How it works

1. Create a screen — it lives in your own space ({you}/Screens/…).
2. Open its display full-screen on the device that should be the screen (TV, beamer, second monitor). The display follows every command instantly.
3. Steer it:
   • 🎛️ the screen's Remote on your phone — present any node/area, one-tap re-present from history;
   • 📺 Present on screen in any node's menu — one tap presents that node (a specific area is picked on the remote);
   • 💬 the Remote Control agent in a thread — "show the claims dashboard", "next slide", "blink";
   • 🔌 from another instance — connect to this instance's MCP endpoint (OAuth as yourself) and steer the screen node in a connecting thread; the /remote skill carries the procedure.
4. 📡 Project a session — instead of pushing pages one at a time, let the screen follow you: open the portal's user menu, pick Project this session, confirm, and every page you open lands on the screen. Opt-in only, a 🔴 dot in the chrome while it runs, one click to stop, and gone when your session ends. See the design.
5. 👉 Point — open the screen's pointer pad, tap on the content, and a laser dot pulses at that spot on every display for a few seconds. Highlight the number you're talking about without leaving your seat.
6. 🔦 Blink — not sure which display you're steering? Blink it: the screen flashes its name full-screen.

Everything the screen shows is this instance's data under your own access rights — a screen presents layout areas of nodes you may read, nothing else, and multiple displays can follow the same screen.

The dynamic node types

• Screen — the shared presentation surface: display + remote as two live views of one node.
• Console — the Start launcher and the one-tap Cast target of the node-menu entry.

See the Guide for the mechanism and the cross-instance channel.

Contents