Note: compression runs fully in-browser by rebuilding the file with optimized object streams and stripping metadata. Actual savings depend on the source — image-heavy PDFs benefit most from a server-side pipeline, which this frontend-only tool intentionally avoids.