How to reduce a PDF’s file size

Start with the kind of PDF you have

A PDF is a container. It can hold real text, embedded fonts, vector lines, and photographs stored as JPEG or other images. Two files with the same number of pages can differ by tens of megabytes because one is a letter and the other is a scan of a letter.

Look at whether you can select a sentence. If you can, the page is mostly text and vectors. If the whole page highlights as one picture, it is a scan or a slide that was exported as images. The right way to shrink it depends on that difference. Compress PDF offers both paths, and it shows the measured size either way.

Repacking a text PDF

Some PDFs are saved without the compact object streams that modern writers use. Copying the same pages into a new file can tidy that structure. Text stays selectable, links inside the page content usually survive, and you have not thrown pixels away.

The saving is often modest, and it can go the wrong way. A file that was already tightly written may come back a little larger, because the new file adds its own structure. That is a real result, not a failure of the measurement. If the number goes up, keep the original.

Repacking will not turn a 40 MB photo book into a 2 MB file. The photographs are still the photographs. Tools such as Ghostscript can recompress those images while leaving the text as text. This site does not ship Ghostscript or MuPDF. Both of the usual builds are AGPL, and a browser page that quietly claims “up to 90% smaller” is usually calling that kind of engine, or it is guessing.

When the pages are really pictures

A scan at 300 dots per inch, stored without much JPEG compression, is large because of the pixels. Rebuilding each page as a JPEG at a lower quality, or at a slightly smaller scale, throws detail away on purpose. The new PDF is a stack of pictures. You cannot select the words, search them, or copy a column of figures.

That trade is reasonable for a signed form you only need to read, or for a slide deck you are about to email. It is a poor trade for a contract you may still edit, or for a statement you need to search. Use the quality slider, look at a page, and only then decide. Scale 0.75 is softer. Scale 1.5 is sharper and often larger than the file you started with.

If you need the pictures themselves, PDF to JPG writes each page as an image without wrapping them back into a PDF. JPG to PDF does the reverse when you already have the photographs.

Other ways the file gets smaller

Delete pages you do not need to send. A twenty-page scan with three relevant pages should become a three-page file. Delete PDF pages and Split PDF both do that on the device.

Do not email the same scan twice inside one PDF, and do not merge a cover sheet that is itself a huge image if a one-page text cover would do. Merge PDF joins what you give it. It does not shrink the pieces.

A password-protected PDF has to be unlocked before any of this works. Unlock it locally. Nothing on this site asks you to upload the file in order to remove the password.

A practical order

If you can select the text, repack first and compare the two sizes. If the file is still too big and the pages are photos or scans, try an image rebuild at quality 70 and scale 1, then look at the page rather than only the number. If it is still too big, drop pages or lower the scale, and accept that the text is no longer text.

If neither number is smaller, the honest outcome is that this browser could not shrink that file without a heavier engine. Keep the original and send a link to a shared drive, or print to a new PDF from the app that created it, which sometimes recompresses images on the way out.