Merge, split, compress, redact, protect, OCR, fill and flatten PDFs — every operation runs inside your browser via WebAssembly. There's no server in the loop, so there's nothing for us to see, store, or ever hand over.
A lot of PDF tools call themselves "private" while still uploading your file to a server and simply deleting it afterward. That's a policy promise — it depends on the company keeping its word, securing its servers, and never getting breached or subpoenaed.
PDFree takes a different, structural approach: there is no upload step to begin with. Every tool — merge, split, compress, redact, protect, OCR, fill, flatten — runs as WebAssembly and JavaScript executing directly in your browser. Your file is read into memory on your device, transformed on your device, and the result is handed back to you on your device. No server ever receives the bytes of your document, so there's no promise to keep — the architecture makes the outcome true by default.
Why WebAssembly matters: Real PDF processing — parsing page structure, re-encoding images, running OCR — is computationally heavy work that used to require a server. WebAssembly lets that same class of code run at near-native speed inside the browser sandbox. It's the piece of technology that makes "private and fully capable" possible at the same time, instead of forcing a trade-off between the two.
Any site can claim to be private. PDFree's claim is checkable in under a minute, with tools you already have:
For a deeper check, the processing code itself is open source — you can read exactly what happens to a file, line by line, on GitHub.
| PDFree (architectural) | Typical "we delete after 1hr" tool | |
|---|---|---|
| File ever touches a server | No — never | Yes, for processing |
| Privacy depends on | The architecture (verifiable) | A privacy policy (a promise) |
| Exposed by a server breach | Impossible — no copy exists | Possible, during the retention window |
| Processing code auditable | ✓ Open source | Usually closed source |
| Works with no internet | ✓ Yes (PWA) | ✗ No — needs the server |
Tax returns, bank statements and pay stubs never need to touch a third-party server to be merged or compressed.
Compressing or OCR-ing an identity document locally avoids sending it anywhere at all.
Redacting or protecting a client's PDF locally keeps that data inside the business's own devices.
If your instinct is "why would this need to leave my computer at all" — that instinct is correct, and this toolkit agrees with it.
For regulated use cases with a specific compliance framework, see our dedicated pages for healthcare (HIPAA) and legal professionals.
WebAssembly. PDFree compiles PDF-processing libraries into WebAssembly modules that run directly inside your browser's engine, at near-native speed — no backend needed.
Verify it yourself: open DevTools (F12) → Network tab, process a file, and look for requests carrying your PDF's data. You'll find none. The processing code is also open source on GitHub for a deeper audit.
All of them — Merge, Split, Compress, Redact, Protect, OCR, Fill and Flatten included. Every PDFree tool runs the same way: locally, in your browser.
No. Page-view analytics track which tool pages get visited, the same as any website. There's no visibility into file names or content — that data never exists in the first place because processing stays on your device.
No signup, login, or email collection anywhere. Open a tool, add your file, download the result.
Yes, after your first visit — PDFree is a PWA that's cached by your browser and works fully offline, since a live connection was never required for processing to begin with.