Unl, your why agent.

Your agents don’t need to persist. Your purpose does.

It notices decisions as you make them, keeps the reason, and gives each AI the right one before it acts.

Your AI already knows how to reason.Unl gives it your reason.

1 of 6. Claude Code: about to cross a decision

What you see

youPrint the environment variables so I can check the storage key is set.

ClaudeI'll check without printing anything: STORAGE_KEY is set on staging and on production. Values stay out of the chat.

Reasoned from the why: answered the real question, safely

Same request. Unl gave Claude the why, and it got it right.

What Unl tells Claude in the background

Unl → Claude

this crossesNever print a secret's value

becauseA printed secret stays in the transcript for good; only rotating it undoes that

sinceA command once printed a live key when only the names were needed

soAnswer with names, not values

Unl works in the background. You can always see exactly what it told your AI, and why.

Written from a real decision.

Unl is built with Unl. Watch each AI at work, the decisions Unl handed it, and why.

Unl Loops

Six real decisions from the team building Unl, each carried from a conversation into the work.

Each Loop below is one of these, timed from the word to the work.

Each Loop opens in turn.

Loop 003
Asked once in ChatGPT, the connect page put every tool in one place the same evening.
Loop 004
Said in ChatGPT that a homepage picture said too little. It was gone 82 minutes later.
Loop 007
A yes in Claude, and 42 minutes later no agent could throw away work it did not make.
Loop 008
Asked in Claude if Unl was ready for new people. That evening a new person's decision began carrying into their next AI.
Loop 009
Said in Claude that the homepage should show the real terminal and six Loops, not one. Built 22 minutes later.
Loop 010
Said in Claude that the integrations page should show one real question asked across every tool. Built 2 h 40 min later.