Why Klarfile does not upload your files
Klarfile is a set of PDF and image tools that run entirely inside the tab of your browser. This page says why that choice was made, and what it costs. The tools are here.
Online PDF services all ask for the same thing: your documents
A contract, a payslip, a medical file, a bank statement, a legal brief: uploaded to a third party, processed on its servers, “deleted within an hour”, allegedly. The terms and conditions point at sub-processors, often outside the European Union, and all you are left with is the provider's word.
For many professions that is not a preference, it is a prohibition: a lawyer, a doctor, an HR department, an accountancy firm, a public body have no business handing case files to a service they have not vetted. And yet the task itself is mundane: put four pages together and take one out.
It does not only apply to documents. “Compress your photo online” means uploading a picture of an identity card, a proof of address, a child, to a service funded by advertising that keeps the file for as long as it likes.
What happens here instead
The work is done by your own processor. The only file that moves is the one you save yourself at the end, from the memory of the tab to your disk.
Cut your network connection once the page has loaded: everything keeps working. It takes ten seconds to see.
The direct consequence: no destination server, therefore no transfer outside the European Union, no sub-processor to vet, no retention period to negotiate, nothing to erase afterwards. What that changes, precisely.
The one request this site makes
There is exactly one: after an export, the page tells a counter how many pages were processed (/compte?p=12). No byte of any document, no file name, no identifier, no cookie. It exists so that the home page can show a total, and it is written in a log that holds a date and a number. What it contains exactly.
In the package sold for intranets, that counter is removed from the build, not disabled. Klarfile on your intranet.
What it does not do
What these tools cannot do, written here rather than discovered along the way.
- It is not a PDF editor. We add a layer on top of the page; we do not rewrite an existing paragraph. A PDF is a frozen layout, and the tools that claim to edit one guess and damage it.
- A scan needs text recognition first. Otherwise it stays an image.
- A watermark is not protection. It is a notice: it shows on a photocopy, it locks nothing.
- A password-protected PDF opens with its own password, which you are asked for when you drop it and which does not leave your tab. Nothing is broken: without it, the document stays shut.
- Very large documents have a limit. Everything fits in the memory of the tab: that is the price of having no server.
Still a question?
Free of charge, cookies, maximum size, protected PDFs, offline use, phones: the ten questions we are asked most. What each tool does not do is written on its own page.