Mermaid MCP Server

@peng-shawn

代码仓库

基本信息

分类

其他

许可证

MIT

传输方式

stdio

作者

peng-shawn

提交者

Shawn Peng

关于 Mermaid MCP Server

A Model Context Protocol (MCP) server that converts Mermaid diagrams to PNG images

Mermaid MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "mermaid-mcp-server": {
      "command": "npx",
      "args": [
        "tsc"
      ]
    }
  }
}

Mermaid MCP Server 工具

工具列表暂未提供

当前收录暂时没有可展示的工具列表，不代表服务没有工具。请查阅项目文档了解支持的能力。

点击上方按钮,在临时沙箱中运行此服务器并实时获取工具列表。

概览

What is Mermaid MCP Server?

Mermaid MCP Server converts Mermaid diagram code into PNG images or SVG files. It uses Puppeteer for headless browser rendering and implements the Model Context Protocol (MCP) so AI assistants and other applications can generate visual diagrams from textual descriptions.

How to use Mermaid MCP Server?

Run the server with npx -y @peng-shawn/mermaid-mcp-server. Configure it as an MCP tool in Claude Desktop, Cursor, or Cline using the provided JSON entries. The server exposes a single generate tool that accepts parameters code, theme, backgroundColor, outputFormat, name, and folder. Set the environment variable CONTENT_IMAGE_SUPPORTED to true (default) for inline images or false to save to disk.

Key features of Mermaid MCP Server

Converts Mermaid code to PNG or SVG
Supports four diagram themes: default, forest, dark, neutral
Customizable background colors
Uses Puppeteer for high‑quality rendering
Implements MCP protocol for seamless AI integration
Returns images inline or saves them to disk

Use cases of Mermaid MCP Server

Generate flowcharts, sequence diagrams, and class diagrams in AI assistants
Create diagrams with specific themes and background colors
Save diagram files to disk when inline images are not supported (e.g., Cursor)
Automate diagram generation in MCP‑compatible tools

常见问题

Does Claude desktop already support Mermaid via canvas?

Yes, but it does not support the `theme` and `backgroundColor` options. A dedicated server makes it easier to create Mermaid diagrams with different MCP clients.

Why do I need to set `CONTENT_IMAGE_SUPPORTED=false` when using with Cursor?

Cursor does not support inline images in responses yet. Setting the variable to `false` saves the diagram to disk and returns the file path.

What are the runtime dependencies?

The server requires Node.js (to run the npx command) and relies on Puppeteer’s bundled Chromium or a system‑installed Chrome/Chromium for headless rendering.

How are generated images returned or stored?

When `CONTENT_IMAGE_SUPPORTED=true` (default), the image or SVG is returned inline in the MCP response. When `false`, the file is saved to the specified folder and the file path is returned.

Which transport does the server use?

The server communicates over stdio using the MCP protocol. It can be tested and debugged with the MCP Inspector (`npx @modelcontextprotocol/inspector`).

其他分类下的更多 MCP 服务器

Glasswarp

See and control a real Windows PC you own — from any MCP client, locally or remotely. Observe (UIA + screenshots), click/type/drag/scroll, launch apps, owner Live View. BYOH: your machine, your key.

Senso MCP for Joomla

sensomedia

Senso MCP turns a Joomla 5 or 6 site into a remote MCP server. It runs inside Joomla itself (no extra server, no Node.js bridge) and is secured with OAuth 2.1 and Joomla's own ACL. The free pack lets an AI assistant crea