Think inside your AI world.
The governed routine: scheduled, but not ungoverned
A time-based loop runs on an interval — /loop locally, /schedule in the cloud as a routine — until you cancel it or the work completes. The governed routine keeps the schedule but refuses the ungoverned merge: each run injects the canon current at that moment and stages its output as a diff against your settled decisions.
A scheduled routine is genuinely useful — a report every Monday, a queue drained overnight. The risk isn’t the schedule; it’s that a run on Wednesday can act on a decision you changed on Tuesday, and merge before you look. The governed routine closes that gap: current canon is injected at execution time, and the output is staged, never merged ungoverned — a diff against settled decisions plus a short digest of what needed a judgement you haven’t made. Unl holds the canon each run reads.
How does a scheduled loop run today?
Fairly described: a time-based loop is triggered by an interval. /loop runs it locally; /schedule moves it to the cloud as a standing routine. It ends when you cancel it or the work is done — a PR merges, a queue empties. For repetitive, low-judgement work this is exactly the right shape, and it’s a real convenience.
The subtlety is timing. A decision you settle on Tuesday should govern Wednesday’s 6am run — but a routine set up last month carries last month’s assumptions unless something feeds it the current ones. And because the run isn’t watched, its output can land before you’ve read it.
What is the governed routine?
Two disciplines on top of the schedule. First, each run injects the canon that is current at execution time, so Tuesday’s decision governs Wednesday’s run rather than a stale copy. Second, the output is staged, not merged: the routine produces a diff against your settled decisions and a digest — “three items handled within canon; one item needed a judgement you haven’t settled, parked for you.”
Nothing that touches an unsettled judgement becomes canon without a gesture. The routine keeps running on its interval; what it can no longer do is quietly merge a change that crosses a decision you never made.
Where does the ratification gesture live in a scheduled routine?
At the staged digest. The run does its work unattended, but its results wait as a reviewable diff against settled decisions; the items that stayed within canon are noted, and the one that needed a fresh judgement is parked for a single ratify-or-discard gesture when you next look. Governance sits at the merge, not at every 6am firing.
Each scheduled run reads current canon and stages a diff against your settled decisions — every review meeting is a manual measured read, done once, up front.
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 a governed routine?
A scheduled loop that injects the canon current at execution time and stages its output as a diff against your settled decisions instead of merging ungoverned. It returns a digest — what was handled within canon, and what needed a judgement you haven’t settled — and nothing touching an unsettled decision becomes canon without a gesture.
Why not just let a scheduled agent merge its work?
Because a routine set up once carries the assumptions it was set up with, and an unwatched merge can cross a decision you changed since. Injecting current canon at each run and staging the output as a diff means a Tuesday decision governs a Wednesday run, and the merge waits for a gesture.
How does the routine know my current decisions?
It reads them at execution time from Unl, so the latest settled decision governs the next run. 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.
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.