What you say, what you show and exactly where you point stay attached in one task packet, ready for your agent.
Building the interaction standard between people and AI agents.
The packet of a real task given by voice — in front of you, as it went out. Click a badge: the exact screen, code and frame it points at opens. Press play: the speech flows word by word. Speech, Events, Context — three faces of the same task.
One spoken session — not a single typed keyword — produced these pages: Cockpit · Dictation · How I work. Want the process itself? Walk the full flow.
Speak it. Capture the evidence. Send the whole task.
Get early access
Enumerate every task as [n/m] and proceed; change the counter when the goal changes. At the end, hand everything back in one pass.