Behavioural data is the rate-limiter for human-centred AI. We fixed the supply.

Situation

Behavioural data underpins every human-facing AI product, research study, and matching system.

Complication

Supply is broken — surveys are noisy, US panels are EU-incompatible, every study starts from zero.

Action

We built a consumer engine: a 60-second daily self-discovery game that quietly produces research-grade longitudinal data, with consent and royalties routed to the user.

Resolution

That data layer powers Neumatics — an EU-native platform of APIs, datasets, and research products buyers can actually license.

Every day, a user spends 60 seconds inside an AI-generated narrative micro-experience. Their forced-choice responses get scored against ten of the most-studied psychological instruments in the world. The user gets instant interpretable feedback. We get a calibrated, longitudinal trace. The game is the product; the dataset is the product. Both sides win.

Alpha users

~20

Sessions logged

200+

Daily cap completion

3 / 3 organic

Time to first score

≈ 60 s

• Limited-risk under EU AI Act. Hiring, clinical, and educational-assessment use cases are explicitly gated. Consumer surface is 18+ only; minors are excluded from trait-scoring distributions.
• No-train guarantee — buyer API request data never trains shared models.
• Per-subject deletion (GDPR DSAR) cascades across Firestore, BigQuery, and trace storage.
• EU region pinning at Enterprise tier (europe-west12 / de-central1).
• SOC 2 Type II on roadmap. HIPAA BAA available at Enterprise. VPC-SC perimeter; CMEK at rest.
• Quarterly methodology releases with human sign-off.

Stage

Closed alpha · public beta in progress

Funding to date

Bootstrapped via freelance UX revenue. No external capital.

Funding sought

Pre-seed (~€2M target) after public beta + first paid design partner.

Use of funds

Commercial co-founder · scientific validity hardening · public beta GTM.

Traction signal

Closed-alpha users complete 3/3 daily Echoes organically — zero incentive.

Investor contact

fotis@nexus-soulmap.com

What does Neumatics do, in one sentence?

We turn 60-second daily self-discovery sessions into EU-compliant longitudinal behavioural data, and sell that data layer as APIs, datasets, and research products.

Who is the buyer?

Three segments. Research teams (academic, pharma, market-research) buying compliant longitudinal samples. AI labs and alignment teams licensing anonymized training datasets. Enterprises licensing the API for personalisation, cultural-fit, or matching products.

How is this not just another personality app?

The consumer app is the acquisition engine, not the product. Underneath sits a calibrated psychometric pipeline — Big Five, HEXACO, attachment, Schwartz values, emotional intelligence, moral foundations — with IRT scoring, conformal CIs, scalar-invariance gating, and per-revision provenance. The methodology is published; see /about/science.

Why EU-native?

EU research buyers — pharma, universities, Horizon-funded consortia, market-research firms under AI-Act scrutiny — cannot legally buy from non-compliant US panels. EU-native compliance is a structural moat, not cosmetic.

What about the EU AI Act?

Neumatics is a limited-risk system by classification. Clinical, hiring, and educational-assessment uses are explicitly gated. The consumer surface excludes minors and high-stakes decisions. Full posture is at /about/data-ethics.

Can my data be used to train AI models?

Only if you explicitly opt in. The buyer API has a no-train guarantee. A separate, withdrawable consent toggle exists for users who choose to license their anonymized profile data for ML training; the dataset goes through aggregation, k-anonymity, and (where needed) synthetic-data generation before licensing. Full reasoning at /about/data-ethics.

Why a 60-second daily session?

Short enough that users come back voluntarily — alpha users complete all three daily Echoes organically, no incentive. Long enough to produce a calibrated longitudinal trace. The session length is the unlock for both consumer retention and data quality.

Where can I see the methodology?

/about/science is the hub — sub-pages cover data layers, persona generation, calibration loop, validation receipts, and the citations roster. /developers/methodology covers the API surface specifically.