What this does

This makes a PDF smaller, in one of two ways, and it is worth understanding which one you're asking for — because most compression services pick for you and don't say.

Tidying up rewrites the file's internal structure: the objects that describe the document get packed together and compressed, instead of being written out one after another. Nothing about the content changes. Text stays selectable, images keep every pixel, and the file is identical to read. The saving is whatever the original was wasting — sometimes a quarter of the file, often a few percent, occasionally nothing at all.

Rebuilding as pictures renders each page and stores it as a photograph of itself. This is where large savings live: a scanned contract can lose 80–95% of its size. The cost is total and permanent — the page is pixels afterwards, so text can't be selected, searched, copied or read aloud by a screen reader.

Which one you want

If your PDF is a scan — a photographed or scanned document, which is already nothing but pictures — rebuilding costs you nothing you had and saves a great deal. That's the case this tool is really for, and when it detects one it says so and picks that mode.

If your PDF has real text — an invoice from software, an exported report, a thesis — rebuilding is usually the wrong answer twice over: you lose the text layer, and the file often gets bigger, because a page of crisp text stored as a photograph takes more room than the letters and instructions that drew it. Tidy up instead, and if that isn't enough, the honest answer is that the file's size is its images and fonts, and a browser can't re-encode those in place without risking the document.

How to use it

Drop the PDF in. It's examined straight away, and the mode is preselected for you.
If you're rebuilding, name a size to aim for — or leave it empty and take the best quality the resolution gives.
Press Compress, then download.

Resolution, quality and the search

Two dials control a rebuild. Resolution decides how much detail each page keeps: 150 dpi reads comfortably on any screen and is the sensible default, 96 is fine for something that will only ever be looked at online, and 300 is for pages that will be printed.Quality is the JPEG setting behind each page, and it's the one the tool works out for you.

When you name a target, it encodes the whole document, measures it, and adjusts — narrowing in on the quality that comes closest to your size without going over. An attempt that's already too big is abandoned as soon as the running total passes the target, rather than finishing for the sake of it, which is what keeps a long document from taking all afternoon. If quality alone can't get there, the resolution comes down and the search starts again.

Is my file private?

Yes. The PDF is read, rendered and rewritten inside this page by your own browser. Nothing is uploaded, there's no queue, and no size limit beyond your own memory. That matters more here than on most tools: the documents people want to shrink are passports, bank statements, contracts and medical records, and the ordinary way to compress one online is to hand it to a stranger's server.

A rebuilt PDF is a new document and carries none of the original's metadata — no title, author or timestamps. Tidying up leaves all of that exactly as it was, since that mode promises to change nothing.

Frequently asked questions

Why did tidying up save nothing?

Because the file was already packed well — most modern software writes efficient PDFs, and there's nothing left to reclaim. That's a real answer, not a failure: the tool tells you rather than handing back an identical file and calling it compression.

Why did rebuilding make my file bigger?

Because it was a text document. A page of text is a short list of instructions; the same page as a photograph is a million pixels. Use tidy-up mode for text documents.

Can it shrink the images inside a PDF without touching the text?

Not here. It's the thing everyone wants — and doing it properly means rewriting image objects in place while matching colour spaces, transparency masks and the several ways a scan can be encoded. Done badly it produces files that look wrong or won't open, and a tool that runs entirely in your browser has no safety net for that. Rather than ship something that quietly corrupts a contract, this offers the two operations it can do correctly.

Will the text still be searchable?

After tidying up, yes — nothing changed. After rebuilding, no: the page is a picture, and running it back through the PDF to Word converter here won't recover the text either, since there's no text left to find. If you need searchable text out of a scan, that's OCR, which this doesn't do.

Does it work on a phone?

Yes, though rendering every page of a long document is real work — a hundred-page scan at 300 dpi will take a while and use a lot of memory. Lower the resolution first if it struggles.