mcp -chart

This MCP server, mcp-server-chart, provides chart generation capabilities powered by AntV. It exposes a set of generate_* chart tools that you can invoke through the MCP protocol to produce various chart types (area, bar, boxplot, column, district-map, dual-axes, fishbone, flowchart, funnel, histogram, line, liquid, mind-map, network, organizational, path-map, pie, pin-map, radar, sankey, scatter, treemap, venn, violin, word-cloud, and spreadsheet/pivot table). You can run the server locally via npx and then reference it in your MCP-enabled client or desktop app integrations (e.g., Claude, VSCode, Cline, Cherry Studio, Cursor). The README shows how to configure the server for desktop apps that support the MCP config format, using an mcpServers entry named mcp-server-chart. Once running, you can request chart data and visuals through the MCP transport (SSE or HTTP) and receive chart configuration or rendered outputs as defined by the server.

Prerequisites:

• Node.js (LTS version) and npm installed on your machine
• Internet access to fetch the package from npm registry

Install options:

1. Run directly with npx (no global install):

npx -y @antv/mcp-server-chart

This will run the MCP server without installing a global package.

2. Install globally (optional) and run locally:

npm install -g @antv/mcp-server-chart
antv-mcp-server-chart  # or the exact start command described in docs

3. Alternative: Use within a project (local installation):

mkdir my-mcp-project
cd my-mcp-project
npm init -y
npm install @antv/mcp-server-chart --save

Then start the server using the package binary or a script defined in package.json.

Once installed, configure your MCP client to connect to the running server according to your environment (Desktop App, HTTP/SSE transport, etc.).

Tips and notes:

• The server exposes many generate_* chart tools (e.g., generate_line_chart, generate_bar_chart, generate_pie_chart, generate_network_graph, generate_treemap_chart, etc.). Each tool corresponds to a specific chart type and data schema; feed it appropriate data to obtain chart configurations or rendered outputs.
• Environment variables mentioned in the MCP chart server docs include VIS_REQUEST_SERVER, SERVICE_ID, and DISABLED_TOOLS. These can be used to customize request routing, track usage, or filter available tools. If deploying privately, consider setting VIS_REQUEST_SERVER to your request handling endpoint and DISABLED_TOOLS to a comma-separated list of tool names to disable.
• Run methods support SSE or HTTP-based transports. Check your client’s MCP integration for how to request and stream results.
• The package name for this Node.js server is @antv/mcp-server-chart; use the npm_package field to reference it in tooling or automation.