Method, in plain language

Adaptive without being mysterious.

A deterministic, versioned flow chooses concrete questions from your contexts and keeps separate kinds of evidence separate.

01

Label current conditions

Sleep, stress, treatment, workload, structure, illness, and how typical the period feels are recorded before symptom questions. They label current scores but never change the trait-oriented Pattern Score.

02

Map situations in your life

Eight lightweight questions establish useful contexts without requiring work, school, driving, a partner, meetings, or any other role.

03

Ask concrete, adaptive questions

The engine samples each core construct, asks a confirmation only when useful, and changes context when an example does not fit.

04

Count relevant events first

Associated emotional and rejection/evaluation profiles are conditioned on a clear 30-day sample, with a 90-day fallback when exposure is sparse.

05

Report scores with their evidence

Bands, gates, coverage, confidence, state labels, event windows, qualifiers, and source distinctions travel together.

What the engine refuses to do

A missing situation never becomes a quiet zero.

These are deliberate safeguards, not edge-case handling.

No opportunity

The situation was not present; choose another context or reduce coverage.

Cannot remember

Offer a recall fallback or retain uncertainty; never rewrite it as no.

No recent event

The associated profile is unavailable for the period—not emotionally low.

Missing observer

Self report stands alone. No observer means no penalty.

Safety response

Interrupt for support when needed, but never add severity points.

Actual adversity

Record mistreatment as context; never pathologize it as a distorted perception.

Ready to see the flow?

Begin privately, with no account and no time limit.

Start privately