ClinicCall AI

Inspiration

Clinics still lose hours to missed calls, long hold times, and manual scheduling. I wanted a receptionist that could triage, find nearby providers, and book appointments without heavy staff involvement.

What it does

ClinicCall AI is a chat‑first receptionist that understands intent, asks clarifying questions, locates nearby providers by ZIP, suggests slots, and confirms bookings. It also handles urgent routing, reschedules, cancellations, and sends confirmation emails.

How we built it

FastAPI + SQLAlchemy + Postgres power the backend. An LLM routes patient intent and guides the flow. Provider data comes from the NPI Registry, with ZIP‑based fallback lookups. The UI is a lightweight HTML/CSS/JS front end focused on conversational booking.

Challenges we ran into

Handling missing details (department, ZIP, reason) without stalling the conversation. Also, matching provider data reliably when ZIPs had no direct results and keeping multi‑step state consistent.

Accomplishments that we're proud of

A working end‑to‑end flow from “I need an appointment” to confirmed booking with email confirmation, plus provider discovery from real-world data.

What we learned

Designing resilient conversational flows, integrating live healthcare data, and managing multi‑turn state in a production‑minded backend.

What's next for ClinicCall AI

Bring back voice calling, add clinic/location‑specific routing, and create a lightweight admin dashboard for availability and staffing.

Built With

• api
• docker
• fastapi
• html/css/javascript
• llm
• npi
• openai
• postgresql
• python
• registry
• routing)
• smtp
• sqlalchemy
• zippopotam.us