mcp-server-chart MCP Server

Overview

The mcp-server-chart project is a TypeScript-based Model Context Protocol server developed by antvis. It enables large language models to handle chart generation and data analysis using AntV visualization tools. The server translates tool calls into structured visual outputs, allowing connected desktop clients, integrated development environments, and web services to render a wide selection of diagrams directly within conversational or analytical workflows.

This server supports over 26 distinct visualization tools covering statistical, hierarchical, relational, and spatial representations. Integrations are available across multiple desktop apps such as Claude, Cursor, Cherry Studio, VSCode, and Cline, as well as cloud platforms like Dify, Aliyun, ModelScope, Glama.ai, and Smithery.ai. It can operate over stdio, HTTP, and Server-Sent Events transports depending on the deployment setup.

What it can do

When a client requests a chart, the server invokes specific tools dedicated to individual chart types. For example, tools exist for basic plots like line, bar, pie, and scatter charts, as well as complex formats like Sankey diagrams, treemaps, violin plots, and liquid charts. It can also generate dynamic structural diagrams such as mind maps, flowcharts, fishbone diagrams, and organizational charts.

For tabular and geographic data, the server offers specialized tools. The spreadsheet tool renders standard tables or pivot tables based on row and value configurations. Spatial features include district maps, path maps, and pin maps. The server processes geographic rendering through the AMap service, which restricts geographic chart output strictly to regions within China.

Users can modify server runtime behavior using environment variables. Setting DISABLED_TOOLS allows specific tools to be excluded via a comma-separated list. Setting VIS_REQUEST_SERVER directs chart generation calls to a custom private deployment endpoint, while SERVICE_ID tracks service records.

Install and configure

These commands were found in the project's own README. We checked each package name against the public registry.

npx skills add antvis/chart-visualization-skills

(skills found on npm, latest 1.7.4)

npm install -g @antv/mcp-server-chart

(@antv/mcp-server-chart found on npm, latest 0.9.10)

npm install

(not a registry package: check the source before running it)

The official MCP Registry lists this server with these packages:

Example configuration from the README (check the project's documentation for the current format):

{
  "mcpServers": {
    "mcp-server-chart": {
      "command": "npx",
      "args": ["-y", "@antv/mcp-server-chart"]
    }
  }
}

Environment variables:

Requirements

The package declares Node.js >=18 in its engines field. The README mentions setup for these clients: Claude Code, Cursor, Cline.

Use cases

Data analysts and developers can use this server to render categorical, continuous, or statistical comparisons directly from query results. By invoking tools for histograms, dual-axes charts, boxplots, or word clouds, users can quickly review data distributions, trends, and key term frequencies without switching to standalone plotting software.

Organizations needing process, hierarchy, or relation mapping can leverage specialized tools like organization charts, network graphs, flowcharts, and Venn diagrams. Teams designing workflows or root-cause analyses can generate structured diagrams like Ishikawa fishbone layouts directly through automated prompt interactions.

This server suits software developers, data analysts, and technical operators who use Model Context Protocol clients and need automated data visualization. It is especially tailored for users leveraging the AntV ecosystem or building interactive AI tools that require dynamic chart and diagram generation.

Limitations and things to check

Geographic visualization tools—including district maps, path maps, and pin maps—rely on the external AMap service. Because of this integration, geographic chart generation is limited strictly to locations within China. Users outside this region or analyzing global spatial data cannot rely on these specific tools.

Prior issue logs reveal past operational errors, such as internal error code negative 32603 during chart generation calls and client connection blocking when multiple clients connect concurrently. Potential users should review deployment configurations and test concurrent access behaviors before introducing the server to critical environments.

Before you connect it:

• Read the code or the documentation of the tools it exposes, and note what each tool can do.
• Grant the smallest permissions you can: limit folders, API scopes and database roles.
• Keep credentials in environment variables or a secrets manager, never in a shared config file.
• Run it in a container or a separate user account if it can execute code or touch files.

Release history

The most recent releases published on GitHub:

• 0.9.10 (2026-02-25): ## What's Changed Move skills to external repository and update README by @Copilot in https://github.com/antvis/mcp-server-chart/pull/263 Fix high CPU usage during concurrent deployments by @Copilot in https://github.com/antvis/mcp-server-chart/pull/273 Fix TypeError in Docker local deployment by…
• 0.9.9 (2026-01-22): ## What's Changed feat: Add tool annotations for improved LLM tool understanding by @triepod-ai in https://github.com/antvis/mcp-server-chart/pull/249 feat: add chart-visualization skill by @hustcc in https://github.com/antvis/mcp-server-chart/pull/255 Update source identifier to…
• 0.9.7 (2025-12-24): ## What's Changed feat: add waterfall by @XinyueDu in https://github.com/antvis/mcp-server-chart/pull/242 feat: add startAtZero by @XinyueDu in https://github.com/antvis/mcp-server-chart/pull/247 feat: update modelcontect version by @XinyueDu in https://github.com/antvis/mcp-server-chart/pull/248…

Known issues and troubleshooting

These are the most discussed threads on the project's issue tracker. They show what other users ran into; read them before you install, and check whether they are still open.

• dify 报错400 (closed, 30 comments)
• 这个mcp两个client同时连接第2个会阻塞 (closed, 19 comments)
• MCP error -32603 error failing to generate chart while calling 'generate_line_chart' tool (closed, 19 comments)
• 本地客户端能正常识别工具，但调用时失败：MCP error -32603 (closed, 16 comments)

Maintenance and community

The repository has 24 contributors, 28 commits in the last 12 months, 7 open issues and pull requests. It is maintained by AntV AI Visualization Team, a GitHub organisation with 127 public repositories and 2,383 followers. GitHub's community checks found no security policy file, no contribution guidelines, no code of conduct. The npm package has 2 maintainers, 2 direct dependencies and was downloaded 5,890,698 times in the last week.

Frequently asked questions

Is mcp-server-chart free to use?

The repository is released under the MIT license.

When was mcp-server-chart last updated?

The last commit was on 2026-08-27, and the project is marked active by our 12-month check.

Which AI clients does mcp-server-chart work with?

Its README gives setup notes for Claude Code, Cursor, Cline. Other MCP-compatible clients may also work, but the project does not say so.

What environment variables are available for configuration?

You can configure VIS_REQUEST_SERVER for custom private deployments, SERVICE_ID for chart generation tracking records, and DISABLED_TOOLS to supply a comma-separated list of tool names to disable.

Which transports and execution methods are supported?

The server supports stdio transport via standard Node package runners, as well as HTTP, Server-Sent Events (SSE), and streamable transports for cloud or remote deployments.

Can geographic charts map locations outside China?

No. Geographic tools such as pin maps, district maps, and path maps utilize the AMap service, which currently supports chart generation exclusively within China.

How can I disable specific chart generation tools?

Provide a comma-separated list of the tool names you wish to turn off using the DISABLED_TOOLS environment variable.

Sources

This page combines data from:

• mcp-server-chart on GitHub
• Official MCP Registry entry (registry.modelcontextprotocol.io)
• npm package skills