pkirimi/mediassist-kenya

A simple medical chatbot built for clinicians to answer simple medical questions # MediAssist — AI Clinical Decision Support System 🏥🤖 > RAG-powered AI assistant that helps clinicians query medical guidelines, drug interactions, and diagnostic protocols in natural language. --- ## 🎯 Problem Clinicians spend valuable minutes searching through dense guidelines and drug databases. In low-resource settings, that delay costs lives. MediAssist lets a doctor ask: *"What's the first-line treatment for pre-eclampsia?"* and get an accurate, cited answer in seconds. --- ## ✨ Features - 📄 **Document ingestion** — upload clinical PDFs, guidelines, protocols - 🔍 **Semantic search** — finds relevant content even with imperfect queries - 🤖 **LLM response generation** — clear answers with source citations - 🧠 **Memory** — maintains context across multi-turn conversations - ⚡ **Optimised for low bandwidth** — minimal token usage, fast responses --- ## 🏗 Architecture ``` PDF Documents → Text Chunking → Embeddings → Pinecone Vector Store ↓ User Query → Embedding → Similarity Search → Relevant Chunks ↓ LLM (GPT / Claude) → Cited Answer ``` --- ## 🛠 Tech Stack | Layer | Technology | |-------|-----------| | LLM | OpenAI GPT-4 / Claude | | Orchestration | LangChain | | Vector DB | Pinecone | | Embeddings | OpenAI text-embedding-3-small | | API | FastAPI | | UI | Streamlit | | Deployment | Hugging Face Spaces | --- ## 🚀 Quick Start ```bash git clone github.com cd mediassist pip install -r requirements.txt cp .env.example .env # add your API keys streamlit run app.py ``` --- ## 📸 Demo Live Demo → --- ## 🙋 Built By **Purity Kirimi** — AI Engineer, Nairobi Kenya LinkedIn · GitHub