Logo Lanfrica

mekonnes/Her-wellness

Domaine:

healthcarenatural language processing

Type de record:

software
Créateur:
mek
Hôte:
AI-powered health education tool for Black and African women # Her Wellness — AI Health Education for Black Women An AI-powered health education tool built specifically for Black and African women. Uses a RAG (Retrieval Augmented Generation) pipeline to provide accurate, culturally aware health information grounded in medically reviewed sources. --- ## Live Demo Live App --- ## What It Does Users select a health topic and ask any question in plain English. The system retrieves relevant information from a curated medical knowledge base and generates a clear, accurate, culturally aware response grounded only in verified sources — never hallucinated. ### Health Topics Covered - **Fibroids** — symptoms, treatment options, why Black women are disproportionately affected - **Maternal Health** — maternal mortality disparities, warning signs, advocacy strategies - **Mental Health** — stigma, barriers to care, finding culturally competent therapists - **Hypertension** — risk factors, prevention, management for Black women - **Nutrition** — culturally relevant dietary guidance, managing chronic conditions through diet - **Navigating Healthcare** — patient rights, self-advocacy, finding culturally competent providers --- ## How It Works User Question → FAISS Vector Search → Retrieve Relevant Document Chunks → Groq LLM → Culturally Aware Answer 1. **Knowledge Base** — 6 text documents written from NIH, CDC, Office on Women's Health, and Black Women's Health Imperative sources 2. **Embeddings** — HuggingFace all-MiniLM-L6-v2 model converts documents into vector representations 3. **FAISS** — vector similarity search retrieves the most relevant document chunks for each question 4. **Groq LLM** — llama-3.3-70b-versatile generates a response grounded only in retrieved context 5. **Streamlit** — clean elegant UI with black and gold design --- ## Why RAG Instead of a General LLM A general LLM like ChatGPT can hallucinate medical information — confidently giving wrong answers. RAG grounds every response in specific verified documen …