Home / How this differs
Where your document gets processed
Almost every online PDF tool uploads your file, does the work on a server, and sends a result back. This one does the same work in your browser, so the file never leaves your device. That single difference explains the account, the page limits, the paid tier and the retention policy — none of which exist here.
This page names no products, on purpose. Prices and limits elsewhere change without notice, and a stale claim about somebody else would undermine the one thing this site is for. What follows is structural, and it holds for any tool that processes your document on its own hardware.
On a server, or on your device
| Processed on a server | Processed on your device | |
|---|---|---|
| Your document is copied to a server | ✓ | — |
| A copy persists in logs and backups | ✓ | — |
| Usage metered, capped, or gated behind an account | ✓ | — |
| Works with the connection unplugged | — | ✓ |
| Batch processing without a paid tier | — | ✓ |
| Runs inside an air-gapped network | — | ✓ |
| The whole editor is one portable file | — | ✓ |
| Reconstructs editable Word and Excel documents | ✓ | — |
| Reads scanned pages with OCR | ✓ | — |
The last two rows are ours to lose. They are on the list because a comparison that only ever wins is an advertisement.
Why hosted tools are built that way
It is a cost decision, not a privacy one. Converting a PDF takes real computing time, and if that computing belongs to a company then the company pays for it. Anything paid for has to be measured, which means an account; anything measured eventually has a limit; and a limit is what a paid tier removes.
Every constraint you meet on those sites follows from the first step. Move the work to the machine that already holds the document and the whole chain unwinds — nothing to upload, nothing to meter, no account, no reason to charge, and no server that could be breached because there is no server.
When a server-based tool is the better choice
Honestly, sometimes.
- Converting to Word or Excel. Rebuilding an editable document from a PDF is genuinely hard and needs more than a browser tab should attempt. It is not offered here.
- Reading scanned pages. OCR is possible locally but heavy, and it has not shipped yet.
- Reflowing a paragraph. You can replace a line of existing text here and it exports as real text, but the surrounding paragraph will not re-wrap around it.
- A very large file on a very small device. A phone has less memory than a server does.
When this is the better choice
- The document is confidential. Client files, medical records, HR documents, anything under privilege or a confidentiality clause. Uploading creates a disclosure; processing locally does not.
- There is no connection. On a plane, in a restricted facility, on an unreliable link.
- You do not want an account. There is none to create.
- You have a folder, not a file. Batch runs on your own processor, so nothing meters it.
- You stamp or number documents repeatedly. Design a stamp once, save a workflow, and send either to a colleague as a file.
What we have chosen not to build
The roadmap lists this in full, and the short version belongs here too. There is no permanent redaction: covering text with a shape hides it visually and leaves it selectable underneath, and until we can remove content and prove it is gone, the answer is no. There is no OCR. Signatures are visual, not cryptographic. Encrypted files are refused rather than quietly mangled.
A tool that lists only its strengths is asking to be trusted. This one would rather be checked.
Try it on something you would not upload.
No account, no limits, and you can watch the network tab while you work.