First run: empty workspace to published post
What does a new user actually do, step by step?
The complete cold start in three lanes — what the operator does, what Cradle does, and what the fleet does.
Drag to pan, scroll to zoom, click an element to trace it.
Open full screenWhat to look for
- The three lanes, and how often work crosses between them.
- The "single use" and "write-only" tags on the two security branches.
- The final edge coming back up from the fleet — that is what makes the post done.
Read the full explanation
This is the journey a real person takes on day one, drawn so you can see all three participants at once. The top lane is the human: sign up, name a workspace, add an agent, add an account, compose, watch. The middle lane is Cradle reacting. The bottom lane is the fleet doing physical work.
The side branches are where the security decisions live. Signing up sends a six-digit code to your email. Adding an agent mints a token that is valid for two hours and works exactly once. Adding a social account writes its password into a vault and never shows it again — not to you, not to an admin, not anywhere in the interface.
Follow the return edge near the end, from the fleet lane back to Cradle. A post is not marked published because the app sent an instruction; it is marked published because a device reported that it happened. Every tick on the onboarding checklist works the same way — it is driven by a real event, never by having visited a screen.
The whole path works without owning a single phone. The agent step accepts the simulator, which registers and reports devices exactly as real hardware would.
Anyone evaluating the product, and anyone writing onboarding or support material.