VoxRelay privacy: local audio processing and model downloads
VoxRelay processes captured playback audio, language detection, transcription, and translation on the local Windows computer. Captured audio is held in memory only for processing and is not sent to Max Power IT or Hugging Face.
The app downloads the selected CUDA or CPU model from a revision-pinned copy on maxpowerit.com and verifies every file with SHA-256. Hugging Face is retained as the backup source. These requests do not include captured audio or subtitle text, but either web server may process ordinary connection metadata such as an IP address, request time, requested file, response status, and user-agent string. VoxRelay does not require an account and contains no advertising, analytics, tracking, or telemetry SDK. Public support posts are voluntary and visible to other visitors, so never include confidential information.