CureWise-AI is a modern, production-grade healthcare data platform that leverages Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and advanced AI/ML techniques to deliver intelligent analytics, medical report parsing, and conversational AI for hospitals, doctors, and patients.
- Project Overview
- Key Features
- Architecture
- Tech Stack
- Directory Structure
- Setup & Installation
- Development & Contribution
- License
CureWise-AI brings together LLMs, RAG, and healthcare data engineering to provide:
- Conversational AI for medical queries, appointment booking, and patient support.
- Automated parsing and structuring of medical documents (e.g., blood reports) using LLMs.
- Retrieval-Augmented Generation (RAG) for context-aware, accurate responses using a vector database (Pinecone).
- AI-powered analytics and disease detection. Analyze patient images and provide a diagnosis.
- Realtime Agents connected with database to provide realtime analytics and insights.
- Seamless integration with hospital operations, doctor scheduling, and patient management.
- LLM-Powered Chatbot: Natural language interface for patients and doctors, powered by Gemini and Llama models.
- RAG System: Combines LLMs with Pinecone vector search for context-rich, accurate answers.
- Medical Report Parsing: Extracts and structures data from blood reports using LlamaParse and Groq.
- Appointment Booking Agent: Intelligent agent for scheduling, rescheduling, and querying appointments.
- Acne & Disease Detection: Image-based analysis using AI models (extendable for other conditions).
- Memory & Context: Maintains chat and report history for personalized, context-aware responses.
- Extensible Tools: Modular agent design for adding new tools and workflows.
[React Frontend] <---> [FastAPI Backend: LLM, RAG, AI] <---> [PostgreSQL, Pinecone]
|
[LangChain, Llama, Groq, Gemini]
- LLM Agents: Orchestrate workflows, parse documents, and answer queries.
- RAG: Uses Pinecone for vector search over medical conversations and documents.
- FastAPI: Exposes REST APIs for all LLM and AI features.
- PostgreSQL: Stores chat, report, and appointment history.
- LLMs: Gemini (Google), Llama (Meta), Groq
- RAG: Pinecone, LangChain, Google Generative AI Embeddings
- Backend: FastAPI, Python, psycopg2
- Frontend: React
- Database: PostgreSQL
- AI/ML: LlamaParse, custom disease detection models
- Containerization: Docker
backend/
│
├── main.py # FastAPI app, API endpoints
├── utils/
│ ├── agents.py # LLM agent orchestration, tools, booking agent
│ ├── pineconeutils.py # RAG, Pinecone vector DB, retrieval chains
│ ├── parser.py # LlamaParse, Groq, report parsing/structuring
│ ├── email.py # Email utilities
│ └── ... # Other utilities
├── models/
│ └── schemas.py # Pydantic schemas
├── config/
│ └── settings.py # API keys, DB config
├── notebooks/
│ ├── RAG_Chatbot_Pinecone.ipynb
│ ├── LlamaParser.ipynb
│ └── ... # Demos and experiments
└── requirements.txt # Python dependencies
- Python 3.9+
- PostgreSQL
- Pinecone account (API key)
- Google Generative AI API key
- Groq API key
- LlamaParse API key
git clone https://github.com/tkarim45/CureWise-AI-Medical-Healthcare.git
cd CureWise-AI-Medical-Healthcaredocker compose up --buildcd backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt- Set API keys and DB credentials in
config/settings.pyor as environment variables:GOOGLE_API_KEYPINECONE_API_KEYGROQ_API_KEYLLAMA_PARSER_API_KEYDB_NAME,DB_USER,DB_PASSWORD,DB_HOST,DB_PORT
uvicorn main:app --reloadcd frontend
npm install
npm startTo set up PostgreSQL for local development, run the following commands:
brew install postgresql
brew services start postgresql
psql -l
createdb curewise
psql -U tkarim45 -d curewise -h localhost- Fork and clone the repo.
- Use feature branches and submit PRs.
- Follow best practices for Python, LLM prompt engineering, and API design.
- Add tests and update documentation for new features.
This project is licensed under the MIT License.
For questions or support, open an issue or contact the maintainer.