PDF Compressor
Compressed document
What this actually does. The document is rebuilt with cross-reference and object streams, unused objects dropped and metadata optionally removed. It does not recompress images or downsample them, because doing that in the browser would need a full rasteriser. Expect a useful reduction on documents heavy with structure, forms or revision history, and almost none on a PDF that is mostly scanned images — the real size in that case is the images, and this tool leaves them untouched rather than quietly degrading them. The exact before and after figures are shown so you can judge whether it was worth it.