home / mcp / mcp demo mcp server
基于模型上下文协议（MCP）的数据服务平台，集成数据可视化、SQL 智能查询和链接管理功能---项目持续更新中...
Configuration
View docs{
"mcpServers": {
"ayi456-visual-mcp": {
"url": "http://mcp.zha-ji.cn/mcp"
}
}
}You have a practical MCP server that connects data sources, runs AI-driven SQL queries, and visualizes results as shareable charts. It supports both online access and local deployment, letting you call it from any MCP-enabled client or CLI to generate charts and drive dashboards.
You connect to the MCP server from an MCP client or AI CLI by pointing it to the MCP endpoint and using the built-in tools to create visualizations from SQL results. The primary HTTP endpoint is available at http://mcp.zha-ji.cn/mcp for remote usage. If you run everything locally, you can start the server and point your client to http://localhost:3000/mcp. Common actions include starting a SQL Chat session, asking the AI to generate SQL queries, validating results, and then turning those results into charts that you can share or add to a dashboard.
Practical usage pattern you can follow in an MCP client:ine-tune your database connection in the SQL Chat area, ask a natural language question like “Show the order count by date for the last 7 days,” let the AI generate and run the SQL, review the results, and then request a visualization. Choose a chart type (Auto works if you are unsure), specify the X axis (usually time or category) and the Y axis (the numeric value), and generate the chart to obtain a shareable link or to drop into a dashboard.
Generates a visualization chart from provided data, schema, and chart type, returning a shareable chart link that can be embedded in dashboards or shared with others.