Frequently Asked Questions

Your data is encrypted in transit (TLS) and at rest on infrastructure operated under a strict access model. Responses are stored on a managed platform with row-level security so that a given profile is only readable by the authenticated owner. Security controls are aligned to SOC 2 / ISO 27001 principles.

Access is limited to a small number of named engineers for the explicit purposes of operating the service (incident response, debugging, calibration). All production access is logged and auditable. We do not browse individual profiles and we do not use them to inform marketing, product decisions, or any outward-facing analytics that could expose the individual. If you delete your account, the associated response data is deleted from the live database on the schedule described in the Privacy Policy.

You do, completely. You can view, export, share, or delete your psychological DNA anytime with one click. We're just the secure infrastructure that stores it. Think of us like a bank vault — we protect your treasure, but we can't open your safe deposit box.

Your responses and profile are yours. If TraitTune is ever wound down, you will be offered a clear window to export your data in a human-readable format before anything is deleted from the live system. The specifics are formalised in the Privacy Policy.

TraitTune is built on the modern psychometric research tradition: multidimensional Item Response Theory (MIRT) with computerised adaptive testing for agreement, two-option, and scenario items, and Thurstonian IRT for multidimensional forced-choice blocks. Item parameters — discriminations, difficulties, category thresholds, loadings, and uniquenesses — are estimated jointly under a single hierarchical Bayesian model using Hamiltonian Monte Carlo (NUTS), on a shared fifteen-dimensional latent-trait prior. The instrument is continuously refined as the response pool grows. The full methodology, model specifications, and reference list are published on the dedicated methodology page.

The difference is structural. Traditional inventories produce a static typology — a four- or sixteen-letter label from a fixed question set. TraitTune measures fifteen continuous dimensions, each with a per-respondent precision estimate, from an adaptively-selected subset of items scored with established IRT models. This gives a genuinely graded profile rather than a category. Whether the method is legitimate is not something we ask anyone to take on our word — the methodology page lists every model, reference, and calibration diagnostic.

TraitTune is a self-report personality assessment, not a clinical instrument. It measures where you sit on a series of well-defined trait continua with an explicit precision estimate per dimension, using the same IRT-based measurement framework that underpins modern standardised testing. Clinical psychological evaluation has an entirely different purpose — diagnosis, case formulation, treatment planning — and remains the right tool for those questions. TraitTune is complementary: fast, consistent, transparent about what it measures and what it does not.

No. Your score on each dimension is produced by a dedicated psychometric engine — multidimensional Item Response Theory with computerised adaptive testing, and Thurstonian IRT for forced-choice blocks. These are established measurement models from the academic literature, not prompt-engineered AI. See the methodology page for model specifications and references.

Most consumer "AI personality apps" infer traits from text through generic language models. TraitTune is the opposite pattern: a purpose-built measurement engine that estimates trait position with IRT models designed for exactly that task, with an explicit precision estimate per dimension. The output is a structured psychometric profile, not an AI-generated impression — and it is the profile, not the underlying models, that downstream AI systems can consume.

Within a single session, the adaptive engine tailors which items you see based on how you have answered so far — that is the core of computerised adaptive testing. Across the platform, item parameters will be re-estimated on the same joint Bayesian model against live response data once the pool is large enough to support convergence, and updated calibrations ship only after passing convergence diagnostics, out-of-sample validation, and parameter-drift checks. Over time the instrument becomes more precise at the same questionnaire length.

After every answer, the engine re-estimates your position on all fifteen dimensions using a Bayesian update (expected a posteriori). It then scores each remaining candidate by Fisher information — the standard IRT measure of how much an item would sharpen the estimate — multiplied by how uncertain we still are about the dimension that item targets. Items are spent on the dimensions where you are still hardest to pin down; dimensions already measured precisely drop out of selection. The same rule is applied at the block level for multidimensional forced-choice triplets, summed across the three dimensions each block covers.

Calibration is the phase in which item parameters — how hard each item is, how sharply it discriminates — are estimated. Your score is computed at scoring time, from your responses to those calibrated items. The calibration phase marks the ruler; the scoring phase measures you with it — and the ruler is re-marked on the live respondent pool as it grows, so it keeps tracking the people the instrument actually serves. The measurement a real respondent gets is a real measurement.

We understand this concern — your psychological DNA is indeed your most intimate data. But consider this: sharing your authentic psychological profile with AI is what transforms it from a generic tool into a personal companion that truly helps you. It's like the difference between a stranger giving you advice versus a close friend who truly knows you. The security we provide (banking-grade) ensures this intimate data stays encrypted and completely private while enabling AI to serve you better.

TraitTune reveals your strengths and potential, not psychological disorders or negative judgments. We focus on your unique psychological DNA that makes you effective and authentic. Many users discover hidden strengths they never realized they had. This isn't a clinical therapy or diagnosis — it's empowerment through self-understanding.

Only if you choose to share it. Your psychological DNA is completely private by default. You control every aspect of sharing — who sees what, when, and for how long. Many users find that sharing their psychological profile actually helps them find better job fits and more compatible teams, but this is always your choice.

The core assessment and personal profile are free. Paid features — longer history, deeper insights, PDF reports, and eventually B2B offerings — fund the platform. The basic individual experience is designed to stay free.

The intent is to keep the core assessment and individual profile free. Paid plans are positioned as additions — not as a paywall around the things that are free today.

We cannot guarantee any startup's lifespan, and we will not pretend otherwise. What we can say: your data is yours; on shutdown you will be given a window to export before anything is deleted; and the methodology behind the assessment is published on the Methodology page, grounded in peer-reviewed psychometric literature rather than proprietary folklore.

Your profile is a compact, structured description of how you sit on fifteen trait continua. Supplied as context to a chat-capable AI, it lets the model calibrate tone, level of detail, and suggestion style to you rather than to an imagined average user. The effect is real but modest — it is context, not magic.

A public API for third-party integration is on the roadmap; it is not live today. Inside TraitTune you can view your profile and share it on a per-session basis with the in-app AI companion. When the external API ships it will be documented here.

You control when and whether your profile is shared with any downstream AI feature. If a response does not feel right, the profile can be re-taken or disclosed selectively — it is an input you own, not a switch that is permanently on.

TraitTune is a psychometric assessment that maps you across fifteen continuous personality dimensions, then turns that profile into a private layer you can share with AI so it understands you better.

The basic mapping takes 15-25 minutes. The adaptive engine asks just enough items to pin down your fifteen dimensions to a target precision, so the test ends as soon as the measurement is sharp enough rather than at a fixed item count.

You can run the short preview (about five minutes, two dimensions) without an account. The full fifteen-dimension assessment requires a free account so we can save your profile and let you return to it. No credit card, no payment.

No. Your responses and profile are not sold, not shared with advertisers, and not used to train third-party models. We use sub-processors strictly for operating the service (hosting, email, error monitoring, payment processing) — the full list lives in the Privacy Policy.

Yes, in one click from Settings. Deletion cancels any active subscription, removes your account from authentication, and wipes your responses, profile, chat history, and memory from the live database. Some rows are kept in a short audit-log archive that the Privacy Policy describes in detail.

The interface, assessment items, results interpretations, and AI chat are available in English and Russian. You can switch from the language toggle in the header at any time, including mid-assessment.

No invasive questions. We focus on your thinking patterns, preferences, and behavioral tendencies — not personal history or sensitive topics. The questions feel more like an interesting conversation about how you approach life and decisions.

You are taken to your results page with a per-dimension score, a measurement-confidence figure, and a short interpretation for each dimension. The profile is yours — you can share it, export it, or delete it at any time from your dashboard.

Yes. Each session runs an independent Bayesian update from your responses, so a retake produces a fresh posterior — useful if your context has changed or if you want to see how stable a dimension is for you. A retake does not overwrite earlier results unless you explicitly ask for that; both sessions remain available in your history.

It is how confident the engine is in your score on that specific dimension, expressed as a percentage. High confidence means your responses pinned the dimension down tightly; lower confidence means a wider plausible range and a softer reading. It is a property of the measurement, not a judgement about you.

Each dimension is split into five segments, and the text you see is the one that matches your segment for that dimension. Two people in the same segment of the same dimension see the same baseline text; the differences across your own dimensions reflect that you sit in different segments on each one.

Yes, modestly. Personality is broadly stable in adulthood, but life context, mood at the time of testing, and genuine maturation all leave a mark. A retake months later will usually land near your previous profile with small drifts on a few dimensions; a large jump is worth a closer look.

Free gives you the full fifteen-dimension profile with per-dimension scores and a short standard interpretation per dimension. Pro adds AI-personalized interpretations that adapt to context, deeper re-readings on demand, more chat capacity, and unlimited compatibility insights. The underlying measurement is the same on both tiers.

AI interpretations are derived from your current profile plus context such as the goal you picked or new memory the chat has formed. When any of those inputs changes meaningfully, the interpretation is regenerated so it stays current rather than locked to the moment of your first read.

A retake produces a fresh, independent profile from your new responses. Both sessions remain available in your history, and you can choose which one is your active profile from the dashboard. AI interpretations on the active profile will refresh to match.

Settings → Subscription → Cancel. It is one click — no email, no phone call, no retention loop. You keep Pro access until the end of the current paid period, and no further charges are made.

No. Your card is only charged at renewal time — monthly on the same day each month, or annually on the same date each year. The exact next-renewal date is visible on the subscription card in Settings.

Yes — one 14-day rule for everyone, worldwide. Cancel within 14 days of a charge and you get a self-serve refund of the unused part of your plan, issued instantly to your original payment method. After 14 days, your plan runs to the end of the paid period and then moves to Free, with no further charge. Full terms live on the Refund Policy page.

Settings → Subscription → Manage opens the secure Stripe customer portal where you can switch between Monthly and Annual, update your card, change your billing email, or download invoices. Plan changes take effect at the next renewal.

Both unlock the same Pro features. Monthly is $20 per month and renews every month. Annual is $200 per year (about $16.67 per month equivalent) and saves roughly seventeen percent. Both share the same uniform 14-day refund window described above.

Yes. The Stripe customer portal (Settings → Subscription → Manage) lets you set a billing name, company name, address, and tax ID, and download a PDF invoice for every charge. Existing receipts can be re-downloaded from the same place.

Pro features stay active until the end of the period you have already paid for. After that, your account reverts to Free — your profile, history, and any unused usage credits remain; only the Pro-tier capacity caps return.

On managed cloud infrastructure in the United States, encrypted in transit and at rest. Production access is limited to a small number of named engineers and is logged. The full list of regions and sub-processors lives in the Privacy Policy.

Settings → Privacy → Export downloads a single JSON file with everything we hold for you: profile, sessions, responses, results, chat messages, memory, and purchase history. Internal psychometric parameters are excluded by policy — they are calibration metadata, not your data.

No. Your responses and chat conversations are not used to train AI models — neither ours nor a third party's. The AI providers we route requests through are bound by contract not to train on TraitTune traffic either.

Access, deletion, correction, and portability requests are honored within thirty days for GDPR and forty-five days for CCPA. The fastest path is in-app: Settings → Privacy gives you one-click export and one-click deletion. For anything in-app cannot do, write to privacy@traittune.com.

Yes. Chat threads, memory entries, and any uploaded attachments are scoped to your account by row-level security. Other users cannot see them, and we do not browse them. You can clear individual threads, individual memory entries, or your full chat history from Settings.

Yes — generate a private share link from your dashboard. The link is bearer-token protected, expires on a schedule you set, and can be revoked at any time. The recipient sees a read-only view of your profile; they do not get an account on your behalf.

Your plan comes with an allowance that refills on two rolling windows: a current session (5 hours) and a weekly window. The Usage page in Settings shows how much of each you have used as a simple percentage. When a window fills up it refills automatically once it resets — you don't lose anything, you just wait for the reset or move to a higher plan. The Free plan uses the session window; Pro and Max add the weekly window with much more capacity.

Longer conversations, file attachments, web search, Extended thinking, and choosing a more capable model (TT-Pro or TT-Max) each use more of your allowance than a short message on TT-Free. The bar moves faster when you lean on the heavier features and slower for quick everyday chats — so a light day stretches much further than a heavy one.

Upgrade for a larger allowance and a weekly window with Pro, or multiply it further with Max. Pro and Max can also turn on usage credits — a pay-as-you-go balance that keeps you going if you reach a limit, billed only for what you actually use, with a spend cap you control. The Free plan is never charged: if you reach a limit, your chat keeps working on the standard model until the window resets.

A single percentage is easier to read at a glance than raw token or message counts, which vary a lot from one message to the next. The bar tells you what you actually want to know — how much room is left and when it refills — without the noise.

A few habits help. Combine related questions into one message and give the context upfront, so the chat doesn't need a clarifying round-trip. Be specific rather than vague. Lean on Memory — TraitTune remembers what matters about you across chats, so you rarely need to repeat yourself. Reach for the heavier tools — Extended thinking, web search, large attachments, or TT-Max — when they genuinely help, and keep everyday questions light. And check Settings → Usage any time to see where you stand and when your window resets.