Herbal Remedies Guide

An AI-powered herbal remedies chatbot and natural medicine research assistant. This RAG (Retrieval-Augmented Generation) system provides evidence-based answers about herbal remedies, medicinal plants, and traditional medicine.

Features

• Natural Language Q&A: Ask questions about herbal remedies and medicinal plants in plain English
• Book-Based Knowledge: Queries The Little Handbook of Natural Remedies via vector embeddings
• Source Citations: Answers include context retrieval and citations from the handbook
• Semantic Search: Uses sentence transformers for context-based information retrieval
• Real-time Responses: WebSocket support for interactive conversations
• LLM-Enhanced: Generates natural language responses with expansion beyond book scope

Payment & Access Control

• Stripe integration for premium access
• Vercel KV storage for user tracking
• Conditional paywall rendering

How It Works

1. Document Ingestion: Processes handbook with hash-based change detection
2. Text Chunking: Splits content into configurable chunks
3. Embedding Generation: Creates semantic vectors using sentence transformers
4. Vector Storage: Saves embeddings as numpy arrays
5. Query Processing: Finds relevant sections using cosine similarity
6. Response Generation: LLM synthesizes answers from retrieved context with citations
7. LLM Expansion: Supplements book knowledge for broader queries

Performance Optimizations

• Lazy loading of ML models
• Vector caching with hash-based validation
• Batch processing for embeddings
• Request validation and rate limiting