Compress PDF

Why is my PDF so large?

A PDF is usually large because it carries photos or scanned pages at a far higher resolution than a screen or printer needs. A phone camera image dropped into a document can be 4000 pixels wide even when it is printed three inches across, and scanners often save every page at 300–600 DPI. Fonts embedded in full, duplicated images, and old revisions left inside the file by editing software add the rest.

Files exported from Word, Google Docs or a browser are mostly text and vector graphics, which are already small, often under 100 KB per page. Those gain little from any compressor, and a big saving on such a file would only come from damaging it. That is why this tool reports an already optimized file instead of forcing a smaller but worse copy.

Which compression level should I choose?

Choose Recommended for most files, Lossless when nothing visible may change, and Extreme only when the smallest possible size matters more than selectable text. Recommended is the default because it combines the structural clean-up of Lossless with photo resizing, which is where most of the saving in scans and photo-heavy reports comes from. The table lists exactly what each level does to the file.

How does Compress PDF work?

Compress PDF works in two passes inside your browser: a structural rewrite at the Lossless and Recommended levels, plus an image pass at Recommended. The structural pass uses qpdf, an open-source PDF toolkit compiled to WebAssembly, which rebuilds the file, packs its internal objects into compressed object streams and drops resources that no page uses.

In the image pass, JPEG photos of 12 KB or more are decoded, scaled and re-encoded with the browser's own image engine. Images that act as masks, use colour keying, CMYK colour, an EXIF rotation tag or unusual encodings are left exactly as they were, because recompressing them in a browser could shift colours or break the page.

Extreme mode renders each page with pdf.js, the viewer engine built by Mozilla, and assembles the page images into a new PDF that keeps the original page sizes. At every level the result is compared with the original, so the download is never the larger of the two, and nothing leaves the page at any point.

How small does a PDF need to be for email?

A PDF needs to be under 25 MB for Gmail and under 20 MB for many other mail services, and smaller is safer. Gmail's 25 MB attachment limit applies to the whole message, and attachments grow by roughly a third when they are encoded for sending, so aim for about 18 MB or less per email. Job portals, visa forms and university systems are usually much stricter, with limits of 1, 2 or 5 MB per document.

When a file is still over the limit after Recommended compression, two options remain. Extreme reduces scanned pages the furthest, at the cost of selectable text. Alternatively, Split PDF divides a long document into parts that each fit, and Remove Pages drops blank sheets, covers and appendices that the recipient does not need, which keeps the remaining pages at full quality.

How do I get the smallest file on any device?

The smallest files come from fixing the source before compressing. If a document will be scanned again, choose a lower scanner resolution such as 150 or 200 DPI in greyscale; that saves more than any later compression. Exporting straight to PDF instead of printing to PDF also helps, because printing can bloat fonts. For a searchable scan that is also small, run Compress first and OCR PDF afterwards.

No app or plug-in is needed. The tool works in current versions of Safari, Chrome, Edge and Firefox on any operating system, including iPhone and iPad. On a Mac this is a practical alternative to the Reduce File Size filter in Preview, which re-renders images aggressively and can leave scans blurry. On phones, compress one large file at a time and keep the tab open until the download button appears, because mobile browsers may pause pages running in the background.