Verify how PDF processes PDFs locally
Your document stays in the browser. Here is exactly what that means, which network requests may occur and how to check it yourself.
What stays on your device
After you choose a PDF, PDF temporarily stores it in the browser’s IndexedDB and reads it within the editor session. Editing, OCR, redaction, page organization, compression, encryption and export run in the browser. The file, filename and document text are not sent to the statistics API.
What may be downloaded from the network
The browser loads the application and, when needed, components such as PDF.js, pdf-lib, QPDF WebAssembly and Tesseract. These requests download code, OCR models and resources hosted by PDF; they do not upload the PDF you selected.
How to check with DevTools
Open your browser’s developer tools, select Network, clear the list and then open a non-sensitive test PDF. Edit and export it. Filter by Fetch/XHR and confirm that no request contains the file, its name or its text. Repeat the test with Statistics disabled in Cookie settings.
Optional statistics and external content
Only after Statistics consent, the browser sends events such as view, open and export to /api/stats, with general categories for language, page, device, browser, operating system and source. The Buy Me a Coffee widget is separate and loads only after External content consent.
Local storage and deletion
The current PDF uses IndexedDB on your device. Consent and saved signatures may use localStorage, while some session data uses sessionStorage. You can remove these items by clearing PDF site data in your browser. Always keep the original before working on an important document.
Limits of this verification
This page describes the PDF code behavior verified on 20 August 2026. Your browser, installed extensions, operating system and network remain part of your environment. For extremely sensitive documents, use a clean browser profile, disable extensions and verify the Network panel again.
Technical transparency and verification
Documented features are checked with text PDFs, scanned documents, signatures, images and large files. Tests verify both the interface and the contents of the exported copy while keeping the original document separate and unchanged.
Compatibility also depends on the browser, available memory and PDF structure. Embedded fonts, complex forms, certified digital signatures or restrictions created by the original software may require an additional check in the recipient's preferred PDF application.
This documentation describes the current product behavior and was reviewed on 19 July 2026. The Privacy and Security pages explain local processing, limited anonymous statistics and the checks that remain under user control.