LifeBridge

Login Page
Attorney Network
Chatbot assistant
Crisis Console
Crisis Corridor
Dashboard
Document Vault
Find Embassy
Help
Knowledge Base
My Cases
Tasks
Timeline

Inspiration

People in crisis — whether fleeing conflict, natural disasters, or personal emergencies — face fragmented, slow, and confusing systems when they need help the most. We were inspired by the real-world gap between those in danger and the resources that could save them. Cross-border mobility during a crisis is a life-or-death challenge, and we wanted to build something that makes navigation, reunification, and safety accessible in real time.

What It Does

LifeBridge is a crisis navigation and reunification platform that moves people in crisis to safer nearby options using:

Live AI Guidance — Powered by DigitalOcean Gradient AI with retrieval-augmented generation (RAG) and traceable tool calls
Safe Haven Search — Verified nearby shelters, services, and refuges with rich metadata
Risk-Aware Route Generation — Routes scored by danger level, border crossings, and safety indicators
Safety Check-Ins — Idempotent check-ins to signal you're okay
Family Reunification Beacons — Create, lookup, and update reunion points for separated families
Nearby Help Matching — Match help requests with volunteers and aid offers in real time
Tracker Workspace — Case continuity with tasks, notes, history, and document storage
Crisis Command Center — A live operations console (/crisis) with runtime health indicators

How We Built It

Architecture:

Frontend: Next.js with Tailwind CSS (responsive, mobile-first design)
Core API: FastAPI with DigitalOcean Gradient AI integration
Tracker API: FastAPI with SQLite for case management and continuity
Docgen Service: FastAPI for document generation
Storage: PostgreSQL + object storage compatible pattern
Deployment: Google Cloud Run (fully containerized)

DigitalOcean Gradient Integration: LifeBridge uses Gradient in live mode through a DigitalOcean Agent endpoint:

Live runtime status: GET /crisis/runtime
Live connectivity probe: GET /crisis/runtime/live-check
Agent orchestration query: POST /crisis/agent/query
Retrieval-aware sources and persisted traces
Runtime modes: live (strict Gradient), mock (local fallback), auto (smart fallback)

Challenges We Ran Into

Designing idempotent check-ins that are reliable even with network instability
Making AI-powered route generation risk-aware without false positives
Ensuring the Gradient AI runtime gracefully falls back to mock mode when connectivity is unavailable
Building a real-time family reunification beacon system that works across borders
Containerizing and orchestrating 4 services (web, api, tracker-api, docgen) for Cloud Run deployment

What We Learned

DigitalOcean Gradient's agent endpoint is powerful for real-time retrieval-augmented AI at scale
Crisis UX demands extreme clarity — every second matters, so UI/UX was ruthlessly simplified
Operational resilience (fallback modes, idempotency) is as important as features in emergency tools
FastAPI + Next.js is an excellent stack for rapid full-stack prototyping with live deployments