RunToolRunHow to Compress a PDF
Select a PDF file
Click the upload area or drop the PDF you want to shrink.
Compression runs on your device
Pages are re-rendered with optimized image encoding directly in your browser — the file is never uploaded.
Download the smaller PDF
Compare the before/after size and save the compressed version.
How this tool actually reduces the size
Each page is rendered to a bitmap at 1.5× its nominal size and re-embedded as a JPEG in a new PDF. That is a blunt instrument, and it is deliberately blunt: it works on every PDF regardless of how the original was produced, and the size reduction is predictable because it depends on the JPEG quality setting rather than on what the document happens to contain.
The consequence is the single most important thing to know before using it: the output is a picture of your document, not the document. Text is no longer text.
What you give up
You cannot select or copy text from the result, search inside it, or have a screen reader read it aloud. Links stop working, form fields become flat images of form fields, and anything that depended on the document's structure is gone. For a scanned document that was already images, none of this is a loss. For a text report, all of it is.
This matters legally and practically more often than people expect. A CV compressed this way will not be parsed by applicant-tracking systems. An invoice will not be read by accounting software that extracts totals. If the file will be processed by a machine rather than only looked at by a person, this is the wrong tool.
Choosing a quality setting
The default sits at 62, which is a deliberate compromise: text edges stay readable on screen while photographs lose enough detail to shrink substantially. Above roughly 80 the saving becomes small and you may as well not compress. Below about 40, JPEG's block structure becomes visible around letterforms and the document starts to look like a bad fax.
Because compression is applied per page as an image, the amount saved depends heavily on content. A page of dense text compresses poorly — sharp black-on-white edges are exactly what JPEG handles worst — while a page dominated by photographs can shrink by ninety percent. A mixed document lands somewhere between, and the total is dominated by whichever pages carry the images.
When a different approach is better
If the PDF is large because it embeds full-resolution camera photographs, resizing those images before building the PDF will beat any amount of compression afterwards, and the text will stay text. If it is large because of embedded fonts, subsetting them is the correct fix. If it is a scan, scanning at 200–300 dpi in greyscale rather than 600 dpi in colour prevents the problem at source.
And if you need the text to remain searchable, use a tool that recompresses embedded images while leaving the document structure intact. That is a more surgical operation, it does not work on every file, and it is not what this tool does.
Privacy and practical limits
Rendering and re-encoding happen on your device; the file is never uploaded. That is the main reason to use a browser tool for a payslip or a signed contract rather than a website that processes uploads on a server you know nothing about.
The cost is memory. Every page becomes a full-resolution bitmap in RAM before it is encoded, so a hundred-page document at 1.5× scale is genuinely demanding. If the tab becomes unresponsive on a very large file, splitting the document and compressing it in parts is more reliable than waiting.