Are online PDF tools safe to use?

Two very different designs

Traditional PDF sites upload your file to their servers, process it there, and send back a result. Your document leaves your device, and how long it stays on theirs depends entirely on their policy and their competence.

Browser-based tools do the work locally. The code that manipulates the PDF is downloaded to your device and runs there, the same way a calculator app runs on your phone. The file never travels.

When the difference actually matters

For a flyer, a menu or a public form, it genuinely does not matter. Uploading a restaurant menu to a server exposes nothing.

For an employment contract, medical results, a bank statement, a national ID, a passport scan, a tenancy agreement, exam papers or anything covered by a confidentiality clause at work, it matters a great deal. In some professions it is not merely unwise but a compliance breach: uploading client files to an unvetted third party can violate legal privilege, medical confidentiality rules or your employer's data policy.

A useful rule: if you would hesitate to email the file to a stranger, do not upload it to one.

How to tell which kind you are using

The offline test is the conclusive one, and it needs no technical knowledge at all. Load the page, switch on airplane mode, and try to do the job. A server-based tool will fail immediately, because it has nowhere to send your file. A local tool carries on as though nothing happened.

This is worth doing once with any tool you plan to trust with sensitive material. A claim in a footer is not evidence; a silent network tab is.

1. 1Open your browser's developer tools and select the Network tab before running the tool.
2. 2Process a file and watch the requests. An upload appears as a large request carrying your file's size — a 10 MB document cannot cross the network invisibly.
3. 3Stronger test: load the page, disconnect from the internet, then use the tool. If it still works, the processing is genuinely local.

What a good privacy policy actually says

If a service does upload, its policy should answer four questions plainly: what is stored, for how long, who can access it, and what happens to backups. Vague reassurance — "we take your privacy seriously", "files are handled securely" — answers none of them.

Look for a specific retention window ("files are deleted within one hour"), a statement about whether staff can access uploads, and a named company with a real address. Be especially wary of a policy that reserves the right to use uploaded content to "improve our services", which can mean almost anything.

"Deleted after processing" and "deleted from backups" are very different promises. Most policies make only the first.

Other warning signs worth noticing

Check for HTTPS in the address bar — without it, your file is readable by anyone on the network path, regardless of what happens at the other end.

Be wary of tools that require an account for basic operations, or that add a watermark and then sell its removal. Both are signs the business model is not what it appears, and a service that monetises frustration is not one to trust with a contract.

Aggressive advertising that opens new windows, downloads that arrive as .exe rather than .pdf, and requests for permissions a PDF tool has no use for are all reasons to close the tab.

The honest exception: some tasks need a server

It would be misleading to claim everything can run locally. Converting a PDF to an editable Word or Excel file requires layout analysis that is genuinely heavy, and most services do that work on a server. The same applies to some advanced OCR and to conversions that depend on licensed fonts or software.

Recognising text is a borderline case. It is normally a server-side service, but the recognition engine can be shipped to the browser and run on your device — which is how the OCR here works, including for Arabic.

So the practical approach is not "never upload", but "know when you are uploading, and decide deliberately". Use a local tool for anything sensitive, and if a task genuinely requires a server, choose a service with a clear retention policy and consider whether that particular document should be involved at all.

A short checklist

1. 1Does it work with the internet disconnected? If yes, your file is safe by construction.
2. 2Is the site on HTTPS?
3. 3Does the privacy policy give a specific retention period?
4. 4Is there a real company name and a working contact address?
5. 5Does it demand an account, or watermark your file and sell the removal?
6. 6For this particular document, would you mind a stranger having a copy?