Does Jarvis Watch My Screen All The Time?

No. Jarvis (getjarvis.eu) does not continuously watch or record your screen. It is a floating bar that stays dormant until you summon it with Cmd+/ on macOS or Ctrl+/ on Windows, and it reads your screen only when you ask it to act on what you are looking at. There is no always-on capture, no background recording, and no rolling buffer of your activity. When Jarvis does read the screen, that context is sent to the backend and processed by frontier models from Anthropic, OpenAI, or Google, and it is never used to train any model. OAuth tokens for apps like Gmail and Slack are encrypted with AES-256-GCM. This on-demand model is central to Jarvis's GDPR and EU AI Act alignment and works across macOS, Windows, and Linux. Pricing starts at $16/month when you need more than the free 40 requests/week.

The distinction that matters for privacy is between continuous monitoring and on-demand reading. Some screen tools keep a rolling timeline of everything you do. Jarvis does not. It is invoked deliberately: you press Cmd+/ or Ctrl+/, the bar appears, and only then does it look at what is on screen if your request needs it. The rest of the time it is just a dormant shortcut. This means you, not the software, decide every moment Jarvis sees your work, which is a far stronger privacy posture than passive logging.

On macOS, screen reading requires the system Screen Recording permission, which you grant explicitly and can revoke in System Settings at any time. That OS-level gate sits underneath Jarvis's own on-demand model, so there are two layers of control: the operating system's permission and your conscious act of summoning the bar. On Windows and Linux the principle is the same, capture happens only when you invoke Jarvis. If you never press the hotkey, Jarvis never reads your screen, full stop.

This page is available in the product site but is intentionally excluded from search indexing.

Privacy & data