ClinicBook - Agentic AI-Powered Healthcare Platform (Production-Grade)

A production-ready, multimodal, AI-powered healthcare platform that transforms appointment booking into an intelligent, conversational workflow.

ClinicBook features a Context-Aware Agentic AI capable of reasoning, speaking, listening, reading, writing, and executing actions across a real clinic management system.

Live Deployment: https://clinicagentbook.onrender.com
Docker Image: https://hub.docker.com/r/honeydoc/clinicbook-repo
CI/CD: GitHub Actions → Docker Hub → Render

---

Key Features

Agentic AI Health Assistant (Powered by Gemini)

A hybrid system combining Agentic Tool Use + Retrieval-Augmented Generation (RAG).

• ** Voice Interface**: Built-in Speech-to-Text lets users speak commands like:

  "Book an appointment", "Create slots for next Monday", "Who is visiting me tomorrow?"

• ** Hybrid RAG System**:

  • Dynamic Data: Reads live data from SQLite (doctors, slots, appointments).
  • Static Knowledge: Answers clinic policy questions from clinic_policies.txt without DB calls.

• ** Context-Aware Reasoning**:

  • Automatically detects User Role (Doctor/Patient)
  • Understands Relative Dates like today, tomorrow, next Monday

• ** Tool-Calling Agent**:

  • Executes DB tools for:
    • Booking appointments
    • Creating slots
    • Marking appointments completed
    • Searching doctors

---

Doctor Portal

• Voice-Controlled Slot Management

• Smart Dashboard

• Ask Natural Questions:

  "Who is visiting me today?"

• Agentic Automation:

  • Create slots
  • Close appointments
  • Manage schedules through chat

• Secure Session Isolation

---

Patient Portal

• Symptom-to-Specialist Mapping

• One-Click Booking

• Ask AI:

  "Which doctor should I visit for chest pain?"

• Transparent Appointment Status & Policies

---

Production Architecture

GitHub Push
↓
GitHub Actions (CI)
↓
Docker Build & Push
↓
Docker Hub
↓
Render Auto Deploy (CD)
↓
Live Production Service

---

Technology Stack

• Backend: Flask (Python)
• AI Engine: Google Gemini 2.5 Flash
• AI Architecture:
  • Agentic Tool Calling
  • Lightweight RAG (Policy Injection)
  • Role-Based System Prompting
• Frontend:
  • Bootstrap 5 + Jinja2
  • Web Speech API (Voice)
• Database: SQLite
• DevOps:
  • Docker
  • GitHub Actions CI/CD
  • Render Image Deployment

---

AI Query Router Logic

The agent uses a Router Architecture:

1. Knowledge Query

   "Do you accept insurance?"

   • Source: clinic_policies.txt
   • Action: RAG Response

2. Data Query

   "Find a cardiologist"

   • Tool: search_doctor_by_specialization
   • Action: DB Read

3. Action Command

   "Book the 10 AM slot"

   • Tool: book_appointment_by_patient
   • Action: DB Write

---

Run Locally

Prerequisites

• Python 3.8+
• Google Gemini API Key

Installation

git clone https://github.com/Mfaj-cod/ClinicBook.git
cd ClinicBook
python -m venv clinic
clinic\Scripts\activate   # Windows
source clinic/bin/activate # Mac/Linux
pip install -r requirements.txt

Configure Environment

Create .env:

GEMINI_API_KEY=your_api_key
GEMINI_MODEL_NAME=gemini-2.5-flash
SECRET_KEY=your_secret_key

RUN: python app.py

Run with Docker

docker pull honeydoc/clinicbook-repo:latest
docker run -p 5000:5000 honeydoc/clinicbook-repo

---

Project Structure

ClinicBook/
├── app.py
├── Dockerfile
├── requirements.txt
├── src/
│   ├── init_db.py
│   ├── seed.py
│   ├── gem.py
│   ├── prompt.py
│   ├── tools_config.json
│   └── doctors_data.py
├── data/
│   ├── clinicBook.db
│   └── clinic_policies.txt
├── templates/
├── static/
└── README.md

Use Cases

• For Clinics:

  • AI receptionist
  • Voice-controlled scheduling
  • Zero-dashboard navigation
  • Automated workflows

• For Patients:

  • Natural language booking
  • Smart doctor discovery
  • Instant answers

Roadmap

Agentic AI with tool calling ✔️

Voice input ✔️

RAG knowledge base ✔️

CI/CD deployment ✔️

Payment gateway

Analytics dashboard

Multi-clinic support

📞 Support

Open an issue in GitHub for support.

Built With Purpose

This project demonstrates real-world Agentic AI, DevOps, MLOps, and production system design.