Free forever · No paywall · No ads
Typology,
measured properly.
Most personality bots hand you four letters and act like that settles it. Persona measures all eight Jungian cognitive functions directly, fits your profile against every valid type — and tells you how confident it actually is.
Live now · 5 frameworks · 13 practitioners · Powered by frontier AI models we pay for ourselves
The method
Most personality tests
quietly cheat.
Three things go wrong in almost every free test on the internet. Persona fixes all three — and the fixes are the entire product.
The problem
One point flips your type
Count T answers, count F answers, take the bigger pile. Someone who is genuinely balanced gets a confident, arbitrary label — and never trusts the result again.
The problem
Agreeable people score high on everything
If you tend to agree with statements, an ordinary test reads that as having every trait at once. Your result is decided by your questionnaire manners, not your mind.
The problem
Ties get guessed
When two types are neck and neck, the test picks one anyway and never mentions it was close. You are handed a coin flip with the confidence of a diagnosis.
What we do
Confidence, stated out loud
Every result carries a real number and its runner-up. "INTJ — 84% confident, next closest INFJ" is honest, and far more useful than a bare label.
What we do
Ipsatized scoring
We subtract your own average across all eight functions before comparing anything. Agreeableness cancels out, and what remains is the shape of your profile — which is what type actually is.
What we do
Tests that check each other
HEXACO's Honesty-Humility and the Dark Triad are strongly related in the published literature. Persona measures both, then tells you whether two independent instruments actually agree about you. Nothing else in this space does that.
What we do
Adaptive tie-breakers
If two types finish close, the test quietly injects extra forced-choice questions aimed at exactly the functions that separate them. It resolves the ambiguity instead of hiding it.
What a result looks like
Eight functions,
measured directly.
No test here infers your functions from four letters — that would just be re-reading the letters back to you. Ni is measured by Ni questions. Then your profile is fitted against all sixteen stacks at once.
The instruments
Five frameworks,
one profile.
Take one or take all of them. Results stack into a single profile that follows you into every server Persona is in. Two more are built and waiting on their question banks.
16 Types Live
All eight cognitive functions, fitted against sixteen stacks with a confidence score and adaptive tie-breakers. The flagship.
Enneagram Live
Core type, wing, instinctual stacking and tritype. Instincts are measured on a separate axis so they can't just echo your type back.
Big Five · SLOAN Live
The five-factor model, and the most widely recognised framework in psychology. Also derives your classical temperament.
HEXACO · Six Factors Live
The model that succeeded the Big Five. Adds Honesty-Humility — the sixth factor five-factor tests have no room for — and cross-checks itself against your other results.
Dark Triad Live
Machiavellianism, Narcissism, Psychopathy — as ordinary traits, banded rather than diagnostic. Most people come out unremarkable, and it says so.
Socionics · Model A Soon
Full Model A with all eight positions, plus intertype relations computed from first principles rather than a copied table.
Attitudinal Psyche Soon
Volition, Logic, Emotion and Physics placed by confidence and process orientation — not naive strength ranking.
The session
Or skip the questionnaire
entirely.
Sit down with one of thirteen practitioners for a real conversation — not a form. Jung, Freud and Adler. Rogers, Frankl and Perls. Ellis and Beck. Nietzsche and Schopenhauer. Myers, Augustinavičiūtė and a trait psychologist who only believes numbers.
Each one works differently, because each one is different. Rogers barely asks questions. Perls won't let you talk about last year. Ellis goes after the belief and not the person. Pick five minutes or forty — the long one has a shape, opens, presses on the thing you keep circling, and lands. They read your test results before you start, and they remember the last time you came. Some end by naming a type. Some refuse to, on principle.
AI portrayals of historical figures — not real people, and not therapy. If you're struggling, findahelpline.com is free and confidential, anywhere in the world.
Shareable
Your profile,
as a card.
All eight functions drawn as a radar rather than a bar chart — extraverted functions on the right, introverted on the left, so an introvert's card leans visibly left from across the room. Your rarity is on it, counted against everyone else who took the same tests. Four themes. It is a PNG, so it leaves Discord.
The AI layer
It already knows
your results.
Ask Persona anything about typology and your profile is injected into the conversation as structured data — your function strengths, your confidence score, what you haven't taken yet. Answers are about you, not about the abstract idea of your type.
We run the most capable frontier models available — not the cheap tier, not a distilled knock-off. Every token is funded entirely by us. No subscription, no credits to buy, no "upgrade to continue".
Depth psychology
“Until you make the unconscious conscious, it will direct your lifeC. G. Jung
and you will call it fate.”
Jung never meant type as a box. He meant it as a map of where your attention goes by default — and what stays in shadow because of it. Persona is built on that reading: a profile is a starting point for noticing yourself, not a verdict you're stuck with.
Reference
Nine commands.
That's the whole bot.
/testsBrowse and take a test. Buttons, 3–5 minutes, only you can see it while you're taking it.
/therapyA conversation instead of a questionnaire. Thirteen practitioners, 5 to 40 minutes. They read your results first and remember the last one.
/cardYour profile as a shareable image — all eight functions as a radar, your rarity, four themes.
/profileYour saved results across every framework. Add a user to view theirs.
/compare @userHow two people's types actually interact — intertype relations, shared functions, predictable friction.
/askAsk anything about typology. Reads your profile first so answers are personal.
/helpA walkthrough of what the bot is and where to start.
/privacySee, export or permanently delete everything stored about you.
/settingsAdmins. Automatic type roles — assign @INTJ or @5w4 the moment someone finishes a test.
Behind it
Fully funded, developed
and designed by one person.
Persona is free because it was built that way on purpose. No investors, no ads, no data sold — the model costs come straight out of pocket.
Creator
polarissoulcode
Everything you see — the scoring engine, the question banks, the artwork and this site — is the work of one person, paid for out of pocket so that it can stay free for everyone.
Ready when you are
Find out what shape
your mind is.
Free, forever. Add it to your server and run /tests.
Takes about ten seconds · You'll need Manage Server on the server
And a distinction that matters: a validated model is not the same as a validated questionnaire. Persona's questions are our own, written to measure those published constructs — they carry no reliability data, no norms and no peer review, and they are not the official instrument for any model. Read every score as indicative, never as a measurement. Persona is built for insight and for arguing about types with your friends. It is not a diagnosis, and nothing here should be used to make a decision about someone's health, career or relationships.