# AntV Chart MCP — Generate Charts from Data

> AntV Chart MCP generates charts via MCP using `npx -y @antv/mcp-server-chart`; verified 4,062★ and supports SSE + streamable transports for 25+ types.

## Install

Merge the JSON below into your `.mcp.json`:

## Quick Use

```bash
# Add to your MCP client config (.mcp.json or equivalent):
# command: npx
# args: ["-y", "@antv/mcp-server-chart"]
npx -y @antv/mcp-server-chart
```

## Intro

AntV Chart MCP generates charts via MCP using `npx -y @antv/mcp-server-chart`; verified 4,062★ and supports SSE + streamable transports for 25+ types.

**Best for:** Turning CSV/JSON into charts during agent runs (reports, dashboards, analysis)

**Works with:** Claude Desktop, VS Code MCP, Cursor/Cline and other MCP clients (npx-based stdio)

**Setup time:** 3-10 minutes

### Key facts (verified)

- GitHub: 4063 stars · 376 forks · pushed 2026-05-06.
- License: MIT · owner avatar + repo URL verified via GitHub API.
- README-backed entrypoint: `npx -y @antv/mcp-server-chart`.

## Main

- Use it as a structured output step: generate a small JSON dataset + chart intent, then call the tool to render consistently.
- Pick transport intentionally: stdio for local clients; SSE/streamable endpoints when you need HTTP access.
- If you need chart-type selection, pair it with the `chart-visualization-skills` repo referenced in README.

### Source-backed notes

- README provides MCP client config examples where the server is launched via `npx -y @antv/mcp-server-chart`.
- README documents running the server with `--transport sse` (endpoint `/sse`) and `--transport streamable` (endpoint `/mcp`).
- README mentions a skill companion repo and claims the skill can choose from 25+ chart types.

### FAQ

- **Is npx required?**: No—README also shows `npm install -g`, but `npx -y` is the lowest-friction start.
- **Which transport should I use?**: Use stdio for local desktop clients; use SSE/streamable for remote HTTP access.
- **Can it pick the chart type for me?**: Use the related skill repo referenced in README for automatic chart selection logic.

## Source & Thanks

> Source: https://github.com/antvis/mcp-server-chart
> License: MIT
> GitHub: ⭐ 4063 · forks 376

Thanks to the upstream maintainers for shipping the original project.

---

<!-- ZH -->

## 快速使用

```bash
# Add to your MCP client config (.mcp.json or equivalent):
# command: npx
# args: ["-y", "@antv/mcp-server-chart"]
npx -y @antv/mcp-server-chart
```

## 简介

AntV Chart MCP 让 agent 通过 MCP 生成图表（`npx -y @antv/mcp-server-chart`）；已验证 4,062★，文档提到 25+ 图表类型，并支持 SSE/streamable 传输。

**适合谁用：** 在 agent 任务中把 CSV/JSON 变成图表（报告/看板/分析）的用户

**适用环境：** Claude Desktop、VS Code MCP、Cursor/Cline 等支持 MCP 的客户端（npx 方式）

**配置耗时：** 3-10 minutes

### 关键事实（已验证）

- GitHub：4063 stars · 376 forks；最近更新 2026-05-06。
- 许可证：MIT；作者头像与仓库链接均已通过 GitHub API 复核。
- README 中可对照的入口命令：`npx -y @antv/mcp-server-chart`。

## 正文

- 把它当作“结构化输出”步骤：先产出小型 JSON 数据集 + 图表意图，再调用 chart 工具生成一致的可视化结果。
- 按场景选择 transport：本地用 stdio；需要 HTTP 访问则用 SSE/streamable。
- 如果想自动选图表类型，可结合 README 提到的 `chart-visualization-skills`。

### 基于来源的要点

- README 提供 MCP 客户端配置示例：通过 `npx -y @antv/mcp-server-chart` 启动 server。
- README 说明 `--transport sse`（`/sse`）与 `--transport streamable`（`/mcp`）两种运行方式。
- README 提到配套 skill 仓库，并声称可覆盖 25+ 图表类型选择。

### 常见问题

- **必须用 npx 吗？**：不必须。README 也给出 `npm install -g`；但 `npx -y` 最省事。
- **transport 选哪个？**：本地桌面客户端选 stdio；需要远程 HTTP 访问则选 SSE/streamable。
- **能自动选图表类型吗？**：README 提到配套 skill 仓库，可用于自动选图逻辑。

## 来源与感谢

> Source: https://github.com/antvis/mcp-server-chart
> License: MIT
> GitHub: ⭐ 4063 · forks 376


---
Source: https://tokrepo.com/en/workflows/antv-chart-mcp-generate-charts-from-data
Author: MCP Hub