Your files never leave your browser.
Unfleece runs every PDF operation locally via WebAssembly. There is no server that receives, stores, or processes your documents — which is exactly why it can be free.
Prove it's private in 10 seconds.
Don't take our word for it. Watch the network while you work — no document is uploaded. If a tool errors and reports are enabled, you may also see the content-blind /api/err request described below. The source is open, so you can read exactly what it does.
Open any tool and press F12 → Network tab.
Run it on a file. Watch the request list.
See no file upload. Turn off Wi-Fi — it still works.
Content-blind error reports
When a tool errors, Unfleece sends a single tiny report so we can find and fix bugs we can't see otherwise. It contains only which tool failed, an error category, the engine, your browser family and language — never your files, file names, error text, or any identifier. It honors your browser's Do‑Not‑Track, and you can turn it off here:
No server, no bill, no paywall
Because all the work happens on your computer, Unfleece has no compute or bandwidth bill — so there's nothing to monetize with a trial trap. The fleeceware sites charge you for something your own device can already do.
What it can and can't do
What works great
- Merge, split, rotate, organize, crop
- Images ↔ PDF, PDF → images / text
- Watermark, page numbers, metadata, sign
- Lossless optimize & compress
Honest caveats
- Very large files are bound by your device memory
- PDF→Text is selectable text only (scans need OCR)
- Compress uses Ghostscript; raster fallback is image-only
- We label every tool for what it actually does
Auditable by design
Unfleece is released under the AGPL-3.0 license. Contributions, audits, and self-hosting are welcome — view the source on GitHub ↗.