LLM Hub brings fast, private AI to Android. Run Gemma-3, Llama-3.2 & more fully on-device with RAG, embeddings, vector search, persistent memory, multimodal support, offline use, GPU acceleration & optional DuckDuckGo search — all privacy-first.