AI Use Disclosure
What we built, and how AI assisted.
Disaster Compass was built by a two-student team during the official competition window. We used AI tools to assist development; all architecture, data design, and final decisions were made by us, and no code or designs were copied.
Tools and what each did
- Lovable
- UI and project scaffolding, edge functions, data to Supabase operations and various functionalities.
- Claude Code
- Complex TypeScript implementation, shared types, rules engines (action / scoring / matching / recovery), react-leaflet map integration, and live-API data layer.
- ChatGPT
- Advised on architecture, provided focused code snippets, and assisted with prose.
- Claude
- Advised on architecture, provided focused code snippets, and assisted with prose.
- FlowScholar
- Task management, task assignment, and understanding workflows with minimal AI assistance.
How AI is used in the product
Rules decide every recommendation — AI does not make safety decisions. AI-style text only phrases the rule output in plain, calm language. Volunteer help always requires human coordinator approval.
Data honesty
Disaster Compass uses real-time map APIs and live data feeds for routing, weather, alerts, and elevation. The app is built on a modern tech stack and pulls from public data sources where available, with sensible fallbacks when connectivity is limited.
Emergency services
Disaster Compass complements — it does not replace — emergency services. Call 911 in an emergency.