How to tell whether a web tool is uploading your file
Two checks that take under a minute and do not require trusting anybody’s privacy policy — plus which operations genuinely need a server and which do not.
4 min read · updated
Why this is worth thirty seconds
Most file tools on the web upload your document to a server, do the work there, and send the result back. That is not sinister — plenty of conversions genuinely require it — but it does mean your file has been copied to a machine you do not control, by a company you may not know, in a country you did not choose.
For a holiday photo, nobody cares. For a payslip, a contract, a medical letter, a passport scan or a document you are about to redact, the calculation is different. And the only way to know which kind of tool you are using is to check, because both kinds look identical.
Check one: watch the network
Every browser ships with a tool that shows exactly what a page sends and receives. It takes longer to describe than to do.
Open developer tools
Press F12 on Windows, or Option-Command-I on a Mac. In Safari, enable the Develop menu in Settings → Advanced first.
Switch to the Network tab and clear it
This lists every request the page makes. Clear the list so you are only looking at what happens next.
Use the tool
Pick your file and run the conversion as normal.
Look at the size column
An upload shows up as a request carrying roughly the size of your file. A 4 MB document cannot be uploaded without about 4 MB going somewhere. If the largest thing that moved is a few kilobytes, your file stayed put.
Check two: turn the internet off
Cruder, faster, and impossible to misread. Load the page, then disconnect — turn off Wi-Fi, or switch on aeroplane mode. Now use the tool.
If it works, nothing was uploaded, because there was nowhere to upload it to. If it stalls, hangs, or reports an error, the work was happening somewhere else.
What “we delete your files” actually promises
Most upload-based tools say files are deleted after an hour, or a day. That is usually true and it is worth something. It is also worth understanding what it is not.
- It is a promise about a future action, not a description of the present. The file is on their server right now.
- It covers the file. It does not necessarily cover backups, logs, caches, or copies made by whatever service they run on.
- It is unverifiable from outside. You cannot check it, which means it rests entirely on trust in the operator.
- It says nothing about who could read the file during the window it exists.
None of that makes upload-based tools unusable. It makes them a judgement call about the specific document, which is a judgement you can only make if you know an upload happened.
What can genuinely run in a browser — and what cannot
A browser is a capable computer, but not everything fits in one. Being honest about the line matters, because a site claiming everything is local is either mistaken or misleading.
| Operation | Can it be local? | Why |
|---|---|---|
| Merging, splitting, reordering PDF pages | Yes | Rearranging pages is structural work on the file, and needs no rendering. |
| Images into a PDF | Yes | The browser already decodes images; assembling them into a document is straightforward. |
| Redacting by flattening pages | Yes | A browser can render a PDF page and rebuild it as an image. |
| Reading text out of a PDF | Yes | Parsing the text layer is ordinary work for a script. |
| Converting an ebook to a PDF with correct typography | Not realistically | It needs a full layout engine driving pagination, hyphenation and fonts — a browser cannot do that to its own pages from inside a tab. |
| Office documents to PDF | Rarely done well | Faithful output depends on fonts and layout rules that live in the original application. |
Be suspicious of any site that claims every one of its tools is local, especially if that list includes ebook or office conversion. It is a sign the claim is marketing rather than architecture.
Frequently asked questions
Does a padlock in the address bar mean my file is private?
No. HTTPS means the connection cannot be read in transit. It says nothing about what the server does with the file once it arrives. An upload over HTTPS is still an upload.
What if the tool works offline but only after I used it once online?
That usually means the page cached its code, which is normal and fine. The test still holds: if the conversion itself completes with no connection, the conversion is not happening on a server.
Is a desktop application always safer?
Not automatically — plenty of desktop software also sends documents to a cloud service. The same two checks work on a desktop app: watch its network activity, or disconnect and see whether it still functions.
I already uploaded something sensitive. What now?
Check whether the service offers an immediate delete and use it, change any credential that appeared in the document, and treat the contents as disclosed when deciding what to do next. Then use a local tool for the replacement.
Related tools
- Redact PDF
Remove text from a PDF for good — then see it proved. Nothing is uploaded, and a black box here is not just a black box.
- Merge PDF
Join several PDFs into one document, in the order you choose — without uploading anything.
- Split PDF
Pull out the pages you need, or break a document into one file per page — without uploading anything.
- JPG to PDF
Turn photos, scans and screenshots into a single PDF — in the order you choose, without uploading anything.