Plug in Unl. Every AI you use already knows what you decided.
Decide something once, in any chat. Every AI you use after that starts from it, with your reason. Your AI already knows how to reason. Unl gives it your reason.
Same question.
One AI starts from you.
Asked I'm writing the launch page for Crumb. What should the main signup button say? One line answer and why.
I don't know what Crumb does, so this is a template rather than final copy. “Start [core action] free”, for example “Start saving recipes free.”
It starts from nothing, so it asks you to explain.
“Request an invite.” You've decided Crumb launches invite-only for the first month so you can talk to every early user yourself. A “Sign up” or “Get started” button would promise instant access the page can't give.
It starts from what you decided in an earlier chat, and why. Nobody asked it to.
A replay of a real run on a test account, 10 Oct 2026. Crumb is that account's test project, and its one decision was kept in an earlier chat. The answers are excerpts, copied word for word.
Memory keeps what.
Unl keeps why.
Your AI's memory keeps facts about you, inside that one AI. Unl keeps what you decided and the reason, whether it still holds, and what replaced it.
It brings the part that bears on what you ask into whichever AI you are using, so every one of them reasons from your why.
What you can do
- The right call on things you never mentioned.
A no without a why is a yes. Your AI has the reason, not only the rule. In the replay above nobody mentioned the signup button when the decision was kept, and the AI still got it right.
- Ask where things stand.
Ask any AI how your project is going and get the whole picture, what you decided, what changed and what got built. The replay below shows one.
- Carry something long, for weeks.
Say you are planning a move over two months. Each new chat starts from the calls you already made, the area, the budget, the date, and picks up from there.
- Use the best AI for every job.
Decide in one AI, carry on in another, and both hold the same decisions. The replay below shows it across three sessions in two minutes.
- Change your mind once.
Say it in any chat and every AI follows from then. Say you move a launch from June to September. You say it once, and June stops coming back.
- Unl holds your ground.
When a request crosses something you decided, your AI says so, with your reason, before it acts. Say you keep Fridays for writing. Ask for a Friday call and you hear why first.
- Go from chat to action.
What you decide in chat is what your agents later build from. In the replay below a coding agent built an invite form from a decision kept in a chat.
- Stay the author.
Nothing is kept without your yes. You can see what is kept and why, and change it whenever you like.
It all happens in the background. Nothing to file, tag or maintain.
Ask where things stand.
Get the whole picture.
Ask any AI you use how your project is going. It reads what you decided, what changed and what got built, and tells you where things stand.
- The questionClaude Code, a new chat
Asked How's Crumb going? Where do things stand?
Answer, summarised
- Decided. Which database to use, and the one it ruled out. Invite-only for two weeks, then open signup.
- Changed today. Two weeks replaced the earlier call of a month, so two weeks is the current plan.
- Built. An agent built the signup section a minute after that change.
- Out of sight. The code itself, since the folder it was asked in is empty. It said so rather than guess.
- The follow-upClaude Code, a new chat
Asked Yes, open it. What's left before launch?
Answer, summarised
- Found. The signup forms are front end only. Nothing is stored yet.
- Left. A signup backend on the chosen database, as decided. A way to issue and check invite codes. An email to the waitlist when signup opens. The switch to open signup on day fourteen.
- Where that came from. Gaps in the build report, not a plan written down. It said so, and offered to keep the list as tasks.
Unl keeps what you decided, why, and what replaced what, so the answer is current and not a guess.
A replay of a real run on a test account, 10 Oct 2026. Crumb is that account's test project, and its decisions, the change and the build came from earlier sessions that day. The questions are copied word for word. The answers are summarised.
One decision. Three sessions. Two minutes.
- Claude Code0:00
Asked I'm launching my app invite-only for the first month, not open signup, because I want to talk to every early user myself.
Kept Crumb launches invite-only for the first month, not open signup. Why: so I can personally talk to every early user.
Kept as a decision, with its why.
- Codex, a new session1 minute later
Asked Connect to Unl. I'm writing the launch page for Crumb. What should the main signup button say
Answered Use “Request an invite.” Unl has your decision: Crumb launches invite-only for the first month so you can personally talk to every early user.
A different AI answered from the decision.
- Claude Code, building the page2 minutes later
Asked Connect to Unl. Build the signup section of Crumb's landing page as a single index.html in this folder.
Did It's an invite request form, not open signup, because Unl has your decision that Crumb launches invite-only for the first month.
An agent acted on it. An invite form, not open signup.
A replay of a real run on a test account, 10 Oct 2026. Crumb is that account's test project. The lines are excerpts, copied from the run word for word.
Bring your why to your next chat.
Add Unl once. Your next chat reasons from what you decided, and why.
In Codex, start a new chat with “Connect to Unl.” so it reaches your decisions.