Batch
Compress 500 PDFs at once
Not one PDF, five hundred. A year of scanned invoices, a disclosure set, an archive that has to fit somewhere smaller. Choose the folder, choose the operation, and come back when it has finished.
Every hosted tool for this is one file at a time, and the reason is not technical. Compression is cheap; receiving five hundred files is not. The megabytes cross somebody’s network, sit on somebody’s disk and appear on somebody’s invoice, which is why the free tier is two files a day and a hundred megabytes, and why nobody has ever offered to take your folder.
With no server in the path, none of that arithmetic exists. There is no per-file ceiling, no combined ceiling, no count limit and no daily quota, because there is nothing to meter. The operation runs on the machine the files are already on.
Practically, that changes the unit of work. The question stops being “which of these files matters enough to bother with” and becomes “which folder”. Point at it, preview the first result to check the quality is what you want, and leave it running.
Start here
- Open the batch runnerChoose a folder, pick compression, preview the first real result, then run every file. On Chrome and Edge it can write the results straight back into a second folder you choose.
- Compress a single PDFThe same operation on one file, when a folder is not what you have.
- The offline-ready versionOpen it once with a connection, then use it with the network switched off — which is also the simplest proof nothing is being uploaded.
What this will not do
Worth reading before you put a client’s file through it.
- — Compression is lossy on images. A scan compressed hard is smaller and blurrier; preview the first result before committing a folder to it.
- — A PDF that is already compressed will not shrink much, and the summary says so rather than pretending otherwise.
- — Writing results straight back into a folder needs Chrome or Edge. Other browsers give the same results as a ZIP instead.
- — The work happens in your browser tab, so a very large folder is limited by your machine’s memory. The page warns you rather than refusing.
Check it rather than trust it
Turn off your wifi and use any tool here. If the work still happens, the file is not going anywhere — that is a proof you can perform in three seconds, and it beats anything we could write. The longer version, with the protocol and the result, is on the proof page.
Everything here is free, MIT-licensed and has no account, quota or watermark. It is maintained by one person, unpaid. If it saves you an afternoon, you can support the work.