A boundary states what you will do about your own participation. An ultimatum sets a final condition on the relationship: meet this or I withdraw everything. The difference is rarely the wording and almost always the structure — whose action the sentence describes, what the 'or else' names, how it is delivered, and whether you mean it. This checker asks five questions about one sentence and sorts its structure. It does not judge whether you should say it, stay, or leave.
You never type the actual sentence — the checker only asks about its structure. Answers live in this page's memory only; refreshing clears them. Nothing is stored or sent.
Read the sentence back. Whose behavior has to change for it to mean something?
Some boundaries have no 'or else' — the statement is the whole sentence.
Delivery is where boundaries most often drift into leverage.
The honest question: what happens if nothing changes?
Scope is a reliable tell.
A boundary keeps the action on your side: 'if you keep driving after drinking, I won't get in the car.' An ultimatum puts a final condition on the relationship itself: 'if you drink again, we're done.' Both are yours to carry out, both can be said firmly, and both can end a relationship — the difference is scope and finality. The boundary withdraws one form of participation; the ultimatum threatens to withdraw all of it. That is why the same sentence can be honest in one delivery and pressure in another: said once, at a calm moment, and meant, a final condition is a real decision; repeated every argument while nothing changes, it becomes leverage.
The checker looks at the parts of the sentence that decide its structure: whose action it describes, what the condition names, how it has been delivered, what actually backs it, and how much of the relationship it puts on the table. Answer for one specific sentence you are considering — not a general grievance. Four answered questions are enough to sort; the fifth sharpens the call.
An ultimatum is honest when it is the announcement of a decision you have already made: specific, said once, in plain words, with full willingness to act. It is leverage when it is repeated to extract compliance you do not actually intend to enforce — and leverage is corrosive in both directions: the person hearing it learns your words are blanks, and you learn to say things you do not mean. If your result is LIKELY_ULTIMATUM, that is not a verdict on you; it is a prompt to either make the decision real or retract the sentence. The full taxonomy, including threats, lives on boundary vs threat vs ultimatum.
It cannot judge whether your sentence is fair, whether your partner will change, or whether you should stay. It does not detect abuse: coercion, monitoring, and fear are safety matters, not structure matters — if they are present, see the safety guide instead of fine-tuning wording. And it only classifies one sentence at a time; for the wider situation, map it on the relationship boundary map.
No. An ultimatum is a final condition on the relationship, and it can be an honest last resort when it is true, specific, said once, and meant — for example, after a long pattern of broken agreements around drinking or debt. What makes it corrosive is repetition without follow-through: a sentence said every argument but never enacted is leverage, and leverage erodes trust in every word you say.
Yes. 'If shouting starts, I'll leave the room' has a condition and a follow-through, and it is still a boundary, because the follow-through is a specific action of yours. The tool distinguishes that from a condition that puts the whole relationship on the table ('or we're done'), which is the ultimatum structure.
Please do not — the checker is designed so you never have to. It asks five multiple-choice questions about the sentence's structure, and nothing you select is stored, saved, or sent anywhere. Refreshing the page clears everything.
Either the answers mixed (some lean boundary, some lean ultimatum) or the sentence has not taken a definite shape — it is rehearsed but not delivered, or its follow-through is undecided. That is actionable: rewrite it so the action is only yours, choose a next step you can carry out alone, and decide what you will do if nothing changes.
An ultimatum withdraws your own participation ('we're done if this continues'); a threat forces compliance through fear of harm ('do what I want or you'll regret it'). The mechanism is the difference: withdrawal versus fear. Both are distinct from a boundary, which never requires the other person to act — see /boundary-vs-threat-vs-ultimatum/ for the full comparison.
The structure check does — whose action, what condition, how delivered, what backs it, what scope. The context differs: with parents or in-laws, the workable move is usually a boundary about your own participation rather than a final condition on the relationship. See /family-interference-boundaries/ for that specific version.