antvis/mcp-server-chart

What it solves

This project provides a way for AI models to generate professional data visualizations. It bridges the gap between raw data and visual representation by allowing LLMs to call specific tools to create a wide variety of charts and diagrams using the AntV library.

How it works

It is implemented as a Model Context Protocol (MCP) server. This means it can be integrated into AI clients (like Claude, Cursor, or VSCode) that support MCP. The server exposes a set of tools (e.g., generate_bar_chart, generate_pie_chart) that the AI can trigger. When the AI decides to a chart is needed, it sends the request to the server, which then uses the AntV ecosystem to render the visualization.

Who it’s for

• AI Application Developers: Those building agents or IDEs that need native chart generation capabilities.
• Data Analysts: Users who with want to integrate data analysis and visualization directly into their AI-powered workflows.
• LLM Users: People using MCP-compatible clients to turn data into visuals instantly.

Highlights

• Extensive Chart Library: Supports over 26 different types of visualizations, including standard charts (bar, line, pie), complex diagrams (Sankey, fishbone, mind maps), and geographic maps (district, pin, and path maps).
• Flexible Deployment: Can be run via npx, Docker, or as a standalone server using SSE or Streamable transport protocols.
• Private Deployment Option: Supports custom chart generation service URLs via the VIS_REQUEST_SERVER environment variable for users who need private infrastructure.
• Skill Integration: Compatible with AI IDEs that support skills (like Claude Code) via the chart-visualization skill for automatic chart selection.