The short version
Check two numbers first: the file size and the page count. If it is size, run compress PDF once and read the saving. If the saving is zero or close, you have a scan and no compressor will help — the fix is sending fewer pages or scanning again at a lower setting. If the file is password protected, or the portal says it cannot open the document, size was never the problem.
Five reasons a PDF gets refused
| Cause | What you see | What to do |
|---|---|---|
| Over the size limit | A stated maximum, or an upload that fails at the end | Compress, then split |
| Too many pages | A page cap in the instructions, or a silent rejection | Split out the pages they asked for |
| Password protected | A password prompt, or a tool that cannot read the file | Remove the password where it was set |
| It is a scan | Huge file, few pages, compressing saves nothing | Rescan lower, or rebuild it from images |
| Damaged or unusual file | The document could not be opened | Open it and export a fresh copy |
For the page count, drop the file on split PDF: it prints the total before you choose anything. The file size is in the file's own properties, or beside the name in any PDF tool here.
A scan is a stack of photographs
Open the document and try to select a word. If the cursor will not grab any text, there is no text in the file. Each page is one large picture a scanner or phone camera produced, and the entire weight of the document lives in those pictures.
That matters because of what a PDF compressor does. The one here rewrites the document structure: compressed object and cross-reference streams, one current revision, and everything the pages point at carried across untouched. Photographs are not re-encoded. On a report from a word processor that removes real weight; on a scan there is nothing to squeeze, so the saving is reported as zero. Large savings on a scan come from re-encoding the page images at lower quality, which is irreversible and can turn a readable scan of a contract into something a clerk sends back.
What works instead, in order of how much it helps:
- Send fewer pages. Ask what they actually need. Three pages out of sixty is a split, and it clears any limit in one step.
- Scan again at a lower setting. Scanners often default higher than a form requires. Dropping the resolution, or scanning greyscale rather than colour, transforms the size.
- Rebuild it. If the scan exists as separate image files, shrink them with compress image or cut their width with resize image, then reassemble with images to PDF. Width is the bigger lever.
Two kinds of PDF password, and only one asks you anything
A PDF can carry a password that stops the file opening at all, and a separate one that leaves it readable but declares restrictions such as no printing. The first stops a portal dead, because the portal has no password to give. The second is subtler: the document opens fine for you, and a tool that reads it strictly still refuses, because the file is encrypted whether or not you were ever prompted.
Bank statements can be a surprise: some are issued locked with a password such as your date of birth, and forwarded without anyone noticing, because the sender's own reader remembers. Save an unprotected copy and upload that. The tools here do not open encrypted files: removing protection belongs in the application that owns the document, not in a converter.
When the document cannot be opened at all
Acrobat states it as the file being either an unsupported type or damaged — one message for two unrelated problems, which is why it explains so little. Mail is a common route for the second: some clients re-encode an attachment in transit, and a PDF that is decoded wrongly at the other end arrives unreadable — older versions of Acrobat's message gave exactly that example. Acrobat also enforces parts of the PDF standard strictly, for security reasons, so a file assembled by unusual software can be refused with nothing wrong with the content. The fix is nearly always to open the document in whatever will open it and export a fresh PDF from there. If it was downloaded, download it again first: a transfer that ended early leaves a file that looks the right size in a folder listing and is not a complete document.
The cases where the obvious fix does not work
- The PDF is digitally signed. Changing the file can invalidate the signature, and a compressor rewrites the whole file. Send the signed file as you received it and solve the size problem by sending fewer documents, not smaller ones.
- They asked for PDF/A. Archival and court filing formats have their own validation, and a file that has been through a general-purpose compressor may no longer pass. Produce PDF/A from the original application.
- It uploaded but came back rejected. Usually a content rule, not a file rule: a missing signature, the wrong form version, pages out of order. Merging and splitting fix the last.
- The limit is not the one they published. Mail systems measure the encoded size, which is larger than the file on your disk, so a document under a stated cap can still bounce. See email attachment too large and file too large.
- It is not really a PDF. A file named
.pdfthat is a photograph of a document gets rejected by anything that checks. Build a real one with images to PDF, or start at unsupported file type. If nothing was said at all, why was my file rejected covers the checks forms run silently.
One last thing about the documents in question. A signed contract, a payslip, a passport scan, a hospital letter: these are what people run through an online compressor, and the files you would least like left on a stranger's disk. Everything here works in the tab instead, and you can check that in a minute. Photos have the same problem in a different shape, in image is the wrong size.
Frequently asked questions
Last reviewed 2 October 2026