Pressure Test

Pressure Test is a mobile-first behavioral scenario game that captures how a person makes decisions under startup pressure, then generates an evidence-based personality artifact from their actions.

This project was built as a demo-ready MVP focused on product taste, full-stack development, and practical behavioral data design.

Why this project matters

Most personality products rely on self-report forms. Pressure Test uses behavioral evidence:

• forced choices in a high-stakes scenario
• hesitation timing and free-text edits
• a live voice interruption (with resilient fallback)
• transparent trait scoring with confidence and contradiction logic

The result is a user experience that feels like a game, but behaves like a personalized signal platform.

What I built

End-to-end product experience

• / polished landing page with clear framing and CTA
• /play one complete playable Tangle (Launch Day Rollback)
• /results/:id behavioral artifact with evidence, contradiction, and growth edge
• /share/:id collectible public share card
• /admin/trait-lab internal analytics view for signal quality and model inspection

Product features

• Progressive timeline narrative and chat-style interlude
• Free-text capture with edit/backspace behavioral signals
• Vapi-powered incoming call interruption with realistic phone UI
• Typed fallback mode if voice fails (never breaks the user flow)
• JSON export of session + artifact data
• Low-confidence warnings and mixed-profile explanations

Highlights

• Typed, modular architecture separating:
  • scenario content
  • scoring logic
  • voice integration
  • UI/presentation
• Versioned scoring/artifact model for future experimentation
• Convex schema + functions designed for production migration
• Local storage shim that mirrors backend entities for zero-friction demoing
• Obsidian-style force graph (D3) in Trait Lab for exploratory analysis

Tech stack

• React + Vite + TypeScript
• Tailwind CSS
• Convex (schema/functions scaffolded)
• Vapi Web SDK
• D3 (graph visualization)

Architecture snapshot

src/
  pages/           Landing, Play, Results, Share, TraitLab
  components/      tangle/, voice/, results/, traitlab/
  lib/
    scenario/      scenario primitives and engine
    scoring/       trait scoring and artifact generation
    voice/         Vapi integration and signal extraction
    traits/        typed trait model and versions
    sessionStore   local data shim matching Convex schema
convex/
  schema + session/voice/artifact functions

Build and run

npm install
npm run dev

For full setup, environment variables, deployment notes, and implementation details, see BUILD.md.

Additional Documentation

• Vapi integration: see https://docs.vapi.ai
• Product + implementation guide: BUILD.md
• Convex backend notes: convex/README.md