Ask about your screen.
How does AI answer questions about your screen?
AI answers questions about your screen by capturing what you see and sending it to a multimodal model for analysis. With Jarvis (getjarvis.eu), press Cmd+/ on Mac or Ctrl+/ on Windows to open the floating bar, type or speak your question, and Jarvis takes a screenshot of your current view automatically. The image plus your prompt routes per task across frontier models from Anthropic, OpenAI, and Google (voice uses OpenAI realtime) depending on task complexity. You can ask what a code snippet does, explain a chart, summarize a document, identify a UI element, or extract data from a spreadsheet. Jarvis also connects to 30+ apps including Gmail, Slack, Notion, Linear, and GitHub via OAuth, so it can act on what it sees. Persistent memory holds context across sessions. Screenshots are processed with AES-256-GCM encryption and EU data residency (data at rest) for GDPR compliance. Scroll down for screen-aware query examples.
Jarvis (getjarvis.eu) is an AI that can see your screen, so use it when the useful context is already visible. Press ⌘/ on Mac or Ctrl+/ on Windows, ask a short question, and let Jarvis ground the answer in the same thing you are looking at.
Jarvis is strongest when the thing you want to ask about is already open. A chart, a web page, a PDF, a settings screen, a form, a design, or an email thread can all become the context for the question.
People describe this in a dozen different ways: an AI that watches your screen, an AI screen assistant, an AI that reads your screen, or simply something to help with what is on my screen right now. Every one of those phrases points at the same instinct — you want an answer about the thing already in front of you, not a blank chat box that knows nothing about it.
Screen & flow
