A map, not a verdict
Every disagreement has a shape. Crux draws it.
Paste a thread, an email chain, two people talking past each other — Crux finds the exact seam. A fact, a value, a word, or nothing real at all. It never tells you who's right, because that's usually the wrong question.
No sides taken. No verdicts issued. Just the shape of the disagreement.
Hover or focus a bridge to see what kind of disagreement it is. This sample shows three of Crux's four types — the fourth, Phantom, shows up in the demo below.
Two ways in
Start from the mess, or start from the summary.
Crux takes a disagreement in whatever shape you already have it — no preferred format, no better answer.
Paste the raw thread
The whole back-and-forth, unedited — texts, email, Slack, whatever it was. Crux reads all of it and works out where the two of you actually stopped agreeing.
Paste two summarized positions
Already know roughly what each side thinks? Give Crux one paragraph per person — no transcript required, just what each of you believes.
Either way in produces the same map. Crux reads for the disagreement, not the format it arrived in.
The mechanism
Four ways to disagree. Crux tells you which one you're having.
Every bridge on the map is exactly one of these four types — named, not scored.
Factual
They believe different things about the world. That's testable, at least in principle.
Values
They weigh the same facts differently. No new evidence resolves that.
Definitional
They're using the same word to mean two different things.
Phantom
They appear to disagree. They actually don't.
Naming the type doesn't settle the argument. It tells you what kind of evidence — or conversation — actually would.
Try it yourself
The remote-work-vs-office thread, mapped.
Crux ships with a built-in demo disagreement, one click to load inside the product. Here, it's already loaded — open any bridge for the exact quotes and the question that would resolve it.
Wants the team back in the office most of the week — thinks proximity is what actually builds trust and catches problems early.
Wants to keep the flexibility that came with going remote — thinks trust comes from output, not attendance.
Rewrite each side at its strongest, and see if the map still looks the same.
Factual Same six months, read two different ways
Person A“Our output tanked when we went remote — the numbers are right there.”
Person B“Output dipped during the transition, then recovered past baseline.”
Would resolve it: are we both looking at the same stretch of time, or two different windows of it?
Values Proximity versus protected focus time
Person A“If people are in the building, mentoring happens by accident. That matters more than a spreadsheet shows.”
Person B“I'd rather protect four hours of deep work than get an accidental hallway conversation.”
Would resolve it: if both were guaranteed, which would each of you actually choose?
Definitional Two meanings of “flexible”
Person A“We already are flexible — everyone can leave early on Fridays.”
Person B“By flexible, I mean I choose which three days I come in. Every week.”
Would resolve it: what would “flexible” have to include for both of you to call it that?
Phantom Arguing against a position no one holds
Person A“Nobody's saying never come in.”
Person B“Nobody's saying five days a week, either.”
Would resolve it: has anyone actually proposed the position the other is arguing against?
Steelmanned: the phantom bridge dropped out — read at full strength, that one wasn't a real disagreement.
Common ground
Both sides agree the current policy is ad hoc, and neither wants to keep relitigating it every quarter.
Dependency-free — no login, no export queue.
Illustrative example, not a real conversation.
Get started
Try it before you argue about it.
Paste your own thread, or start from the demo above. Either way, you'll see exactly where the disagreement is — not who wins it.