PDF Compressor — Shrink Scanned and Image-Heavy PDFs

What this compressor really does — and what it doesn’t

There is no single thing called “compressing a PDF.” A PDF’s size can come from several places — embedded fonts, vector drawings, metadata, and above all images — and different tools attack different parts. It is worth being precise about which lever this particular tool pulls, because that determines exactly when it helps and when it cannot.

This tool works by rasterising: it renders each page to a canvas, converts that canvas to a JPEG at the quality you choose, and rebuilds the document from those JPEGs. For a page that is essentially a photograph already — a scan, a phone snap of a receipt, a graphic-heavy slide — this is a big win, because a high-resolution scan re-encoded as a sensible-quality JPEG is dramatically smaller than the original, with no visible difference at reading size. That is the job this tool is genuinely good at, and where the 50–90% savings come from.

The flip side is the part most compressors quietly get wrong. If a page is real digital text, that text is already stored extremely compactly — a whole page of writing can be a few kilobytes. Rasterise it and you replace those few kilobytes with a full-page image weighing hundreds of kilobytes, so the file gets larger, and the text stops being selectable or searchable into the bargain. A naïve tool ships that bigger file and still labels it “compressed.” This one does not: if the rasterised result is not actually smaller, it throws the result away and keeps your original untouched, telling you plainly that there was nothing to shrink.

Text becomes a picture — the trade you are making

The single most important thing to understand before compressing is that, on any file this tool does shrink, the page’s text is converted into an image. After compression you cannot highlight a sentence, copy a paragraph, or use Ctrl-F to find a word, and a screen reader sees a picture rather than words. The page looks the same to your eye; it is simply no longer “live” text underneath.

For the documents this tool targets, that costs you nothing, because a scanned page never had live text in the first place — it was always just pixels arranged to look like letters. Compressing a 12 MB scan down to 1 MB loses no capability you had. But if you were to compress a born-digital report or an invoice exported straight from software, you would be trading away genuinely useful text, which is why the tool declines to do it when there is no size benefit to justify the loss. If you need both a smaller file and selectable text, image compression is the wrong instrument; the answer there is a PDF optimiser that recompresses only the embedded images while leaving the text layer alone, which is a different kind of tool than this one.

How the quality slider actually behaves

The slider controls two knobs together, and knowing what each does helps you land on the right setting quickly. The first is the render scale — how many pixels each page is drawn at before it becomes a JPEG. Toward the smallest-file end the pages are rendered at close to their natural size; toward the best-quality end they are rendered larger, capturing finer detail. The second is JPEG quality — how aggressively the resulting image is compressed, on the usual scale where lower means smaller and blockier.

Because both move at once, the ends of the slider are meaningfully different. At maximum compression a dense scan can shrink enormously, but very fine print or a thin signature may soften. At minimum compression the pages stay crisp and close to the original, with a more modest size saving. There is no universally correct setting — it depends on whether the document will be skim-read on a screen or printed and scrutinised. A good habit is to start in the middle, check the downloaded file, and only slide toward quality if something important looks soft. Since everything runs locally and instantly, re-running with a different setting costs nothing.

When it keeps your original instead

You may upload a PDF, click compress, and get told the file was kept unchanged. That is a deliberate outcome, not an error. Before handing anything back, the tool compares the rasterised result against the file you started with; if the “compressed” version is not actually smaller, it discards it and returns your original bytes exactly as they were. The reasoning is simple: a compressor that sometimes makes files bigger is worse than useless, because you can no longer trust its output without checking every time.

In practice this is the tool’s honest way of telling you that your PDF is already efficient. Text-based documents, well-optimised exports, and files that have already been through a compressor will usually come back untouched. When that happens, there is no clever setting that will help — the size is not in images, so there is nothing for an image compressor to squeeze. If the file is still larger than you need for email or an upload, the lever to reach for is not more compression but a different approach entirely, such as removing unused pages or splitting the document.

Encrypted files, and where the work happens

A password-protected PDF cannot be compressed while it is still locked, because its page contents are scrambled until the correct password unlocks them and there is nothing readable to render. Rather than fail with a cryptic message, the tool stops as soon as it meets an encrypted file and names it, pointing you to remove the password first with the Unlock PDF tool. Compress the unlocked copy afterwards.

Everything here runs on your own machine. The PDF is read into browser memory, its pages are rendered on a local canvas, and the smaller file is built and offered straight back as a download — no byte is sent to a ToolMintX server. Even the rendering engine that draws the pages is served from this site itself rather than fetched from a public CDN, so compressing a document makes no outside request at all. For files as sensitive as scanned contracts, tax paperwork, or identity documents, that is the point: the work never leaves your browser.

Working out what is making your PDF big

Before you compress anything, it is worth a moment’s diagnosis, because the right fix depends entirely on where the weight is. There are really only a few culprits. The most common by far is images — scans, photos, or high-resolution graphics — and that is precisely what this tool addresses. A single full-page 300-DPI colour scan can be several megabytes on its own, so a document that is a stack of them balloons quickly.

The other contributors are usually small by comparison. Embedded fonts add a fixed cost — typically tens to a few hundred kilobytes total, no matter how many pages use them — so they rarely dominate. Vector graphics, the lines and shapes in diagrams and logos, are stored as instructions rather than pixels and are almost always compact. And there is a little structural overhead in every PDF that you cannot meaningfully reduce. A quick tell: divide the file size by the page count. If a text-heavy report works out to a handful of kilobytes per page, it is already lean and this tool will keep it unchanged. If it is hundreds of kilobytes or more per page, the weight is in images and there is real room to shrink.

That mental check saves time. It tells you in advance whether to expect a big saving or to hear that the file was already efficient, and it points you at the right tool when compression is not the answer — fewer pages, or lighter source images, rather than harder compression of something that has nothing left to give.

Related tools

If your PDF is really a stack of photos, compressing the images before you build the PDF often works better — the image compressor handles that. When a file is large because it simply has too many pages, the Split PDF tool lets you keep only what you need. And if the file is locked, clear its password with Unlock PDF before compressing.