MCP FastAPI Chart Server
An MCP (Model Context Protocol) server that interacts with a FastAPI service to generate matplotlib charts.
Overview
This MCP server provides tools to:
- Generate charts using your FastAPI chart generation service
- Check the status of your FastAPI server
- Return chart images directly to the client
Prerequisites
- Python 3.8+
- Your FastAPI server running on
localhost:8000with the/plotendpoint
Installation
- Install the package:
pip install -e .
Configuration
For Claude Desktop
Add the following to your Claude Desktop configuration file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"fastapi-chart": {
"command": "python",
"args": ["-m", "mcp_fastapi_chart.server"]
}
}
}
For Other MCP Clients
Configure your MCP client to run: ``bash python -m mcp_fastapi_chart.server ``
Available Tools
1. generate_chart
Generate a chart using the FastAPI chart generation service.
Parameters:
chart_type(string): Type of chart to generate - "line", "bar", "scatter", or "pie"title(string): Main title of the chartsubtitle(string): Subtitle of the chartx_label(string): Label for the X-axisy_label(string): Label for the Y-axisdata(object): Chart data with labels and valueslabels(array of strings): Labels for data pointsvalues(array of numbers): Values for data points
Example Usage: ``json { "chart_type": "line", "title": "Asylum Applications Trend", "subtitle": "Monthly applications received (2023-2024)", "x_label": "Month", "y_label": "Applications Received", "data": { "labels": ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"], "values": [15400, 16800, 14200, 18900, 17600, 19500, 20300, 18700, 17400, 16200, 15800, 14900] } } ``
2. check_fastapi_status
Check if the FastAPI server is running and accessible.
Parameters: None
FastAPI Server Requirements
Your FastAPI server should have an endpoint at /plot that accepts POST requests with the following JSON structure:
{
"chart_type": "string",
"title": "string",
"subtitle": "string",
"data": {
"labels": ["string"],
"values": [number]
},
"x_label": "string",
"y_label": "string"
}
And should return a PNG image as the response.
Development
To run the server directly for testing:
python -m mcp_fastapi_chart.server
Cloud Deployment
When you move your FastAPI server to the cloud, update the fastapi_base_url in mcp_fastapi_chart/server.py:
self.fastapi_base_url = "https://your-cloud-domain.com"
Troubleshooting
- Connection Errors: Make sure your FastAPI server is running on
localhost:8000 - Chart Generation Errors: Verify your FastAPI
/plotendpoint is working correctly - MCP Connection Issues: Check your Claude Desktop configuration file syntax
License
MIT












