We are a multilingual self-service testing platform. We believe good psychometric language belongs in the hands of anyone who wants to understand their own mind better — and that the online testing market is still dominated by vibes and shallow gamification. We try to be the serious option that doesn't sacrifice usability.
These are not wall values. Each one is encoded in product behavior — database rules, API contracts, publication gates — and verifiable in code.
Each test declares its base model, authors, references, and DOI where applicable. For models with protected IP, we use "supervised reinterpretation", never disguised copying. The Constructor classifies each model in 4 risk colors (green, yellow, orange, red) and blocks publication where legality is uncertain.
In B2B context, anonymous aggregates with N≥5 are enforced by the platform, not a PDF promise. Whitelabel admins cannot see individual data without explicit employee consent. In sensitive tests, there is prior warning and the option to answer outside the company context.
The report structure is a fixed template. AI only fills textual content, always with identification of the model used and its version, in the report footer. Subsequent changes to the prompt or model do not affect already-generated reports, ensuring forensic immutability.
Before starting (modal with acceptance), at the top and footer of the report, and on the back of the certificate. Adapted by jurisdiction in countries where "psychology" is a regulated profession (BR, PT, ES, FR and others). The canonical text lives at /scientific-disclaimer.
NOESIS started in 2025 as an internal HR tool. External demand grew quickly: companies requesting whitelabels, independent professionals requesting access to reports without building their own infrastructure. In 2026 we opened as a multilingual platform.
HR tool for internal use: Big Five + Schwartz applied in quarterly pulse surveys.
Three companies requesting their own panel with branding, aggregates, and direct contracting. The B2B version is born.
B2C launch with 12 tests in PT, EN, and ES. Test Constructor with 4 IP risk colors. Affiliates program in beta.
Corporate whitelabels and Duo pairing in production. Integrative Chat crossing multiple tests from the same person.
Snapshot from May 2026. Updated monthly in the transparency report.
90+ scientific tests
Curated scientific instruments, science-based. Continuously expanding.
6 languages
PT, EN, ES, FR, DE, IT. Each test is translated with local scientific approval before publishing.
NOESIS operates with 14 people on the internal team. Sensitive decisions (catalog, AI, IP, disclaimer) go through three external committees composed of credentialed researchers.
Engineering, design, data science, and operations. Lean structure because scientific content and multilingual copy are done by credentialed partners, not by generalists with a prompt.
Each sensitive decision goes through one of the three committees. Opinions are public when the decision affects users, and archived when it affects only internal operations.
5 researchers. Approve new catalog tests, define risk colors, validate authorized reinterpretations.
3 ML specialists + 2 in applied ethics. Audit canonical prompts, model versions, and report template changes.
4 lawyers (BR, PT, ES, FR). Adapt disclaimer, adjust terms by country, and monitor data protection regulations.
We respond within 3 business days. For product technical issues, use the support channel.
