Compress PDF on Linux

Files save to ~/Downloads or wherever your browser is configured to put them, as a normal file owned by you with standard permissions. Nothing needs root, and nothing is written outside your home directory.

The tool itself behaves identically regardless of what you are running it on. The same options, the same validation, the same progress reporting and the same output — a file produced on Linux is byte-for-byte equivalent to one produced anywhere else, because the same code does the work in both cases.

That consistency is worth something practical. Start a job on Linux and finish it on a different device, or hand the page to a colleague on another platform, and neither of you has to learn a different interface or account for a difference in behaviour.

The Lossless level loads the document with pdf-lib and re-saves it using cross-reference object streams. This packs the file's internal structure more tightly without altering any page content — useful for documents with a lot of structural bloat, neutral for everything else.

The other levels take a different route. Each page is rendered to a canvas with pdf.js at a chosen scale, encoded as a JPEG at a chosen quality, and embedded into a fresh PDF at the original page dimensions. Target size mode walks through six scale-and-quality pairs from 1.5x at 82 per cent quality down to 0.55x at 34 per cent, stopping at the first one that meets your number. Whatever route is taken, the final output is compared against the plain lossless rewrite and the smaller of the two wins.

Your files are not uploaded. When you add a PDF file, your browser reads it from your device into the memory of the tab you are looking at, and the compress pdf operation runs there using your own processor. Nothing is transmitted to any server at any point.

That means there is no file for anyone to store, no retention policy to trust, and no deletion schedule to take on faith. The document exists only in the memory of an open browser tab, and it is released the moment you close it, reload the page, or press start over.

The site uses standard anonymous analytics to count page views and how often each tool is used. Those events record which tool ran and whether it succeeded — never a file name, never file contents, and never anything that could identify a document or the person processing it.

No account is required, so no personal information is collected. There is no profile, no history and nothing tied to an identity, because no identity was ever created.

The strongest security property this tool has is structural: your file never travels anywhere, so there is no transmission to intercept and no server-side copy to breach. This is not a policy that could change — it is a consequence of where the code runs.

You can verify it in about thirty seconds. Open your browser's developer tools with F12, select the Network tab, and use the tool. You will see the page's scripts load — including the PDF libraries, which are fetched on demand the first time a tool runs — and then nothing further. No request carrying your document is ever made.

For a stronger test, disconnect from the network once the page has loaded. The tool will keep working exactly as before, which is only possible because nothing is being sent.

Uploaded files are validated in the browser before anything runs: the extension must match what the tool accepts, the file must not be empty, and it must be within the size limit. Malformed documents produce a clear error and a retry option rather than a silent failure.

Being clear about limits is more useful than a feature list that quietly omits them. Any level other than Lossless converts page text into an image, which removes text selection, copy and paste, and searchability. Compression also cannot reduce a file below the size its content fundamentally requires.

If you need capabilities beyond that — redaction, PDF/A conformance, OCR, or scripted batch processing across thousands of documents — dedicated desktop software genuinely earns its licence fee, and this is not trying to replace it. For the specific job of producing a compressed PDF, everything you need is on this page.