Verify local PDF processing
Browser tools process document contents on your device. You can inspect network requests and test a public sample offline; engine downloads and site visits still use the network.
Try it with a harmless sample
- Open Merge PDF and download the sample files.
- Open your browser's developer tools and select Network. Let page assets finish loading.
- Clear the request list, add the samples, merge and download the result.
- Inspect any requests and their payloads. Document contents should not be sent.
- Switch the network panel to Offline and repeat after all required assets have loaded.
Opening a page successfully is not proof that every operation works offline. OCR and encryption load additional assets on first use. Check the particular tool and browser you plan to use.
Capabilities and practical limits
| Operation | Processing and downloads | Important limit |
|---|---|---|
| Merge, split and organize | JavaScript PDF libraries on your device | Memory limits; document-level structures may not carry over |
| Compress to size | Pages rendered and rebuilt locally | Image-based output; small targets may not be possible |
| OCR and searchable PDF | Tesseract engine, worker and language assets load before recognition | Memory, language and scan quality affect results |
| Password protection | qpdf WebAssembly loads before local encryption | Requires WebAssembly; retain your password |
| Read aloud | Local device voices only | A compatible installed voice is required; remote voices are excluded |
| Office conversion guides | Instructions for separate software | Those applications have their own requirements and privacy behavior |
Browser support
Use an up-to-date browser with JavaScript enabled. OCR and encryption also need WebAssembly and workers. Asset loading and worker behavior can differ on mobile devices and when opening files directly from disk; use a local HTTP server for a downloaded site copy. Test a small sample before a large document.
What the site can still receive
Your browser requests pages and program files from the hosting provider. Those requests include normal network information such as an IP address and user agent. This differs from uploading document contents. Read the privacy policy.
Inspect the implementation
Core PDF operations · Additional operations · Navigation and consent behavior
Libraries: pdf-lib, PDF.js, Tesseract.js, and qpdf. These projects are independent of PDF Toolbox.