vcenter

This MCP server provides a focused interface for managing vCenter environments through the Model Context Protocol. It enables both information retrieval about virtual machines on ESXi hosts and operational actions such as migrating VMs between hosts or on a host-wide scale. The server integrates with vCenter by utilizing the configured credentials and exposes actions via the MCP client layer, allowing automated workflows and toolchains to query VM status, perform migrations, and manage host state.

Using this server, you can create a VM by specifying host or cluster details, migrate single or multiple VMs to a target host, list VM information for a given host, and control host power states. The provided tooling is designed to be composable with MCP clients, so you can script migrations, gather summaries, and drive complex maintenance tasks from your existing automation pipelines.

Prerequisites:

• Python 3.8+ installed on the host
• Git available to clone the repository
• Access to a running vCenter instance with a valid account

Installation steps:

1. Clone the repository (adjust URL as needed): git clone https://gitee.com/rooky-top/vcenter-mcp-server.git cd vcenter-mcp-server

2. Install dependencies: pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple

3. Configure vCenter connection: Create or edit the .env file with vCenter details: VCENTER_HOST=your_vcenter_host VCENTER_USER=administrator@vsphere.local VCENTER_PASSWORD=your_password

4. Run the MCP server (via mcp-proxy in SSE mode as recommended): mcp-proxy --sse-host=0.0.0.0 --sse-port=8080 uv run vMotion_server.py

5. Configure the MCP client to point to this server: { "mcpServers": { "vCenter-mcp": { "url": "http://localhost:8080/sse" } } }

Tips and common issues:

• Ensure the vCenter API credentials have the necessary permissions for VM migration and folder/host view operations.
• If migrations fail, check that the target host is reachable and has enough resources; also verify network connectivity and datastore permissions.
• When running via mcp-proxy SSE, ensure the port is accessible from the MCP client environment.
• Keep the Python environment isolated (virtualenv) to avoid dependency conflicts.
• For security, avoid placing plaintext passwords in logs; consider environment-based secret management where supported by your deployment platform.