VoiceMedi AI 🩺🤖

This repo is a fork for publication, to distinguish it from the original thesis repo.
Any code changes after the thesis are done here.

An Edge-AI Chatbot for Symptom Extraction and Appointment Pre-Fill

VoiceMedi AI is an open-source healthcare chatbot. It helps users describe symptoms naturally, predicts conditions using machine learning, and shares structured appointment data. This prototype is a core module in a broader Edge-AI pipeline to automate clinical intake workflows.

---

🧠 Key Features

• 🔍 Symptom-Based Condition Prediction – Machine learning-based diagnosis using natural language input
• 💬 Conversational Interface – Interprets user input via Ollama's Chat completion API
• 📦 Modular Design – Can be extended to cover other medical domains and scheduling use cases
• ⚙️ Modern Stack – FastAPI backend, React frontend, containerized with Docker
• 🧩 Edge-AI Ready – Completely offline designed with edge deployment scenarios in mind

---

🚀 Getting Started (with Docker)

Prerequisites:

• Ollama installed

• Docker & Docker Compose installed

• Environment variables configured:

  • backend/.env:

    OLLAMA_URL=http://host.docker.internal:11434 ## incase of local deployment, use localhost
    DEBUG=True  
    

  • frontend/.env:

    BACKEND_URL=http://localhost:8000  
    

---

🦙 Ollama Setup (for LLaMA 3.2:1b Model Inference)

VoiceMedi AI uses Ollama to run lightweight LLMs like LLaMA 3.2:1b locally or on edge devices.

🛠️ Step 1: Install Ollama

Linux

curl -fsSL https://ollama.com/install.sh | sh

Mac

brew install ollama

Windows

Download the official installer: 👉 ollama

📥 Step 2: Download the LLaMA 3.2:1b Model

ollama pull llama3:2.1b

This will download and configure the model locally for inference.

🚀 Step 3: Run the Model

ollama run llama3:2.1b

Run the Project:

# Clone the repository
git clone https://github.com/Prof-it/voice_chatbot.git
cd voice_chatbot

# Build and launch containers
docker-compose up --build

---

🧪 Local Development Setup

Backend (FastAPI)

cd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --reload

Frontend (React)

cd frontend
npm install
npm start

---

🖥️ System Architecture

---