MCP 服务器

highchart-mcp-server

Enables AI assistants and clients to generate Highcharts-based charts with schema validation, export to PNG/SVG/PDF, and natural language to chart conversion.

README

Highcharts MCP Server

A Model Context Protocol (MCP) server designed to generate Highcharts‑based charts and visualizations in a production‑ready, AI‑enhanced, and validated manner. This server can be integrated with any MCP‑capable AI client (such as Claude, ChatGPT, Cursor, Dify, VSCode, etc.) to automate chart generation, validation, and export workflows.

Overview

The Highcharts MCP Server provides:

Chart generation for all common Highcharts types
Schema validation of chart configurations
Production grade features for scalability, security, and monitoring
AI‑assisted functionality, such as natural language chart generation
CLI and SDK tooling for developers

It acts as an MCP tool suite exposing Highcharts chart generation endpoints that can be invoked by AI assistants or custom clients.

Features
Architecture
Installation
AI Client Integration
Usage
Validation System
Production Readiness
AI & LLM Integration
Tooling & Developer Experience
Analytics & Monitoring
Contributing
License

1. Features

Core Chart Capabilities

The MCP server supports generating charts using Highcharts configurations:

Line, Column, Bar
Area, Scatter, Pie
Gauge, Treemap, Heatmap
Financial charts (OHLC, Candlestick)
Specialized chart types (Sankey, Network, Timeline)
Export to PNG, SVG, PDF, and interactive HTML bundles (Concepts based on MCP chart servers and Highcharts Node export support)([highcharts.com][2])

Schema Validation

All incoming chart configuration requests are validated against JSON schema or Zod schemas:

Type enforcement for chart options
Validation of data arrays, axis definitions, series structure
Detailed and structured error responses on invalid configs (Validation approach inspired by existing MCP chart servers)([GitHub][3])

Production‑Grade Reliability

Scalable deployment via Docker, Kubernetes
Multi‑transport support: HTTP(S), SSE, Streamable HTTP, STDIO
Health check endpoints
High availability and load balancing
Configurable caching and rate limits

Security & Access Control

API key, OAuth2, and JWT authentication
RBAC (Role‑based access control)
Rate limiting and quotas
Audit logging for compliance

AI & LLM Enhancements

Natural language to chart config translation
Chart type suggestions based on dataset or analytics question
Auto‑correction of invalid configs using AI
Integration hooks for LLM workflows

Export & Output

Static and interactive chart outputs
Batch and scheduled export capabilities
Link generation for shareable chart assets

Monitoring & Observability

Prometheus / Grafana integration
Structured logging
Performance metrics
Error reporting dashboards

Tooling & SDKs

CLI tooling for quick local operations
JavaScript and Python SDKs
Interactive schema documentation
Playground UI for testing configurations

2. Architecture

The server consists of:

Transport layer — Handles different MCP transports (HTTP/SSE, streamable, STDIO)
Schema validation layer — Ensures chart configs are correct
Chart renderer — Uses Highcharts engine (Node export or server rendering)
AI processing layer — Optional natural language preprocessing
Output services — Export and sharing

The server uses a modular structure to isolate responsibilities and enable plug‑ins/extensions.

3. Installation

Prerequisites

Node.js 18+
Docker (optional)
Kubernetes (optional)

Local Setup

Clone the repository:

git clone https://your-repo-url
cd highcharts-mcp-server
npm install
npm run build
npm start

Docker

Build the Docker image:

docker build -t highcharts-mcp-server .
docker run -p 8080:8080 highcharts-mcp-server

4. AI Client Integration

This MCP server works with any MCP-capable AI client. Below are configuration examples for common clients.

Claude Desktop

Add to your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "highchart-mcp-server": {
      "command": "node",
      "args": ["/absolute/path/to/highchart-mcp-server/dist/index.js"],
      "env": {
        "TRANSPORT": "stdio",
        "LOG_LEVEL": "info"
      }
    }
  }
}

Cursor / VS Code

Add to .cursor/mcp.json or .vscode/mcp.json in your project root:

{
  "mcpServers": {
    "highchart-mcp-server": {
      "command": "node",
      "args": ["/absolute/path/to/highchart-mcp-server/dist/index.js"],
      "env": {
        "TRANSPORT": "stdio",
        "LOG_LEVEL": "info"
      }
    }
  }
}

Streamable HTTP (Network Clients)

For clients that connect over the network, start the server with the HTTP transport:

TRANSPORT=streamable-http PORT=3000 node dist/index.js

Then connect your MCP client to http://localhost:3000/mcp.

Available Tools

Tool	Description
`create_chart`	Generate a Highcharts config from structured input (type, title, series, categories). Best for simple charts with guided parameters.
`render_chart`	Render a chart from a full Highcharts configuration object. Accepts any valid Highcharts options — best for advanced/custom charts.

Basic Chart Call

Chart generation is done by invoking MCP tools with chart config objects. For example:

{
  "tool": "create_chart",
  "input": {
    "type": "line",
    "data": {
      "categories": ["Jan", "Feb", "Mar"],
      "series": [
        {"name": "Sales", "data": [10, 15, 20]}
      ]
    },
    "options": {}
  }
}