MillFile exists because of a bug I hit while building it. Most PDF compressors shrink a file by rendering every page to an image. That works - and it quietly deletes the text layer. On a 6-page text PDF I measured it: 30.5 KB out, and zero extractable characters. The document still looks right on screen. It is now a photograph of words. So MillFile repacks the document structure first and only re-renders pages when it has to, and only as far as it must. Same file, 8.9 KB, and all 26,652 characters still selectable. It runs entirely in the browser - no upload, no account, no watermark, no daily quota - and the source is MIT open source, so the compression step is auditable rather than a claim. One question I would like to know: when you compress a PDF, what do you actually check afterwards?

MillFile

Compress a PDF to the exact size a form asks for

Compress a PDF to the exact size a form asks for