The gesture
The context you never chose to keep
AI memory mostly fills itself: it watches your chats and decides what’s worth keeping. It’s convenient, and it means the context steering your model is made of things you never actually chose — extracted, not ratified.
A memory that populates automatically holds a picture the tool assembled from whatever came up. Handy as acquaintance — and genuinely fine for that. But nothing in it carries a moment where you said keep this, which is exactly what makes a decision something the model can act on as your settled position. Unl holds only what you ratified.
Who chose what’s in there?
When a memory fills itself, the context steering your model was chosen by the tool, from the flow of conversation. It’s a reasonable guess at what mattered, and as background acquaintance it’s fine. But it’s collected, not committed to — and some of what gets collected you’d disown the moment you saw it stated back as a standing fact.
Extracted vs ratified
There’s a difference between context that’s about you, scraped from what you said, and context that’s yours, put there by a deliberate gesture. The first is a decent guess at what was important; the second is what you actually settled. Only one of them you’d stand behind without re-checking — and only one is safe for the model to treat as a decision.
Why the gesture is load-bearing
The gesture is what makes a decision authoritative. Because you ratified it, the model can reason from it as your settled position, and you can recognise it, stand behind it, or override it. Without the gesture there’s no author — just a store of plausible extractions, some of which you’d disown — and a model has no way to tell which is which.
Foil, not enemy
For an autonomous agent keeping its own state, auto-capture is the right call — the agent is its own author, and a machine-assembled record serves it well. The mismatch is the human environment. Here you are the author, and context you didn’t choose doesn’t speak for you — so the same convenience that helps the agent quietly steers you from things you never ratified.
What ratified context looks like
Unl holds only what you put there on purpose: a decision saved because you said keep it, with its reasoning, superseded when you say so. You command what’s held — nothing lands or leaves without your word — which is what makes it context you’d actually trust the model to act on. Tuned for Claude, Claude Code, ChatGPT & Cursor at launch, extending across the AI ecosystem. Connects anywhere MCP does.
Auto-captured memory is a guess at what mattered; ratified context is what you actually chose to keep — the gesture is what makes it yours and lets the model act on it as your settled position, so the answer comes back measured against what you already decided, and why.
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
Does AI memory keep things I didn't ask it to?
Often, yes — many memory features populate automatically, extracting what seems worth keeping from your chats. It's convenient, but it means the context steering the model is assembled by the tool rather than chosen by you.
What's wrong with automatic memory?
Nothing, as convenient acquaintance. The gap is authorship: context you didn't choose can't reliably speak for you, and some of it you'd disown on sight. For decisions you want the model to act on, the choosing gesture is the point.
What does 'ratified' context mean?
Context put there by a deliberate gesture — you said keep this — rather than scraped from the flow. Ratified context is your settled position, so the model can reason from it and you can recognise and override it. It's the difference between yours and merely about you.
Do I have to save everything by hand?
Only what you want the model to treat as settled. The point isn't more effort — it's that the decisions steering your AI are ones you chose, held on your word and retired on your word, so you stay in command of what it acts on.
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.