title	Medical Chatbot
emoji	🩺
colorFrom	blue
colorTo	green
sdk	gradio
sdk_version	5.44.1
app_file	app_gradio.py
pinned	false
license	mit

Medical Chatbot with RAG

A Retrieval-Augmented Generation (RAG) chatbot that answers medical questions using LangChain + FAISS.

Live Demo

How It Works

1. PDF → Chunks — Medical PDF is split into ~500-char chunks
2. Chunks → Embeddings — Each chunk is embedded with all-MiniLM-L6-v2
3. Embeddings → FAISS Index — Stored for fast similarity search
4. Query → Retrieve → Generate — User question retrieves top-4 chunks, LLM generates answer

Local Setup

git clone https://github.com/AyushDocs/medical_chatbot_rag.git
cd medical_chatbot_rag
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Build FAISS index (place PDF in data/ first)
python store_index.py

# Run Flask app
python app.py

Models

Component	Model	Size
Embeddings	sentence-transformers/all-MiniLM-L6-v2	80MB
LLM	google/flan-t5-base	~250M params