An international research team, led by Hao Li, distinguished professor at Tohoku University's Advanced Institute for Materials Research, has unveiled DigMethpy, an artificial intelligence platform engineered to accelerate the discovery of molten catalysts for methane pyrolysis. This innovation targets a pivotal pathway for sustainable hydrogen production, offering a method to split methane into hydrogen and solid carbon while bypassing direct carbon dioxide emissions. A primary obstacle in methane pyrolysis involves navigating the extensive chemical design space of molten media. DigMethpy resolves this by synthesizing scientific literature, experimental data, computational simulations, machine learning models, and large language models into a cohesive discovery framework. The platform functions through a closed-loop workflow that automates information gathering, predicts promising catalyst candidates, and refines recommendations based on validation feedback. Its database comprises over 40,000 curated data points extracted from more than 500 scientific publications and computational records covering molten metals, alloys, salts, and mixed systems. Utilizing this integrated architecture, the researchers characterized essential chemical properties governing catalyst performance, such as atomic charge descriptors, diffusion dynamics, and hydrogen adsorption traits. These findings facilitated the design of highly active multicomponent molten alloy catalysts. The methodology enables scientists to optimize the utilization of voluminous scientific data while substantially curtailing the resources required for material discovery. DigMethpy represents a significant stride toward autonomous catalyst discovery by unifying experimental knowledge with advanced computational intelligence, thereby supporting more efficient scientific decision-making. The study appears in the journal AI Agents, in which Li serves as founding editor. The research team plans to expand the platform's data repository, enhance predictive accuracy, and engineer autonomous multi-agent systems to advance next-generation catalyst development.