IDE and client configuration#

PyFluent-MCP can be integrated with multiple MCP-compatible tools. This page explains configuration for the most popular clients.

Claude Code#

Claude Code includes MCP support. You can add PyFluent-MCP using the command-line tool.

Set up PyFluent-MCP globally#

Configure PyFluent-MCP for all your Claude Code projects:

claude mcp add --transport stdio --scope user pyfluent -- uvx --from git+https://github.com/ansys/pyfluent-mcp ansys-fluent-mcp

Key features

  • Uses STDIO transport by default (local integration).

  • Uses uvx for automatic fetching from GitHub.

  • Requires no manual management of configuration files.

  • Provides full MCP protocol support.

Documentation

See Claude Code MCP installation documentation.

Visual Studio Code#

Visual Studio Code integrates MCP servers through the Copilot extension using a JSON configuration file.

Start quickly from GitHub#

Add this code to the .vscode/mcp.json file in your project directory:

{
  "servers": {
    "pyfluent": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/ansys/pyfluent-mcp",
        "ansys-fluent-mcp"
      ]
    }
  }
}

Features

  • Uses STDIO transport (recommended for local development).

  • Fetches the latest version from GitHub.

  • Requires uvx to be installed on your system.

Set up for local development#

Use this code for development or testing with local source code:

{
  "servers": {
    "pyfluent": {
      "type": "stdio",
      "command": "./.venv/Scripts/python",
      "args": ["-m", "ansys.fluent.mcp"],
      "env": {
        "FLUIDS_MCP_LOG_LEVEL": "DEBUG"
      }
    }
  }
}

On macOS or Linux, replace ./.venv/Scripts/python with ./.venv/bin/python.

Features

  • Uses a local Python virtual environment.

  • Enables debug logging for troubleshooting.

  • Works well for development and testing.

  • Requires pip install -e ".[pyfluent]" in your virtual environment.

Claude Desktop#

Add PyFluent-MCP to the Claude Desktop configuration file.

On macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

On Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "pyfluent": {
      "command": "ansys-fluent-mcp",
      "args": []
    }
  }
}

Ensure ansys-fluent-mcp is on your PATH (installed via pip) or provide the full path to the executable in your virtual environment.

Cursor#

Cursor supports MCP servers through its settings UI or a project-level configuration file. Add PyFluent-MCP using the same STDIO command as other clients:

{
  "mcpServers": {
    "pyfluent": {
      "command": "ansys-fluent-mcp",
      "args": []
    }
  }
}

For HTTP transport (for example, when running the server separately):

ansys-fluent-mcp --transport http --host 127.0.0.1 --port 8000

Then configure your client to connect to http://127.0.0.1:8000/mcp using streamable HTTP transport (refer to your client’s MCP documentation for the exact JSON schema).

Environment variables#

You can pass environment variables through the MCP client configuration. Common variables include:

Variable

Purpose

FLUIDS_MCP_LOG_LEVEL

Server log level (default INFO)

FLUIDS_MCP_HTTP_TIMEOUT

Timeout for outbound HTTP operations

FLUIDS_MCP_CA_BUNDLE

Path to a PEM CA bundle for corporate or self-signed TLS environments

FLUIDS_MCP_VERIFY_TLS

Enable or disable TLS verification (default enabled)

FLUIDS_MCP_SETTINGS_JSON

External file for overriding the bundled settings schema

For the full list, see Configuration.

Next steps#