Visa applications, university portals, insurance claims, tax filings, court submissions. Somewhere in all of them is a number — 2 MB, 5 MB, sometimes a cruel 1 MB — and a file that will not fit under it.
Almost every guide on this subject says the same thing: run it through a compressor. That works for one kind of PDF and does nothing at all for another, and the difference is easy to see once somebody tells you what to look at.
First: which kind of PDF do you have?
There are two, and they behave completely differently.
Open the file and try to select a line of text with your mouse.
| What happens | What you have | Can it be compressed? |
|---|---|---|
| The text highlights, word by word | A text PDF | Barely. It is already small |
| Nothing highlights, or the whole page selects as one block | A scanned PDF — pictures of paper | Yes, enormously |
That is the whole diagnosis, and it decides everything that follows.
If it is a text PDF
A twenty-page contract exported from Word is usually somewhere between 100 and 400 KB, because text in a PDF is stored as instructions — set this word, in this font, at this position — not as pixels. There is almost nothing to squeeze.
If a text PDF is nonetheless enormous, one of three things is inside it, and the fix depends on which:
- Embedded fonts. A PDF can carry entire typefaces. Subsetting them — keeping only the characters actually used — is normally automatic, but exports from some design software do not do it, and a handful of full font families runs to several megabytes.
- A logo or a signature at photographic resolution. One 4000-pixel PNG on a letterhead, repeated on every page, is a common and invisible cause.
- Hidden layers or attachments. Files exported from CAD or design tools can carry the whole editable original inside the PDF.
General-purpose compression will not touch any of these meaningfully. Re-exporting from the original document, with images downsampled first, will.
If it is a scanned PDF
This is where the megabytes are, and where compressing genuinely works.
A scanned page is a photograph of a piece of paper. Scanners and phone apps often default to 600 dpi, which for an A4 page is about 35 million pixels — for a document whose entire content is black text on white.
| Scan resolution | A4 page becomes | Good enough for |
|---|---|---|
| 600 dpi | ~35 megapixels | Archival reproduction, fine detail, tiny print |
| 300 dpi | ~8.7 megapixels | The standard for documents. What most authorities ask for |
| 200 dpi | ~3.9 megapixels | Perfectly legible text. Usually indistinguishable on screen |
| 150 dpi | ~2.2 megapixels | Readable, and visibly soft if anybody zooms in |
Going from 600 to 300 dpi cuts the pixel count by three quarters. That single change usually solves the problem on its own.
The order that works
Compressing is the last step, not the first.
1. Re-scan if you still can. Set the scanner or the phone app to 300 dpi and, if the option exists, black and white or greyscale rather than colour. A greyscale scan of a black-and-white document is a third of the size and loses nothing that was there.
2. Remove the pages nobody asked for. Blank backs of double-sided scans, the cover sheet, the terms and conditions the form did not request. This is the cheapest megabyte you will ever save and most people skip it.
3. Then compress. What a PDF compressor actually does is find the images inside, reduce their resolution, and re-encode them as JPEG at a lower quality. It does not compress "the PDF" as a single blob — it goes after the pictures, because that is where the size is.
4. Read it before you send it. Open the compressed file at full size and check the smallest print on the page — a reference number, a date stamp, a signature. Compression takes exactly the fine detail that official documents are checked against.
The trap: do not compress an identity document too far
Worth saying plainly, because it costs people weeks.
Passport pages, ID cards, bank statements and certificates are read by a human — or increasingly by a machine — looking at small printed detail. A compressed scan that looks fine at screen size can be rejected as illegible when somebody zooms in on the document number, and the application goes back to the start of the queue.
If a document is being used as proof of who you are, keep it at 300 dpi and find the space somewhere else: fewer pages, greyscale instead of colour, or splitting the submission if the form allows it.
When the limit is genuinely impossible
Sometimes 1 MB for a twelve-page colour scan simply cannot be met at a legible quality. Three real options, in the order worth trying:
- Greyscale. If the document is not colour-dependent, this alone is often a two-thirds saving and costs nothing legally.
- Split the submission. Many forms have several upload fields, each with its own limit, and people fight to fit everything into the first one.
- Ask. The office behind the form has heard this before, and a great many of them accept documents by email when the portal will not take them. It is the option nobody tries first and the one that most often works.
Why this should not need uploading anywhere
Notice what is in these files. Passports. Bank statements. Medical letters. Signed contracts. The documents people need to shrink are, almost by definition, the ones that should not be handed to an unknown company.
The usual way to compress a PDF online is to upload it to a website, wait for a server to process it, and download the result — which means a complete copy of your identity documents now sits on somebody else's disk, under a retention policy you did not read.
None of it is necessary. A browser can rewrite a PDF perfectly well on its own, which is how the compressor on this site does it: the file is read from your disk, changed in memory, and saved back. It never leaves the machine, and that is also why there is no size limit and no queue.