This project is a modular, multilingual AI assistant that combines retrieval-augmented generation (RAG), knowledge graphs, and dynamic domain orchestration to deliver intelligent, contextual, and personalized responses. Users can interact through text or voice, with support for Indonesian and English. The system detects the user’s intent and domain—food, travel, marketplace, or general—then builds a context-rich prompt using the user’s preferences, recent behavior, and relevant documents retrieved via semantic search (FAISS). A Groq-hosted LLaMA model generates responses with fallback logic per domain. The assistant evolves with user interactions, stores feedback, and enables visual graph exploration for transparency.
Category tags: