← Back to all guides

How to Clean Up, Shrink, and Email a Scanned Document Without It Looking Like a Fax From 1995

Published October 20, 2026

A step-by-step workflow for turning a phone-camera scan of a paper document into a clean, small, searchable PDF that's actually pleasant to open — using a background cleanup pass, OCR, and compression, in that order.

Most "scanned" documents today aren't scanned at all — they're photos taken with a phone, usually at a slight angle, under mixed lighting, with a visible shadow from whoever's holding the camera. The result is a PDF that looks gray and muddy, weighs 8-15 MB because it's really just a giant JPEG wrapped in a PDF container, and can't be searched or copied from because there's no actual text in it, just pixels.

None of that is necessary. With three tools used in the right order, you can turn that photo into something that looks like it came out of a real flatbed scanner — white background, crisp black text, a file size small enough to attach to an email without hitting a size limit, and real selectable/searchable text underneath. Here's the order that actually works, and why the order matters.

Step 1: Clean the background before you do anything else

It's tempting to jump straight to compressing the file, since file size is usually the thing that's actually causing a problem (bounced emails, upload limits). Resist that urge. If you compress a photo that still has a gray, shadowed background, you'll just get a smaller gray, shadowed file — compression doesn't fix lighting problems, it just makes them blurrier.

Start with PDF Background Remover. Feed it the raw scan-photo PDF (or a photo saved as a PDF via your phone's share sheet). It analyzes each page, figures out what's actual document content versus what's uneven lighting or paper texture, and flattens the background to clean white. This single step does more for how "official" a document looks than almost anything else you can do to it — a document on a pure white background reads as a real scan, while the same document on a gray gradient background reads as a phone photo, even if the text itself is identical.

If you only have an image file at this stage (not a PDF yet) — say, a .jpg straight off your camera roll — do the cleanup at the image level instead. The general-purpose Background Remover under the image tools works the same way: upload the photo, let it strip out the uneven background, download the cleaned PNG, and then move on to Step 1.5 below before continuing to OCR.

Step 1.5: Get the image into a PDF, if it isn't already

If you cleaned up a standalone image rather than an existing PDF, use Image to PDF to wrap it into a PDF container. You can drop in multiple pages here too, which matters if you're scanning a multi-page document one photo at a time — stack the photos in order and it'll produce a single multi-page PDF instead of making you send five separate attachments.

Step 2: Make the text actually readable by software, not just by eyes

At this point you have a clean-looking PDF, but it's still just a picture of text, not text. That matters more than it might seem: a picture-of-text PDF can't be searched with Ctrl+F, can't have its text copied and pasted into another document, and — if anyone needs to run it through a document management system, a compliance check, or just wants to quote a sentence from it — it's basically a dead end.

Run the file through PDF OCR. Optical character recognition scans the image of each page and generates an invisible, selectable text layer that sits exactly on top of the visible text — so the page still looks identical, but now you (or anyone you send it to) can select a paragraph, copy it, search for a word across a 40-page contract, or have a screen reader read it aloud. This step works dramatically better on a page that's already been cleaned up in Step 1 — OCR engines get confused by shadows and gradients, misreading characters far more often on a muddy scan than on a crisp white one. This is the real reason the order matters: clean first, then recognize.

Step 3: Now, and only now, compress it

With a clean, OCR'd PDF in hand, run it through PDF Compress. Because the background is now flat white instead of a noisy photographic gradient, the compression algorithm has a much easier job — flat colors compress far more efficiently than photographic noise, so you'll typically see a bigger size reduction here than you would compressing the original muddy photo, on top of the fact that the result also looks sharper.

A realistic before/after for a 3-page phone-scanned document: the raw photo-PDF export often lands around 9-12 MB. After background cleanup, OCR, and compression (in that order), the same three pages typically come out between 400 KB and 1.2 MB — small enough to attach to virtually any email service without triggering a size warning, and small enough that the recipient's inbox won't choke on it either.

A quick note on what to do if the document also needs a signature or a password

Two common follow-ups once the document is clean:

  • If you need to sign it before sending, do that after compression, using PDF Sign — adding a signature is a small, lightweight edit that won't meaningfully change the file size, so there's no benefit to doing it earlier in the pipeline, and doing it last means you're signing the final, clean version rather than a messy draft.
  • If the document is sensitive (a contract, a medical form, a signed agreement) and you want to restrict who can open it, add a password with PDF Add Password as the very last step, after signing. Encrypting the file first and then trying to run OCR or compression on it afterward generally won't work, since the tool can't read an encrypted file's contents without the password.

Common pitfalls

  • Compressing before cleaning the background. As covered above, this just locks in the gray/shadowed look at a smaller size — it doesn't fix it. Always clean first.
  • Running OCR on a rotated or upside-down page. If a page came in sideways, straighten it before running OCR — a tilted page dramatically increases misreads. Most phone scanning apps have a rotate option before export; if yours doesn't, you can generally rotate within your PDF editor before moving to the OCR step.
  • Expecting OCR to be 100% perfect on handwriting. OCR is very reliable on printed or typed text. Handwritten notes are a much harder problem for any OCR engine, doitium's included, so don't be surprised if a handwritten signature block or a scribbled margin note doesn't come through as selectable text — the printed body of the document will still work fine.

The short version

  1. Clean the background — PDF Background Remover (or the image-level Background Remover if you're still working with a raw photo).
  2. If you started from a photo, wrap it into a PDF with Image to PDF.
  3. Make the text searchable with PDF OCR.
  4. Shrink the file with PDF Compress.
  5. Sign and/or password-protect last, if needed, with PDF Sign and PDF Add Password.

Every one of these runs entirely in your browser — the document never leaves your device or touches a server, which matters if what you're scanning is a lease, a medical form, or anything else you'd rather not upload to a random website.

← Back to all guides