🩺 Medical Assistant (RAG-based Chatbot)

🚀 Live Demo:
👉 https://medical-assistant-w9bt2juw8kghmos5pwg6qy.streamlit.app/

A professional medical question-answering chatbot built using LangChain, Pinecone, Hugging Face embeddings, and Streamlit.
The assistant answers only from verified medical documents using Retrieval-Augmented Generation (RAG).

---

✨ Features

• 🩺 Medical-only responses from trusted documents
• 🔍 Semantic search using Pinecone Vector DB
• 🧠 HuggingFace MiniLM embeddings
• 💬 Clean Streamlit chat interface
• 🚪 Exit commands supported (exit, stop, over)
• 🧹 Clear chat option

---

🏗️ Architecture

PDF Documents
↓
Cleaning
↓
Chunking
↓
Embeddings (MiniLM)
↓
Pinecone Vector DB
↓
Retriever
↓
LLM
↓
Streamlit UI

---

---

▶️ Run Locally

git clone https://github.com/pranotosh2/Medical-Assistant.git
cd Medical-Assistant
pip install -r requirements.txt
python ingest.py 
streamlit run app.py

---

🔐 Environment Variables

Create a .env file:

PINECONE_API_KEY="your_pinecone_key"
HUGGINGFACEHUB_API_TOKEN="your_huggingface_token"

---

🛡️ Medical Disclaimer

This application is for educational purposes only.
It is not a substitute for professional medical advice, diagnosis, or treatment.

---

👨‍💻 Author

Pranotosh Mandal
GitHub: https://github.com/pranotosh2