Think inside your AI world.
The authored proactive: policy you pre-authored
A proactive loop is the highest-autonomy shape: it fires on an event or a schedule with no human in real time, composing scheduling, goals and dynamic workflows. The authored proactive gives it the strongest frame — you pre-author the policy as canon, and anything that touches an unsettled judgment lands in a ratification inbox rather than executing.
Proactive autonomy is where the aftermath cost concentrates, because no one is watching any single round. The authored proactive answers the hardest version of the question — where is the ratification gesture when the human isn’t watching at all? — by moving it to two places: a policy you pre-author as canon (what may be handled, what must be escalated), and an asynchronous ratification inbox for anything outside it. Unl holds the policy and the running ledger of what the rounds did.
How does a proactive loop work?
Fairly described: a proactive loop is triggered by an event or a schedule with no human in the moment. It composes the other shapes — a schedule to fire it, a goal to bound each task, dynamic workflows to carry it out. Each task exits on its goal; the routine runs until it’s disabled. For well-bounded, low-judgement work this is powerful, and it’s the natural end of the delegation ladder.
It is also where a settled decision is most exposed, precisely because no one reviews any single round in real time. Without a frame, the aftermath here isn’t a phase you enter once; it’s a backlog that accumulates quietly between the rounds and the next time you look.
What is the authored proactive?
The strongest-autonomy shape gets the strongest authorship. You pre-author the policy as ratified canon: what may be handled autonomously, and what must be escalated. Anything inside policy executes; anything touching an unsettled or high-judgement decision lands in a ratification inbox instead of acting. And the aftermath becomes a standing surface — a running ledger of what the autonomous rounds did, drift flags against canon, and one-gesture recovery.
So the human isn’t watching every round, and doesn’t need to: the policy they authored governs the routine ones, and the queue catches the rest for a gesture when they return.
Where does the ratification gesture live when no one is watching?
In two places, both moved off the live moment: the pre-authored policy that decides what may run unattended, and the asynchronous ratification inbox that holds anything outside it. Between them sits the aftermath ledger — not an omitted phase you discover the hard way, but a standing surface where drift is flagged against canon and recovery is a single gesture.
You pre-author the policy as canon; anything touching an unsettled judgment waits in a ratification inbox — the aftermath is a designed surface, not a backlog you discover.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. The reach lane is live: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
Read further
Questions people ask
What is the authored proactive?
The proactive loop — event-triggered, no human in real time — framed by a policy you pre-author as ratified canon. What’s inside policy executes; anything touching an unsettled or high-judgement decision lands in a ratification inbox rather than acting, and a standing ledger flags drift against canon for one-gesture recovery.
How can an unwatched agent stay governed?
By moving the gesture off the live round: you author the policy up front (what may run, what must escalate), and everything outside it queues for ratification. The human doesn’t review each round; the authored policy governs the routine ones and the inbox catches the exceptions.
Isn't a ratification inbox just another queue to clear?
It’s the phase the loop diagrams leave off, made explicit and cheap: only the items that touched an unsettled judgement land there, each with its divergence named, so clearing it is a gesture per item rather than a reconstruction. Unl holds the policy and the ledger; the reach lane is live, PostHog first wired.
What this is
Think inside your AI world — you stay in command
Unlimitless (Unl to friends) holds what you've settled, reads what your tools are showing, and catches what's changed out in the world — and hands your AI whatever bears on the work, the moment it's needed, without you asking. The right thing, in front of the model, unprompted, with you in command of the call. So you keep moving toward what you set out to build, on top of everything you've already decided.
It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector. Quick to connect, in a couple of steps.
Unlimitless is open now to invited Alpha. Apply for the Beta waitlist to come in ahead of the full launch:
Alpha is invite-only · free at launch.