intellicia-public/parastore

Parastore lets you sketch a retail floor plan, generate a batch of LLM personas with configurable demographics and preferences, then run them through the store to observe browsing and purchasing behavior. The isometric 3D visualization is not just cosmetic — it makes the simulation readable at a glance, which matters when you are running dozens of persona variants. What separates this from a spreadsheet of synthetic survey responses is the spatial and sequential element: personas react to placement, proximity, and shelf layout, not just product attributes. That is a genuinely novel hook for anyone doing early-stage positioning work, store-format testing, or training data generation for retail models. The open-source repo means you can wire in your own persona templates or swap the LLM backend. Honest reservation: this is a research sandbox, not a production analytics tool — expect rough edges and no SLA. -> Best for: indie hacker or ML researcher doing synthetic consumer research or retail simulation work