on save.
Diskretio sidesteps that hide-and-seek: every page is rendered and a new PDF is built from page images plus a searchable text layer. Author, title, subject, keywords, creator and producer are empty. Whatever was not visible on the page cannot come along.