syedabraash

Healthcare Voice AI Assistant — AI skill for Claude Code

AI community

Voice AI agent for patient intake — Vapi + Claude + FastAPI + SQLite.

How to install Healthcare Voice AI Assistant

This entry records only its repository, not the path inside it, so there is no exact command to give. Open syedabraash/healthcare-voice-ai-assistant and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Healthcare Voice AI Assistant does

Voice AI agent for patient intake — Vapi + Claude + FastAPI + SQLite. Collects demographics over a phone call, validates and persists via a REST API, includes a review dashboard. Backend and voice-agent config are complete and unit-tested; not yet provisioned on a live number (see README).

Alternatives in AI

AI Trading Agent — Multi-symbol AI trading agent — Claude AI + FastAPI + Next.js + MetaTrader 5 72 ★
Language Command — Sets the output language for all system responses 46 ★
Multi AI Chat Skill — Free multi-AI research assistant: your agent consults 6 top web AIs in parallel, cross-validates their answers 38 ★

README

Voice AI Patient Registration

A voice agent that answers a real phone number, conversationally registers a new patient, persists the record, and exposes it through a REST API + small dashboard.

**Status: code complete, not yet deployed to a live number.** See [Known Limitations](#known-limitations--trade-offs) — I built and unit-tested the entire backend, the Vapi assistant config, and the tool-call webhook adapter, but provisioning an actual phone number requires a Vapi/Twilio account (billing + identity signup) that I don't have access to. Deployment steps below take about 15-20 minutes.

Architecture

Caller ──dials──▶ Vapi (telephony + STT/TTS + LLM orchestration)
                        │
                        │ tool calls (lookupPatientByPhone / createPatient / updatePatient)
                        ▼
              POST /vapi/webhook  ──┐
                                     │  app/vapi_webhook.py (adapter layer)
                                     ▼
                              app/crud.py (service layer)
                                     │
                                     ▼
                          SQLite database (app/models.py)
                                     ▲
                                     │
                    GET/POST/PUT/DELETE /patients  ◀── REST clients / dashboard
                              (app/main.py)

**Separation of concerns**, deliberately:

app/main.py — the public REST API. Knows nothing about Vapi.
app/vapi_webhook.py — translates Vapi's tool-call envelope into calls against the same service layer the REST API uses. If you swapped Vapi for Retell or a raw Twilio+GPT stack tomorrow, only this file would change.
app/crud.py — the actual data operations, shared by both entry points.
app/schemas.py — all validation lives here (Pydantic), enforced server-side regardless of what the voice agent thinks is valid.
vapi/system_prompt.md + vapi/assistant_config.json — the conversatio