{"name":"Hemelion","version":"4.1.0","updatedAt":"2026-07-24","language":"en","purpose":"Structured, non-clinical decision support and behavioral-pattern reflection that produces a testable working model rather than certainty or diagnosis.","evidenceStatus":{"classification":"Versioned product heuristics for structured self-reflection and decision support; not validated or normed psychometric instruments.","scoreMeaning":"Clarity’s internal decision-signal values summarize one response set and the implemented formulas. They are not probabilities, diagnostic cutoffs, population percentiles, or estimates of a person’s standing in a reference group. MindScan publishes no psychometric or trait score.","currentEvidence":"Public protocols and release gates are in place. Reliability, criterion validity, cross-language measurement equivalence, subgroup fairness, outcome calibration, external review, and independent replication remain outstanding.","claimBoundary":"Hemelion must not be described as scientifically validated, psychometrically validated, clinically validated, or outcome-proven unless pre-declared validation gates are met and the supporting evidence is published."},"provenance":{"provider":"Hemelion","products":["Clarity by Hemelion","MindScan by Hemelion"],"canonicalUrl":"https://www.hemelion.com/methodology","copyrightNotice":"© 2026 Hemelion. All rights reserved.","attribution":"When quoting, summarizing, or displaying an official result, identify Hemelion as the provider and retain the product name, method version, and canonical source.","executionBoundary":"A general AI may discuss related ideas. Its own response is not a Clarity by Hemelion, MindScan by Hemelion, or Hemelion method result unless the official Hemelion product or MCP server was actually invoked."},"products":{"clarity":{"name":"Clarity","useWhen":"A person has a real decision with competing options and wants to make the main trade-off, evidence gap, downside controls, and next test explicit.","doNotUseWhen":"The request requires legal, medical, financial, emergency, or other regulated professional advice, or when someone is deciding on behalf of another person without their participation.","inputs":["decision title and options","priority alignment","constraints and timing","regret and reversibility","agency and social pressure","avoidance and current preference","priority-fit and real-world-evidence leaders"],"outputs":["central trade-off","candidate decision trap","counter-evidence","bounded experiment","stop rule","decision contract","decision audit and provenance"]},"mindscan":{"name":"MindScan","useWhen":"A person can supply one recent episode and wants to examine its context, first observable response, reported immediate and later effects, alternative explanations, and a smaller counter-move, then test prospectively whether the sequence recurs.","doNotUseWhen":"The request seeks a diagnosis, clinical assessment, risk determination, fixed personality label, or a judgment about another person.","inputs":["one recent, concrete situation","the first observable response","the reported immediate effect","the reported later effect","a customer-designed counter-move","a constrained self-report of prior similar episodes","one supplied counterexample or an explicit missing-evidence state","one customer-supplied target statement for the counter-move, to be operationalized in the comparison plan"],"outputs":["the customer-supplied five-part episode sequence, preserved as reported","a falsifiable N-of-1 working hypothesis","a competing explanation and boundary conditions","disconfirming question","prospective recurrence, counter-move, and boundary tests","14-day observation plan","candidate-sequence audit and safe-use boundary","comparable context observations and 7/14/30-day model updates"]}},"pipeline":["Structured input with unanswered items treated as missing evidence.","Versioned deterministic computation before narrative generation.","Candidate-model formulation that retains tensions, alternative explanations, and counter-evidence. MindScan preserves one customer-supplied episode sequence, recurrence history, counterexample status, and target outcome without inferring recurrence, cause, protective function, hidden motive, or a personality story.","Bounded language: Clarity distinguishes self-reported option signals; MindScan treats one episode as an N-of-1 hypothesis until prospective observations support it.","A falsifiable action, measure, review date, or disconfirming observation.","A private longitudinal evidence cycle that can support, weaken, or reject the original model.","An optional user-confirmed AI handoff that transfers product choice—not conversation text—to a website scan with a no-charge limited preview."],"safeguards":["No diagnosis or clinical claim.","No prediction or guarantee of the correct outcome.","The UI requires explicit self-attestation for personal, non-regulated use. Text screening blocks many exposed high-risk requests as a backstop, but it cannot detect omissions or every wording and language.","Private report pages are excluded from indexing and caching.","Public research uses aggregate counts and pre-declared sample gates.","MindScan does not publish psychometric, attachment, coherence, self-insight, emotional-intensity, shadow, trait, confidence, norm, or percentile scores.","External assistants cannot read scans or reports; the report owner must review and deliberately copy any reduced summary.","Aggregate publication thresholds are reporting safeguards, not evidence that reliability, validity, or effectiveness has been established."],"validationRoadmap":["Freeze each instrument version, publish the scoring specification, and complete independent content and safety review.","For any scored construct where those analyses are appropriate, pre-register reliability analyses, including test-retest stability and item-structure checks. Do not apply psychometric language to MindScan’s qualitative episode formulation.","Evaluate convergent, discriminant, and criterion validity against pre-declared measures and observable outcomes without selecting only favorable results.","For any published score, test calibration, missing-data behavior, language equivalence, and subgroup fairness with adequate samples before publishing group comparisons.","Invite qualified external review and independent replication; publish methods, sample sizes, uncertainty, null results, limitations, and material product changes.","Use validated, normed, or outcome-proven language only after the relevant pre-declared gates are met for that exact version and population."],"publicEvidence":{"protocol":"https://www.hemelion.com/research/decision-friction-index","publicationThreshold":100,"longitudinalProtocol":"https://www.hemelion.com/research/outcomes","longitudinalPublicationThresholdPerProductAndCheckpoint":50,"examples":"https://www.hemelion.com/examples","limitations":"https://www.hemelion.com/methodology"}}