How to Compress a PDF Without Uploading It Anywhere

If you've ever Googled "compress PDF online," you've seen the same tools appear, such as iLovePDF and Smallpdf. They work, and both say in their own terms that they process your file on their servers, so it has to be uploaded first.

For a PDF of your coffee shop menu, that's probably fine. For a contract, tax return, medical record, or anything covered by an NDA, it's worth pausing.

What actually happens when you "compress PDF online"

When you drag your file into an online compressor, the usual path is:

1. Your browser uploads the PDF to the service.
2. The service compresses it on its own servers.
3. Your browser downloads the result.
4. The service deletes your file after a stated period.

That last period is the exposure window. iLovePDF's terms say files are deleted within two hours of processing, and Smallpdf says most of its tools delete files after one hour (both checked on their own pages on 15 October 2026). Those are reasonable policies, but your document still sits on someone else's server for that time, and you are relying on their security and their word.

How Stax compresses a PDF in your browser

Stax PDF Compressor does the work on your device with two open-source libraries, pdf-lib and PDF.js. You choose a target size, such as 100 KB, 200 KB, 500 KB, 1 MB or 2 MB, or type your own between 20 KB and 20,000 KB. Then:

1. A lossless re-save. The tool rewrites the PDF with a tidier internal structure and strips metadata, leaving every page untouched. If the result already fits your target, you get that file and the text stays selectable.
2. If it does not fit, pages become images. Each page is drawn once, re-encoded as a JPEG, and the tool searches for the best mix of image quality and resolution that fits under your limit. It then rebuilds the PDF at the original page size and checks the real file size before offering the download.
3. If it still cannot fit, it says so. A long document of dense colour pages can be too big for a tiny limit even at low quality. The tool shows the smallest size it reached instead of pretending.

The important trade-off: when step 2 is needed, the output looks like the original but the text is part of a picture, so it can no longer be selected, copied or searched. There is also a "Keep text (lossless)" mode that only does step 1, with modest savings. Always keep your original file.

How to check that nothing is uploaded

Open your browser's developer tools (F12), go to the Network tab, filter for POST requests, and compress a file. An upload would show up as a request about the size of your PDF. You will see page files load and some requests to analytics and advertising scripts (the privacy page lists them), but nothing that carries your document. The same test works on any tool, ours or anyone else's. Two quick tests show whether an online tool uploads your file walks through it step by step.

What to expect from different PDFs

How much a PDF shrinks depends on what is inside it:

• Scans and phone photos saved as PDF are already images, so turning pages into optimised images loses nothing structurally and can shrink them a lot.
• Text-based PDFs (contracts, reports, bank statements) are often small already. The lossless re-save usually saves only a little, and reaching a very small limit means turning the pages into images.
• PDFs full of high-resolution images can shrink well because the image data is the bulk of the file; compress the source photos first with the Image Compressor if you still have them.
• Vector-heavy PDFs (charts, diagrams) are inherently small, and compression has little effect on them. If one is large, the size usually comes from images placed in the layout.

We do not quote a typical percentage, because it varies far too much between files. The tool shows the real before and after size for yours.

Step-by-step: compress a PDF without uploading

1. Open stax.tools/pdf-compressor.
2. Drag your PDF onto the drop zone, or click to select it.
3. Choose a target size, or switch to Keep text (lossless).
4. Press Compress and check the result panel: it says whether the target was reached and shows the approximate resolution used.
5. Download the new file, named after your original plus the target (for example form-200kb.pdf). Your original is never changed.

No signup, no email, no upload. How long it takes depends on your device and the page count.

What about very large PDFs?

The limit is your device's memory, not a server's. Because every page is drawn in your browser, we suggest splitting documents of more than about 60 pages with the PDF Splitter and compressing the parts. Password-protected PDFs have to be unlocked first. For multi-gigabyte archives, a local command-line tool such as Ghostscript (free, open source) is more appropriate:

# Ghostscript one-liner for heavy compression
gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 \
   -dPDFSETTINGS=/ebook -dNOPAUSE -dQUIET -dBATCH \
   -sOutputFile=compressed.pdf input.pdf

/ebook targets 150 DPI images, which is good for screen reading. Use /printer for 300 DPI or /prepress for 300 DPI with colour preservation.

Quick comparison

Understanding what makes PDFs large in the first place

Before compressing, it helps to know what is eating the file size. A PDF is a container that can hold several kinds of content.

Embedded images are usually the biggest contributor. A scanned document at 300 DPI creates a lot of image data, and each embedded image is stored as a separate stream.

Embedded fonts matter too. A PDF that embeds a whole font rather than only the characters it uses carries the unused characters as well, which can add a few hundred kilobytes for each font.

Edit history and annotations can accumulate. Some programs append changes to a PDF rather than rewriting it, so a file edited many times can carry content that is no longer visible.

Metadata such as the author, creation date and software name is small but present in almost every PDF.

PDF compression strategies for different document types

Scanned documents

Scans are images wrapped in a PDF, so the most effective lever is resolution. As a rule of thumb, around 150 DPI is usually enough to read comfortably on a screen, and 200 to 300 DPI is better if the document will be printed or archived. The tool's target-size mode chooses a resolution for you and shows the approximate value it used.

Office-generated PDFs (Word, Google Docs, PowerPoint exports)

These contain formatted text, embedded fonts and often some images. They are often small already, so use the lossless mode first and see what it saves. If a portal demands a smaller file than that, the target-size mode will get there by turning pages into images.

Government forms, bank statements and property documents

Forms, certificates and statements are often generated with several embedded fonts, and scanned property documents can run to tens of megabytes. Pick the limit the portal asks for, check the preview of the result for legibility, and split long files into parts if the portal allows several uploads.

Verifying compression quality after download

After downloading, check three things:

1. Is the text still selectable? Click and drag across text in a PDF viewer. If you used target-size mode and pages were turned into images, it will not be, and that is expected.
2. Are the images and small print legible? Zoom to 100% and check that numbers, stamps and signatures are still readable.
3. Are all pages present? Scroll through, or run the file through PDF Page Info to check the page count and sizes.

When to use cloud tools instead

In-browser compression is not the right answer every time:

• You need the smallest possible file from an image-heavy PDF. Server tools and desktop tools such as Ghostscript can be tuned further than a browser tool can, so they may produce a smaller file.
• You are processing thousands of files. Server automation handles volume that a browser tab can't.
• Compression has to happen inside a business process. An automated pipeline calls for a server-side tool.
• The document is not sensitive. A restaurant menu or public brochure is fine in any tool.

The bottom line

For sensitive documents, compressing in the browser is a sound choice, not because cloud tools are untrustworthy, but because skipping the upload skips the exposure. For everyday files where privacy isn't a concern, any of the tools above works.

Try the Stax PDF Compressor: it is free, runs in your browser and needs no installation. For long documents, combine it with the PDF Splitter to work in smaller parts, and use the PDF Merger first if you want one small file from several documents (see our guide to merging PDFs without uploading). The same approach works for video: see how to compress a video without uploading it.