Neumatics

About Neumatics

Neumatics builds SoulMap — a live self-discovery app. Every day, you spend sixty seconds inside an AI-generated narrative micro-experience (an Echo); your forced-choice responses feed a personal map of twenty-nine trait dimensions across seven construct families, named for the constructs behind ten of the most-studied psychometric instruments in the world. No questionnaire is administered, the readings carry their uncertainty openly — and every validation result behind them so far comes from synthetic populations, not measured humans; the science leads with that boundary.

SoulMap is usable today: sign up, play an Echo, read your first trait readings within a minute. The readings are interpretable and honest about their confidence — every score carries reliability metadata, and the app says so plainly when the engine is not yet confident about a dimension.

The problem we are addressing

The consumer side of personality tech is a wasteland of pseudo-science: one-off quizzes, static labels, no validity, no consent hygiene. Meanwhile the scientific side is stuck with self-report questionnaires that are noisy, easy to game, and blind to change over time — the replication crisis is, at root, a data-quality problem.

SoulMap's answer is behavioural measurement through play, repeated daily:

  1. Behaviour over self-report. Forced choices inside a narrative reveal how you actually decide — not how you describe yourself on a Likert scale.
  2. Longitudinal over one-off. A daily 60-second session builds an evidence trail over time: traits, states, drift, and contradiction — things no single quiz can see.
  3. Measurement discipline. Deterministic scoring of recorded choices, uncertainty stated on every reading, pre-registered validation gates on synthetic populations, and published methodology — including published corrections when our own audit overturned earlier claims. When a reading is uncertain, it says so.

The science and the architecture, in receipts

We publish what we measure and how we measure it:

  • The science — How we map 29 trait dimensions without questionnaires: the synthetic-persona calibration programme, what it has and has not shown, and the 2026-08-19 audit that reset our published claims.
  • Cloud infrastructure — Architecture, operations, cost engineering.
  • Data ethics — Consent, GDPR, the EU AI Act posture, and what we will never do with your data.

What comes next — and why it isn't live yet

The engine beneath SoulMap is designed to carry more than the consumer app: a psychometric inference API for researchers, consented research-data products with royalties settled back to contributing users, and matching surfaces built on validated traits.

None of that is live, and none of it will be until the inference engine completes its validation programme — large-scale calibration and peer-reviewed evidence that the measurements hold up. We think the order matters: measurement first, products on top of proven measurement. SoulMap already gives its users real, interpretable readings today; the engine's next level of authority is what unlocks everything else. Until then, SoulMap is the product, and the platform ambitions stay on this page — as a design, not a promise.

Team

Fotis Dovas — Founder & Product Architect. MSc Products & Systems Design Engineering (University of the Aegean, 2019). Three full product rewrites of SoulMap in eight months — the current stack (Next.js, Firebase, Python 2nd-gen Cloud Run, Gemini, WebGL2) designed, built, and operated solo. Prior: core production team on a viral EU-wide Greek social-media app (2024). LinkedIn.

Scientific review. Construct validity is reviewed at every major iteration by a licensed clinical psychologist — MSc Applied Behaviour Analysis (Queen's Belfast), MMPI-2 certified, twenty years in private practice.

Contact