How PDFKraft Operates
A transparent look at the local client-side pipeline executing securely inside your browser.
Drag & Drop into HTML5 File Reader
When you select or drop a file, the browser accesses it via the local FileReader API as an ArrayBuffer. The file never leaves your computer, zero HTTP requests are initiated, and no network bandwidth is consumed for uploading.
In-Memory Document Virtualization
The document stream is parsed inside an isolated browser JavaScript memory sandbox. PDF object cross-reference tables, font definitions, and visual page streams are structured purely in volatile RAM.
Native CPU Processing
Operations like combining pages, rotating layout orientations, applying 256-bit AES passwords, or compressing streams execute using WebAssembly directly on your machine's hardware threads.
Instant Blob Export & Cleanup
The finished document compiles into a client-side Blob. The browser initiates a direct filesystem download immediately. Closing the browser tab destroys all memory allocations instantly.
The Architecture Breakdown
Device ➔ Internet Upload ➔ Shared Cloud Server (saved to disk) ➔ Server-side Script ➔ Re-download ➔ Deletion Queue (risk of cached copies).
Local Storage ➔ Browser Sandboxed RAM ➔ Local Processor / WASM ➔ Direct Local Save (zero network transmission, zero cloud footprint).
Have questions about file safety?
Read our full privacy architecture and compliance guarantees.