How to make a PDF smaller for email and upload limits

Compressing a PDF means re-encoding what is inside it, and there are only two honest ways to do that. Lossless optimisation rewrites the file's internal plumbing — object streams, duplicated resources — without touching a single letter or pixel. Strong compression re-renders every page as a JPEG image at a lower resolution and rebuilds the document around those pictures. The first is safe and gives little; the second gives a lot and costs you the text layer. Whatever a tool calls its sliders, one of those two things is happening.

The quickest route here: open Compress PDF, drop the file, pick Strong (pages → images) or Lossless (keep text), and the tool shows the old size, the new size and the percentage saved before you download anything. It runs in your browser, so a payslip, a medical letter or a passport scan stays on your device. If a method cannot actually make your file smaller, it says so rather than handing you a heavier file.

The three ways to shrink a PDF

Method	Good for	Watch out for
A browser tool like this one	Any device, nothing to install, files stay local	Very large scans depend on your device's memory
Desktop software (Acrobat Pro, Ghostscript, LibreOffice)	The same job every week, a whole folder at once	Installation, a paid tier for Acrobat, a command line for Ghostscript
Built-in OS features	One quick job on a Mac	macOS Preview only; Windows, iOS and Android have nothing for this

On macOS, Preview can do it: File → Export…, then open the Quartz Filter menu and choose the "Reduce File Size" filter. It is a blunt instrument with fixed settings, but it is free and already installed. On Windows there is no built-in PDF compressor: "Microsoft Print to PDF" re-prints a document with no quality setting and often returns a file of a similar size. On iPhone, nothing in Files or Books compresses a PDF; if the document came from your own camera, re-scanning it (Notes → new note → camera icon → Scan Documents) gives a far lighter file than a photo. On Android it is the same story: Google Drive's scanner writes a reasonable PDF, but there is no way to shrink one you already have.

For scripted work, Ghostscript is the reference: gs -sDEVICE=pdfwrite -dPDFSETTINGS=/ebook -o out.pdf in.pdf downsamples images towards 150 DPI, and /screen aims at 72 DPI. Acrobat Pro's "Reduced Size PDF" and PDF Optimizer put a dialog in front of the same controls.

Where the weight actually is

A page of text costs a few kilobytes. Fonts are shared across the whole document, and the letters are just instructions to draw glyphs. Images are a different order of magnitude: one A4 page scanned in colour at 600 DPI is 4961 × 7016 pixels, about 100 MB of raw pixel data before any compression. The same page in greyscale at 200 DPI holds 27 times less data before compression even starts.

So the first question is not "which level?" but "is this PDF mostly text or mostly images?" Try to select a sentence with the mouse. If individual words highlight, there is a text layer and lossless mode is the honest option. If you can only drag a rectangle across the page, it is a scan — which is what strong compression was built for.

What compression does to your file

• Strong mode rasterises the pages. Text can no longer be selected, copied or searched, and a screen reader cannot read the document. That matters for a contract, a CV that goes through an applicant-tracking system, or anything someone may need to quote.
• Lossless mode changes nothing visible — same pages, same fonts, same selectable text — which is also why its gains are modest, typically 0–15%.
• Quality does not come back. Re-encoding is one-way, so keep the original.
• Small print goes first. Footnotes, stamps, handwritten signatures and thin table rules are where 80 DPI shows.
• Metadata depends on the mode. Lossless mode keeps the document as it was, author and dates included; strong mode builds a new file around the page images, so those fields are not carried over. PDF metadata shows you what is actually left.

In practice, that gives four common cases: scanned pages and photos of documents, strong mode at 110 or 150 DPI; a presentation exported to PDF, strong at 150 DPI after checking the small print on charts; a contract typed on a computer, lossless; a mixed report, lossless and then fix the photos at the source.

How to reach an exact target

Forms usually state a ceiling — 2 MB, 5 MB, 10 MB — and email providers commonly reject attachments over 20–25 MB. Work downwards instead of guessing:

1. Try the recommended level (110 DPI) and look at the size you get.
2. Still too big? Drop one step (150 → 110 → 80 DPI) and compare a page of small print at 100 % zoom before you accept it.
3. If the scan is in colour but the document is black text on white, rescanning in greyscale at 200 DPI beats any compression setting.
4. If the PDF was built from photos, compress them first with Compress Image — 1600–2000 px wide is plenty — then rebuild it with JPG to PDF.
5. If nothing fits and the form takes several files, Split PDF is the honest answer.

When compression is not the answer

Some files have nothing left to remove: a 300 KB text report is already near its floor, and both modes will say so. Three cases call for a different approach:

• Encrypted files. A password-protected PDF cannot be read page by page until it is decrypted. Remove the password with Unlock PDF if you know it, compress, then protect the result again.
• Documents that must stay searchable. A contract a reviewer will search through should not be flattened into pictures. Send it in parts instead.
• A single oversized image. One full-page photograph in an otherwise small report is better fixed in the source document than by re-rendering forty pages.

The guide My file is too big to upload goes through the limits you are most likely to hit.

Common problems, and what they mean

"This PDF is already well optimised." Lossless mode found nothing worth rewriting. That message is accurate, not a failure — try strong mode if the file is image-heavy, and keep the original if it is not.

The compressed file is bigger than the original. Re-encoding an already heavily compressed JPEG adds bytes rather than removing them. The tool stops and keeps your original in that case.

The text is now blurry at 100 % zoom. You went one level too far. Redo it from the original at 110 or 150 DPI; sharpness cannot be restored afterwards.

I can no longer search the document. Strong mode turned the pages into images. Go back to the original and use lossless mode, or accept the trade for a document nobody needs to search.

A 200-page scan freezes my phone. Every page is rendered on your device. Split the file first with Split PDF, or use a computer.

Questions people ask

Is 150 DPI enough for printing? For office printing of text, yes. For photographs or anything a print shop produces, 300 DPI remains the usual floor — DPI explained works through the arithmetic.

Are my documents uploaded to a server? Not with this tool. Compression runs in the browser on your own device. For any service that does upload, check where files are processed and how long they are kept before sending anything sensitive.

Does compressing change the page count or the page size? No. Pages keep their dimensions and their order; only what is drawn on them is re-encoded.

Can I compress a scanned PDF and still search it? Not here — this tool does not do OCR. Recognise the text with an OCR program first, then compress in lossless mode.

Should I compress before or after merging? After. Merge with Merge PDF first, then compress once, so you judge the real final size.

Next steps

• Send the document in parts: Split PDF, and the guide How to split a PDF
• Shrink the pictures before they become a PDF: Compress image
• Turn pages into images on purpose: PDF to JPG