Provides natural language control over SUMO traffic simulations through 26 tools grouped into scenario generation, policy experimentation, result analysis, visualization, and utilities. You can extract OpenStreetMap data, generate networks and traffic demand, apply interventions like lane reductions or signal timing changes, then convert XML outputs to SQLite for cross-scenario analysis. Built for policy questions that need simulation backing: road closures, demand shifts, signal optimization. Includes tools for network statistics, routing, coordinate validation, and rendering heatmaps of per-edge metrics. The web dashboard shows geospatial overlays and trip replay, but the MCP layer handles the simulation orchestration and data pipeline from OSM download through route generation to HTML reports.
claude mcp add --transport stdio io.github.mw-jeong-agentsumo-mcp uvx agentsumo-mcp