Project Deep Dive
MediCall
MediCall is a voice-based healthcare coordination system that performs multilingual intake, understands insurance constraints, and routes users to appropriate in-network care. It uses multi-agent workflows for triage, matching, and scheduling, grounded by RAG over structured benefits data (copay, coinsurance, deductibles). It also includes an RBAC-secured dashboard for clinic/hospital operations and real-time data integration.
Thesis
Multilingual real-time voice AI that routes patients to in-network care using insurance-grounded retrieval and automated scheduling workflows.
Key capabilities
- Handles multilingual voice intake and routes users to appropriate in-network care.
- Runs multi-agent workflows for insurance NLP, triage, hospital matching, and scheduling.
- Retrieves copay/coinsurance/deductibles via RAG over structured benefits data.
- Provides real-time operational dashboards (patients, prescriptions, appointments).
- Uses custom API routes + React hooks for dynamic data fetching in the UI.
- Secures clinic/hospital tooling with RBAC via Auth0.
Stack
Twilio, ElevenLabs, n8n, MongoDB Vector Search, Next.js, Auth0, OpenCage, Google Maps, Retell AI, LangGraph, MongoDB, OpenAI Speech API
Add richer write-up, architecture diagrams, and production links here as you publish them.