A PDF is usually large because of what's inside it, not because of the format. Scanned pages hold full-resolution photographs of paper; exported decks carry the original artwork behind every cropped image. Compression works by re-encoding those images at a resolution that still looks right on screen, and by discarding data no reader needs — duplicate fonts, unused objects, revision history.
Text is left alone. Words stay selectable and searchable at every level here, so a compressed contract is still a contract you can copy from. If a file is mostly text and already small, this tool will tell you so rather than shaving off a pointless 2%.
Everything runs in your browser. Your file is never uploaded, never queued on a server, and never deleted from one — because it was never there.