Rehearsal

Rehearsal

Rehearsal is a voice-first deliberate-practice lab for difficult workplace conversations, built especially for the moment new managers know a conversation has to happen but do not have a safe place to practise the exact turn where it can go wrong.

Core loop

Attempt → evidence → rewind → retry → compare.

Instead of giving generic advice, Rehearsal puts the user inside a structured scenario, tracks observable communication behaviors, identifies the exact turn where the attempt broke down, and lets the user rewind and retry from that moment. The first demo scenario is pushing back on an unrealistic deadline.

Practice, not advice

Rehearsal evaluates only observable behaviors such as acknowledging the other person's concern, stating a clear position, giving a concrete reason, and proposing a next step. It does not make unsupported judgments about personality, confidence, honesty, competence, anxiety, or mental state.

Scenarios have goals, counterpart constraints, explicit rubrics, deterministic state transitions, retry checkpoints, and before/after evidence. The model layer can generate counterpart dialogue and explain verified evidence, but it cannot directly manufacture rubric scores. The key experience is safe deliberate practice before the real conversation: retry the exact moment that broke down, then compare the observable evidence before and after.

RevenueCat

RevenueCat powers the Pro entitlement. The free product keeps the complete starter practice/retry loop; Pro expands repetition with unlimited analyses, advanced scenario packs, custom scenarios, longitudinal skill history, and before/after exports. Purchase and restore flows both derive premium access from RevenueCat CustomerInfo, and the integration fails closed when configuration, an active entitlement, or an offering is missing.

Development targets RevenueCat Test Store so the final native demo can show a real entitlement transition without requiring a paid store account. The actual Test Store project ID and live purchase/restore transition are still being verified and will not be claimed until proven.

Shipaton targets

Primary submission path: Next Gen Award. Additional strategic fit: Influencer Award — Career Coaching, RevenueCat Design Award, and HAMM Award.

Current verified build status

Implemented in source: deterministic scenario/rubric engine, rewind checkpoints, before/after scoring, constrained coaching adapter, Expo/React Native mobile UI, RevenueCat entitlement/purchase/restore adapter, fail-closed entitlement tests, native iOS build/capture automation, exact 1024×1024 app icon, and exact 1179×2556 frameless screenshot asset.

Remaining before final submission: verify the real RevenueCat Test Store project and project ID, exercise the actual native purchase/restore entitlement transition, assemble and publicly host the ≤2-minute demo, attach final Devpost media, validate the student's academic email for Next Gen eligibility, and run the final compliance audit.

Built With

• expo.io
• react-native
• react-native-purchases
• revenuecat
• typescript
• vitest