Mesa

Mesa is a Python-based agent-based modeling framework developed at the Santa Fe Institute, designed for researchers who need the flexibility of a general-purpose programming language without sacrificing the accessibility of a visual modeling environment. Unlike NetLogo, which isolates modelers from the underlying code, Mesa exposes the full Python stack: modelers write standard Python classes for agents, environments, and schedules, and can integrate with any scientific Python library. Mesa also provides a built-in browser-based visualization server that replicates NetLogo's real-time parameter manipulation, but the visualization is a convenience layer rather than the primary interface. The tradeoff is explicit: Mesa demands programming literacy but rewards it with unlimited extensibility. It is the tool for researchers who have outgrown NetLogo's toy-simulation aesthetic but still need to communicate their models to audiences who cannot read code. The danger of Mesa is that it occupies a precarious middle ground: too code-heavy for novices, too slow for high-performance computing, and too tied to Python's ecosystem to be language-agnostic.