Otto only reads your screen if you switch it on — offline, as text, with its own expiry. On request, it additionally looks at an image if you allow it. And only in the autopilot level does it operate a web page itself, in its own hidden window, with three spots it never clicks.
Contents
Docs overviewScreen & autopilot
Two ways of looking, one principle
Otto can look at what is on the screen in two quite different ways — and in the autopilot level it can even operate a web page itself. This page describes both, and where they draw the line.
What belongs together here
- Screen memory
- Off by default. Switched on, Otto reads along on a fixed beat, stores only text, and skips locked-out spots.
- Look on request
- A second, separate permission: on “what am I looking at here?” Otto takes a picture of every screen.
- Operating a browser
- Opening, reading, clicking, filling in, submitting — at every autonomy level, with rising approval classes.
- Pointer control (autopilot)
- Otto sees a picture of its own window and clicks a spot — only at the autopilot level, never at the user’s real pointer.
In this section
Connections — the catalogue of every channel How it works — the three autonomy levels in detail