Moshimo Town

Rehearse one change before rearranging your coffee shop. Moshimo Town is a visual workflow sandbox for small coffee-shop owners and planners. Arrange five stations, run a simulated morning, then ask an AI companion to suggest one improvement. Before and after share the same customers, assumptions and clock, so you can see how a changed layout affects staff walking and cups served. Recommendations remain unapplied until you accept them. Watch both layouts side by side, inspect the changed routes, and decide whether to keep the move or try again. In the recorded demo, two accepted Muse recommendations reduced simulated walking from 617 to 514 meters, while cups served increased from 56 to 57. These are partial improvements in one synthetic scenario, not real-store savings or completion of the introductory challenge. We began with a geographic shop-opening concept, then narrowed the scope because location data could not support credible demand forecasts. The resulting experience focuses on a concrete, testable decision: where to move one piece of equipment. Its intended benefit is helping owners explore workflow tradeoffs before physically rearranging a shop; practical impact remains to be tested with owners. Live app: https://moshimo-town.studio-gennai1.workers.dev No login is needed for manual rehearsal or live Muse experiments using listed improvement preferences. Optional reviewer access enables short written preferences. The application does not persist prompts, layouts or model results. Cloudflare processes network/security data and Meta processes live inputs.