Safety & respect
Consent, boundaries, autonomy, conflict safety, and freedom from coercive control.
Methodology
A transparent sorting system for a complicated human choice
Dating Standards Test uses deterministic rules—not a black-box model—to build every Keep, Discuss, and Flex map and select the paid plan’s lead conversation. Optional AI personalization can explain why those finished patterns connect, but it cannot change a category, priority, or safety rule.
Four domains, three core questions plus two optional refinements each
Communication & reliability
Follow-through, repair, stated intentions, sustainable contact, and curiosity.
Future compatibility
Relationship structure, parenting, place, money, work, home labor, and care.
Preferences & lifestyle
Social rhythm, attraction, interests, status proxies, and routines that can often flex.
The sorting rules
- SAFE-OVERRIDE-001: a protected safety condition always lands in Keep, even when the selected answer says flexibility.
- CORE-KEEP-001: a non-safety standard marked essential and reciprocal lands in Keep.
- RECIPROCITY-DISCUSS-001: an essential standard with a self-identified growth gap moves to Discuss.
- CONTEXT-DISCUSS-001: a context-dependent answer lands in Discuss.
- PREFERENCE-FLEX-001: only a non-safety item explicitly marked flexible can enter Flex.
The engine is deterministic. The 12-question core set and the optional 20-question refined set each produce the same categories from the same answers under rule version dst-1.1.0. No model profiles the user, and no hidden score ranks their standards.
The paid Next Conversation Plan is also rule-led: the engine selects the lead conversation, follow-ups, response signals, backup scripts, and next steps. A supporting AI reading runs only afterward. It receives a minimized brief of the already-sorted priorities, domain anchors, reciprocity notes, and pattern notes, and may connect those ideas in plain language. Its output is rejected if it changes the three priority standards, their order, or the lead conversation.
Our safety promise
Safety rules are never treated as flexible
We test every possible answer combination for the five protected questions. If our checks ever allow one into Flex, we pause the tool and fix it before anyone receives that result.
Read safety resourcesWhat this tool cannot tell you
It cannot diagnose a person, certify a relationship as safe, predict compatibility, replace lived context, or decide whether you should stay or leave.
Categories describe how today's answers were sorted; they do not establish facts about another person. Context, access needs, culture, disability, and changing life circumstances may all affect a real conversation.
The supporting pattern reading can also be incomplete or mistaken. Treat its wording as a reflection aid, not as evidence about another person or a substitute for professional support.
Versioning and corrections
Every report names the guide version used to sort it. Corrections apply to new results; old reports are not silently rewritten. If a safety rule fails, we pause the tool and fix it before anyone receives that result.