Every "compress PDF online free" tool asks you to do the same thing: pick a file, upload it, wait, download the result. It feels harmless. But think about what actually happened to that file. A document that existed only on your laptop was copied to a machine you do not control, written to disk somewhere, passed through load balancers and logs, and handed to whatever subprocessors that vendor uses. For a meme, who cares. For a signed contract, a payslip, a passport scan, or a medical report, that is a real disclosure - and a privacy policy is a promise, not a technical guarantee.
There is a better default. With WebAssembly you can compress PDF to 200KB, convert JPG to PDF, turn HEIC to JPG, merge PDF files, or resize photo to 20KB with no upload at all - every byte stays on your device.
The upload model leaks by design
The problem is not that these services are malicious. Most are not. The problem is structural:
- Retention windows. Files are "deleted after an hour" - but backups, object-storage versioning, and CDN caches all have their own lifecycles.
- Breach surface. A file that never leaves your device cannot be in someone's dump.
- Compliance drag. Under GDPR, HIPAA, or an ordinary enterprise DPA, sending a client document to a random third party is a processing event you have to justify.
- Trust cannot be audited. You cannot verify a server deleted anything. Ever.
What WebAssembly changed
The old excuse for uploading was performance. Parsing a PDF, re-encoding a JPEG, or rasterizing pages is CPU-heavy work, and a decade ago JavaScript alone was too slow to do it comfortably on a big file. That excuse is gone.
WebAssembly lets battle-tested native libraries - the same C/C++ codecs and PDF engines servers use - compile to a compact binary that runs in the browser sandbox at near-native speed. Pair it with a few modern platform APIs and the whole pipeline moves client-side:
- Web Workers keep the heavy loop off the main thread, so the UI never freezes.
- SIMD and threads in Wasm bring image encoding within a small factor of native.
- File System Access API and streams let you handle files bigger than memory without a round trip.
- Blob plus URL.createObjectURL hands the user a download that was generated locally.
The result: merging twenty PDFs or compressing a 40 MB scan happens in a couple of seconds, with zero bytes on the wire.
How to compress PDF to 200KB (and merge PDF or unlock PDF) locally
Hitting a hard size cap means re-encoding the embedded images, downsampling scans, and pruning unused objects - all of which a Wasm PDF engine does in memory. The same pipeline handles merge PDF for a multi-document application, or unlock PDF for a statement whose password you already know. Because it is local, the most sensitive file you own never becomes an upload.
JPG to PDF, HEIC to JPG, and how to resize photo to 20KB
iPhone photos arrive as HEIC and half the world's forms reject them, so HEIC to JPG is the most common conversion there is - and it is pure computation libheif-in-Wasm does in your tab. Stitching scans as JPG to PDF for a form works the same way, and when a portal demands you resize photo to 20KB, a client-side compress image online free encoder hits the target without shipping your ID anywhere.
WiFi QR code generator: keep the password in your DOM
A guest-WiFi QR encodes your SSID and password in plain text. A server-side WiFi QR code generator means typing that password into someone else's form. Generate it in the page instead and the credential never leaves your browser.
The side benefits nobody advertises
Privacy is the headline, but local processing wins on other axes too. There is no upload wait, so a 100 MB file is instant instead of bandwidth-bound. It works offline, on a plane or a locked-down network. There are no per-file quotas, because you are spending your own CPU, not someone's egress bill. And hosting is trivially cheap - static files on a CDN - which is exactly why these tools can stay free without monetizing your documents.
How to verify a tool actually does this
Do not take a landing page's word for it. Open DevTools, go to the Network tab, run a conversion, and watch. If your file is genuinely processed locally you will see the .wasm module download once and then nothing - no multi-megabyte POST. Better still, load the page, go offline, and try again. A real client-side tool keeps working.
FAQ
Can a browser really compress PDF to 200KB? Yes - a Wasm PDF engine re-encodes images and strips unused objects in memory, no server round trip.
Is this compress PDF online free with no account? Yes. There is no per-file server cost, so there is nothing to meter.
Do JPG to PDF and HEIC to JPG upload my photos? No. Both run locally under the no upload model; check the Network tab.
Can I merge PDF and unlock PDF client-side? Yes, for documents you own and whose password you know.
What is the fastest way to resize photo to 20KB? A client-side compress image online free encoder that targets the byte size directly.
That is the model behind PixelAndPDF: convert, compress, merge, and edit PDFs and images entirely in the browser, with the file never leaving your device. Run the Network-tab test on it yourself.
The default should flip. For file conversion, "we never see your file" is now an engineering fact you can ship, not a marketing claim someone has to trust.
