The short version

Drop the PDF and press the button. The tool works in three stages, cheapest first, and stops at the first one that gets the file under 25 MB.

Images are recompressed in place first, so text, links and forms stay exactly as they were. Only when a file still will not fit are the pages rebuilt as images, and even then an invisible text layer keeps the words searchable. A file already under the cap is handed back untouched.

Twenty-five megabytes holds a hundred and fifty pages scanned in colour, or six hundred pages of ordinary text. If your document is nowhere near either of those lengths and still passes the cap, the scanner is the reason, and compression is the second-best answer to that.

Where the 25 MB limit comes from

Who asks for itPlaces that cap uploads at 25 MB
  • Mail services quoting a 25 MB attachment cap25 MB stated
  • File transfer and sharing services25 MB free tier
  • Large filing and submission portals25 MB per bundle
  • Records requests and disclosure sets25 MB per batch
  • Digitisation and archive deposits20 to 50 MB

Twenty-five megabytes is the most quoted attachment limit on the web, and it is very often not the limit that applies. Mail attachments are re-encoded in transit and grow by about a third, so a service advertising 25 MB accepts a file of roughly eighteen. If email is where this file is going, compressing a PDF for Gmail works to the real ceiling instead of the advertised one.

How the tool hits 25 MB

1

Add the PDF

Drop it or browse. It is read into browser memory; nothing is uploaded, so a slow connection makes no difference.

2

Stages one and two: tidy, then the images

Unused objects, thumbnails and stale metadata go first, which costs nothing at all. Then the pictures inside the file are recompressed in place, starting at 200 ppi, while your text, links and form fields are left exactly as they were.

3

Stage three, only when it is needed

If the images were not where the weight was, the pages are rebuilt as pictures with an invisible text layer behind them, so the words stay searchable. It probes at 150 dpi, works out the resolution that should fit, and will not go below 60 dpi.

4

Download, and read the result line

It shows the final size and which stage produced it. If none fitted, you still get the smallest attempt, labelled honestly, rather than a file that silently missed the target.

The image ladderSeven rungs, from 200 ppi down to 60
200q82150q75120q66100q5880q5072q4260q36Rung 1: images barely touched (ppi above, JPEG quality below)Rung 7: visibly softYour text, links and form fields are untouched at every rung
Stage two walks down this ladder, recompressing the pictures inside the file and leaving everything else alone. It stops at the first rung under 300 KB, and abandons the ladder early when the images are not what is making the file big.
Stages
Three

Structure, then the images in place, then the pages as pictures. It stops at the first stage that fits.

Text after compression
Kept

Untouched at the first two stages. If the pages do get rebuilt, an invisible text layer keeps them searchable.

Bytes uploaded
0

The IDs, forms and statements people compress for portals never leave the device.

What fits in 25 MB, and what does not

The arithmetic is worth doing before anything else. Text costs about forty kilobytes a page, greyscale scanning a hundred and thirty, and colour scanning two hundred and forty. If your page count multiplied by the right figure lands nowhere near the size of your file, the file was scanned at a resolution far above any of these.

Try it

Will your file fit in 25 MB?

Enter the page count and what the pages contain. The estimate shows the attempt that is likely to succeed and what that attempt looks like.

Typical sizesCommon documents against the 25 MB line
150-page greyscale scanabout 20 MB100-page colour scanabout 24 MB80 full-page photosabout 30 MB130-page colour scanabout 31 MB25 MB
Estimates at sensible scanning resolutions. A document far heavier than these figures was scanned at a resolution nobody needed.

When it cannot reach 25 MB

Some documents will not fit legibly, and the tool says so rather than pretending. When the result line reports a miss, three moves usually solve it:

  • Split first. Use Split PDF to break the file into parts and compress each. Most portals accept several attachments even when each one is capped at 25 MB.
  • Rescan the source. A fresh 150 dpi greyscale scan compresses far better than a colour scan that has already been through one compressor. Every pass costs quality, and the first pass should start from the cleanest source.
  • Check the real limit. Portals sometimes state one number and accept another. Read the fine print, or try a slightly larger file before compressing harder than you need to.
One pass, from the original
Compressing an already-compressed file compounds the quality loss without much further saving. If the first result is not right, run the original again rather than the output. And keep the original: the compressed copy is for the portal, not for your records.

A 25 MB text document is a scan of a text document

The difference between the two is enormous and completely invisible on screen. A page of text stored as text is a list of characters and positions, and it weighs a few kilobytes. The same page stored as a picture of itself is several million pixels, and at 600 dpi in colour it weighs around a megabyte. Both look like a document. Only one behaves like one.

The same page, two waysWhy a scan costs a hundred times more
Stored as text8 KBcharacters and positionssearchable · selectable · sharp at any zoomScanned at 600 dpi980 KB34 million subpixels of papernot searchable · blurs when compressed
One A4 page, and the words are identical. The storage differs by a factor of a hundred and twenty because the second version records the paper as well as the words, at a resolution chosen by a default setting.

When that is the situation, compression is damage control rather than a fix. The tool above will get the file under 25 MB, and at this size it usually does so without anything looking obviously wrong. But it is reducing the resolution of a photograph of text, and every pass in that direction makes the smallest characters a little less certain.

Two better options exist, depending on what you have. If you still have access to the source document, export a fresh PDF from it and the problem disappears: the file will be a fraction of the size and the text will be text. If you only have the scan, run it through OCR first. That does not shrink the file by itself, but it puts a real text layer behind the images, which means the document becomes searchable and quotable, and any compression that follows costs you image quality rather than the content itself.

If you have to rescan, the settings that matter are resolution and colour. Three hundred dots per inch is the standard for archival scanning and is more than enough for anything that will be read on a screen. Greyscale rather than colour halves the size again with no loss on a black-and-white original. A hundred and fifty pages scanned that way come in comfortably under the cap with no compression at all.

Where people use the 25 MB target

Typical jobsWhat gets compressed to this size
  • Scanned reports, books and back catalogues100 to 200 pages
  • Disclosure and records-request batchesany length
  • Historic filings pulled from an archive50 to 300 pages
  • Photograph sets bundled as one document40 to 100 pages

Why this runs in your browser

Disclosure batches, archive pulls and scanned case files are among the most sensitive documents anybody handles, and they are also the largest. Uploading twenty-five megabytes of them to a website in order to make them smaller is both the slowest option and the worst one.

Two ways to compressOne of them never transmits the document
Your devicePDFin memoryBrowser enginecompressDownload0 bytes leavethe machineTypical online toolPDFon diskuploadServercopies filedownloadA full copy now sits on hardware you do notcontrol, under a policy you cannot audit
PDF.js renders the pages and pdf-lib rebuilds the file inside the tab. The document is read into memory, compressed there, and offered as a download.

The tool is right above

Everything on this page runs in your browser. Scroll up, drop the file, and keep it on your machine.

Jump to the tool

Common questions

The questions people have with a 25 MB cap in front of them.

01Why is my scanned document so large?

Because it stores a picture of each page rather than its text. At 600 dpi in colour, one A4 page costs roughly a megabyte on its own.

02Is compressing it the right fix?

It works, but it is second best. Exporting a fresh PDF from the source document, or running OCR before compressing, keeps far more of the document intact.

03What scanner settings should I use instead?

Three hundred dots per inch, and greyscale unless the colour carries information. That is enough for anything read on screen and a fraction of the size.

04Does OCR make the file smaller?

Not by itself. It adds a text layer so the document becomes searchable, which means later compression costs image quality rather than content.

05Is 25 MB the real limit for email?

Usually not. Attachments are re-encoded in transit and grow by about a third, so a service advertising 25 MB accepts roughly eighteen.

06What fits in 25 MB?

A hundred and fifty greyscale pages, a hundred colour pages, or several hundred pages of ordinary text.

07Are my files uploaded anywhere?

No. Rendering and compression run in your browser, which also avoids a twenty-five megabyte upload on a slow connection.

08Is it free?

Free. Every tool, unlimited use, no signup.