StormBridge AI is an international hackathon MVP for weather intelligence, disaster preparedness, AI risk analysis, community reports, and offline accessibility.

• Supabase is already connected to the existing contactfeed project.
• The StormBridge tables are already created in Supabase.
• Open-Meteo works without an API key.
• NVIDIA is configured through server-side environment variables.
• The app keeps working with deterministic fallback agents if any provider fails.

master orchestrator -> weather intelligence -> disaster response -> community intelligence -> offline accessibility -> notification and escalation -> translation

The chain lives in src/lib/agents/chain.ts and always returns the required JSON shape.

The local .env.local is already populated for this workspace. For Vercel, add:

NEXT_PUBLIC_SUPABASE_URL=
NEXT_PUBLIC_SUPABASE_ANON_KEY=
SUPABASE_SERVICE_ROLE_KEY=
NVIDIA_API_KEY=
NVIDIA_RISK_API_KEY=
NVIDIA_EMBEDDING_API_KEY=
NVIDIA_VISION_API_KEY=
NVIDIA_BASE_URL=https://integrate.api.nvidia.com/v1
NVIDIA_RISK_MODEL=meta/llama-3.1-70b-instruct
NVIDIA_EMBEDDING_MODEL=nvidia/nv-embedqa-e5-v5
NVIDIA_VISION_MODEL=meta/llama-3.2-90b-vision-instruct

npm install
npm run dev

Open http://localhost:3000.

• /: premium product entry page.
• /check: location input, user type selector, weather risk result.
• /report: community hazard report form.
• /dashboard: responder dashboard showing alerts and reports.
• /offline: last saved alert and emergency checklist.

• POST /api/weather
• POST /api/analyze-risk
• GET /api/reports
• POST /api/reports
• GET /api/alerts
• GET /api/integrations/status

{
  "location": "",
  "risk_level": "Low | Medium | High | Critical",
  "main_threat": "",
  "summary": "",
  "recommended_actions": [],
  "offline_message": "",
  "alert_message": "",
  "affected_groups": [],
  "confidence": "Low | Medium | High"
}

1. Push the repo to GitHub.
2. Import it in Vercel.
3. Add the environment variables above.
4. Deploy and test /check, /report, /dashboard, and /offline.

• Risk guidance uses the NVIDIA risk model key.
• Embeddings and vision keys are separated so future report clustering and image/satellite analysis do not block the core chain.
• Supabase stores alerts, reports, and user location checks.
• localStorage keeps the latest alert available offline.