Compress PDF
Shrink PDF file size in your browser with three quality presets. Runs on your device — nothing is uploaded, nothing is stored.
Drop PDFs here or click to select
Support for multiple files • Up to 50MB each
Compress PDF reduces a document's file size so it clears the limits that keep bouncing it back — the attachment ceiling on an email, the upload cap on a job-application or government portal, the size a messaging app will accept. Instead of re-exporting from the original app, you shrink the finished PDF directly.
There is one compression method and it's worth understanding before you use it: every page is rendered to an image, re-encoded as a JPEG, and placed back on a page of the original dimensions. Three presets pick the trade-off — High quality, Balanced, or Small file — trading render resolution and JPEG quality against size. For a scanned document, which is already a picture, that costs you almost nothing. For a crisp text PDF it costs you the text layer.
The whole process runs in your browser, so a confidential invoice or contract is never uploaded to a compression server. The tool shows the before and after sizes, so you can try a preset, look at the numbers, and pick a stronger one if it still doesn't fit.
Why this runs on your device
You rarely compress a PDF for your own benefit — you compress it because someone else's upload form has a limit. Which means the file in question is usually a passport scan, a payslip, a medical report, a signed lease or a tax return, and the alternative is putting that document through a stranger's processing queue purely to shave a few megabytes off it. Doing it locally also skips two transfers: you don't push 40 MB of scans up a slow connection just to pull a smaller file back down.
- Open the PDF you need to shrink.
- Pick High quality, Balanced, or Small file.
- Compress and wait for the pages to be re-encoded.
- Check the before/after sizes and that it's still legible, then download.
- Rasterizing is the only mode. Once compressed, the text can't be selected, copied or searched, screen readers can no longer read the document, and links, form fields and annotations on the page are gone. There is no option to keep the text layer, and no OCR to put one back.
- On a lean text or vector PDF the output can be larger than the input — a page of type becomes a full-page photograph. The tool still reports that as a success and simply shows a negative reduction percentage, so check the before/after numbers rather than assuming it worked.
- Three fixed presets, nothing else. There's no target file size, no custom quality or DPI control, and no image-only recompression — the Ghostscript-style server tools that shrink embedded photos while leaving the text layer intact are doing something this tool cannot do.
- One file at a time, on the page's main thread, with no progress bar and no cancel button. A long document will lock the tab up while it works. Transparency is flattened too, since every page is re-encoded as JPEG, and password-protected PDFs are rejected outright.