mcp-server-chart

Why we included this project

LLM agents and chat assistants that need to turn raw data into readable charts can use this MCP server, which exposes more than two dozen visualization tools built on the AntV stack. Each chart type is a separate tool, from bars and scatter plots to Sankey diagrams, boxplots, treemaps, maps, and pivot tables, so a model can pick the right visual for the question at hand instead of hand-rolling matplotlib code. Teams already on MCP clients such as Claude Desktop, Cursor, or Cherry Studio can wire it in with a short config, and it plugs into platforms like Dify for agent-driven reporting. Because it returns rendered charts rather than code snippets, analysts and developers get a finished visual in a consistent style without having to run generated plotting code. The bundled data-analysis skills also make it handy for quick exploratory plots before you move on to deeper analysis.