TypeScript MCP (Model Context Protocol) server that exposes 25+ AntV-powered chart-generation tools over stdio, SSE or streamable transports.
https://github.com/antvis/mcp-server-chartYour AI assistant can now create production-ready visualizations on demand—no context switching, no chart libraries, no manual coding. The AntV MCP Server Chart puts 25+ professional chart types directly into your AI workflow.
You're deep in analysis mode, data scattered across your terminal, when you need a quick visualization. The usual routine: save data to CSV, fire up a notebook, import matplotlib/D3/whatever, write chart code, tweak styling, export image. Five minutes later, you've lost your train of thought.
This MCP server eliminates that friction entirely. Your AI assistant becomes your visualization engine, generating everything from basic bar charts to complex Sankey diagrams without you writing a single line of chart code.
Built on AntV's battle-tested visualization engine, you get immediate access to:
Standard Charts: Line, bar, column, area, pie, scatter plots
Advanced Analytics: Box plots, violin plots, histograms, radar charts
Flow & Hierarchy: Sankey diagrams, treemaps, org charts, mind maps
Specialized: Funnel charts, liquid gauges, word clouds, Venn diagrams
Geographic: District maps, path maps, pin maps (China-focused)
Each chart type handles data validation, styling, and export automatically. No configuration files, no theme setup, no dependency management.
Data Analysis Sessions: "Show me a correlation matrix of these metrics as a scatter plot" → instant visualization, no notebook required
Report Generation: "Create a funnel chart for our conversion data" → professional chart ready for slides
Debugging Performance: "Generate a line chart of these API response times over the last week" → immediate visual debugging
Team Presentations: "Make a Sankey diagram showing our user flow" → publication-ready diagram in seconds
The server works with Claude Desktop, VSCode with Cline, Cursor, and any MCP-compatible AI client. Install once, use everywhere.
This isn't a toy visualization tool. AntV powers major enterprise dashboards, and you're getting the same rendering engine. Charts export as high-quality images ready for presentations, documentation, or web applications.
Want customization? The server accepts styling options, custom color schemes, and data transformations. Need private deployment? Point it to your own rendering service.
Add to your MCP configuration:
{
"mcpServers": {
"mcp-server-chart": {
"command": "npx",
"args": ["-y", "@antv/mcp-server-chart"]
}
}
}
That's it. Your AI assistant now has visualization superpowers.
For teams requiring private deployment, the server supports custom rendering endpoints. Deploy AntV's GPT-Vis-SSR service internally and configure the VIS_REQUEST_SERVER environment variable.
The server includes specialized tools for complex visualizations most developers avoid coding manually:
These typically require specialized libraries and significant setup time. Now they're a single prompt away.
Your AI workflow just got a massive productivity boost. No more breaking flow to generate visualizations. No more settling for text-only analysis when a chart would tell the story better.
Start generating charts that actually look professional, without the professional development time.