Nutanix released an open-source Model Context Protocol (MCP) server for its Nutanix Cloud Platform (NCP) on Aug. 10, giving AI assistants such as GitHub Copilot, Claude Code and Cursor a controlled path into infrastructure management through the Nutanix Prism V4 API. The server translates natural-language requests into API operations, letting operations teams automate cloud-management tasks while keeping existing access policies in place. The initial v0.8 release is a Technical Preview and is not supported for production workloads.
How the MCP server works
The server sits between AI tools and Nutanix infrastructure, using the Prism V4 API Gateway for execution, governance and security. Connected AI assistants operate within the role-based permissions assigned to their Prism Central identities, so they don't get unrestricted access to the platform. The server defaults to read-only enforcement, blocking non-GET operations unless administrators explicitly enable write access.
Nutanix also documents auditing of operations, throttling, metering and asynchronous task management for longer-running infrastructure actions. The Model Context Protocol server requires Python 3.11 or later and connects to Prism Central over its V4 APIs. It dynamically discovers available API namespaces on the connected deployment, so its tools reflect the platform version in use.
What the server exposes
According to Nutanix, the server exposes more than 1,000 V4 API operations and can provide AI tools with API schemas, code samples and permission information. Developers can use that data to generate scripts in Python, Go, Java, JavaScript, PowerShell, curl and REST/JSON-compatible formats.
"By creating a secure gateway between AI tools and our platform, we are giving customers the confidence to safely use AI to operate and govern their hybrid multicloud environments," said Thomas Cornely, executive vice president of Product Management at Nutanix.
For Microsoft-centric teams, GitHub Copilot is among the assistants Nutanix identified as compatible. GitHub documents MCP support for Copilot Chat in Visual Studio Code and Visual Studio, though organizations using Copilot Business or Copilot Enterprise must enable the appropriate MCP policy before connecting servers. Nutanix's current integration docs include setup examples for Cursor and Claude, while GitHub Copilot-specific configuration isn't documented yet.
Why this matters for operations
The read-only default is the detail that matters most for operations teams evaluating this tool. It means an AI assistant can inspect system health, surface diagnostics and prepare automation workflows without the ability to change anything until an administrator flips the switch. That's a practical middle ground between letting AI touch production infrastructure and keeping it entirely out of the loop. The open-source nature of the v0.8 release also means teams can test it in non-production environments and review exactly what it does before deciding whether it earns a place in their toolchain. For operations professionals exploring AI for Operations, this is a concrete example of how AI assistants can fit into existing governance structures rather than bypassing them.
Your membership also unlocks:
