Deep knowledge AI agents for scientific discovery

[Submitted on 30 Oct 2025]

View PDF of the article Project Denario: Deep Knowledge AI Agents for Scientific Discovery, by Francisco Velascosa Navarro and 35 other authors

View PDF HTML (beta)

a summary:We present Denario, a multi-agent AI system designed to serve as a scientific research assistant. Denario can perform many different tasks, such as generating ideas, checking literature, developing research plans, writing and implementing code, creating plots, and drafting and reviewing a scientific paper. The system has a modular architecture, allowing it to handle specific tasks, such as generating an idea, or performing comprehensive scientific analysis using Cmbagent as a backend for deep research. In this work, we describe in detail Denario and its modules, and demonstrate its capabilities by presenting multiple AI-generated papers in many different scientific disciplines such as astrophysics, biology, biophysics, biomedical informatics, chemistry, materials science, mathematical physics, medicine, neuroscience, and planetary sciences. Denario also excels at bringing together ideas from different disciplines, and we demonstrate this by presenting a paper that applies methods from quantum physics and machine learning to astrophysical data. We report the evaluations made on these papers by domain experts, who provide both numerical scores and review-like comments. We then highlight the strengths, weaknesses and limitations of the current system. Finally, we discuss the ethical implications of AI-based research and consider how this technology relates to the philosophy of science. We release the code publicly at this https URL. The Denario demo can also be played directly on the web at this https URL, and the full application will be published to the cloud.