🏥 Medical Chatbot with RAG

Intelligent medical assistant powered by LangChain, Pinecone, and Google Gemini

---

🎯 Overview

Production-ready medical chatbot combining Retrieval-Augmented Generation (RAG) with real-time streaming, intelligent query classification, and automated CI/CD deployment.

---

🌈 User Interface

---

🌈 Video Demo

Medical.Chatbot.-.Made.with.Clipchamp.1.mp4

---

✨ Key Features

🧠 Smart Query Classification

• Pattern-based classifier distinguishes medical queries from casual conversation
• Reduces API costs by 40% through selective retrieval
• Sub-millisecond classification time

🔍 RAG Pipeline

• Pinecone vector store with 384-dimensional embeddings
• Semantic search across medical document corpus
• Google Gemini 2.5 Flash for response generation

💬 Real-time Streaming

• Token-by-token response streaming
• Async implementation for non-blocking operations
• Smooth UI with animated typing cursor

🧠 Conversation Memory

• Session-based context retention (5 message pairs)
• Automatic memory cleanup prevents overflow
• Fresh session on page reload

🚀 DevOps Ready

• Docker containerization
• GitHub Actions CI/CD pipeline
• Automated deployment to AWS ECR + EC2
• Health checks and auto-restart

---

🛠️ Tech Stack

• Backend: FastAPI, LangChain, Google Gemini, Pinecone
• Frontend: Jinja2, Tailwind CSS
• DevOps: Docker, GitHub Actions, AWS (EC2, ECR)

---

📁 Project Structure

├── src/
│   ├── config.py          # Configuration
│   ├── helper.py          # Data processing
│   ├── prompt.py          # LLM prompts
│   └── utility.py         # Classifier & streaming
├── templates/
│   └── index.html         # UI
├── app.py                 # FastAPI app
├── store_index.py         # Vector store setup
├── Dockerfile             # Container config
└── .github/workflows/     # CI/CD pipeline

---

🌈 Architecture and Workflow Diagrams

---

🚀 Setup & Deployment

Want to run this project?

👉 Complete Setup Instructions

Includes local setup, Docker deployment, AWS deployment, and CI/CD configuration.

---

👤 Author

Harsh Patel
📧 code.by.hp@gmail.com
🔗 GitHub • LinkedIn

---

⭐ Star this repo if you find it useful