Skip to content

Curaway Documentation

AI-powered cross-border medical travel coordination — connecting patients from the US, UK, and UAE with accredited healthcare providers in India, Turkey, Thailand, and Spain.

:material-api: Live API Docs :material-web: Launch App

Technology Stack

Layer Technologies
Backend FastAPI (Python 3.12+), LangGraph, LangChain
Frontend Next.js, React, TypeScript (Vercel)
Databases PostgreSQL (Railway), Neo4j Aura, Qdrant Cloud
Cache & Events Upstash Redis, Upstash QStash
Storage Cloudflare R2 (S3-compatible)
Auth Clerk (JWT, RBAC, multi-tenant orgs)
Feature Flags Flagsmith
LLM Inference Claude Haiku 4.5, Claude Sonnet 4.6, GPT-4o mini
Observability Langfuse (LLM traces), PostHog (user behavior)
Email Resend
Hosting Railway Pro ($20/mo), Vercel (free)

Total infrastructure cost: ~$20/month (21 of 25 services on free tiers)

Platform Flow

graph LR
    A[Patient uploads<br/>medical report] --> B[AI extracts<br/>clinical data]
    B --> C[Structured FHIR<br/>record created]
    C --> D[Provider + Doctor<br/>matching]
    D --> E[Explainable<br/>reasoning]
    E --> F[Consent &<br/>forwarding]

Key Numbers

Metric Value
PostgreSQL tables 22+
Providers seeded 42 (8 countries)
Doctors seeded 8 (mixed completeness)
Neo4j nodes 158+
Neo4j relationships 549+
Qdrant vectors 124+ (4 collections)
API endpoints 50+
Test coverage 100+ tests
LLM agents 4 (Clinical Context, Intake, Match, Explanation)
Resource URL
Live API services.curaway.ai
Swagger Docs services.curaway.ai/docs
App app.curaway.ai
Backend Repo github.com/whoinsane/curaway
Frontend Repo github.com/whoinsane/curaway-health-navigator