Autonomous Universal Research Assistant (AURA): Agentic AI meets nanoHUB's FAIR Workflows and Data

Agentic artificial intelligence (AI) is poised to redefine how science is conducted, automating not just data analysis but the entire research lifecycle, from hypothesis generation to validation. Yet most current AI agents remain domain-bound, tailored to specific applications such as materials synthesis or quantum chemistry, limiting their applicability and impact. We introduce AURA, a multi-agent, autonomous research assistant that integrates a large language model (LLM) with nanoHUB's open scientific ecosystem, encompassing over 340 FAIR-compliant simulation tools and 1.6 million data points contributed by its users. Unlike prior systems, AURA dynamically discovers and orchestrates simulations and data through runtime metadata reasoning, enabling seamless execution of research plans that span multiple disciplines. As nanoHUB expands, AURA automatically integrates newly published simulation tools and associated datasets, learning their services and requirements from their metadata. Through case studies in protein folding, high-pressure materials properties, and the prediction of melting temperatures of high-entropy alloys, we demonstrate AURA's capacity for scientific ideation and planning, adaptive workflow composition, and autonomous execution. Since simulation results are automatically stored for every simulation executed in nanoHUB, AURA naturally learns from prior work, including negative results, to optimize simulation setups. AURA reproduced a recently published result in minutes, a 10x speedup over advanced users with computational and domain expertise. AURA marks a step toward general-purpose, scalable, and expandable AI research assistants capable of accelerating discovery in the physical and life sciences.