Healin is a futuristic, intelligent medical chatbot powered by Google Gemini and Pinecone vector embeddings. It uses a robust internal RAG (Retrieval-Augmented Generation) pipeline to fetch domain-specific medical knowledge to ensure accurate and context-aware responses.
- Cyberpunk UI: A beautifully stylized hacker-aesthetic chat interface.
- RAG Architecture: Uses Pinecone for rapid document retrieval alongside Hugging Face embeddings.
- Powered by Gemini: Fully integrated with the Google gemini-1.5-flash generative AI framework for instantaneous intelligence.
Follow these step-by-step instructions to run the application completely offline on your own machine.
Make sure you have Python 3.10+ installed on your computer.
Open your terminal (PowerShell or Command Prompt) and navigate into the project folder. Create and activate an isolated Python environment so your global packages don't conflict:
Create the environment:
powershell python -m venv venv2
Activate the environment:
powershell .\venv2\Scripts\activate
(You should see (venv2) appear at the start of your terminal line).
With your environment active, install all required packages:
powershell pip install -r requirements.txt
Create a file named literally .env inside the main project folder. Add your API keys inside this file:
ini PINECONE_API_KEY="your_pinecone_key" HUGGINGFACEHUB_API_TOKEN="your_huggingface_token_for_embeddings" GOOGLE_API_KEY="your_google_gemini_api_key"
If this is your first time setting up the project and Pinecone requires the medical documents, run the index script to build your vector database:
powershell python store_index.py
Boot up the main Flask backend:
powershell python app.py
Open your favorite web browser (Chrome, Edge, Safari) and navigate to: 👉 http://127.0.0.1:8080
Note: All old boilerplate AWS, Docker, and deployment legacy artifacts have been cleanly removed to focus strictly on an optimized, fast offline RAG implementation.