PDFGym
EN

Compress large
PDF

or drop one here
Compression level
Target size
Choose a PDF file first.

A huge PDF can shrink right here in your browser, with no upload and no account, but the tab doing the work has a memory ceiling you should respect. This page shows how to prepare a monster file, when to split it into parts first, and what to do when the run stalls halfway.

Running a monster file through this page

  1. Load the file on a desktop

    Press Choose a PDF or drop the file onto the page. Use a desktop browser with free memory for anything over a few dozen megabytes; the whole job happens in this tab, and phones run out of room first.

  2. Set a realistic ceiling

    Pick a level such as Average, or type an exact cap in KB or MB that sits below the current size. For a first pass on a giant scan, aim high, for example a tenth of the original, and tighten only if the result still will not send.

  3. Press Compress and give it time

    Big files take minutes, not seconds, because every method builds its candidate inside the tab. Keep the tab in front and plugged into power; closing it throws the work away.

  4. Check the numbers before downloading

    The result states the new size, the saving and the original size, beside a preview of the pages, with the method-by-method report waiting in the Details panel. Press Download PDF only when the small print still reads well at full zoom.

The full story of the four methods, and why some documents fold to a tenth while others barely move, lives in how it works. To see which chapters carry the weight before you start, open size breakdown.

The 300 MB bundle on a Sunday night

Daniel is selling his late father's house, and the solicitor wants the whole paper trail as PDFs: deeds, surveys, planning letters, twenty years of receipts. His scanner app produced house-sale-bundle.pdf, 240 pages at 310 MB, and the solicitor's portal stops at 25 MB per file. One file, one cap, no chance in a single pass.

He does not fight the whole bundle at once. First he strips the forty blank versos the duplex scan left behind, which costs nothing in quality. Then he splits the bundle by folder, deeds, surveys, receipts, and runs each part toward its own target. Five uploads of 8 to 15 MB each pass on the first try, and every page stays sharp because each part got a generous budget.

That order, strip, split, then shrink, beats any single heroic run. A 240 page file squeezed to one target starves every page; five parts at comfortable targets keep the small print legible and survive mail and portal caps the whole never would.

The ceiling is your tab, not your target

Nothing here uploads, which is the privacy win and also the constraint: the original file, its parsed form and each candidate the engine assembles all share the working memory of this tab. That memory is bounded, and an oversized document can use it up. No setting raises it.

When the tab dies mid run, that is the ceiling talking. Nothing leaked, because nothing ever left the machine; only the attempt is lost. Close other tabs, retry on a machine with more memory, and if it dies again, split the document with Split PDF and compress the parts one by one. Parts also download faster and reopen faster for whoever receives them.

As a rule of thumb, desktops with free memory handle far more than phones. If the file matters and the phone is your only device, borrow a computer for ten minutes rather than fighting the small screen and the small memory together.

Strip and split before you squeeze

Scanners are generous with pages nobody needs: cover sheets, separator pages, blank backs. In a 200 page archive, thirty or forty of those are ordinary, and removing them hands every remaining page a bigger share of the target budget. Organize pages lets you see the thumbnails and drop the dead weight in one pass.

Then split by meaning, not by page count: chapters, exhibits, years. Split PDF cuts the file, and each part can aim at its own sensible number, for instance 10 MB for a photo heavy exhibit or a tighter cap for pure text. When the document must stay in one piece, that same 10 MB page is the right starting point, and the large file guide walks the whole workflow slowly.

One more lever people forget: colour. A colour scan of black and white paper costs roughly triple a greyscale scan before compression even starts. If the paper is still on your desk, rescan text pages in greyscale; if not, the image pass here does the equivalent after the fact.

Three ways big runs fail, and the fix for each

A frozen tab near the end means memory ran out. Retry with other tabs closed, on a desktop, with the file split in half. A result far above the target means the pages cannot honestly fit: raise the cap, or split so each part gets the full budget. A result that looks soft means the per page budget was too thin: fewer pages per run, or a roomier number.

Watch also for the password trap. A bundle that asks for a password on open resists the rebuild tricks, so the size barely moves. Unlock PDF opens it first, then compression works on the unlocked copy.

Finally, do not judge the run by the first preview alone. Zoom to full size and read the smallest line on the densest page, a footnote or a receipt total. If that line reads cleanly, the whole file is fine.

When a browser is the wrong place for the job

Honesty first: a server side tool or a desktop app compresses giant archives faster and without a tab ceiling, because it is not holding everything in one browser tab. If the file is hundreds of pages that must stay whole, or the deadline is an hour away and the tab keeps dying, use installed software for this one job.

The price of that speed is the upload: your document travels to someone else's machine. For deeds, medical files, legal exhibits or anything with signatures and ID numbers, the slower local run here is usually the right trade. For a public brochure with no secrets, pick whatever finishes first.

Quick answers about huge PDFs

How big a file can this page take?
There is no fixed cap; the limit is the tab memory on your machine. Desktops with free memory manage far more than phones. When a run dies, splitting the file is more reliable than retrying it whole.
Does a failed run expose my document?
No. The file never left your machine, so a crash costs you only the attempt. Close other tabs or split the file and try again.
Should text heavy archives split too?
Often they do not need to. Text and fonts collapse well under the lossless methods, so try the whole file first at a roomy target. Split when the file is scan heavy or the tab struggles.
Does the page order stay the same?
Yes. Pages keep their order and their count; the output is reopened and the page count is compared with the original before download. Only images, fonts and structural overhead are rewritten.
What output name will I get?
The original name with -compressed added, for example bundle-compressed.pdf. Rename it per part, bundle-deeds.pdf, before uploading so the recipient can tell the parts apart.

Companion tools for oversized documents

Split, inspect, repair or convert around the compression run. All of them run in the tab like this one, with no account.

Continue with a calmer file