How it works

Honest limitation: this performs structural optimization only (object streams, metadata removal) — it does not re-encode or downsample embedded images, which is where most real PDF size reduction comes from.
Works best on PDFs with lots of repeated objects or bloated internal structure. PDFs already dominated by large photos will see minimal change — sometimes under 5%.
Metadata removal strips title, author, subject, keywords, and producer fields — a small size win, and a privacy win if you're sharing a document externally.
For heavy image-based compression, a server-side tool with image re-encoding is genuinely necessary — that's intentionally not what this tool does, and we'd rather say so than overpromise.
A short ad plays before your download starts — it helps keep this tool free. Processing itself happens instantly; the ad is just a brief pause before the file downloads.
All processing happens locally in your browser using pdf-lib — your files are never uploaded anywhere.