Shrink your PDF until it will actually send — sized to clear Gmail and Outlook attachment limits with room to spare.
Email attachment limits are smaller than they appear. Gmail advertises 25MB and Outlook 20MB, but neither figure is the size of the file you can attach. Email cannot carry raw binary data, so attachments are re-encoded into text before sending, and that encoding inflates them by roughly a third.
So a 25MB Gmail limit is really about 18MB of actual file, and Outlook’s 20MB is closer to 14MB. This is why a 22MB PDF bounces from an account that claims to allow 25MB — a genuinely baffling failure if you do not know about the encoding.
This page targets 18MB, which clears Gmail with room to spare. If your recipient is on Outlook or a corporate mail server, set a custom target of around 14MB instead — company mail systems are frequently stricter than the public services, and some cap attachments at 10MB.
Worth remembering that the limit applies to the whole message, so two 15MB attachments will fail even though each one is individually fine.
Almost always, images. A PDF holding nothing but text is remarkably small — a hundred pages of plain text might be under a megabyte, because the text is stored as characters and font instructions rather than as a picture of words.
The moment scanned pages or photographs enter the document, that changes. A phone camera photograph of a single page can be three or four megabytes on its own, and a scanner set to 600 dpi produces files of a similar size per page. Ten scanned pages and you are comfortably past most upload limits.
This is why the same page count can produce wildly different file sizes, and why compressing a text-only PDF achieves almost nothing while compressing a scanned one can cut it by ninety per cent. It also explains the usual surprise: a document that looks identical to another is ten times the size, because one was typed and the other photographed.
Most tools give you three vague presets — low, medium, high — and leave you to guess which one lands under your limit. That turns into a cycle of compress, check, try again, which is maddening when a portal has already rejected you once.
This one works backwards from the number instead. It compresses, measures the result, and adjusts, closing in on the target across several passes until it finds the highest quality that still fits. You get a file just under the limit rather than one needlessly far below it, which matters because every megabyte you give away is detail you did not need to lose.
If the first result is not what you wanted you can set your own figure, in kilobytes, and run it again.
The documents people compress are rarely trivial. They are contracts, medical records, bank statements, immigration paperwork and identity documents. Sending those to a stranger’s server to save a few megabytes is a poor trade, however reputable the site looks and whatever its deletion policy claims.
This page reads your PDF in your browser, rewrites it in memory and hands it back. Nothing is transmitted, nothing is stored, and there is no server log with your document in it, because there is no server involved. Once the page has loaded you can disconnect from the internet entirely and it will still work — which is a straightforward way to check the claim rather than take it on trust.
Drop your PDF above and press Compress. The file is reduced until it fits under the limit, then offered back as a download. You can set a different target if your form wants something else.
No. The whole thing runs inside your browser, so your document never leaves your computer. That matters for contracts, bank statements and identity papers.
Yes. Text and vector content are left alone; the savings come from the images inside the document, which are re-encoded at a lower resolution. A scanned page will look slightly softer, but it stays legible.
If the file is already under the limit, it is handed back untouched rather than being re-compressed. Re-compressing an already small file usually makes it larger, not smaller.
Some documents cannot physically get that small — a hundred pages of high-resolution scans has a floor. When that happens you get the smallest version achievable, with its real size shown, so you can decide whether to split the document instead.
No hard limit. Because the work happens on your own machine, the practical ceiling is your device’s memory. Very large scanned documents on an older phone may be slow.