Construct a population
Combine editable market profiles with psychometric-style statements and personal details. Seeded generation creates fictional respondents; optional market shares come from your own evidence.
Desktop studio / Synthetic market exploration
Explore a market.
Question a concept.
Build fictional respondent populations, ask structured questions through Jev, and inspect what changes across concepts and segments. A local workspace for developing research hypotheses.
Explore the method01 / The method
From an idea
to a better research question.
Typed decisions through Jev.
Analysis in a local desktop workspace.
Run an experiment
Supply context and structured questions. Jev Score handles ordered scales; Jev Choice handles nominal choices. Results include typed decisions and option probabilities.
Compare and inspect
Explore distributions, segment differences, combined metrics, and perceptual maps. Compare concepts and trace results back to individual synthetic personas.
Desktop application in development
A studio for hypotheses.
JevPersonas is built with Tauri, React, and TypeScript. Workspaces live locally and can be exported as portable JSON; research results can also be exported as CSV. Running a study sends its persona context and questions to the selected model provider.
These are model responses to fictional contexts. Use them to explore concepts and improve real-world research, rather than treating them as observed consumer demand or a representative survey.
More about Mariano