About FileCrisp

What this is

FileCrisp is a set of 16tools for images and PDFs — compressing, converting, resizing, cropping, merging, scanning and signing. All of them run entirely inside your browser using WebAssembly. Your files are read from your own disk into your own device’s memory, and they are never transmitted anywhere.

Why that decision drives everything else

Almost every other tool in this category uploads your file, processes it on a server, and sends it back. That works, but it means your documents sit on hardware you do not control, and it means every file you process costs the operator bandwidth and CPU. That cost is why those services cap free tiers, add watermarks, or ask you to sign up.

Processing on your device costs us nothing per file. So there is no daily limit, no watermark, no account, and no file-size ceiling other than what your own device can hold in memory. The privacy is a consequence of the architecture rather than a policy we ask you to trust — you can open your browser’s network tab while you work and watch it make no requests at all.

What it means in practice

• Large files do not need uploading first, so a 50 MB PDF starts immediately rather than after a long wait on a slow connection.
• The limits are your device’s, which is why a phone can handle less than a laptop. You are told before anything starts rather than having the tab crash.
• Some things are harder this way. PDF compression, for instance, is weaker on text-only documents than tools built on server-side engines we cannot use. That trade-off is explained on the Compress PDF page.

How to check any of this

Every site in this category says it respects your privacy, so the words carry no information. This one is unusual in that the claim is checkable in about thirty seconds, without taking anyone’s word for it:

• Open your browser’s developer tools and switch to the Network tab. Drop a file into any tool and run it. You will see requests for the page and, on the first run, for the codec — and none carrying your file.
• Run a file through once, then turn off your connection and run another. It still works, because nothing was going to be sent anyway. The first run is the one that needs a connection, and only to fetch the codec.
• Every tool prints what it measured beneath the result: how many requests were made while it ran, and how many of those went anywhere other than this site. The answer to the second is zero.

What it is built with

Images are decoded and encoded by jSquash, which wraps the same encoders the rest of the industry uses — mozjpeg, libwebp and oxipng — compiled to WebAssembly. PDFs are assembled with pdf-lib and rendered with Mozilla’s pdf.js. The work happens in a Web Worker so the page stays responsive while a large file is processed.

Licensing shapes what is possible here. The two strongest PDF compression engines, Ghostscript and MuPDF, are AGPL-licensed, which is incompatible with a closed-source site funded by advertising. Choosing permissive libraries instead is the direct reason compression on text-only PDFs is weaker than some competitors — a real trade-off, stated rather than hidden.

What it does not do

• It does not store your files. There is nowhere to store them: they are never sent anywhere, and closing the tab discards them.
• It does not read text out of scans. Making a scan searchable needs OCR, which this does not do — a scanned page comes back as a picture of a page, as it went in.
• It cannot open encrypted PDFs. A password-protected file is rejected with a clear message rather than producing something broken.
• Analytics record page views and performance, never file contents. Your files cannot appear in them, because they never leave your device to be measured.

How it is paid for

At the moment, it is not — the site earns nothing. The intention is advertising placed in the explanatory content below each tool, never inside the workspace and never between you and your file. There is no paid tier and none is planned, because processing on your device costs almost nothing to run and there is nothing to meter.

Who builds it

FileCrisp is built and maintained by Navjot Suman. If something is broken, wrong, or missing, saying so is genuinely useful — get in touch on the contact page.