Make scanned PDFs tiny — and keep every word searchable. 100% in your browser.
No. Everything runs inside your browser — the PDF never leaves your device, and there's no server.
Scanned or photographed documents usually shrink 3–5×, while keeping the original look and staying fully selectable and searchable.
Scans and photos of text documents. PDFs that are already digital text won't shrink much.
It fetches the compression and text-recognition engines once. Your browser caches them, so later runs are instant — and it then works offline.
129. It auto-detects the document's language when you compress and lets you override.
Yes, and open source (AGPL-3.0).
Each page is rebuilt as MRC (Mixed Raster Content): a crisp 1-bit text mask (CCITT G4) over two heavily-compressed JPEG layers — the text-colour foreground and the paper background — plus an invisible OCR text layer for search.
It's all WebAssembly, client-side:
Open the browser console to watch the per-page pipeline. The MRC segmentation is derived from the Internet Archive's archive-pdf-tools.