Six different lenses on personality. Each stands on its own. Together they form a key that's yours alone.
No single model captures everything. The Big Five captures broad traits. The Enneagram captures motivation. Cognitive functions capture information processing. The zodiac captures archetypal patterns. Each is true, but partial.
Each system runs end-to-end on its own terms. No shared questions, no borrowed scoring, no collapsed dimensions. This means more questions overall, but each result is genuinely what that system claims to measure.
No account required. No email needed. Your responses stay on your device. Nothing is sent to any server. This is static software — no backend, no database, no tracking.
What it measures: Five broad traits — Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. The most researched model in personality psychology.
Format: Adaptive testing, typically 30-50 items.
Epistemics: Typology — based on established trait theory.
Best for: Understanding your broad personality profile in terms validated by decades of research.
What it measures: Four dichotomies producing a 4-letter type (e.g., INFP). A generic implementation inspired by the MBTI framework.
Format: Fixed form, 40 items.
Epistemics: Typology — categorical types.
Best for: Understanding your preferences in how you direct energy, take in information, make decisions, and approach the outside world.
What it measures: Nine personality types distinguished by core motivations and fears, not just behaviors.
Format: Fixed form, typically 40-50 items.
Epistemics: Typology — distinct types with motivational structure.
Best for: Understanding what drives you at a deeper level — your core fears, desires, and automatic patterns.
What it measures: The twelve astrological signs as personality archetypes, scored based on your responses rather than birth date.
Format: Fixed form with specialized scoring.
Epistemics: Typology — archetypal categories.
Best for: Exploring personality through an ancient symbolic lens, divorced from astronomical claims.
What it measures: Eight cognitive functions (Se, Si, Ne, Ni, Te, Ti, Fe, Fi) scored as continuous dimensions — no type reduction.
Format: Fixed form, 40 items.
Epistemics: Typology — Jungian function theory.
Best for: Understanding how you process information and make decisions across all eight function-attitudes.
What it measures: Your leading cognitive function-attitude, reported only when the data shows clear separation.
Format: Fixed form with ranked scoring.
Epistemics: Typology — hierarchical function model.
Best for: Identifying which cognitive function you lead with, if one clearly emerges.
For the Big Five, we use Item Response Theory (IRT) to select the most informative next question based on your current estimated trait level. This typically reduces test length by 30-50% while maintaining or improving precision.
Other systems use fixed forms with carefully balanced item selection. Items are ordered to avoid response sets and maintain engagement while ensuring psychometric quality.
Every result shows its uncertainty explicitly. We use a stanine scale (9 levels) and only show a digit when the standard error supports it (SE < 0.5). Otherwise, we show ?.
Each system adds another "cut" to your key — the combination is uniquely yours.
Your personality key is the collection of your results across all six systems. Think of it like a physical key with multiple cuts — each system adds another dimension.
Example key progression:
O? C? E? A? N? · ? · ???? · vir→? · ±? O? C? E? A? N? · ? · I??P · vir→? · ±.62 O9 C6 E3 A6 N6 · 5w4 · IN?P · vir→gem · ±.46 What improves isn't more digits — it's fewer question marks. The uncertainty (±) shrinks as you answer more.
All six reference packs are currently marked as "piloted" rather than "validated." This means:
The IRT parameters are generated deterministically from item IDs so the scoring contract can be exercised and reproduced exactly. Every one carries calibration.illustrative: true.
We could have filled the calibration table with plausible values and labeled it as estimated from a sample of twenty thousand. That is the failure this specification exists to prevent.
The Big Five pack will move to empirical parameters once we have response data from live traffic. The architecture supports this — swapping illustrative parameters for real ones changes data, not code.
This requires either: (1) a public response dataset for GRM calibration, or (2) bootstrapping from live traffic (which would require a backend for response collection).
These assessments are for self-discovery and reflection only. They are not diagnostic tools and should not be used for clinical decision-making.
Personality tests are poor predictors of job performance and can introduce bias. Do not use these results for hiring, promotion, or any employment decisions.
These results are meant for your own self-understanding. Using them to make decisions about other people — even with their results — risks oversimplification and stereotyping.
Personality models reflect the cultures they were developed in. The Big Five has cross-cultural support; other systems have more limited validation across cultures.
MBTI® is a trademark of The Myers-Briggs Company. Our four-letter pack ships under a generic ID and is a distinct implementation, not the MBTI instrument.
What this means: Your personality data never leaves your device unless you explicitly export and share it. There is no central database of personality profiles.
Built with Astro + React islands. The entire site compiles to static HTML/CSS/JS — no server-side rendering, no origin logic beyond a static file server.
The scoring engine (`@pk/engine`) has zero Node.js dependencies. It runs identically in Node, in the browser, or in a Worker.
Each personality system is a "pack" — a self-contained unit with its own manifest, item bank, anchor sets, norms, result schema, content pack, and correspondence map.
155 checks across 6 test suites run on every change. The specification comes with tests that prove its claims.
Not for clinical, selection, hiring, or any high-stakes decision use.
These assessments are tools for self-discovery and reflection only. They are not diagnostic instruments, not validated for personnel selection, and not appropriate for making decisions about other people.
The reference packs are currently in "piloted" status — psychometric parameters are illustrative, not empirically validated. Results should be interpreted as provisional and used for personal insight only.
Start with any system. Your results will accumulate as you go.