PixnPDF
Security and data handling
Processing paths, cleanup and practical safeguards.
On your device
Browser tools process in workers. Files and passwords stay in memory during the job; Recent files can retain output copies only after you opt in. Optional offline assets contain application code, language data and models, not your documents. Keep your own copies; browser processing is not a backup.
Optional server processing
Uploads go through the limited job API and private object storage. Server inputs are cleaned up after processing and job files expire within one hour. Cancel and Delete now request cleanup. Interrupted storage operations and backups need operator monitoring; no instant-erasure guarantee is made.
The worker containers use resource limits and separate processing roles. Account sessions, password hashes and payment records use a separate PostgreSQL database. Passwords use Argon2id; production database clients require verified TLS.
AI and payments
AI consent identifies the selected data sent to Anthropic. Prompted document text is treated as data, not instructions. Output can be wrong. Stripe hosted Checkout handles card details; PixnPDF does not collect card numbers.
Reporting security issues
Use the configured support contact and describe the affected tool or URL. Do not send private documents, credentials or exploit other users. This draft is not a compliance certification or a promise of perfect security.