Skip to content
PDF privacy guide · offline testing

What does it really prove when a PDF tool works offline?

Disconnecting from the Internet is one of the most useful ways to investigate how a web PDF tool processes documents. But the conclusion depends on exactly what was loaded before disconnecting.

The short conclusion

If you fully load a tool, disconnect before selecting the PDF, then complete the operation and obtain a valid result, that strongly supports the conclusion that the tested workflow does not need to send the document to a server. It does not automatically prove that the whole application can cold-start offline.

Two different tests that should not be confused

Test A · preload, then disconnect

  1. Open the tool online.
  2. Wait for the application and required components to load.
  3. Disconnect the network.
  4. Only then select the PDF.
  5. Perform the operation.
  6. Verify that the result is valid.

This tests whether the operation requires remote communication after the application is already available in the browser.

Test B · fully offline cold start

  1. Close the page or browser.
  2. Keep the device offline.
  3. Try to open the application from scratch.
  4. Observe which resources remain cached.
  5. If the application starts, select the PDF and perform the operation.

This additionally tests whether the application itself can load without Internet access. It is a different claim.

Why a locally processing tool may still need Internet initially

  • Initial HTML may come from a server.
  • Processing JavaScript may download when the page opens.
  • Fonts, translations, icons or other resources may not yet be cached.
  • Browser cache may have been cleared or invalidated.
  • Some resources may load only when a particular feature is opened.

Cache does not turn remote processing into local processing

Cached application code explains why an interface can remain available after disconnecting. The critical question is what happens when the PDF is selected after the network is gone.

Test order is essential

  1. Use a synthetic PDF.
  2. Load the tool before selecting the document.
  3. Wait for loading to finish.
  4. Disconnect.
  5. Verify that connectivity is really gone.
  6. Only then select the PDF.
  7. Perform a real operation.
  8. Verify the output file.

Tests that can lead to misleading conclusions

Selecting the PDF before disconnecting The document might already have been uploaded.
Disconnecting after pressing Process A remote request may already have started.
Only checking whether the interface remains visible The interface can be cached while processing still requires a server.
Not validating the output A download button alone does not prove successful processing.
Generalising from one feature Offline merge does not prove compression, OCR or signing uses the same architecture.

What the test can strongly support

If the PDF is selected after disconnection and a valid result is produced without restoring connectivity, that execution did not require sending the document to a server to complete the tested operation.

What the test does NOT prove

  • That every tool on the platform is local.
  • That the application can cold-start offline.
  • That no telemetry exists when online.
  • That future versions use the same architecture.
  • That every browser, file size and feature behaves identically.
  • That no transfer occurs at other moments outside the tested processing flow.

Combining offline testing with Network inspection is stronger

Offline testing asks whether communication is required. Online Network inspection shows which requests actually occur. Together they provide stronger evidence.

What we tested in PDFPrivado

In PDFPrivado's published audit, each tool was preloaded online, the network was disabled and only then were synthetic documents selected. All 14 tested tools completed their workflow and produced valid results. In the equivalent online run, no document transfer was observed during processing. A fully offline cold start was not tested.

Precision matters

The correct claim is therefore not that PDFPrivado always starts completely offline, but that the 14 audited operations worked after preloading and disconnecting before document selection.

See methodology and evidence in Privacy Lab →

Related guides

How to tell if a PDF tool uploads your documents → Local vs server PDF processing → HTTPS does not mean your PDF is not uploaded → Telemetry, analytics and document transfer → Back to the Guides Center

Frequently asked questions

If a tool works after disconnecting, does that mean local processing?

It is strong evidence for the tested operation if the PDF was selected after disconnecting and a valid result was produced.

Is working offline the same as starting offline?

No. An application may continue after its resources are loaded while being unable to cold-start without Internet.

Do I need to clear browser cache to test local processing?

Not necessarily. Clearing cache tests offline startup. To test whether processing needs a server, preload first and disconnect before selecting the PDF.

Does one offline test describe the whole website?

No. Conclusions should be limited to the tested feature, version, browser and conditions.

Published and verified: September 7, 2026