PDFtoAudio is designed with privacy as a foundational principle.
- Local Document Storage: All uploaded documents, extracted text files, transcripts, bookmarks, notes, and study materials are stored locally on your device within the application storage directory.
- No Secret Telemetry: Document text and private files are never transmitted to external analytics or third-party servers without explicit user setup.
- Neural Voice Stream Privacy: Voice synthesis uses local TTS synthesis pipelines or direct HTTPS streams. No user document content is logged or persisted by external analytics trackers.
- Path Traversal Prevention: Filenames and paths are sanitized before processing.
- File Upload Limits: Document uploads are constrained to a maximum size limit of 50MB and maximum 500 pages per document.
- Temporary File Cleanup: Audio stream buffers and temporary conversion files are automatically purged after generation.
- Never commit API keys or secret tokens to source code. Use environment variables as shown in
.env.example.