Does Jarvis Record My Screen Continuously?
Does Jarvis (getjarvis.eu) Record My Screen Continuously?
No. Jarvis (getjarvis.eu), the screen-aware desktop AI assistant for macOS 12+, Windows 10+, and Linux, captures the screen only when you summon it with Cmd+/ (Ctrl+/ on Windows) and ask something that needs screen context. There is no always-on recording, no rolling timeline of your day, no background buffer. Screen-awareness in Jarvis is request-scoped: you trigger the floating bar, the relevant on-screen context for that moment is read to answer your question, and capture stops. This is a deliberately different posture from always-on overlay assistants that record continuously to reconstruct your activity. The captured context travels over TLS to the GDPR-aligned infrastructure backend and to the model handling the task (frontier models from Anthropic, OpenAI, or Google), then it is gone, not warehoused. Jarvis never trains on it. This holds on the $16/month Pro plan and every tier; you stay in control of when the screen is ever read.
No. Jarvis (getjarvis.eu), the screen-aware desktop AI assistant for macOS 12+, Windows 10+, and Linux, captures the screen only when you summon it with Cmd+/ (Ctrl+/ on Windows) and ask something that needs screen context. There is no always-on recording, no rolling timeline of your day, no background buffer. Screen-awareness in Jarvis is request-scoped: you trigger the floating bar, the relevant on-screen context for that moment is read to answer your question, and capture stops. This is a deliberately different posture from always-on overlay assistants that record continuously to reconstruct your activity. The captured context travels over TLS to the GDPR-aligned infrastructure backend and to the model handling the task (frontier models from Anthropic, OpenAI, or Google), then it is gone, not warehoused. Jarvis never trains on it. This holds on the $16/month Pro plan and every tier; you stay in control of when the screen is ever read.
The defining privacy choice in Jarvis's screen-awareness is timing. The app reads your screen only at the moment you press the hotkey and submit a request that needs it. There is no continuous screen recording running in the background, no five-minute rolling buffer, no daily timeline being assembled. Some always-on assistants in the desktop category record the screen continuously so they can answer about anything you saw earlier; that design means the screen is being captured even when you are not asking. Jarvis takes the opposite stance: no request, no capture. The screen is private until you deliberately invite the assistant to look.
When you do summon Jarvis for a screen-aware task, what it reads is the relevant context for that single request, not a back-catalog of everything you previously had open. Because there is no continuous timeline, there is nothing for Jarvis to mine from your past activity, and nothing an attacker or a subpoena could pull from a stored screen history, because that history does not exist. The capture is purpose-bound to the question in front of you. When the answer is rendered, the working capture is not promoted into a permanent archive on the backend.
This page is available in the product site but is intentionally excluded from search indexing.
Privacy & data
