A situation we handle
When your AI remembers what you said — not what you decided
It brings up the old discussion, word for word — then re-argues the conclusion you reached in it. Maddening, but not mysterious: a transcript can't tell what stands from what was said. Only a settled decision with its why can.
Happy to. One thought first — I remember we discussed annual vs monthly billing a few weeks back. Have you considered leading with monthly? It usually converts better, so it might be worth revisiting before we finalise the copy…
It remembers the discussion happened — and re-opens it, because a transcript can't say where it landed.
Working from your settled call — annual-first, decided three weeks ago for cash-flow predictability. Here's the tightened copy on that basis…
The decision arrives with its why — so the model builds on it instead of re-arguing it.
Same request, two outcomes. The difference isn't the model — it's the unit it was handed.
The unit failing, not the model
Recall was never the problem. Standing is.
Memory did its job: the discussion came back. What it couldn't supply is the one thing that matters — that the discussion ended, in a decision, for reasons. To a transcript, your settled call and the three options you rejected on the way to it are the same kind of text.
So the model, quite reasonably, treats the question as open. It isn't being disobedient; it's reasoning over the unit it was given — and the unit says topic, not verdict.
What changes with the right unit
A decision, ratified on your word, carrying its reasoning.
Unl holds the settled point itself — annual-first, for cash-flow predictability — and serves it as unprompted relevant context when the topic bears on the turn. Superseded when your thinking moves, so what arrives is always the call that stands. Tuned for Claude, Claude Code, ChatGPT & Cursor at launch, extending across the AI ecosystem. Connects anywhere MCP does.
Give the model the decision with the reasoning and it stops re-arguing it. Reasoning agents respect a why — that's what they're built to build from.
Questions people ask
Why does my AI re-open decisions I've already made?
Because what it holds is conversation, not conclusions. A transcript can't distinguish the option you settled on from the options you discussed and rejected — they're all just things you said. Without a unit that says this one stands, and here's why, the model treats the question as open.
Isn't that just the model being weak?
No — a capable model reasoning over a transcript should treat unresolved-looking questions as open. The fix isn't a smarter model; it's a better unit: the settled decision with its reasoning, served at the moment it bears.
How does Unl know a decision stands?
Because you said so — decisions enter on your word, not by scraping the chat. And when your thinking moves, you supersede: the old call retires visibly with its history kept, so what surfaces is always current.
Does this work in my coding tools too?
Yes — the same settled decisions serve across Claude, Claude Code, ChatGPT and Cursor over MCP. One place your calls live; every tool inherits them.
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.