Why a scanned PDF barely gets smaller in Folio

A scan is a photograph of paper stored inside a PDF. Folio rewrites structure. It does not shrink those photographs and it does not OCR.

Open splitTry compress anyway

People bring a 40 MB colour scan to Compress and expect a 2 MB file. Online tools that deliver that usually re-JPEG the page on a server. Folio will not do that. The bargain would change: quality loss, and usually an upload. IDs, medical pages, payslips and marked-up contracts do not belong on that server.

Use this page to decide whether split, a new photo, or a different app is the honest next step. Merge will not make a scan lighter. Compress on High, Balanced or Maximum will not invent a codec. Gmail caps are a separate job — that is the email size guide. Chat sends are the WhatsApp guide.

Confirm it is a scan

Try to select a word. If you cannot, the page is an image. Phone scanner apps — Notes, Adobe Scan, CamScanner exports, a Files PDF from the iPhone camera — are usually images. A Word “Save as PDF” is usually text plus fonts. A hybrid exists: a digital letterhead with a photographed appendix. Selectable header, dead-heavy last pages. Split the appendix. Compress will not rescue those pages.

Printed “300 dpi colour” in a scanner dialog is a warning, not a quality badge for email. Each page is a picture. Twenty pages is twenty pictures. The PDF wrapper is a box. Folio can tidy the box. It cannot shrink the pictures.

What Folio’s compressor actually does to a scan

It copies pages into a new PDF. It can drop leftover metadata and pack object streams. The heavy image XObjects stay. That is why before and after look alike. The FAQ line “scanned PDF barely gets smaller” is not a bug report. It is the product boundary. Read how to compress without uploading if you still have a digital export in the same folder.

pdf-lib, the library in this tab, is a page copier. It does not run a JPEG encoder on embedded images. It does not downsample. Sites that advertise “up to 90% smaller” on a scan are almost always rasterizing on a machine you do not control. Folio will not fake that number.

If you only need some pages, split

A 40-page scan where the portal wants pages 1 and 40 is a split job. Range mode on Split builds one PDF of those pages. The extract is smaller because it has fewer photographs, not because we compressed the pixels. That is the honest saving. Step-by-step: how to split a PDF in the browser.

After the extract you can merge those pages with a one-page cover letter. Do not merge the full 40-page scan with the letter and then ask Gmail why it bounced. Chat apps are even less patient — see WhatsApp.

If you need every page, recapture or use a scan tool

Photograph the paper again at a lower resolution, or use the scanner app’s own “document” mode. Black-and-white text at 150 dpi emails. Colour photos of paper at 300 dpi do not. A desktop tool that downsamples images will see the file because it runs as an app you installed. Folio will not downsample in this tab. That is not laziness. It is the line that keeps this site from becoming an upload farm.

Files over 80 MB are refused before parse. A phone tab will die anyway. Use a laptop for a 60-page colour stack, then send the result. Closing Safari discards the bytes in memory. Keep the original on disk.

OCR is a different product

OCR turns a picture of words into selectable text. It is useful. It is also the kind of job that often leaves the device. Folio will not run OCR in this tab and will not send the scan to an OCR API. If you already have a searchable PDF, you can merge or split it here like any other digital file. Searchable text is not the same as a smaller file. A searchable scan can still be 40 MB of pictures plus a text layer.

Do not upload “just this once” if it is private

IDs, medical pages, payslips, visas, and marked-up contracts do not belong on a conversion server. If the scan is a public brochure, an upload compressor is your choice, not Folio’s lecture. The operator inbox is on Contact. Do not attach the scan in the first mail.

Password-protected scans are refused. Unlock them in the scanner app or Acrobat, then drop the unprotected copy. We will not take the password. That refusal is documented in the FAQ so a reviewer does not have to guess.

Notes, Adobe Scan, CamScanner, and the iPhone camera

Apple Notes “Scan Documents” usually writes an image PDF. Adobe Scan and CamScanner do the same unless you paid for a text / OCR export, which is a different file and a different privacy bargain. The iPhone camera roll shared as PDF is still a photograph. None of those become “text plus fonts” because you dropped them here. Recapture in the scanner app’s document preset at a lower resolution, or split to the pages a portal actually named.

A 12-megapixel photo of an ID, saved as a one-page PDF, is often 8–20 MB by itself. The portal that said “PDF, 10 MB” will refuse it. Compress will not save you. Photograph the card farther away, in document mode, in good light, without the tablecloth. Then, if the portal wants the ID next to a lease, merge the smaller ID with the lease. Do not merge first and hope the packet shrinks.

A diploma, a transcript, a 10 MB slot

HR and school portals often show one upload field. Five files feel like a merge job — and they are, after each piece fits. A colour photograph of a diploma is usually the piece that does not fit. Compressing the cover letter is theatre. Split is only useful if they asked for one page of the transcript. Honest path: recapture the diploma smaller, then merge. That packet is written out under use cases.

A searchable scan (OCR already run elsewhere) can still be 40 MB of pictures plus a text layer. Folio will merge or split that file like any other digital PDF. It will not throw the pictures away. Searchable is not the same as small.

After you have a smaller file, keep the master

Print shops and archives want the original. Email wants the extract. Folio is not a vault. Download and file both copies yourself. A shop that asked for the scan will reject a mashed JPEG pretending to be the same PDF. Send them the master. Send mail the extract.

Medical pages, payslips, visas

Those files are why this site exists as a tab instead of an upload. A hospital portal that asked for “the discharge summary, PDF, 5 MB” will not be happier if you sent the summary through a random compressor first. Recapture in document mode, or split to the page they named. Do not OCR a medical page on a website you found in an ad. Folio will not do it either. If a workplace asked for a payslip plus an ID in one field, recapture the ID first, then merge. That packet is written out under use cases.

Colour versus grayscale matters at capture time, not here. A colour photo of a black letterhead is still a photograph. Switching Folio from Balanced to Maximum will not turn it grayscale. The scanner app’s “black and white document” preset is the right room.

Open splitTry compress anyway