Healthcare AI Voice Agent – AI-Powered Medical Assistant & Voice Triage

Healthcare AI Voice Agent is a full-stack healthcare application that uses AI-powered voice interactions for patient triage, symptom analysis, specialist matching, and appointment booking. Built with FastAPI, PostgreSQL, React, LangGraph, and leading AI models (Gemini, GPT-4).

Author & Contact

For questions, collaboration, or support, reach out via the links above or open an issue on GitHub.

Table of Contents

• What Is Healthcare AI Voice Agent?
• Key Features
• System Architecture & User Flow
• Technology Stack
• Project Structure
• Quick Start
• API Keys & Configuration
• Running the Application
• Troubleshooting
• License

What Is Healthcare AI Voice Agent?

This healthcare AI assistant streamlines the patient journey by providing:

• Intelligent voice triage – Natural-language symptom collection and analysis via voice
• AI-powered symptom analysis – Symptom normalization and specialist recommendation (Gemini, GPT-4, LangGraph)
• Smart doctor matching – Healthcare provider lookup by specialization (PostgreSQL)
• Appointment booking – Scheduling with optional payment (Razorpay)
• Voice-enabled UI – React frontend with Web Speech API for a conversational experience

Ideal for medical triage, symptom checker projects, healthcare chatbots, and AI voice assistants in healthcare.

Key Features

System Architecture & User Flow

Voice UI (React) → Auth (FastAPI + PostgreSQL) → Voice symptoms (Web Speech API)
    → AI analysis (LangGraph + Gemini/GPT-4) → Specialist lookup (DB)
    → Doctor list & slots → Appointment booking → Payment (Razorpay) → Done

Component Flow

1. Authentication – Email/password → FastAPI /login → sp_login_user → JWT/session
2. Voice pipeline – Microphone → Web Speech API → text → phrase array → LangGraph
3. AI workflow – Raw symptoms → Gemini/GPT-4 → normalized symptoms → specialist mapping → recommendations
4. Data – FastAPI → PostgreSQL stored procedures → JSON → frontend
5. Booking – Doctor + slot → patient details → slot check → sp_create_appointment → payment

Technical Layers

Technology Stack

Project Structure

Healthcare-AI-Voice-agent/
├── backend/                 # FastAPI backend
│   ├── main.py             # API entry, routes
│   ├── config.py           # Configuration
│   ├── db.py               # DB connection
│   ├── models.py           # Pydantic models
│   ├── langgraph_llm_agents.py  # LangGraph AI workflow
│   └── requitements.txt    # Python dependencies
├── src/                    # React frontend (Vite)
│   ├── App.jsx, main.jsx, Root.jsx
│   ├── components/         # Assistant, Dashboard, Recommendation, etc.
│   ├── context/            # UserContext
│   └── gemini.js           # Gemini client
├── sql/
│   ├── schema.sql          # Tables (users, patients, doctors, appointments, etc.)
│   └── functions/          # Stored procedures
│       ├── sp_login_user.sql
│       ├── sp_get_patient_details.sql
│       ├── sp_get_specialists.sql
│       ├── sp_get_doctors_by_specialists.sql
│       ├── sp_get_medical_history.sql
│       ├── sp_get_patient_id.sql
│       └── sp_create_appointment.sql
├── public/
├── index.html
├── package.json            # Frontend deps
├── vite.config.js
└── README.md

Quick Start

Prerequisites

• Python 3.8+
• Node.js 16+
• PostgreSQL 12+
• Git

1. Database setup

psql -U postgres -c "CREATE DATABASE healthcare;"
psql -U postgres -c "CREATE USER fastapi_user WITH PASSWORD 'your_secure_password';"
psql -U postgres -c "GRANT ALL PRIVILEGES ON DATABASE healthcare TO fastapi_user;"

Apply schema and functions (from project root):

psql -U fastapi_user -d healthcare -f sql/schema.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_login_user.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_get_patient_details.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_get_specialists.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_get_doctors_by_specialists.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_get_medical_history.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_get_patient_id.sql
psql -U fastapi_user -d healthcare -f sql/functions/sp_create_appointment.sql

2. Backend setup

cd backend
python -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate
pip install -r requitements.txt

Create backend/.env (see backend/.env.example):

DB_HOST=localhost
DB_PORT=5432
DB_USER=fastapi_user
DB_PASSWORD=your_secure_password
DB_NAME=healthcare
FRONTEND_ORIGIN=http://localhost:5173
OPENAI_API_KEY=your_openai_api_key
GEMINI_API_KEY=your_gemini_api_key

Start backend:

uvicorn main:app --reload --host 0.0.0.0 --port 8000

3. Frontend setup

From project root:

npm install

Create .env.local (see .env.local.example):

VITE_BACKEND_URL=http://localhost:8000
VITE_GEMINI_API_KEY=your_gemini_api_key
VITE_RAZORPAY_KEY_ID=your_razorpay_key_id

Start frontend:

npm run dev

Open http://localhost:5173.

API Keys & Configuration

• OpenAI – platform.openai.com → API Keys → add to backend/.env
• Google Gemini – Google AI Studio → API key → backend .env and frontend .env.local
• Razorpay – Razorpay Dashboard → Test mode → API keys → frontend .env.local

Running the Application

1. Start PostgreSQL.
2. Backend: cd backend && uvicorn main:app --reload --port 8000
3. Frontend: from root, npm run dev
4. Visit http://localhost:5173

Health check: curl http://localhost:8000/

Troubleshooting

License

This project is licensed under the MIT License. See LICENSE for details.

Healthcare AI Voice Agent – built for better healthcare accessibility.
Author: KuchikiRenji · Contact · Discord: kuchiki_renji

For GitHub (Description & Topics)

Use the following in your repo About section:

Description (short):

Full-stack healthcare AI voice assistant: voice triage, symptom analysis, specialist matching & appointment booking. React, FastAPI, LangGraph, PostgreSQL.

Topics (add these in GitHub → About → Topics):

healthcare
healthcare-ai
voice-assistant
medical-assistant
symptom-checker
patient-triage
langgraph
fastapi
react
postgresql
openai
gemini
razorpay
full-stack
ai-agents