home / mcp / cvp mcp server

CVP MCP Server

Provides a local MCP server to access CloudVision data and extend with custom tooling via Claude Desktop.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "noredistribution-mcp-cvp-fun": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "fastmcp",
        "fastmcp",
        "run",
        "/home/ansible/mcp-cvp-fun/mcp_server_rest.py"
      ],
      "env": {
        "CVP": "YOUR_CVP_URL",
        "CVPTOKEN": "YOUR_CVPTOKEN"
      }
    }
  }
}

This MCP server lets you connect Claude Desktop to a dedicated MCP endpoint so you can query connectivity data, create tags for Studios and dashboards, and extend capabilities with custom API calls. It runs as a local process and is designed to be used with a client that supports MCP tooling and the @mcp.tool decorator for extending functionality.

How to use

Install and run the MCP server alongside your MCP client. You will configure the client to talk to this local MCP endpoint and then issue prompts to retrieve connectivity data or trigger actions via the MCP tooling interface.

Practical usage patterns you can perform once the server is running include: - Request connectivity monitor data for your environment - Create tags for Studios and dashboards to organize your MCP workspace - Extend with additional API calls by decorating functions with @mcp.tool() as needed by your workflow

How to install

Prerequisites: ensure you have a runtime capable of executing the MCP server configuration and the client you plan to use. In this example setup, you will configure Claude Desktop to run a local MCP server instance.

1. Create a .env file with your CloudVision token and CloudVision URL.

2. Download Claude Desktop.

3. In Claude's developer settings update the claude_desktop_config.json to point at your MCP server as shown in the snippet.

4. Reload Claude Desktop and start prompting to begin using the MCP server.

Configuration and usage notes

Environment variables you will typically set are: - CVPTOKEN: CloudVision service account token - CVP: CloudVision endpoint URL These variables enable the MCP server to authenticate to your CloudVision environment and access needed data.

The MCP server entry shown in the JSON configuration demonstrates how to run the local process via the uv runtime, with arguments that specify using fastmcp and pointing to the Python script that implements the MCP server.

Example usage flow includes starting Claude Desktop with the provided configuration, then issuing prompts to interact with the MCP server for connectivity metrics and tag management.

Troubleshooting and notes

If you encounter authentication issues, verify that CVPTOKEN has the correct permissions for the CloudVision API and that CVP points to the correct CloudVision endpoint. Ensure Claude Desktop is restarted after updating claude_desktop_config.json.

When adding new API calls via the @mcp.tool() decorator, follow the existing pattern to expose new endpoints and test them from your MCP client to confirm proper integration.

Examples and starter config

The following configuration snippet shows how Claude Desktop should be configured to run the MCP server locally. Copy this into claude_desktop_config.json under the mcpServers section.

{
  "mcpServers": {
    "CVP MCP Server": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "fastmcp",
        "fastmcp",
        "run",
        "/home/ansible/mcp-cvp-fun/mcp_server_rest.py"
      ]
    }
  }
}